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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
99318390-0651-435a-a0c2-55192d7e4702 | dual-variational-generation-for-low-shot-1 | null | null | http://papers.nips.cc/paper/8535-dual-variational-generation-for-low-shot-heterogeneous-face-recognition | http://papers.nips.cc/paper/8535-dual-variational-generation-for-low-shot-heterogeneous-face-recognition.pdf | Dual Variational Generation for Low Shot Heterogeneous Face Recognition | Heterogeneous Face Recognition (HFR) is a challenging issue because of the large domain discrepancy and a lack of heterogeneous data. This paper considers HFR as a dual generation problem, and proposes a novel Dual Variational Generation (DVG) framework. It generates large-scale new paired heterogeneous images with the same identity from noise, for the sake of reducing the domain gap of HFR. Specifically, we first introduce a dual variational autoencoder to represent a joint distribution of paired heterogeneous images. Then, in order to ensure the identity consistency of the generated paired heterogeneous images, we impose a distribution alignment in the latent space and a pairwise identity preserving in the image space. Moreover, the HFR network reduces the domain discrepancy by constraining the pairwise feature distances between the generated paired heterogeneous images. Extensive experiments on four HFR databases show that our method can significantly improve state-of-the-art results. When using the generated paired images for training, our method gains more than 18\% True Positive Rate improvements over the baseline model when False Positive Rate is at $10^{-5}$. | ['Huaibo Huang', 'Yibo Hu', 'Xiang Wu', 'Ran He', 'Chaoyou Fu'] | 2019-12-01 | null | null | null | neurips-2019-12 | ['heterogeneous-face-recognition'] | ['computer-vision'] | [ 1.07636392e-01 -1.35853386e-03 1.77403539e-02 -2.85318017e-01
-1.19780588e+00 -2.74830818e-01 4.24568981e-01 -8.27106416e-01
-3.12505066e-02 8.74890566e-01 1.26995638e-01 2.74405599e-01
-3.47352251e-02 -7.85960317e-01 -7.91536331e-01 -1.06920898e+00
4.98607159e-01 5.68790019e-01 -2.22940132e-01 -1.85206756e-01
-1.18266627e-01 2.00170904e-01 -1.67778170e+00 3.48052025e-01
7.99308658e-01 1.15841770e+00 2.70729661e-01 1.64375138e-02
1.45860389e-01 5.05045474e-01 -7.20059335e-01 -7.38590658e-01
6.07544243e-01 -5.72154164e-01 -3.23554844e-01 3.06928605e-01
6.28346324e-01 -3.07441682e-01 -5.72963893e-01 1.40221810e+00
8.56436193e-01 7.46794045e-02 7.33482897e-01 -1.60500407e+00
-1.18157375e+00 3.03364545e-01 -7.42745697e-01 -2.71710008e-01
2.23655418e-01 4.09500860e-02 7.18239486e-01 -1.18454731e+00
9.29839373e-01 1.40912700e+00 4.85150158e-01 9.61290240e-01
-1.41266429e+00 -1.08833957e+00 1.27471268e-01 3.62029135e-01
-1.89287210e+00 -4.15715009e-01 8.74352694e-01 -4.62910086e-01
4.02305990e-01 1.90355569e-01 3.28298002e-01 1.46603596e+00
-2.22114280e-01 6.18059933e-01 1.00128317e+00 -1.78101271e-01
1.69820443e-01 1.24856636e-01 -3.61983895e-01 5.07278204e-01
1.85033724e-01 -2.41455156e-02 -6.26050830e-01 -3.29135805e-01
7.89312184e-01 9.57215279e-02 -4.15508211e-01 -3.66463631e-01
-1.09653163e+00 1.01868832e+00 1.01913176e-01 1.27978534e-01
-4.24597770e-01 -3.31709564e-01 1.42111644e-01 -1.02815859e-03
5.09526134e-01 -4.93613370e-02 4.85078022e-02 4.74568903e-01
-7.40008295e-01 3.32564205e-01 4.34210658e-01 1.04873383e+00
7.62192905e-01 1.07299618e-01 -5.82004964e-01 1.16334474e+00
4.63045329e-01 7.75834143e-01 4.43318993e-01 -1.27482927e+00
6.84686422e-01 3.60251337e-01 7.88898021e-02 -1.15850067e+00
4.19507414e-01 -3.05501878e-01 -1.20246673e+00 8.31687003e-02
1.68206781e-01 -3.98151092e-02 -8.99854541e-01 2.02478909e+00
5.11002123e-01 4.56012636e-01 3.49790096e-01 7.79939532e-01
8.89625490e-01 8.41971993e-01 -1.94176257e-01 -5.28435230e-01
1.10508788e+00 -7.68391728e-01 -8.91525567e-01 6.66498467e-02
-1.25941917e-01 -8.37662995e-01 7.84049630e-01 2.34203532e-01
-1.08058858e+00 -6.86931610e-01 -9.82460976e-01 6.84172334e-03
1.09280497e-01 2.16073364e-01 2.68504629e-03 5.73035181e-01
-9.31592584e-01 3.27394783e-01 -3.33366990e-01 1.85075819e-01
6.12545788e-01 3.13439935e-01 -6.86441839e-01 -5.73498130e-01
-1.24421096e+00 3.30934852e-01 8.85701105e-02 2.78665245e-01
-1.04121625e+00 -9.94814634e-01 -8.75384867e-01 -1.50229782e-01
2.18751192e-01 -7.38260150e-01 7.42404163e-01 -7.19544888e-01
-1.39647663e+00 9.49969411e-01 -2.46833220e-01 3.26954015e-02
9.53908384e-01 3.07341784e-01 -5.12042522e-01 3.49520184e-02
2.86585510e-01 6.36521161e-01 1.02019048e+00 -1.60214245e+00
-5.32868028e-01 -6.94867671e-01 -4.11067635e-01 5.39432876e-02
-3.18488866e-01 -3.01640481e-01 -8.49421859e-01 -9.00131285e-01
1.18442900e-01 -1.13734019e+00 -5.08189909e-02 -1.94640040e-01
-2.33204693e-01 -2.46018261e-01 8.69065285e-01 -9.42446589e-01
8.48564446e-01 -2.44815040e+00 3.55112910e-01 2.87469208e-01
3.40102524e-01 1.09206803e-01 -4.70075190e-01 -7.76337683e-02
-1.19448200e-01 -4.55924496e-02 -2.92187691e-01 -4.55358714e-01
1.98848218e-01 1.22306772e-01 -3.22103202e-01 4.65003818e-01
7.33614415e-02 6.11995518e-01 -6.71741843e-01 -5.21990478e-01
-4.14997377e-02 9.80232060e-01 -5.97722471e-01 4.59630787e-01
-1.20627701e-01 4.73249882e-01 -2.78386801e-01 5.33389449e-01
1.25983381e+00 -1.74410120e-01 3.63379866e-01 -5.39609492e-01
2.89388448e-01 -5.74167907e-01 -1.39052534e+00 1.46788669e+00
-2.34440431e-01 4.04368222e-01 4.87921797e-02 -9.37146068e-01
1.08043909e+00 5.16666055e-01 6.32786810e-01 -9.31500554e-01
4.31739576e-02 2.65297383e-01 -2.74792761e-01 -3.28362882e-01
2.24637851e-01 -3.35496455e-01 1.03533827e-01 1.77223071e-01
1.00374952e-01 1.98894545e-01 4.53633517e-02 5.50273098e-02
6.34453475e-01 5.57524003e-02 -1.30218700e-01 -9.44274291e-02
5.28213382e-01 -4.47883844e-01 1.17667532e+00 5.12744486e-01
-2.97392845e-01 1.10493982e+00 4.58540708e-01 -2.91665584e-01
-1.16348553e+00 -1.43301654e+00 -1.64695323e-01 5.58705509e-01
3.08888018e-01 -2.77694017e-02 -8.37179601e-01 -6.55735672e-01
7.44829625e-02 3.41256082e-01 -8.78570080e-01 -1.66637450e-01
-4.62299168e-01 -9.77788687e-01 4.26902533e-01 5.70572734e-01
9.50002611e-01 -7.82514453e-01 4.09373015e-01 -1.13887146e-01
-7.31497347e-01 -1.14988041e+00 -5.50903618e-01 -5.85740566e-01
-3.57929498e-01 -8.55854452e-01 -1.28944468e+00 -1.14114213e+00
8.88319552e-01 3.01648736e-01 8.52895319e-01 -2.68106580e-01
-3.92137319e-01 1.13096550e-01 -1.86170980e-01 1.21287264e-01
-3.31886321e-01 -4.18935269e-01 1.33980736e-01 5.07890224e-01
3.25809002e-01 -4.07125264e-01 -6.29221499e-01 5.67582071e-01
-8.64514709e-01 -8.50035399e-02 4.79435921e-01 1.31017709e+00
9.79232788e-01 -9.56256222e-03 5.59595168e-01 -8.06017458e-01
3.93880725e-01 -5.13223588e-01 -5.01293361e-01 5.46682179e-01
-3.48140955e-01 -3.25716883e-02 5.33838809e-01 -5.85132599e-01
-1.31933212e+00 -7.18505979e-02 -1.00454181e-01 -1.02982485e+00
2.15321425e-02 -6.18489413e-03 -7.50366807e-01 -5.10248244e-02
5.04804850e-01 3.88392687e-01 9.70767960e-02 -2.88111567e-01
2.69943655e-01 7.97048450e-01 5.31486869e-01 -6.45451069e-01
9.01141942e-01 6.40446961e-01 -1.49256334e-01 -7.11472154e-01
-5.35127163e-01 -1.09536953e-01 -2.16574624e-01 -1.42494021e-02
7.92333186e-01 -1.27728367e+00 -7.81540394e-01 5.66291928e-01
-9.61805046e-01 -8.09347555e-02 -7.17296824e-02 3.95401508e-01
-4.91964638e-01 3.20025355e-01 -4.21577841e-01 -5.60481012e-01
-1.86327010e-01 -1.42289078e+00 1.16555262e+00 1.27369419e-01
3.13716829e-01 -5.10135829e-01 -5.07761650e-02 7.21003771e-01
1.23204276e-01 2.94149101e-01 7.56057739e-01 -2.32531935e-01
-5.80704331e-01 -4.21573333e-02 -3.30006361e-01 6.37303293e-01
1.91253111e-01 -1.53480664e-01 -1.02343166e+00 -4.46425974e-01
5.13386389e-04 -1.49547338e-01 7.12529242e-01 3.01312447e-01
1.26161218e+00 -2.98709244e-01 -2.56715745e-01 6.70645416e-01
1.38846695e+00 4.42407966e-01 8.68550837e-01 -7.12484345e-02
8.28166783e-01 8.34851444e-01 5.42112052e-01 5.70426583e-01
2.55569011e-01 8.07700455e-01 1.13195360e-01 8.49368051e-02
-3.07430595e-01 -4.07942891e-01 3.04338127e-01 5.89251876e-01
-3.35016064e-02 -2.33485430e-01 -6.31410360e-01 5.47192454e-01
-1.77103555e+00 -1.20838904e+00 3.52810681e-01 2.27436185e+00
7.68432677e-01 -5.53193688e-01 -8.08775946e-02 -1.72892958e-01
1.36288881e+00 3.59004363e-02 -6.25103712e-01 2.92047590e-01
-4.09854323e-01 -4.08445299e-02 1.12820014e-01 5.73694348e-01
-7.42064834e-01 6.95047498e-01 5.85477161e+00 1.10929894e+00
-9.56232131e-01 2.74242133e-01 9.16312933e-01 -2.07396805e-01
-5.82014382e-01 -4.16844964e-01 -9.89078581e-01 6.91809058e-01
3.22567493e-01 -2.18392283e-01 5.62690794e-01 8.16542804e-01
-1.64441973e-01 5.48720717e-01 -8.40630591e-01 1.61785662e+00
5.36355317e-01 -1.33818221e+00 3.75387698e-01 2.54507810e-01
1.14455175e+00 -3.73142570e-01 6.36601090e-01 3.26796860e-01
3.36045831e-01 -9.27249253e-01 7.07250118e-01 3.18847179e-01
9.11657691e-01 -9.87049103e-01 7.52315462e-01 2.86292043e-02
-1.10851848e+00 -8.96820612e-03 -5.49017072e-01 6.24600708e-01
-3.74348164e-02 5.37513554e-01 -4.11826938e-01 6.32002652e-01
8.43335748e-01 5.88440061e-01 -4.66858476e-01 4.22153592e-01
6.48845732e-02 1.36560202e-01 -8.83442014e-02 5.26204407e-01
-3.19973022e-01 -4.76638377e-01 4.49843407e-01 5.42644203e-01
5.25887966e-01 3.56225669e-02 6.47604018e-02 9.39039826e-01
-5.03432035e-01 9.34429094e-02 -6.49369717e-01 1.63831353e-01
6.67860091e-01 9.73270237e-01 -1.46086127e-01 -2.91642547e-01
-2.61674345e-01 1.27776420e+00 1.98598981e-01 6.08524919e-01
-9.80090618e-01 -2.21603289e-01 7.00173616e-01 -2.00108737e-01
4.09802765e-01 2.82311678e-01 -1.70435365e-02 -1.41999948e+00
4.46904808e-01 -1.16112292e+00 3.35456371e-01 -6.45320237e-01
-1.57624769e+00 8.25388372e-01 -1.79371551e-01 -1.31492150e+00
-3.52772444e-01 -3.77301097e-01 -1.79076701e-01 9.99122441e-01
-1.18825960e+00 -1.30466294e+00 -5.59782207e-01 8.55341971e-01
2.02119797e-01 -5.62525272e-01 6.87500000e-01 6.98857367e-01
-7.68495739e-01 1.07430339e+00 3.40774894e-01 2.62283266e-01
7.95472980e-01 -7.86762357e-01 1.40208825e-01 7.33306766e-01
-4.98476140e-02 6.69760168e-01 5.12975216e-01 -6.68525994e-01
-1.32014573e+00 -1.26891840e+00 5.22611499e-01 -2.53168613e-01
7.54440725e-02 -4.29747522e-01 -8.15644920e-01 7.15714931e-01
1.02609098e-01 4.86323565e-01 9.01591837e-01 -1.76691979e-01
-8.05225790e-01 -3.87677908e-01 -1.51555097e+00 7.27311909e-01
1.23885691e+00 -5.91334820e-01 -3.21814746e-01 3.65408547e-02
7.08689988e-01 -1.79157957e-01 -1.10337925e+00 6.66561127e-01
7.75884390e-01 -8.83567691e-01 1.07213163e+00 -3.16944510e-01
3.74331385e-01 -5.31348646e-01 -7.20776796e-01 -1.23412490e+00
-4.13548917e-01 -4.08496022e-01 1.19814321e-01 1.59039772e+00
1.85575455e-01 -6.78587019e-01 8.45847964e-01 4.26983505e-01
1.10431604e-01 -4.31419879e-01 -1.02872312e+00 -8.94411683e-01
8.33826363e-02 3.81912403e-02 7.26399958e-01 9.29232955e-01
-6.16100729e-01 1.28233671e-01 -5.07892847e-01 2.99567312e-01
1.09029198e+00 1.76626489e-01 6.93619609e-01 -9.44978476e-01
-3.63629520e-01 -4.94988561e-02 -4.47468966e-01 -7.41005898e-01
5.76298296e-01 -6.67273223e-01 1.74917951e-01 -1.12065744e+00
5.69230914e-01 -3.21175218e-01 -2.51246899e-01 2.51658171e-01
-1.92411333e-01 7.30971813e-01 2.46712312e-01 1.68935761e-01
-4.21997130e-01 1.03482044e+00 1.39518797e+00 -3.67877811e-01
7.34810252e-03 -3.80052418e-01 -7.29584694e-01 3.47350031e-01
4.45751250e-01 -1.99437380e-01 -4.77543354e-01 -5.46896398e-01
-2.13403448e-01 1.75281063e-01 2.32821509e-01 -9.00557578e-01
1.45475203e-02 -2.63589829e-01 6.00320041e-01 -5.25720119e-01
4.34196383e-01 -6.58513844e-01 5.93175113e-01 1.72321439e-01
-2.08076268e-01 -7.79350773e-02 1.87695008e-02 6.48022115e-01
-5.15699148e-01 2.49821976e-01 1.16395521e+00 3.47832702e-02
-6.28969789e-01 6.65166259e-01 1.72660977e-01 1.27530292e-01
1.28755546e+00 -6.14995472e-02 -3.42151970e-01 -1.56907409e-01
-6.73950255e-01 8.30338597e-02 4.84040976e-01 6.57353044e-01
8.49418819e-01 -1.86494589e+00 -9.92299795e-01 5.79559803e-01
2.91779190e-01 2.29664892e-02 9.67487693e-01 4.09062505e-01
-2.81046242e-01 3.10030617e-02 -3.63732517e-01 -5.68168163e-01
-1.39656603e+00 5.73159635e-01 3.06935996e-01 -2.41983086e-02
-5.14293849e-01 9.07090425e-01 6.24161899e-01 -4.51685429e-01
1.87118605e-01 5.44373751e-01 9.18687880e-03 1.89069957e-01
7.63774633e-01 4.04409409e-01 -5.79684377e-02 -9.68465447e-01
-2.88964838e-01 6.36963069e-01 -1.76884234e-01 -3.42869997e-01
1.26970041e+00 -1.48660421e-01 -1.73378989e-01 -2.34294217e-02
1.58208787e+00 2.99352519e-02 -1.49915612e+00 -1.37924165e-01
-6.78413093e-01 -8.36593986e-01 -1.30255550e-01 -4.33878958e-01
-1.44948196e+00 7.89528847e-01 9.69488502e-01 -2.78894603e-01
1.10688579e+00 -1.04893371e-01 7.30876923e-01 -6.67569861e-02
3.75953764e-01 -1.02147496e+00 3.33381563e-01 5.91633208e-02
1.05010939e+00 -1.32265091e+00 -3.13634515e-01 -5.35549402e-01
-7.80642390e-01 5.65953314e-01 8.12983751e-01 -6.47091120e-02
5.96713781e-01 -7.09298700e-02 1.39314756e-01 2.59289861e-01
-5.21520555e-01 4.41957414e-02 1.31894752e-01 9.86067891e-01
9.61644948e-02 3.25788930e-02 -1.81007206e-01 6.46297812e-01
5.94910532e-02 -6.97592273e-02 1.48193315e-01 4.62143153e-01
1.66700929e-01 -1.25995016e+00 -5.59383273e-01 1.27800271e-01
-4.37885195e-01 2.01636851e-01 -7.47355446e-02 5.51408708e-01
4.32778418e-01 9.02433276e-01 9.28827971e-02 -6.31488919e-01
3.21929991e-01 -1.24550179e-01 7.12853074e-01 -4.19403136e-01
-1.96288358e-02 1.28636539e-01 -4.43801433e-01 -3.90261054e-01
-9.54638347e-02 -7.05240726e-01 -8.82847309e-01 -4.11996931e-01
-2.54384987e-02 2.91299820e-01 3.79837334e-01 7.27333724e-01
6.82921946e-01 3.05289418e-01 9.90815103e-01 -6.97244585e-01
-6.15750849e-01 -8.30319822e-01 -8.41753542e-01 9.46412921e-01
2.12737247e-01 -8.46169293e-01 -4.44141507e-01 2.03861564e-01] | [13.037351608276367, 0.25512441992759705] |
8557b5f1-306a-4597-986a-67e16c995318 | blind-acoustic-room-parameter-estimation | 2303.07449 | null | https://arxiv.org/abs/2303.07449v1 | https://arxiv.org/pdf/2303.07449v1.pdf | Blind Acoustic Room Parameter Estimation Using Phase Features | Modeling room acoustics in a field setting involves some degree of blind parameter estimation from noisy and reverberant audio. Modern approaches leverage convolutional neural networks (CNNs) in tandem with time-frequency representation. Using short-time Fourier transforms to develop these spectrogram-like features has shown promising results, but this method implicitly discards a significant amount of audio information in the phase domain. Inspired by recent works in speech enhancement, we propose utilizing novel phase-related features to extend recent approaches to blindly estimate the so-called "reverberation fingerprint" parameters, namely, volume and RT60. The addition of these features is shown to outperform existing methods that rely solely on magnitude-based spectral features across a wide range of acoustics spaces. We evaluate the effectiveness of the deployment of these novel features in both single-parameter and multi-parameter estimation strategies, using a novel dataset that consists of publicly available room impulse responses (RIRs), synthesized RIRs, and in-house measurements of real acoustic spaces. | ['Wenyu Jin', 'Adib Mehrabi', 'Christopher Ick'] | 2023-03-13 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 2.01898023e-01 -6.01586640e-01 7.57706821e-01 -4.40225393e-01
-1.40526867e+00 -6.90986276e-01 5.11638224e-01 1.36065319e-01
-4.44059879e-01 4.76001263e-01 8.42065215e-01 -1.91417947e-01
-4.38052446e-01 -4.76487130e-01 -5.85553586e-01 -8.33782375e-01
-4.98214006e-01 -5.01230657e-01 -3.45770985e-01 -4.51922148e-01
1.47835031e-01 3.56848150e-01 -1.67193663e+00 4.54991236e-02
5.37723362e-01 1.25184464e+00 -1.16598336e-02 1.02586412e+00
4.37253863e-01 5.68322599e-01 -9.01778460e-01 1.98502585e-01
3.39689314e-01 -1.65540222e-02 -3.16683739e-01 -4.15635705e-01
6.82133019e-01 -3.72408837e-01 -6.01884961e-01 6.39252961e-01
1.19062841e+00 6.43571258e-01 4.24685150e-01 -6.75320685e-01
-5.04370272e-01 5.39305389e-01 4.31706086e-02 2.58023739e-01
5.24970651e-01 1.67426825e-01 1.03603518e+00 -8.69931400e-01
-2.08419204e-01 1.00733483e+00 1.35154402e+00 7.88419619e-02
-1.18309879e+00 -7.18959153e-01 -2.16851354e-01 1.66240428e-02
-1.21898663e+00 -9.86674368e-01 9.81077373e-01 -3.47258478e-01
1.21303105e+00 5.16696751e-01 2.88922787e-01 1.27433443e+00
-2.65214831e-01 3.57583582e-01 1.23761070e+00 -5.50037265e-01
2.53032655e-01 -1.86871037e-01 8.39689597e-02 4.17793304e-01
-1.87267512e-01 5.74165881e-01 -7.48172224e-01 -4.21959490e-01
5.07415771e-01 -3.49503368e-01 -8.88284683e-01 -1.00781567e-01
-1.36657965e+00 5.96277356e-01 5.36839843e-01 3.80712003e-01
-1.45413250e-01 4.38467860e-01 3.20827752e-01 2.21451506e-01
6.14790916e-01 8.46187711e-01 -5.02212703e-01 -1.64573148e-01
-1.12348270e+00 3.37259650e-01 9.57613170e-01 4.75698560e-01
6.78781748e-01 6.13417685e-01 -3.73370945e-01 1.24613845e+00
4.53242004e-01 6.82437301e-01 4.68264461e-01 -9.92594004e-01
4.22632128e-01 -3.89800280e-01 4.10143286e-01 -8.77923012e-01
-7.42446005e-01 -7.35500455e-01 -8.34288657e-01 -3.04459091e-02
3.34591478e-01 -8.13211203e-02 -1.11205208e+00 1.94246972e+00
1.87386215e-01 3.17483813e-01 -1.23752818e-01 1.12212992e+00
9.68911231e-01 5.10789633e-01 -2.42716014e-01 1.66251302e-01
1.17019284e+00 -9.26067829e-01 -8.09300721e-01 -2.41701454e-01
1.02620512e-01 -9.42675650e-01 9.66739058e-01 6.40673816e-01
-8.09860170e-01 -8.51011336e-01 -1.39339745e+00 4.89439070e-02
-4.65396523e-01 1.48780107e-01 4.51179802e-01 1.20645559e+00
-1.31217217e+00 8.32230449e-01 -5.76310337e-01 2.52937526e-01
1.21387392e-01 3.59977931e-01 -3.19831222e-01 -8.24663490e-02
-1.31076944e+00 4.39601541e-01 -3.43229860e-01 4.89911556e-01
-1.20038557e+00 -1.17806959e+00 -1.06429827e+00 2.13067368e-01
2.91594304e-02 -5.53455770e-01 1.36280382e+00 -4.45215672e-01
-1.54491091e+00 2.61446595e-01 1.06830401e-02 -4.57422584e-01
4.02449012e-01 -6.51676893e-01 -7.83353567e-01 1.51332796e-01
-2.28582144e-01 -3.48503329e-02 1.35391653e+00 -1.35575533e+00
-4.14613299e-02 -1.13815390e-01 1.09931314e-02 -5.77793680e-02
-5.81552148e-01 -9.34275910e-02 6.70040399e-02 -9.73589957e-01
2.68024415e-01 -6.85117126e-01 -2.21394792e-01 -3.29175174e-01
-5.21228492e-01 3.42984021e-01 3.63925844e-01 -1.20115507e+00
1.11555040e+00 -2.28790712e+00 -3.42363745e-01 2.60531664e-01
3.96336108e-01 2.57186323e-01 -2.12382242e-01 4.77001369e-01
-3.62304986e-01 -1.24192007e-01 -4.34370697e-01 -8.40900898e-01
3.56202513e-01 -4.84953225e-01 -4.32996124e-01 7.46864438e-01
1.29776269e-01 3.60481590e-01 -6.74472988e-01 6.70030862e-02
5.58827817e-01 1.05283558e+00 -6.94862545e-01 3.51599157e-01
4.35300350e-01 7.19840884e-01 -6.82000667e-02 4.65561658e-01
9.05114532e-01 1.45814478e-01 -2.28201404e-01 -5.23439407e-01
-2.32531890e-01 8.51190805e-01 -1.15795004e+00 1.76573944e+00
-1.09670019e+00 9.37294722e-01 5.57574272e-01 -7.90023744e-01
8.69703948e-01 6.27571642e-01 4.50853735e-01 -6.67552531e-01
5.08749969e-02 3.95582438e-01 -1.79910541e-01 -4.67591822e-01
6.38072789e-01 -2.50647357e-03 1.37149721e-01 4.25462425e-01
3.57945889e-01 -3.50440949e-01 -3.18192273e-01 -2.11109698e-01
1.39383531e+00 3.70833203e-02 -5.78748016e-03 -1.94935456e-01
6.28977418e-01 -7.78443575e-01 9.16396976e-02 8.94634843e-01
-3.17668498e-01 1.11970079e+00 -4.73115265e-01 -1.11144237e-01
-9.80000854e-01 -1.07674468e+00 -1.33793384e-01 1.29754102e+00
-4.97908294e-01 -3.86923850e-01 -7.19728649e-01 -2.39806715e-02
-2.16878741e-03 5.74704587e-01 -7.04831541e-01 -1.92723379e-01
-9.67881262e-01 -7.26255596e-01 8.32938433e-01 4.59973663e-01
2.67089099e-01 -7.50443995e-01 -3.19569588e-01 3.58426094e-01
-4.12480652e-01 -1.19873953e+00 -4.69118118e-01 7.19619989e-01
-3.65821362e-01 -5.87845683e-01 -9.28506553e-01 -5.19543588e-01
6.27715513e-02 4.41044509e-01 1.06042111e+00 -1.59055352e-01
-4.67337370e-01 8.53387773e-01 -4.27698046e-01 -4.19968724e-01
-3.57275307e-01 -2.03254774e-01 3.50179225e-01 1.44880459e-01
-2.45871469e-01 -9.72126126e-01 -9.32125628e-01 1.66649193e-01
-6.08746469e-01 -5.89784503e-01 3.33424121e-01 6.71709597e-01
8.18977728e-02 -1.89874813e-01 7.38664210e-01 -4.85489853e-02
6.36741281e-01 -2.66390234e-01 -3.55152667e-01 -1.40386909e-01
-3.56360614e-01 8.63268971e-03 9.00453508e-01 -4.77629513e-01
-1.05328751e+00 -3.41471076e-01 -5.24968863e-01 -3.95896047e-01
-2.92570919e-01 2.23110825e-01 -9.00096744e-02 -2.55953759e-01
8.07908118e-01 1.73360035e-01 -4.54198480e-01 -7.94034719e-01
9.73989740e-02 7.74562776e-01 7.32445896e-01 -4.79994684e-01
1.11615300e+00 4.88926858e-01 -6.73782155e-02 -9.60193932e-01
-7.84070134e-01 -9.18527424e-01 -1.67713746e-01 4.22271155e-03
5.51155150e-01 -1.12027860e+00 -7.40514755e-01 5.31462789e-01
-9.40268338e-01 -3.74100387e-01 -2.56733626e-01 9.15569901e-01
-6.22213066e-01 3.94603163e-01 -5.05496860e-01 -1.16383040e+00
-4.73032892e-01 -1.01751447e+00 1.26210129e+00 -7.77271017e-02
-9.78699625e-02 -8.64556134e-01 2.74846822e-01 4.75139886e-01
1.02724230e+00 -3.19690965e-02 5.44365406e-01 -6.17717028e-01
-4.31051254e-01 -2.83992827e-01 4.96746860e-02 6.95627332e-01
2.12618232e-01 -4.21793342e-01 -2.02638793e+00 -4.28929597e-01
4.67730939e-01 -2.31634080e-01 1.17837417e+00 7.86905885e-01
1.22016799e+00 -1.26889795e-01 1.90346614e-01 1.09772515e+00
1.25232053e+00 -1.09027103e-01 5.63659549e-01 2.56972551e-01
7.58262575e-01 5.27076364e-01 2.24091768e-01 6.60018861e-01
1.62715912e-01 4.50650185e-01 4.68970478e-01 -2.09411830e-01
-4.16619599e-01 -4.22528051e-02 2.95864373e-01 1.14628792e+00
-2.09849387e-01 -2.66251355e-01 -7.07356989e-01 5.58740437e-01
-1.04885161e+00 -7.45612383e-01 2.91772038e-01 2.31212807e+00
8.54379177e-01 -2.87340462e-01 -3.10209244e-01 7.14811921e-01
4.27741259e-01 5.93397558e-01 -6.05797917e-02 -1.51765153e-01
-1.03030637e-01 7.24889934e-01 5.90666115e-01 4.97990012e-01
-1.47097707e+00 2.36409962e-01 6.70021439e+00 5.41267157e-01
-1.17767692e+00 9.06023756e-02 3.26609284e-01 3.51254307e-02
-3.86816025e-01 -3.48426163e-01 -4.48478132e-01 -2.38448307e-02
1.33117938e+00 4.59766597e-01 9.17926490e-01 4.63484615e-01
3.13786685e-01 1.85253143e-01 -1.07397676e+00 1.03499281e+00
1.55944839e-01 -1.09787595e+00 -6.57490671e-01 -1.72956407e-01
4.64470774e-01 1.20646037e-01 6.95978284e-01 3.24434131e-01
3.40250731e-02 -1.42353725e+00 9.21002805e-01 5.77081084e-01
7.79159904e-01 -5.01624942e-01 4.54286307e-01 -1.26287147e-01
-1.41709256e+00 -4.22574967e-01 -4.85405698e-02 1.71453670e-01
-5.67809157e-02 6.70834899e-01 -1.18129337e+00 6.48713410e-01
1.15347159e+00 3.84852529e-01 -4.79331017e-01 1.34538758e+00
9.25687924e-02 9.84640419e-01 -4.45599228e-01 3.47478032e-01
8.77601281e-02 1.65421799e-01 7.12886810e-01 1.40042007e+00
4.55109358e-01 -3.25626045e-01 -3.73937696e-01 8.30052793e-01
-9.19062570e-02 -1.53187945e-01 -2.39461184e-01 1.86466984e-02
4.91300255e-01 1.34848428e+00 -2.61405617e-01 1.00926131e-01
-1.49658561e-01 4.36956793e-01 -3.79165143e-01 7.75009811e-01
-7.62268424e-01 -5.64385116e-01 7.48644888e-01 -9.55486000e-02
6.02678180e-01 -6.02757871e-01 -1.81547739e-02 -7.02078819e-01
9.04648677e-02 -1.02785039e+00 -2.10535601e-01 -7.27528036e-01
-1.28171062e+00 6.07847393e-01 -2.29243845e-01 -1.41406131e+00
-3.05491984e-01 -6.05028450e-01 -5.73514044e-01 1.17773938e+00
-1.97945249e+00 -1.13286817e+00 -5.34223437e-01 6.63682878e-01
4.76929218e-01 -1.89046618e-02 1.16158640e+00 6.97705090e-01
-2.34474868e-01 7.56524742e-01 3.76789898e-01 2.54009604e-01
1.07293332e+00 -1.30787265e+00 7.26431847e-01 6.60249233e-01
1.14474192e-01 8.63086224e-01 1.07503414e+00 -7.79669126e-03
-1.44652367e+00 -1.02580845e+00 4.09838110e-01 -4.01893198e-01
5.32479405e-01 -7.02710688e-01 -7.86958933e-01 3.84190619e-01
8.04962218e-03 3.68297249e-01 8.63206387e-01 1.45490438e-01
-7.73197949e-01 -2.00296730e-01 -1.00210738e+00 2.31852129e-01
8.18352163e-01 -9.25696671e-01 -4.30907488e-01 2.06218690e-01
9.96738255e-01 -3.21639985e-01 -9.09364104e-01 4.14697677e-01
6.20005965e-01 -1.09446466e+00 1.73780036e+00 -1.95178967e-02
9.47483554e-02 -2.04189733e-01 -7.22326458e-01 -1.55635214e+00
-2.00513139e-01 -1.04698813e+00 -2.41917163e-01 1.18846631e+00
2.09365949e-01 -6.90845788e-01 1.72348320e-01 5.22137508e-02
-7.49454021e-01 -2.46847034e-01 -1.08198619e+00 -7.10797489e-01
-1.65497124e-01 -9.12890017e-01 9.50568974e-01 7.59823561e-01
-6.44807696e-01 1.33973351e-02 -4.98789489e-01 5.46610713e-01
7.57545114e-01 -2.35414967e-01 6.36839509e-01 -1.16954398e+00
-4.90840256e-01 -1.63931087e-01 -1.54295802e-01 -7.95511305e-01
7.70036206e-02 -4.69781876e-01 6.55225933e-01 -1.31878138e+00
-6.42086744e-01 -6.86198771e-01 -7.18072653e-01 1.09645769e-01
-3.83467562e-02 3.56779933e-01 -6.77062795e-02 -2.08267957e-01
-1.94229409e-01 7.01242149e-01 9.56484318e-01 -2.11725488e-01
-1.34395555e-01 2.71387637e-01 -4.00025606e-01 7.36696720e-01
6.54647470e-01 -2.59258538e-01 -1.65239736e-01 -4.08142447e-01
1.39524773e-01 6.24465011e-02 7.82376766e-01 -1.59289384e+00
1.58838212e-01 3.16343129e-01 6.21544480e-01 -4.44140404e-01
1.01206660e+00 -7.08265305e-01 -2.63443470e-01 -4.92751412e-02
-4.56593841e-01 -2.32602805e-01 4.46665347e-01 7.27034986e-01
-3.34268630e-01 1.56716049e-01 5.54175317e-01 6.89174756e-02
-4.77797359e-01 1.31235514e-02 -2.16669574e-01 -2.44149119e-02
-6.72656149e-02 -9.56680849e-02 -2.83087045e-01 -6.46338463e-01
-5.28008819e-01 -5.45393407e-01 -1.29220784e-01 4.14570808e-01
6.03958011e-01 -1.03671563e+00 -6.33370280e-01 4.19668436e-01
3.89371742e-03 -3.04717034e-01 7.16487110e-01 9.56382334e-01
-3.36339444e-01 5.44993699e-01 9.19870883e-02 -6.13991439e-01
-9.40989673e-01 3.13751280e-01 6.63723230e-01 -4.13797423e-02
-5.05391061e-01 1.20911205e+00 1.63044736e-01 -7.95869172e-01
4.09634441e-01 -6.22383833e-01 -3.20887864e-01 2.46473644e-02
5.65757513e-01 4.13485259e-01 6.90081477e-01 -7.33774185e-01
-4.13481116e-01 4.61879611e-01 4.76539403e-01 -4.91490662e-01
1.59316707e+00 -3.74649674e-01 1.48266315e-01 5.72921097e-01
1.63846171e+00 6.28367126e-01 -1.26322377e+00 -3.27399641e-01
-2.96494633e-01 -5.42251050e-01 4.23210323e-01 -9.76652861e-01
-8.28616321e-01 1.07323575e+00 1.02021122e+00 2.88689047e-01
1.32711065e+00 -4.78544652e-01 8.50296259e-01 3.83553714e-01
2.09910899e-01 -1.02265346e+00 3.97107750e-01 5.15308440e-01
9.53757823e-01 -1.24850380e+00 -2.86140218e-02 -1.25736669e-01
-6.69686645e-02 1.10928214e+00 2.80614682e-02 -3.74183506e-02
9.22375321e-01 4.18406636e-01 2.69240856e-01 1.13656022e-01
-2.92951822e-01 -1.24987185e-01 5.49760699e-01 9.65757370e-01
6.10633433e-01 1.31180985e-02 7.22738862e-01 7.61277139e-01
-8.58842075e-01 -5.84174991e-01 2.23670110e-01 7.94552565e-01
-3.76501173e-01 -6.54340625e-01 -9.66643274e-01 1.83521807e-01
-6.51165843e-01 -5.36123872e-01 3.16969275e-01 3.95714074e-01
-1.19548351e-01 1.45198870e+00 -9.39791054e-02 -5.49258351e-01
3.09513599e-01 -3.02421534e-03 4.68072504e-01 -2.61926413e-01
-1.03769588e+00 2.60901541e-01 2.12239116e-01 -6.40728593e-01
-4.95039493e-01 -3.80378842e-01 -8.25753689e-01 -5.64703494e-02
-4.14108485e-01 -7.27586523e-02 1.25863099e+00 6.92199349e-01
-2.07686536e-02 1.19521630e+00 9.38096941e-01 -1.36592567e+00
-8.73346984e-01 -1.22733939e+00 -5.56806922e-01 1.83900639e-01
1.38320374e+00 -4.92480397e-01 -8.37496936e-01 -1.77589118e-01] | [15.172245025634766, 5.730571269989014] |
3a226431-ef1a-4e7c-98f4-5eb49990e6dc | compressed-sensing-constant-modulus | 2110.03385 | null | https://arxiv.org/abs/2110.03385v1 | https://arxiv.org/pdf/2110.03385v1.pdf | Compressed Sensing Constant Modulus Constrained Projection Matrix Design and High-Resolution DoA Estimation Methods | This paper proposes a compressed sensing-based high-resolution direction-of-arrival estimation method called gradient orthogonal matching pursuit (GOMP). It contains two main steps: a sparse coding approximation step using the well-known OMP method and a sequential iterative refinement step using a newly proposed gradient-descent method. To enhance the recoverability, we further propose an efficient projection matrix design method, which considers the constant modulus constraints imposed by the projection matrix hardware components. Simulation results show the effectiveness of the proposed methods as compared to benchmark algorithms. | ['Martin Haardt', 'Khaled Ardah'] | 2021-10-07 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 5.54339945e-01 -4.12451714e-01 5.59618399e-02 -5.23212738e-02
-7.52190650e-01 1.30194113e-01 2.94048309e-01 -1.90482914e-01
-2.52492219e-01 5.33464193e-01 5.32281041e-01 -3.66553903e-01
-1.87354118e-01 -3.70942146e-01 -4.44620103e-01 -6.85525239e-01
-3.30266953e-01 -5.51311933e-02 2.64524043e-01 -9.84108821e-02
5.70087552e-01 4.41698492e-01 -9.19927239e-01 -4.20912467e-02
8.14969182e-01 1.17405403e+00 2.03139439e-01 4.12318796e-01
2.89464206e-01 8.36670637e-01 1.86334223e-01 -4.19729622e-03
8.31985712e-01 -4.31837291e-01 5.98059259e-02 3.33783746e-01
3.71666849e-02 -5.02512217e-01 -6.26230657e-01 1.05719817e+00
5.14144123e-01 3.94179702e-01 3.21611553e-01 -7.69082963e-01
-2.38141179e-01 2.89787084e-01 -1.05505371e+00 3.75021607e-01
6.66732490e-01 -4.60829765e-01 5.01760364e-01 -1.51416945e+00
3.83066893e-01 7.87985325e-01 1.13863957e+00 -1.77101627e-01
-8.22070181e-01 -7.52957642e-01 -4.54745978e-01 3.74007404e-01
-1.88736582e+00 -8.29826236e-01 1.22642279e+00 -3.40664059e-01
6.82301044e-01 2.07464561e-01 4.49351490e-01 1.52407095e-01
2.06958637e-01 6.23349488e-01 1.14466906e+00 -7.61393666e-01
3.92951608e-01 -1.94552243e-01 6.56348541e-02 7.04091907e-01
5.80795884e-01 2.99703330e-01 -5.86441696e-01 -6.99522793e-01
8.63562107e-01 3.00758421e-01 -7.19031990e-01 -5.57547808e-01
-1.27119446e+00 8.75044644e-01 1.21086843e-01 2.37330854e-01
-9.13415253e-01 2.92576216e-02 -4.54735830e-02 1.26244649e-02
2.35345840e-01 -1.38278492e-02 1.81762353e-01 2.21198618e-01
-1.72913313e+00 1.56044379e-01 8.75558615e-01 9.01292026e-01
6.98478401e-01 4.46058065e-01 7.40607455e-02 6.53601468e-01
8.28885257e-01 7.04793632e-01 5.03027439e-01 -8.45067680e-01
4.84256744e-01 -5.62656187e-02 2.59449333e-01 -1.75729311e+00
-2.69069672e-01 -6.24121487e-01 -1.04028904e+00 -1.50693730e-01
-1.45940527e-01 -3.61455709e-01 -4.42781180e-01 1.08372056e+00
5.53106010e-01 8.29182208e-01 2.13220909e-01 1.22213912e+00
4.58954006e-01 7.38875449e-01 -6.52976215e-01 -7.06602037e-01
9.20444191e-01 -8.71302009e-01 -9.50103104e-01 -1.24079794e-01
-6.43015280e-02 -1.09098887e+00 1.34452492e-01 4.18324709e-01
-1.12930393e+00 -2.56247073e-01 -1.52485561e+00 4.68446732e-01
5.76488435e-01 2.39809245e-01 6.78664148e-01 6.70212209e-01
-7.31525540e-01 3.94685328e-01 -8.16722631e-01 9.63794142e-02
1.38998449e-01 1.31292179e-01 -2.59168893e-01 -5.04887104e-01
-7.65986204e-01 3.63554239e-01 7.83425421e-02 8.62159729e-02
-4.38340247e-01 -5.54546833e-01 -8.67662668e-01 1.53184295e-01
1.98602021e-01 -4.55662280e-01 7.49555647e-01 -3.29096407e-01
-1.61298764e+00 2.89373845e-01 -5.99938273e-01 -5.79704344e-01
1.00059986e-01 -3.73981267e-01 -5.58479965e-01 6.20897949e-01
-3.32806557e-02 -2.68644750e-01 1.26549995e+00 -9.72979069e-01
-5.51258683e-01 -3.54921848e-01 -5.80881238e-01 2.05056548e-01
-1.97879016e-01 -1.49535174e-02 -4.79254842e-01 -9.58873391e-01
1.21574426e+00 -8.61604273e-01 -6.83941185e-01 -4.50703166e-02
-1.76083431e-01 7.09035814e-01 5.92637002e-01 -8.67505372e-01
1.56686997e+00 -2.29974103e+00 -5.46537638e-02 6.79899573e-01
3.31160784e-01 -3.79499868e-02 1.14158474e-01 6.50108218e-01
-7.86396116e-03 -8.18013012e-01 -5.46124279e-01 -2.11113676e-01
-4.77310389e-01 -3.19002837e-01 -3.79589230e-01 1.08045018e+00
-4.64765519e-01 1.64614692e-01 -6.09736562e-01 -2.15184331e-01
2.00096801e-01 6.91725254e-01 -6.91022635e-01 1.18503153e-01
6.65364981e-01 4.12397444e-01 -4.48773533e-01 7.38393426e-01
1.20089555e+00 -1.91957355e-01 1.21286400e-01 -4.32542175e-01
-5.45766354e-01 1.39644012e-01 -2.03351641e+00 1.64963222e+00
-1.40335634e-01 4.78132516e-01 6.37715816e-01 -1.19784415e+00
1.16949832e+00 4.95569646e-01 8.75599682e-01 -3.93550605e-01
1.30074799e-01 5.85898638e-01 -2.47472733e-01 -2.31094822e-01
6.01108432e-01 -3.58901858e-01 5.82605898e-01 4.14163023e-01
-4.65536177e-01 3.24394256e-02 -5.20442203e-02 2.78802514e-01
1.13221920e+00 -2.31708229e-01 6.00999713e-01 -5.29028535e-01
8.75796854e-01 8.34130589e-03 8.90303850e-01 5.04361093e-01
-1.69321150e-01 6.87579572e-01 -4.66804504e-01 -4.71773863e-01
-8.34688902e-01 -5.75813472e-01 -3.50393683e-01 3.61628979e-01
3.94868433e-01 -4.40678567e-01 -2.63045162e-01 -5.56652136e-02
-3.84149514e-02 3.78167480e-01 1.58417057e-02 2.89270341e-01
-8.43816400e-01 -7.69256949e-01 5.63521460e-02 1.62217751e-01
7.03543007e-01 -3.40597838e-01 -7.95838892e-01 3.72695863e-01
-1.54828548e-01 -1.17822134e+00 -4.97915536e-01 -1.67419538e-01
-9.90407050e-01 -7.83594370e-01 -1.03191042e+00 -7.09974349e-01
6.78078413e-01 1.06638408e+00 5.57667196e-01 -8.46589077e-03
9.68601257e-02 4.11949933e-01 -4.93040681e-01 1.46344528e-01
2.27548137e-01 -6.37034118e-01 3.00970316e-01 6.56247735e-01
1.65013582e-01 -1.07121205e+00 -7.69013584e-01 1.68503687e-01
-5.35186708e-01 6.85489327e-02 9.74128187e-01 7.15970993e-01
9.75298285e-01 1.43978760e-01 4.93619367e-02 -6.47333384e-01
6.38328969e-01 -5.80669343e-01 -7.90863276e-01 -8.26032460e-02
-7.98148453e-01 -2.38453485e-02 4.29011911e-01 -3.22675139e-01
-8.81022334e-01 3.75767142e-01 -5.08066900e-02 -5.48519552e-01
3.40467572e-01 9.87245321e-01 1.82222486e-01 -8.82189274e-01
4.86387044e-01 9.04747009e-01 -1.84123024e-01 -5.73055565e-01
9.64791253e-02 6.61809683e-01 7.11273611e-01 -3.94726723e-01
1.24464524e+00 7.44666576e-01 3.61845523e-01 -1.28799713e+00
-3.26841533e-01 -1.04360485e+00 -2.27075756e-01 1.27221808e-01
1.73117802e-01 -1.07669771e+00 -1.82135746e-01 1.48197100e-01
-1.02000713e+00 2.59092689e-01 3.81894857e-02 1.23367000e+00
-3.43856007e-01 9.52585876e-01 -6.43408120e-01 -7.76497602e-01
-6.52843595e-01 -8.46708059e-01 6.53579473e-01 -9.34852883e-02
2.56687403e-01 -6.08421326e-01 5.04457712e-01 3.76064144e-02
7.66085744e-01 1.41839519e-01 1.15955740e-01 -2.18114287e-01
-6.53656721e-01 -4.17072713e-01 -2.60705054e-01 -1.94997471e-02
-1.80168837e-01 -7.19266236e-01 -1.80058762e-01 -6.04865491e-01
6.79121733e-01 2.48865485e-01 5.30187905e-01 6.76102519e-01
5.83908737e-01 -5.24050355e-01 -4.68593657e-01 1.35528815e+00
1.85889435e+00 5.55776395e-02 7.99789190e-01 3.79275113e-01
5.40610194e-01 4.70403098e-02 6.48533285e-01 1.04744518e+00
1.88223831e-02 7.17652619e-01 7.13490788e-03 1.46436185e-01
-7.43667735e-03 -8.75593349e-02 4.22957949e-02 1.37615216e+00
-5.69391996e-02 2.72983402e-01 -7.39224732e-01 4.86747354e-01
-1.74204314e+00 -1.01220655e+00 -4.88275588e-01 2.39782143e+00
4.27702785e-01 -2.11728409e-01 -3.35110933e-01 4.67350006e-01
4.75430608e-01 2.91729927e-01 -3.31319839e-01 -4.16849405e-02
3.07067893e-02 3.63870889e-01 9.19294357e-01 6.52459443e-01
-8.64435911e-01 3.01601559e-01 6.87227488e+00 7.20359087e-01
-1.07525456e+00 1.87152579e-01 -1.84311271e-01 4.57445502e-01
-4.12719578e-01 2.85545498e-01 -8.00851166e-01 3.27409685e-01
6.19880736e-01 -2.74820060e-01 5.09753108e-01 8.15870047e-01
3.02788407e-01 2.64869165e-02 -2.71435499e-01 1.45149338e+00
4.06235307e-01 -1.56322908e+00 -1.72063917e-01 -9.78400856e-02
8.06664884e-01 2.74003744e-02 -1.75287932e-01 -1.28339171e-01
-3.80487055e-01 -3.95408481e-01 5.84994495e-01 4.08242673e-01
7.10208774e-01 -5.05061209e-01 4.52450305e-01 4.08537239e-01
-1.57824647e+00 -9.23494436e-03 -5.53711176e-01 -3.24060857e-01
6.72681153e-01 1.17733228e+00 -6.04952931e-01 4.44001734e-01
2.86135912e-01 6.12894893e-01 1.48565307e-01 1.56918788e+00
-2.35792086e-01 9.29557025e-01 -6.61946595e-01 4.08364683e-01
1.03043929e-01 -5.51940560e-01 1.18574667e+00 1.21537399e+00
8.40352774e-01 7.21839488e-01 4.47994441e-01 3.91729623e-01
3.44413310e-01 4.06983227e-01 -5.19139528e-01 3.00948262e-01
6.66009068e-01 1.12654126e+00 -4.91687775e-01 -1.77377358e-01
-7.63508856e-01 9.50725079e-01 -4.34886396e-01 4.41508830e-01
-5.02179801e-01 -5.62715173e-01 1.19604975e-01 5.56410015e-01
7.44523227e-01 -6.78174317e-01 -3.68225098e-01 -1.41983795e+00
7.76167214e-02 -8.99056613e-01 2.28242055e-01 -4.54102248e-01
-7.96200931e-01 5.28089464e-01 -1.30032763e-01 -1.62243855e+00
-2.34666333e-01 -5.08403517e-02 -4.93302733e-01 9.15595889e-01
-1.72592700e+00 -8.18991959e-01 -2.68003196e-01 9.51579034e-01
4.67226028e-01 -4.10966724e-01 5.34394562e-01 5.85159361e-01
-2.51406282e-01 4.30515349e-01 4.70428586e-01 -2.25523859e-01
1.63560972e-01 -3.88281077e-01 -2.52216253e-02 1.61199558e+00
-3.32097933e-02 7.71196604e-01 9.16284144e-01 -7.86574900e-01
-1.83013928e+00 -6.23999774e-01 8.85664463e-01 7.04648674e-01
5.28407812e-01 6.97092488e-02 -8.48284125e-01 6.73591673e-01
-1.49777858e-03 2.21813306e-01 7.77686954e-01 -3.20180058e-01
-1.68925583e-01 -2.70931460e-02 -1.24741757e+00 1.53450489e-01
5.63545346e-01 -2.94648767e-01 -7.01504469e-01 2.62134850e-01
4.07999665e-01 -6.80238485e-01 -5.33436298e-01 6.52389824e-01
6.46298170e-01 -9.32551801e-01 1.19283020e+00 2.31139794e-01
1.43194452e-01 -6.25669420e-01 -8.82695258e-01 -7.56309748e-01
-9.22581911e-01 -1.07418287e+00 -6.33360386e-01 4.83156592e-01
4.76737581e-02 -6.55680776e-01 9.40216064e-01 4.14115071e-01
-1.60269529e-01 -6.96778476e-01 -8.01058531e-01 -6.66969836e-01
-8.84160459e-01 -4.67537642e-01 2.72264004e-01 1.11423934e+00
1.36825934e-01 4.64135140e-01 -7.67191172e-01 6.91998959e-01
1.31517816e+00 3.11109126e-01 5.98909378e-01 -9.15504158e-01
-8.51519287e-01 1.69742048e-01 -4.38857824e-01 -1.57963252e+00
-4.87456173e-01 -6.17009103e-01 1.00992471e-01 -1.29883671e+00
1.14499658e-01 -6.48802876e-01 -3.27863485e-01 -2.58293033e-01
-6.49068207e-02 2.83829004e-01 9.35532078e-02 7.11297095e-01
-4.42799032e-01 5.26089787e-01 7.66534746e-01 1.62895128e-01
-4.60614324e-01 2.41906062e-01 -4.62849438e-01 7.46944904e-01
4.45535243e-01 -6.26582801e-01 -3.69430989e-01 -4.86498415e-01
6.24570921e-02 6.97221279e-01 -1.36442944e-01 -1.66200089e+00
6.30138159e-01 -1.15086168e-01 7.17418343e-02 -8.16080868e-01
4.24386948e-01 -1.01731873e+00 4.69496280e-01 7.49334574e-01
5.43573685e-02 9.10476223e-02 5.50881308e-03 7.69052148e-01
-3.78127098e-01 -3.75590473e-01 7.78723240e-01 1.90293670e-01
-7.68401086e-01 2.81827569e-01 -3.88921320e-01 -3.28374863e-01
7.69925416e-01 -3.73689860e-01 3.17978442e-01 -6.26196682e-01
-3.35789859e-01 -2.76288301e-01 2.67984241e-01 -4.81658936e-01
1.28533506e+00 -1.41118062e+00 -1.17947531e+00 5.70254147e-01
-1.85615361e-01 -5.57335973e-01 2.46740073e-01 1.27524257e+00
-8.24804783e-01 4.67733890e-01 -9.97670218e-02 -4.85874504e-01
-1.03615522e+00 5.40284455e-01 -1.81154072e-01 -3.37386250e-01
-1.04677796e+00 7.88198888e-01 -1.74613595e-01 -9.73291248e-02
-2.40239166e-02 3.50899488e-01 -1.58549011e-01 -6.82713926e-01
1.07416666e+00 6.61498725e-01 -6.17952365e-03 -8.91938686e-01
-3.81703168e-01 8.46774280e-01 3.52265686e-01 -4.15885329e-01
1.41567969e+00 -4.01738912e-01 -2.20867798e-01 -1.77822992e-01
1.19753373e+00 7.55201817e-01 -9.85341191e-01 -5.50281942e-01
-2.67262727e-01 -1.02882850e+00 6.70790613e-01 -8.39082226e-02
-1.08863878e+00 4.87944603e-01 6.57201350e-01 -3.27808201e-01
1.33812714e+00 -6.79252326e-01 1.28841221e+00 3.93127650e-01
7.74614036e-01 -6.14761353e-01 -2.96526164e-01 3.98411959e-01
8.20702791e-01 -9.50704038e-01 6.98138833e-01 -6.85387433e-01
-1.23622954e-01 1.02462173e+00 -1.87579513e-01 -5.83936274e-01
9.99449253e-01 2.90059268e-01 -2.34040335e-01 -8.81611183e-02
-2.65320271e-01 1.18603371e-01 4.65730458e-01 5.05903006e-01
3.02787811e-01 -6.57961965e-02 -1.03516364e+00 4.61370885e-01
1.89946115e-01 2.62860894e-01 4.99154508e-01 1.02977276e+00
-8.56899142e-01 -8.65985811e-01 -9.82013166e-01 3.46056610e-01
-4.92228270e-01 -4.32633609e-01 4.15527195e-01 2.54273444e-01
-3.67514938e-01 9.27659571e-01 -3.83274496e-01 -5.22130847e-01
6.88815117e-02 -4.85299706e-01 3.45294893e-01 -3.29699486e-01
7.37874657e-02 4.90272969e-01 -1.02173597e-01 -7.92983294e-01
-4.34573144e-01 -7.63684571e-01 -1.13861275e+00 -3.61849397e-01
-4.27301466e-01 5.63251078e-01 7.98516870e-01 7.13661075e-01
4.49849159e-01 -9.72078145e-02 8.87560248e-01 -1.05704165e+00
-5.87650537e-01 -5.77571750e-01 -8.24225545e-01 1.80267364e-01
2.86962867e-01 -4.97481167e-01 -5.36544919e-01 -4.18870188e-02] | [6.512970924377441, 1.3813589811325073] |
36d6ccad-e823-435d-92ee-297deecb1dab | w-talc-weakly-supervised-temporal-activity | 1807.10418 | null | http://arxiv.org/abs/1807.10418v3 | http://arxiv.org/pdf/1807.10418v3.pdf | W-TALC: Weakly-supervised Temporal Activity Localization and Classification | Most activity localization methods in the literature suffer from the burden
of frame-wise annotation requirement. Learning from weak labels may be a
potential solution towards reducing such manual labeling effort. Recent years
have witnessed a substantial influx of tagged videos on the Internet, which can
serve as a rich source of weakly-supervised training data. Specifically, the
correlations between videos with similar tags can be utilized to temporally
localize the activities. Towards this goal, we present W-TALC, a
Weakly-supervised Temporal Activity Localization and Classification framework
using only video-level labels. The proposed network can be divided into two
sub-networks, namely the Two-Stream based feature extractor network and a
weakly-supervised module, which we learn by optimizing two complimentary loss
functions. Qualitative and quantitative results on two challenging datasets -
Thumos14 and ActivityNet1.2, demonstrate that the proposed method is able to
detect activities at a fine granularity and achieve better performance than
current state-of-the-art methods. | ['Amit K. Roy-Chowdhury', 'Sujoy Paul', 'Sourya Roy'] | 2018-07-27 | w-talc-weakly-supervised-temporal-activity-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Sujoy_Paul_W-TALC_Weakly-supervised_Temporal_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Sujoy_Paul_W-TALC_Weakly-supervised_Temporal_ECCV_2018_paper.pdf | eccv-2018-9 | ['weakly-supervised-action-localization'] | ['computer-vision'] | [ 3.09300482e-01 -2.92565435e-01 -8.81738842e-01 -2.83047736e-01
-9.08140242e-01 -5.70657730e-01 7.12890148e-01 1.35276869e-01
-5.94200730e-01 7.62801766e-01 4.06486481e-01 1.87677085e-01
9.60827842e-02 -2.67037332e-01 -4.65702832e-01 -9.02284682e-01
-5.31479061e-01 -1.42194659e-01 5.97622752e-01 4.05806124e-01
-5.29538766e-02 2.21358255e-01 -1.38221407e+00 4.43903029e-01
4.53663647e-01 1.33959603e+00 9.62512121e-02 2.67273933e-01
1.15674309e-01 1.56840599e+00 -2.36223713e-01 1.28553733e-01
-1.21080764e-02 -4.74783421e-01 -9.14489686e-01 3.32885295e-01
3.37437093e-01 -2.23097861e-01 -5.60732543e-01 6.15769804e-01
2.62310386e-01 3.10690433e-01 2.38915250e-01 -1.32024038e+00
-5.61527768e-03 4.64802027e-01 -5.16375721e-01 4.76042390e-01
5.03503501e-01 1.83610860e-02 1.01251757e+00 -9.05346274e-01
5.10631382e-01 8.08973551e-01 6.86621904e-01 3.48657548e-01
-1.12965548e+00 -5.84485054e-01 3.19247097e-01 4.22486514e-01
-1.63867795e+00 -5.30784130e-01 7.66186714e-01 -5.22590935e-01
8.97011638e-01 -5.36764413e-02 6.55741274e-01 1.42102301e+00
-2.08006263e-01 1.10252845e+00 1.06348169e+00 -1.28887221e-01
2.39829302e-01 -2.15480745e-01 -2.13190466e-01 1.00811672e+00
-8.99012610e-02 -2.01594189e-01 -1.18755066e+00 -4.99105230e-02
6.66440487e-01 1.25674665e-01 -4.05244529e-01 -5.20617485e-01
-1.55783069e+00 3.91043037e-01 2.23110914e-01 5.30731857e-01
-2.47438937e-01 5.16168952e-01 5.98532081e-01 -3.77381854e-02
8.02455902e-01 1.67485029e-01 -4.31801111e-01 -6.96219742e-01
-1.23197711e+00 -1.50139362e-01 6.13480210e-01 7.98289955e-01
6.89485013e-01 -5.77965789e-02 -4.32279050e-01 7.48093724e-01
2.79811382e-01 1.04764983e-01 4.40351337e-01 -1.08966112e+00
5.86117566e-01 4.55139428e-01 1.64550513e-01 -7.31684387e-01
-4.08575177e-01 -4.84239042e-01 -5.41654885e-01 -1.96656346e-01
5.27148426e-01 -6.05920218e-02 -6.17233217e-01 1.83562624e+00
1.89245343e-01 9.34938073e-01 -3.33665460e-01 9.37723577e-01
4.55656737e-01 6.16558135e-01 3.09159636e-01 -3.24573606e-01
1.14418602e+00 -1.36486220e+00 -8.17583561e-01 -1.94823951e-01
8.99435997e-01 -4.33524430e-01 1.01719010e+00 3.13496232e-01
-1.10522342e+00 -5.11816144e-01 -1.05922246e+00 1.92196965e-01
-2.05484822e-01 3.56203943e-01 7.73240149e-01 3.00157636e-01
-9.57108557e-01 6.31790578e-01 -1.41234767e+00 -4.88750428e-01
8.41744602e-01 1.60717621e-01 -3.86807650e-01 -6.14529438e-02
-8.45781744e-01 5.25349736e-01 4.45055008e-01 5.34744747e-02
-1.35369611e+00 -5.04296720e-01 -9.96220350e-01 1.61823899e-01
8.24754655e-01 -1.36925593e-01 1.18899179e+00 -9.21639919e-01
-1.30485463e+00 8.12439442e-01 -3.03801596e-01 -5.98133743e-01
6.45337820e-01 -4.48108107e-01 -4.12893951e-01 4.36302215e-01
3.45715702e-01 5.06040573e-01 6.35537565e-01 -8.80843401e-01
-1.05274725e+00 2.64182035e-02 1.57125860e-01 1.96364328e-01
-6.31874025e-01 7.65798464e-02 -9.38798130e-01 -8.86793137e-01
-1.38663292e-01 -8.41722786e-01 1.27626471e-02 2.69300312e-01
-1.26955450e-01 -4.29544508e-01 9.95809913e-01 -3.60564142e-01
1.30471087e+00 -2.08444548e+00 -5.57116494e-02 -8.33857730e-02
2.21545696e-01 1.89759865e-01 -1.20440975e-01 2.96816438e-01
1.08800918e-01 -1.59990087e-01 -1.07327342e-01 -6.04614079e-01
-1.30857781e-01 2.10829809e-01 6.24658875e-02 8.04478943e-01
2.90112972e-01 8.75349879e-01 -1.47420979e+00 -6.81781888e-01
2.20241889e-01 4.30959821e-01 -2.51332372e-01 3.35684925e-01
-2.60811299e-01 6.67802989e-01 -4.01138544e-01 8.71778786e-01
1.00816354e-01 -5.65229893e-01 2.70760775e-01 -2.71821827e-01
-1.84242293e-01 2.21848428e-01 -8.60821426e-01 2.24518442e+00
-4.25240517e-01 9.13917422e-01 -1.08800128e-01 -1.38488090e+00
4.64428633e-01 4.67713207e-01 1.23276901e+00 -5.75305641e-01
5.48422076e-02 3.66473347e-02 -3.49031538e-01 -6.63024604e-01
1.18384987e-01 1.86754525e-01 -2.13153824e-01 5.13050616e-01
5.86285114e-01 6.82861626e-01 6.61611557e-01 1.56844437e-01
1.60937440e+00 7.35836327e-01 2.25568876e-01 -1.34261459e-01
7.40468562e-01 -3.08505118e-01 8.79544675e-01 6.99450016e-01
-5.03428102e-01 3.13448519e-01 3.70704025e-01 -3.86903167e-01
-5.58758199e-01 -1.08079910e+00 1.53133109e-01 1.48575854e+00
3.09185982e-01 -7.10647821e-01 -5.93564808e-01 -1.08168364e+00
-4.24332470e-01 -2.75518857e-02 -4.83802140e-01 -1.53324351e-01
-6.18025899e-01 -4.61012483e-01 8.16378176e-01 9.36154783e-01
7.63607860e-01 -9.81260955e-01 -6.07478201e-01 2.10464403e-01
-6.09871745e-01 -1.76349986e+00 -5.75102627e-01 4.43925709e-01
-7.70928383e-01 -1.18185222e+00 -7.21058309e-01 -9.17951822e-01
5.72527885e-01 4.63424534e-01 1.13315642e+00 3.72147784e-02
-7.00253481e-03 5.56774914e-01 -5.35055161e-01 9.02200639e-02
2.73010761e-01 2.32926473e-01 1.46572337e-01 4.27163422e-01
5.09886324e-01 -6.99811637e-01 -6.97995842e-01 5.45175314e-01
-7.04428971e-01 -7.23169697e-03 4.64825422e-01 6.02822006e-01
7.47194171e-01 1.65161416e-02 5.06917000e-01 -6.85690999e-01
1.23260409e-01 -6.65387690e-01 -4.47162330e-01 2.04347819e-01
-3.78711998e-01 -1.71093717e-02 5.49305618e-01 -5.21011949e-01
-1.16273940e+00 4.97168809e-01 1.34449020e-01 -4.27290887e-01
-1.80537924e-01 5.89084268e-01 -2.52429366e-01 -1.92927331e-01
4.42428440e-01 3.02478760e-01 -5.49935162e-01 -4.57507670e-01
2.34765977e-01 4.02277648e-01 8.03632736e-01 -6.29582644e-01
6.55293822e-01 8.74537230e-01 -9.22384392e-03 -7.57986546e-01
-1.39538753e+00 -9.98756289e-01 -7.53905356e-01 -6.87799037e-01
9.61156785e-01 -1.22596383e+00 -5.28609753e-01 5.16220629e-01
-8.64655375e-01 -6.61632419e-01 -3.29594910e-01 6.70646608e-01
-7.58351982e-01 3.86125118e-01 -6.85304940e-01 -6.95599735e-01
1.76336721e-01 -6.91268981e-01 1.12801135e+00 1.08936854e-01
-1.85782492e-01 -1.10914850e+00 2.29925707e-01 4.91325647e-01
1.22354172e-01 3.88632149e-01 2.25132704e-01 -5.50864279e-01
-6.73783183e-01 -2.33144283e-01 -1.45350039e-01 3.91932070e-01
2.59799808e-01 -3.70575309e-01 -1.07533967e+00 -3.58412802e-01
-2.63396710e-01 -6.81493044e-01 9.64764178e-01 2.08575353e-01
1.46343577e+00 -5.30208498e-02 -4.54715490e-01 7.77415693e-01
1.16231668e+00 5.86017445e-02 3.45948696e-01 1.91775188e-01
8.53889525e-01 2.59588689e-01 8.32804680e-01 5.96425176e-01
3.06106746e-01 8.93268228e-01 3.45540106e-01 -2.12976500e-01
-1.14653930e-01 -4.32942390e-01 5.59190273e-01 7.91529298e-01
-4.16527629e-01 -3.11270833e-01 -7.95710027e-01 6.26058161e-01
-2.41216373e+00 -1.18598020e+00 2.50232935e-01 1.91683960e+00
5.79920530e-01 2.46436328e-01 3.31347197e-01 2.03126833e-01
5.05618095e-01 6.80756748e-01 -4.01496589e-01 5.02535582e-01
-4.36286703e-02 -3.91071923e-02 4.03019398e-01 5.19849323e-02
-1.70143807e+00 8.72539401e-01 5.81809425e+00 8.21015239e-01
-9.07660544e-01 4.28738892e-01 3.63428354e-01 -4.03962672e-01
4.62438315e-01 -6.93753883e-02 -4.81599361e-01 7.92378485e-01
1.01295590e+00 1.34612218e-01 2.54309505e-01 7.95748472e-01
7.15654373e-01 -2.87045956e-01 -1.32635808e+00 1.17291963e+00
1.89948693e-01 -1.44701087e+00 -3.62742603e-01 8.32151994e-03
7.15669274e-01 1.56812683e-01 -2.67723531e-01 3.69738281e-01
3.55601721e-02 -7.21510947e-01 8.49571645e-01 5.25135994e-01
7.46399164e-01 -5.41532278e-01 8.68279219e-01 3.31479549e-01
-1.95791793e+00 -1.27161592e-01 9.94282812e-02 -1.36914298e-01
4.89373684e-01 4.21280473e-01 -3.20520252e-01 3.40977669e-01
9.91420567e-01 1.52687693e+00 -4.20011669e-01 1.09143555e+00
-3.47759694e-01 9.36648428e-01 -1.31460622e-01 2.01603726e-01
5.08461893e-01 -3.20848152e-02 1.55355573e-01 1.54215777e+00
1.68508098e-01 3.52145433e-02 7.74536431e-01 3.73137534e-01
-4.25161690e-01 -1.95741236e-01 -3.61769110e-01 -3.25909644e-01
3.18376213e-01 1.32779586e+00 -1.07304180e+00 -3.84203345e-01
-8.64580750e-01 1.18610191e+00 4.64381397e-01 4.03550982e-01
-1.10479569e+00 -1.10267133e-01 5.81997454e-01 2.21417144e-01
1.30437672e-01 -4.48682696e-01 2.84377098e-01 -1.50911903e+00
1.37565032e-01 -5.59781075e-01 5.68639398e-01 -6.98295712e-01
-9.51151550e-01 2.10254878e-01 8.73057917e-03 -1.57644987e+00
-2.20948085e-02 -3.39536339e-01 -4.19918746e-01 2.98247576e-01
-1.58368027e+00 -1.16528547e+00 -6.33663177e-01 7.44985580e-01
8.29831898e-01 -6.55761883e-02 4.92222875e-01 6.81226194e-01
-6.04231536e-01 3.87548894e-01 -6.07859269e-02 4.29817230e-01
7.53374696e-01 -1.16108370e+00 -1.40645146e-01 1.15105438e+00
6.47492468e-01 2.37349316e-01 1.93788648e-01 -3.49777341e-01
-1.10080206e+00 -1.48616803e+00 8.05348158e-01 -3.74172419e-01
7.91479886e-01 -5.70136964e-01 -7.49262393e-01 5.95699370e-01
-1.02241017e-01 6.75151408e-01 8.32828283e-01 -1.60776988e-01
-2.36361906e-01 -1.68682769e-01 -7.38929272e-01 3.37743282e-01
1.57468867e+00 -8.82470846e-01 -3.37555557e-01 4.06763881e-01
4.49591815e-01 -2.15209112e-01 -9.39389467e-01 3.49568963e-01
5.57171822e-01 -8.40418220e-01 8.93122077e-01 -4.55887467e-01
1.99980259e-01 -5.12472272e-01 5.47320098e-02 -9.35400248e-01
-2.91143417e-01 -7.37488806e-01 -7.81043291e-01 1.24690557e+00
7.28161931e-02 -9.18447226e-02 1.26276040e+00 1.03792414e-01
-1.51879966e-01 -7.80827522e-01 -8.81865323e-01 -1.07528663e+00
-7.43493617e-01 -6.63174331e-01 -2.19514631e-02 9.45037067e-01
2.86123693e-01 3.93490076e-01 -7.05536783e-01 -3.84490341e-02
5.88743746e-01 -2.49950230e-01 4.69349772e-01 -1.05497563e+00
-2.13793427e-01 -2.47541681e-01 -5.92541516e-01 -1.55261540e+00
4.39539105e-01 -7.79369712e-01 3.14960003e-01 -1.32152247e+00
3.10973495e-01 -2.24505782e-01 -7.75996625e-01 7.54872859e-01
8.33089724e-02 6.27152741e-01 1.32360095e-02 3.41580033e-01
-1.47156525e+00 6.01315439e-01 8.15789759e-01 -1.93311721e-01
-6.43104166e-02 5.32670598e-03 -1.69498891e-01 1.10220838e+00
6.16972148e-01 -5.61380506e-01 -6.45098984e-01 -1.75248727e-01
2.91272737e-02 4.59242117e-04 3.40563446e-01 -1.32464564e+00
2.96738178e-01 -2.19124451e-01 2.07341999e-01 -5.03728151e-01
4.55431223e-01 -9.10013914e-01 -2.03459367e-01 3.51311088e-01
-5.37851095e-01 -3.01572382e-01 -1.81334466e-01 9.57234204e-01
-5.36175013e-01 6.32183775e-02 6.51576698e-01 -7.57249147e-02
-1.04146194e+00 6.32694066e-01 -5.82017243e-01 2.84924567e-01
1.07049119e+00 -1.57478005e-01 -2.04236746e-01 -4.08150673e-01
-7.19142735e-01 3.41413707e-01 2.40277022e-01 4.91478533e-01
3.80425990e-01 -1.61319935e+00 -3.30658197e-01 -2.07150027e-01
4.25385207e-01 -2.27325246e-01 -2.10400205e-03 1.14018881e+00
-2.31460035e-01 5.06829679e-01 5.66991307e-02 -7.42877543e-01
-1.20136094e+00 3.75141859e-01 3.01136255e-01 -5.45648873e-01
-4.67009813e-01 6.92331314e-01 1.97958685e-02 1.53709456e-01
6.36425257e-01 -2.98656553e-01 -1.75635025e-01 1.59613952e-01
5.80417216e-01 6.13341868e-01 -1.20786071e-01 -8.63220394e-01
-6.56397521e-01 4.14613932e-01 2.83465654e-01 -1.43337781e-02
1.33796656e+00 -2.65376866e-01 2.41408646e-01 5.40906668e-01
1.29390144e+00 -3.57856572e-01 -1.84324396e+00 -5.74612856e-01
4.36700553e-01 -4.39058989e-01 3.69964577e-02 -5.19669116e-01
-1.12023103e+00 8.71619165e-01 6.72521591e-01 1.49870619e-01
1.18046606e+00 2.55555928e-01 7.77357697e-01 4.33980435e-01
6.15519404e-01 -1.30434740e+00 6.62582755e-01 3.96103829e-01
2.86549747e-01 -1.30975807e+00 -9.90057141e-02 -2.96541065e-01
-4.13613081e-01 7.80629218e-01 6.74744248e-01 5.18755056e-02
5.73111773e-01 2.17655197e-01 -2.03340307e-01 -1.62427813e-01
-7.09915936e-01 -5.11188269e-01 3.91046345e-01 6.11286104e-01
5.95777214e-01 -2.98898548e-01 -2.66284376e-01 5.13723671e-01
6.70605838e-01 3.50421309e-01 1.43149048e-01 1.21458495e+00
-3.67304742e-01 -9.00257826e-01 4.17081552e-04 3.52165252e-01
-7.34422028e-01 1.70744702e-01 -1.06974714e-01 5.30789554e-01
4.18694109e-01 1.04718304e+00 -1.04277078e-02 -3.54953945e-01
1.74344599e-01 -1.90855134e-02 3.81935775e-01 -6.24844253e-01
-2.58301139e-01 1.82880655e-01 2.76372045e-01 -1.25178802e+00
-1.36801183e+00 -6.39063954e-01 -1.05285358e+00 1.72477096e-01
-1.61045298e-01 2.09893152e-01 2.13429797e-02 1.06488252e+00
1.43251941e-01 5.20039439e-01 5.47903180e-01 -1.14815998e+00
-5.34093194e-02 -7.83264458e-01 -5.86927533e-01 5.80585718e-01
2.81971484e-01 -8.90469491e-01 -2.90245742e-01 6.91615582e-01] | [8.411432266235352, 0.6449281573295593] |
750cc919-4b94-4d59-9cf9-519ede077ee3 | render-for-cnn-viewpoint-estimation-in-images | 1505.05641 | null | http://arxiv.org/abs/1505.05641v1 | http://arxiv.org/pdf/1505.05641v1.pdf | Render for CNN: Viewpoint Estimation in Images Using CNNs Trained with Rendered 3D Model Views | Object viewpoint estimation from 2D images is an essential task in computer
vision. However, two issues hinder its progress: scarcity of training data with
viewpoint annotations, and a lack of powerful features. Inspired by the growing
availability of 3D models, we propose a framework to address both issues by
combining render-based image synthesis and CNNs. We believe that 3D models have
the potential in generating a large number of images of high variation, which
can be well exploited by deep CNN with a high learning capacity. Towards this
goal, we propose a scalable and overfit-resistant image synthesis pipeline,
together with a novel CNN specifically tailored for the viewpoint estimation
task. Experimentally, we show that the viewpoint estimation from our pipeline
can significantly outperform state-of-the-art methods on PASCAL 3D+ benchmark. | ['Hao Su', 'Yangyan Li', 'Leonidas Guibas', 'Charles R. Qi'] | 2015-05-21 | render-for-cnn-viewpoint-estimation-in-images-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Su_Render_for_CNN_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Su_Render_for_CNN_ICCV_2015_paper.pdf | iccv-2015-12 | ['viewpoint-estimation'] | ['computer-vision'] | [ 1.54117182e-01 4.35153022e-02 2.37380683e-01 -3.29291463e-01
-6.73873365e-01 -5.34019530e-01 8.85836244e-01 -4.91196930e-01
-8.15529823e-02 2.12788567e-01 5.53154498e-02 -1.23428740e-01
4.58219409e-01 -7.54717469e-01 -1.07982063e+00 -4.80834067e-01
4.64123577e-01 3.92040730e-01 4.79751527e-01 -2.34022111e-01
3.09135616e-01 8.75884831e-01 -1.78620672e+00 2.68998682e-01
5.10599077e-01 1.18204415e+00 3.03419232e-01 3.43416154e-01
-7.28025436e-02 7.67783582e-01 -5.04614651e-01 -3.93113822e-01
5.94801426e-01 -2.90545493e-01 -5.28517067e-01 3.23155224e-01
9.53985274e-01 -8.60353410e-01 -3.46809298e-01 7.80722857e-01
6.46882057e-01 -1.38853416e-01 5.58073878e-01 -1.19739795e+00
-5.00876546e-01 -3.37610990e-02 -4.83168721e-01 -7.32218400e-02
1.39872521e-01 3.95672560e-01 7.10987866e-01 -1.07679617e+00
9.03133929e-01 1.28049731e+00 5.80995917e-01 6.87461734e-01
-1.12036073e+00 -3.14061552e-01 1.93208069e-01 7.93618187e-02
-1.22726679e+00 -5.11468530e-01 1.08784056e+00 -5.50110221e-01
1.23670995e+00 8.39865133e-02 8.23608577e-01 1.22876251e+00
-1.26530230e-01 1.02418339e+00 1.07823253e+00 -4.17517632e-01
2.37682045e-01 -6.02713041e-02 -6.52825832e-01 7.07255185e-01
-7.83945620e-02 2.79564291e-01 -6.05019093e-01 3.77384275e-01
1.13263094e+00 -1.50534406e-01 -1.37295634e-01 -8.89335930e-01
-1.25211895e+00 6.93881154e-01 5.30961335e-01 -1.86999977e-01
-2.62191653e-01 2.69166470e-01 2.54213244e-01 -8.68730396e-02
7.87183166e-01 5.74047744e-01 -5.16025305e-01 8.24743733e-02
-9.74006057e-01 6.41308725e-01 4.31985766e-01 1.28277147e+00
7.08223045e-01 1.43682003e-01 -1.52900249e-01 5.84709585e-01
2.23244175e-01 4.84194994e-01 -2.00157072e-02 -1.12703443e+00
2.38736346e-01 7.51495242e-01 2.11929902e-01 -8.56423318e-01
-3.67552757e-01 -4.85326201e-01 -6.60829008e-01 5.34845710e-01
4.65303868e-01 4.42874640e-01 -8.63568664e-01 1.57981467e+00
5.43242216e-01 1.51694134e-01 -1.92540884e-01 1.06786740e+00
1.01255357e+00 3.54016960e-01 -3.47858310e-01 1.46721780e-01
1.21809125e+00 -1.19655180e+00 -4.36236888e-01 -2.19664991e-01
3.40476900e-01 -8.73205245e-01 9.91904259e-01 4.31864351e-01
-1.34707665e+00 -7.34697580e-01 -9.61436987e-01 -5.13750792e-01
-9.98379439e-02 2.59137869e-01 7.21948087e-01 4.96692210e-01
-1.12254274e+00 3.93885881e-01 -8.20298254e-01 -2.45246887e-01
7.78296113e-01 6.22548051e-02 -3.64200622e-01 -2.22274259e-01
-5.87132275e-01 9.99154329e-01 1.72839463e-01 1.41303137e-01
-1.06823659e+00 -8.30244184e-01 -9.99955833e-01 -2.09431112e-01
3.08982760e-01 -1.09699166e+00 1.18913603e+00 -6.00190520e-01
-1.70288348e+00 1.29194617e+00 5.52715473e-02 -4.51438576e-01
9.65205073e-01 -2.56031752e-01 2.19599858e-01 1.16480239e-01
-5.44561669e-02 1.14395249e+00 1.02298522e+00 -1.34923315e+00
-5.96058846e-01 -5.39217830e-01 2.04908028e-01 1.99519336e-01
-2.65635669e-01 -1.12682141e-01 -6.41952813e-01 -4.58516121e-01
1.04510747e-01 -9.50441718e-01 -2.60181189e-01 5.13912857e-01
-3.48201722e-01 -1.65383279e-01 8.10890615e-01 -4.94629890e-01
4.64850783e-01 -2.00638700e+00 2.17440784e-01 -4.44124550e-01
2.81235486e-01 4.58565176e-01 -1.77879035e-01 1.58640340e-01
2.27837726e-01 -9.66504961e-02 1.61735751e-02 -6.48082316e-01
8.87071639e-02 4.33221422e-02 -3.48451197e-01 4.92038667e-01
7.04671025e-01 1.07174349e+00 -9.26969826e-01 -3.06686431e-01
7.55858243e-01 6.76226318e-01 -8.59584808e-01 4.86855775e-01
-6.39463246e-01 7.58127391e-01 -2.53511876e-01 6.10109270e-01
9.77022886e-01 -3.32137316e-01 -6.57874942e-02 -4.84991729e-01
-3.02164614e-01 3.97301286e-01 -9.63738680e-01 2.04523873e+00
-5.20277321e-01 6.43567860e-01 -2.30274856e-01 -9.44251955e-01
1.05035174e+00 1.60744533e-01 3.51499885e-01 -6.58714175e-01
2.38004208e-01 2.54441291e-01 -3.18044215e-01 -4.45699543e-01
4.13094640e-01 6.29447028e-02 3.06111902e-01 8.93966332e-02
2.93968379e-01 -8.77894700e-01 1.78179704e-02 -1.69408008e-01
9.40402746e-01 7.88700044e-01 1.76871702e-01 -3.33615281e-02
3.59594822e-01 -3.17621529e-02 3.87859434e-01 3.88992041e-01
-2.23586783e-01 1.02762759e+00 4.04040337e-01 -9.03198063e-01
-1.40577745e+00 -9.41430509e-01 -1.14700057e-01 5.91749370e-01
1.62923887e-01 -2.55635321e-01 -5.86965859e-01 -7.75292695e-01
-4.13988195e-02 5.27126729e-01 -5.73632598e-01 -6.96080597e-03
-6.36376619e-01 -2.33899832e-01 2.90918738e-01 6.21013880e-01
7.32397556e-01 -8.63020837e-01 -7.70678282e-01 1.04860000e-01
-1.66810319e-01 -1.79231215e+00 -1.53379679e-01 -1.48612976e-01
-8.90999079e-01 -8.45724821e-01 -8.06632936e-01 -7.40216613e-01
6.21907234e-01 6.37378335e-01 1.55469191e+00 -9.69558358e-02
-2.89348811e-01 2.71945000e-01 -2.12154746e-01 -6.47529364e-01
-5.18668532e-01 1.75483719e-01 -2.22441599e-01 -1.10223547e-01
1.48557916e-01 -6.95534647e-01 -9.97186005e-01 4.06109869e-01
-8.86738420e-01 6.57805026e-01 6.19440317e-01 5.42115748e-01
7.15870142e-01 -3.44568789e-01 4.15609151e-01 -5.99956930e-01
-1.41166359e-01 3.40127014e-02 -1.01445639e+00 -5.91374747e-02
-3.03576410e-01 -9.41429660e-03 4.88735080e-01 -1.54856175e-01
-1.07298863e+00 6.21780515e-01 -3.87100995e-01 -6.62604749e-01
-3.86864185e-01 -2.09446684e-01 -3.51184636e-01 -2.78631091e-01
7.95096755e-01 2.12275222e-01 8.32464993e-02 -5.63751101e-01
6.05147123e-01 3.76100659e-01 4.11433727e-01 -2.78109044e-01
8.59026015e-01 8.87621522e-01 2.73254126e-01 -7.23839223e-01
-1.07636702e+00 -3.96470934e-01 -7.35280395e-01 -2.78198421e-01
1.03040600e+00 -1.29767275e+00 -4.92802620e-01 6.26757860e-01
-1.48118353e+00 -3.78978312e-01 -2.78695285e-01 3.08144838e-01
-9.35475290e-01 1.69325396e-01 -3.37997824e-01 -5.70853651e-01
-1.72054991e-01 -1.42493224e+00 1.67675328e+00 1.61291361e-02
2.90542301e-02 -7.56935179e-01 -1.11209847e-01 7.02511251e-01
3.64345133e-01 4.51288015e-01 6.25912189e-01 -3.29637021e-01
-1.21991062e+00 -1.21703379e-01 -4.80788529e-01 5.83810925e-01
-2.76061650e-02 -7.42954761e-02 -1.36169672e+00 -1.64031610e-01
1.93422455e-02 -5.65455496e-01 7.62216926e-01 2.84758210e-01
1.27204919e+00 1.01022027e-01 -1.74057513e-01 9.47865963e-01
1.28802776e+00 -4.27500010e-01 6.83057547e-01 2.33199775e-01
9.81615603e-01 5.49006879e-01 4.40906733e-01 4.01638806e-01
4.79018450e-01 1.00867617e+00 9.31670964e-01 -8.05237815e-02
-6.48763657e-01 -3.99475753e-01 -4.20342609e-02 6.79795802e-01
-2.01563880e-01 -1.80376485e-01 -7.88801908e-01 5.39991856e-01
-1.62205815e+00 -7.66136885e-01 -1.94072798e-01 2.16734099e+00
7.44730055e-01 1.37553766e-01 1.18334806e-02 -1.22877732e-02
2.61142731e-01 2.53943950e-01 -5.61505139e-01 -1.17032126e-01
-1.45604938e-01 -3.32499035e-02 2.86647141e-01 1.74616307e-01
-1.16876793e+00 1.08874738e+00 6.40827370e+00 7.47743011e-01
-1.22996783e+00 -1.51466042e-01 6.50902867e-01 8.71615950e-03
-3.72354418e-01 -1.36362642e-01 -1.03496766e+00 2.13922113e-01
3.89626086e-01 2.65648693e-01 3.09245944e-01 1.13989520e+00
8.83813426e-02 2.29737446e-01 -1.31020224e+00 1.19320047e+00
4.09831226e-01 -1.65878391e+00 1.12706579e-01 1.68274999e-01
1.09622121e+00 1.71088085e-01 6.80217519e-02 1.35861516e-01
7.08803460e-02 -8.57934296e-01 9.64920163e-01 1.72494441e-01
8.06969404e-01 -4.32352722e-01 3.28159213e-01 4.93340790e-01
-9.25646365e-01 2.14826986e-01 -5.20683408e-01 -7.28586912e-02
1.17972299e-01 6.66435540e-01 -1.01286876e+00 5.39570928e-01
7.38803804e-01 9.72647250e-01 -6.30535483e-01 1.12033808e+00
-4.37240690e-01 1.31253988e-01 -1.86254472e-01 1.54250100e-01
5.89734018e-02 -5.71647696e-02 4.91749913e-01 8.10946047e-01
3.96040469e-01 -2.49733821e-01 -2.40170974e-02 1.15617454e+00
-2.76975423e-01 -6.34363666e-02 -8.90510023e-01 2.31358677e-01
2.26915583e-01 1.41338766e+00 -7.17147470e-01 -1.84528038e-01
-5.49063444e-01 1.04957235e+00 5.58608890e-01 7.52137303e-02
-8.49406958e-01 2.61510879e-01 5.32237530e-01 2.24296331e-01
5.41311622e-01 -4.04290050e-01 -2.60410368e-01 -1.26627696e+00
1.87769338e-01 -9.52554047e-01 -2.42968500e-01 -1.11534548e+00
-1.27514708e+00 6.29065394e-01 -5.41235022e-02 -1.50753975e+00
-3.01271975e-01 -9.45959389e-01 -1.95966572e-01 6.87662601e-01
-1.88322830e+00 -1.60753083e+00 -6.51603639e-01 3.18792045e-01
9.04800117e-01 -7.59765282e-02 6.72872007e-01 1.26235113e-01
-1.26269877e-01 3.00086826e-01 -3.15112174e-01 -1.02957390e-01
6.68180823e-01 -1.22799575e+00 9.55816567e-01 6.90468371e-01
2.93206125e-01 2.54846245e-01 5.43931007e-01 -1.97688594e-01
-1.57417619e+00 -1.23200572e+00 7.37611711e-01 -9.41543102e-01
1.02084972e-01 -6.15043104e-01 -6.27608418e-01 6.11102819e-01
3.89631018e-02 4.25873846e-01 2.83432245e-01 -3.34392250e-01
-5.74490964e-01 -7.45530659e-03 -9.72494662e-01 8.44353795e-01
1.39624393e+00 -4.67139602e-01 -3.81519616e-01 5.72619200e-01
7.82726705e-01 -8.20338726e-01 -6.89342499e-01 5.76536298e-01
6.05206847e-01 -1.40052557e+00 1.34707272e+00 -2.16136768e-01
8.19780707e-01 -1.67248756e-01 -2.47057438e-01 -1.25827444e+00
-9.99480858e-02 -5.40060759e-01 -1.55190885e-01 1.06565619e+00
5.78057580e-02 -3.43856424e-01 8.81574750e-01 4.85782027e-01
-4.31127906e-01 -7.66888976e-01 -7.98875690e-01 -8.21017385e-01
1.80401132e-01 -3.49643320e-01 7.41174936e-01 6.67375267e-01
-7.22504854e-01 3.25220764e-01 -4.68480587e-01 -9.85108912e-02
6.38182759e-01 2.29738995e-01 1.39779508e+00 -1.19591534e+00
-2.61611223e-01 -4.52645808e-01 -6.97212338e-01 -1.47025335e+00
1.50600493e-01 -6.72594011e-01 1.15183756e-01 -1.52483404e+00
2.52197534e-02 -3.20750684e-01 1.52856693e-01 2.24332884e-01
-5.34493439e-02 6.93582714e-01 1.78416252e-01 1.17788248e-01
-5.41964412e-01 7.15560615e-01 1.79121673e+00 8.22638869e-02
6.68900758e-02 1.19500443e-01 -4.59003180e-01 8.58652890e-01
5.46740055e-01 -1.69976994e-01 -5.04016459e-01 -8.05244386e-01
4.05977786e-01 -1.24233343e-01 6.60191715e-01 -8.97950828e-01
-1.60964802e-01 1.34276813e-02 5.24944246e-01 -8.76323462e-01
6.80125237e-01 -7.78937221e-01 -6.41231686e-02 1.40888408e-01
-2.29156628e-01 -1.48624659e-01 2.18457773e-01 4.68214393e-01
-3.83830667e-02 8.64785239e-02 7.45795012e-01 -3.36269110e-01
-6.80054724e-01 5.53270638e-01 8.32784548e-02 -5.69259841e-03
8.52460504e-01 -2.00510323e-01 -3.53383303e-01 -2.54970759e-01
-2.46744007e-01 -1.39848694e-01 8.36091518e-01 5.79088271e-01
8.00084293e-01 -1.51536918e+00 -7.65773773e-01 5.06256878e-01
4.42756414e-01 5.42658806e-01 2.18104690e-01 6.11547172e-01
-8.63173902e-01 2.93453693e-01 -3.61731350e-01 -9.92364228e-01
-1.00445950e+00 6.60029948e-01 3.87643546e-01 -8.45222473e-02
-7.42303967e-01 8.92603040e-01 5.26673377e-01 -7.53685474e-01
1.63098261e-01 -4.88514811e-01 1.09328426e-01 -3.28810751e-01
4.82075989e-01 1.33270293e-01 3.43983769e-01 -5.13770342e-01
-2.30571628e-01 6.85533166e-01 7.14580417e-02 4.25901711e-02
1.55223620e+00 2.45313495e-02 6.76901639e-02 1.60639018e-01
1.30619729e+00 -2.13822275e-01 -1.88178861e+00 -2.31707707e-01
-3.46557349e-01 -8.90307844e-01 2.05585763e-01 -5.77459812e-01
-1.02330923e+00 1.18206966e+00 4.58659649e-01 7.99707472e-02
1.01789451e+00 1.09234704e-02 8.20594251e-01 2.31930077e-01
4.07624424e-01 -8.66573751e-01 5.10868490e-01 4.29054856e-01
1.25113201e+00 -1.57657552e+00 5.91725893e-02 -6.33495986e-01
-3.67297143e-01 1.00829327e+00 7.43632197e-01 -3.43042463e-01
4.71479088e-01 1.75895989e-01 8.94691050e-02 -1.05495006e-01
-6.78083479e-01 -3.12587440e-01 4.67807710e-01 8.97851288e-01
3.87986958e-01 -3.93000543e-01 1.12614639e-01 1.85344338e-01
-1.61204040e-01 -3.59435827e-02 4.00564373e-01 7.10632503e-01
-1.88596487e-01 -1.02255261e+00 -4.56428230e-02 -1.09101444e-01
-2.22134873e-01 -2.44534500e-02 -3.65413129e-01 8.28084350e-01
1.61466300e-01 5.78840196e-01 5.92567660e-02 -1.70011982e-01
4.85989958e-01 -9.18392837e-02 9.95500147e-01 -5.95802605e-01
-2.96068758e-01 1.56738743e-01 -8.77294242e-02 -7.71954715e-01
-5.33599496e-01 -5.45809388e-01 -5.99250674e-01 -2.23492570e-02
-2.79868484e-01 -7.43409157e-01 9.07349944e-01 8.58220398e-01
5.55167615e-01 5.69254100e-01 7.70401657e-01 -1.44906032e+00
-6.44949257e-01 -7.84587681e-01 -1.85492232e-01 5.51600158e-01
1.42020822e-01 -7.42369115e-01 -2.81315088e-01 1.42488748e-01] | [8.384620666503906, -2.8108129501342773] |
b6c96444-c165-4fa7-b537-1bede4c2f910 | taiwanese-accented-mandarin-and-english-multi | null | null | https://aclanthology.org/2022.rocling-1.6 | https://aclanthology.org/2022.rocling-1.6.pdf | Taiwanese-Accented Mandarin and English Multi-Speaker Talking-Face Synthesis System | This paper proposes a multi-speaker talking-face synthesis system. The system incorporates voice cloning and lip-syncing technology to achieve text-to-talking-face generation by acquiring audio and video clips of any speaker and using zero-shot transfer learning. In addition, we used open-source corpora to train several Taiwanese-accented models and proposed using Mandarin Phonetic Symbols (Bopomofo) as the character embedding of the synthesizer to improve the system’s ability to synthesize Chinese-English code-switched sentences. Through our system, users can create rich applications. Also, the research on this technology is novel in the audiovisual speech synthesis field. | ['Chun-Hsin Wu', 'Kai-Chun Liao', 'Cho-Chun Hsieh', 'Jian-Peng Liao', 'Chia-Hsuan Lin'] | null | null | null | null | rocling-2022-11 | ['talking-face-generation', 'face-generation', 'voice-cloning'] | ['computer-vision', 'computer-vision', 'speech'] | [ 1.48728877e-01 1.75147921e-01 -1.31579846e-01 -4.75287616e-01
-1.07272816e+00 -2.68572569e-01 3.12768370e-01 -1.23356700e+00
1.61226526e-01 7.56626189e-01 5.87437451e-01 -2.98682034e-01
6.01679146e-01 -4.17327404e-01 -5.50624371e-01 -6.26213431e-01
5.01346946e-01 -4.43286598e-02 -6.82577640e-02 -3.62699240e-01
1.24706887e-01 3.71068090e-01 -1.66759109e+00 6.92222118e-01
6.34327114e-01 6.87264025e-01 4.71773416e-01 9.43847120e-01
-4.58495319e-01 6.87690496e-01 -7.17785358e-01 -5.75563312e-01
2.67713945e-02 -6.99535429e-01 -6.33609235e-01 1.13702454e-01
4.65929478e-01 -5.90256751e-01 -3.82885039e-01 8.74098122e-01
9.35454309e-01 6.94461018e-02 5.81846356e-01 -1.21981764e+00
-1.01445520e+00 6.83353782e-01 -1.52129054e-01 -1.75579876e-01
6.88031673e-01 2.66401052e-01 5.82201362e-01 -1.04691041e+00
4.23120886e-01 1.62749553e+00 3.92087609e-01 1.23490191e+00
-6.56326413e-01 -1.15680182e+00 -4.58530307e-01 2.89381146e-01
-1.69717300e+00 -1.33628392e+00 7.60815144e-01 -2.68670768e-01
9.03307855e-01 3.25217694e-01 2.12098122e-01 1.30420721e+00
1.50011316e-01 4.60293740e-01 7.62237430e-01 -6.27057552e-01
-1.53091177e-01 5.54903269e-01 -6.31677270e-01 6.75769806e-01
-8.11634600e-01 4.02712785e-02 -6.98344171e-01 1.79085985e-01
7.43085384e-01 -6.83922470e-01 -1.94645643e-01 5.05930603e-01
-9.64809895e-01 8.83865416e-01 -1.57288209e-01 4.02395695e-01
-4.60371673e-02 -9.23790336e-02 3.67740452e-01 2.77188122e-01
3.02882195e-01 -1.87256053e-01 -5.24810851e-02 -2.45205492e-01
-9.94262218e-01 -1.44300848e-01 6.85271502e-01 1.39797318e+00
2.32653514e-01 8.16246033e-01 -2.82797426e-01 1.32737160e+00
3.90716583e-01 6.56618178e-01 9.37857330e-01 -1.07593441e+00
3.48961204e-01 -2.66399086e-01 -1.43971711e-01 -4.55521137e-01
3.25507522e-01 4.17403400e-01 -4.45638686e-01 -4.41680588e-02
-7.91716799e-02 -7.09920645e-01 -7.59078503e-01 1.53131461e+00
1.90154448e-01 5.15242875e-01 3.24205816e-01 6.61348224e-01
1.22375500e+00 1.24193239e+00 -1.73123479e-01 -4.91809666e-01
1.20989478e+00 -1.26943219e+00 -1.35512030e+00 3.27805914e-02
7.71600381e-02 -1.06011689e+00 1.39331341e+00 1.77277282e-01
-1.19246578e+00 -1.05028963e+00 -8.62491846e-01 -4.14377898e-02
-1.44803107e-01 1.61835253e-01 7.37013593e-02 1.17086482e+00
-1.29788303e+00 -7.43598817e-03 -1.35884851e-01 -5.32692596e-02
2.01990545e-01 2.75469661e-01 -3.76602232e-01 2.15411067e-01
-1.33287966e+00 7.80188620e-01 1.64295003e-01 -2.56717086e-01
-9.41658258e-01 -4.53194618e-01 -9.39629495e-01 1.65187836e-01
2.47457381e-02 -3.26425612e-01 1.59462130e+00 -1.28146231e+00
-2.67375541e+00 6.69578433e-01 -7.00399578e-01 4.05869670e-02
1.88426301e-01 1.65267494e-02 -1.10834694e+00 4.15082365e-01
-2.00059295e-01 9.13953364e-01 1.39235282e+00 -1.00474286e+00
-5.83686113e-01 1.88558802e-01 -4.97580200e-01 3.34953040e-01
-4.62132066e-01 6.67472601e-01 -4.00118500e-01 -7.55893171e-01
-5.28765202e-01 -6.50822341e-01 5.24198771e-01 -4.85444330e-02
-2.01006383e-01 -2.85083234e-01 1.16211033e+00 -1.00502110e+00
1.18215835e+00 -2.56811047e+00 -3.41919414e-03 -2.74924248e-01
-4.57122892e-01 4.08647209e-01 -2.78344452e-01 3.50088120e-01
3.50570902e-02 1.44853458e-01 4.68126982e-02 -4.17485744e-01
-2.23676533e-01 1.27253696e-01 -4.29054141e-01 8.90063271e-02
2.06816196e-01 6.88467443e-01 -4.25761133e-01 -9.07559872e-01
3.98394078e-01 7.23334908e-01 -5.87942243e-01 4.63266879e-01
-2.18664836e-02 5.09268701e-01 -3.40573341e-02 5.76256931e-01
4.80009854e-01 5.06312668e-01 -1.54468700e-01 1.40498072e-01
-3.02783728e-01 7.71602690e-02 -1.07106602e+00 1.59148264e+00
-8.87227476e-01 7.81732380e-01 4.90044266e-01 -5.03345728e-01
1.11458635e+00 1.04107130e+00 1.86070189e-01 -2.62467265e-01
2.49962866e-01 1.95140049e-01 -1.45964742e-01 -9.61805344e-01
5.06125271e-01 -5.26996791e-01 2.83752531e-01 -3.95719297e-02
4.68974173e-01 -5.65612793e-01 -2.32893109e-01 -2.36484081e-01
1.14272550e-01 1.08670719e-01 -5.87831438e-02 -8.50569755e-02
1.02284729e+00 -6.57013774e-01 4.28814262e-01 8.79229754e-02
-2.65261978e-01 6.73002005e-01 4.39216495e-02 4.42949682e-01
-9.31560159e-01 -1.23345840e+00 -1.60976931e-01 1.34774530e+00
-4.65240955e-01 -1.86124772e-01 -1.21090627e+00 -1.70046121e-01
-3.57065082e-01 1.08137429e+00 1.15984403e-01 -9.58695561e-02
-4.95473802e-01 2.49004140e-01 9.86216605e-01 2.24598944e-01
3.72824430e-01 -1.52231622e+00 3.11426818e-01 2.34007314e-01
-5.31959951e-01 -1.16250968e+00 -1.26322460e+00 -6.87790036e-01
-2.74118990e-01 -3.92888397e-01 -1.16813755e+00 -1.57926536e+00
1.58688858e-01 6.27821013e-02 3.50394607e-01 -5.01101494e-01
-1.18173428e-01 2.56665617e-01 -2.92960733e-01 -5.88438272e-01
-9.59981441e-01 -1.54771298e-01 3.47236246e-01 4.09806758e-01
3.15473258e-01 -2.79443830e-01 -3.36337984e-02 3.95169824e-01
-5.48684895e-01 3.42551544e-02 2.24755839e-01 6.62048936e-01
3.69784124e-02 -2.21867085e-01 1.23827648e+00 -1.47685140e-01
1.01747000e+00 -1.59771979e-01 -4.60653126e-01 3.07902277e-01
-6.25285506e-02 -3.54700208e-01 8.37556362e-01 -6.75494075e-01
-1.70013881e+00 1.64126158e-02 -7.42387116e-01 -5.96316218e-01
-3.54137152e-01 -3.84676456e-03 -5.66309571e-01 -1.13965832e-01
3.18573296e-01 5.71056902e-01 4.28762704e-01 -2.04482600e-01
6.64701998e-01 1.90012002e+00 8.13158333e-01 -2.07526967e-01
3.94628048e-01 -5.08146584e-02 -6.93046808e-01 -1.68819523e+00
4.84876968e-02 -1.89506367e-01 -3.11506718e-01 -5.64816773e-01
1.11078107e+00 -1.10612559e+00 -1.02071023e+00 8.21936190e-01
-1.30585444e+00 -2.39564955e-01 -4.03007381e-02 6.77994132e-01
-8.15770507e-01 2.88677275e-01 -8.23565960e-01 -1.03422904e+00
-3.93629640e-01 -1.30347419e+00 9.99018252e-01 4.34507489e-01
-1.25172794e-01 -6.90510035e-01 -3.89728397e-02 6.45269930e-01
5.90698481e-01 -6.05371714e-01 5.67497373e-01 -3.37305129e-01
-2.58310169e-01 1.72060803e-01 1.41757309e-01 8.40619147e-01
4.67866361e-01 3.78902793e-01 -1.36274946e+00 -1.23569071e-01
1.13296114e-01 -3.39506358e-01 1.98122054e-01 5.31225979e-01
1.15111339e+00 -6.35383487e-01 -1.33400321e-01 5.87105334e-01
8.35316837e-01 7.08841085e-01 8.82565916e-01 -5.17043710e-01
4.71285254e-01 7.29116440e-01 1.94069639e-01 2.72310853e-01
3.49696010e-01 7.64715374e-01 -3.02073479e-01 7.21943974e-02
-5.45821130e-01 -5.13198733e-01 8.11134219e-01 1.37307787e+00
1.87310308e-01 -2.55870610e-01 -5.34274578e-01 4.48278219e-01
-1.02086616e+00 -1.28313220e+00 1.38773695e-01 1.89551771e+00
1.23994207e+00 -2.88204521e-01 1.01181924e-01 -7.10541308e-02
1.22846603e+00 9.51741934e-02 -2.26377264e-01 -8.67867529e-01
-6.56045079e-02 5.43516457e-01 7.00048804e-02 1.01039815e+00
-7.27762699e-01 1.35368705e+00 7.33518982e+00 1.14431286e+00
-1.56852651e+00 2.54819036e-01 4.08623308e-01 -6.92203715e-02
-1.98357970e-01 -4.83648866e-01 -9.76513803e-01 6.47780657e-01
1.50124836e+00 -5.28273642e-01 7.86820352e-01 8.17631125e-01
5.23077011e-01 2.59631395e-01 -7.02841461e-01 1.21042860e+00
5.23569465e-01 -1.23354578e+00 5.99841438e-02 -2.15215519e-01
6.29391193e-01 -3.89407456e-01 2.74099320e-01 4.17231798e-01
-1.89162076e-01 -1.21398437e+00 7.07988322e-01 4.40403730e-01
1.43475294e+00 -7.45962799e-01 3.05306762e-01 2.26165771e-01
-1.22496533e+00 -9.78056788e-02 -2.10350811e-01 2.31738761e-01
4.92402524e-01 -2.10825637e-01 -1.23050439e+00 2.70027548e-01
3.78262579e-01 2.21782029e-01 -1.06494524e-01 7.44622052e-01
6.49137190e-03 7.35799193e-01 -3.22056673e-02 -2.11626217e-01
-1.48609132e-01 -2.11822733e-01 3.55786324e-01 1.25862575e+00
7.68467963e-01 6.85327873e-02 -2.29495451e-01 5.93262196e-01
-2.56747931e-01 4.00333792e-01 -7.54258275e-01 -2.14882761e-01
5.43820441e-01 1.01908970e+00 -5.11525683e-02 -3.41349065e-01
-6.03892267e-01 1.11105061e+00 -4.41857070e-01 4.36034083e-01
-9.06080723e-01 -9.00918961e-01 6.69467390e-01 2.25975830e-02
2.87179321e-01 -2.49570012e-01 -1.34040713e-01 -9.23214316e-01
-3.51827294e-01 -1.11430466e+00 -2.79148251e-01 -1.22763538e+00
-1.04492033e+00 7.42264986e-01 -7.89703727e-02 -8.98750782e-01
-6.58556700e-01 -5.10121882e-01 -8.70964050e-01 1.33838820e+00
-1.25375259e+00 -1.23414087e+00 8.73158351e-02 9.15326834e-01
1.18731022e+00 -7.85197735e-01 1.05148089e+00 5.34542620e-01
-4.08388078e-01 8.89752924e-01 9.16340426e-02 1.28152624e-01
1.06774724e+00 -5.16179025e-01 4.06564891e-01 4.65003252e-01
-1.10448964e-01 3.63311857e-01 4.24484760e-01 -4.69145775e-01
-1.09892166e+00 -8.55414033e-01 1.02047646e+00 -3.15185785e-02
3.89278024e-01 -4.97586876e-01 -7.75485575e-01 5.42212903e-01
9.01447773e-01 -4.64998901e-01 9.55376804e-01 -4.65111107e-01
1.15233185e-02 -2.75979280e-01 -1.38237631e+00 7.57746160e-01
6.46147490e-01 -9.53001559e-01 -6.96362555e-01 1.40907407e-01
9.78373230e-01 -1.43967867e-01 -6.20930016e-01 2.99923066e-02
6.68495834e-01 -5.80548882e-01 7.45556533e-01 -3.56614798e-01
2.63714433e-01 6.74968660e-02 -2.20128402e-01 -1.35499668e+00
-7.05082491e-02 -9.96024728e-01 1.95190772e-01 1.71201992e+00
4.29983765e-01 -5.38682818e-01 5.18672705e-01 1.81463212e-01
-3.66313577e-01 -4.81614843e-02 -1.12293041e+00 -6.36134624e-01
2.59812236e-01 -1.72633901e-01 7.54588366e-01 7.81707764e-01
2.29052424e-01 4.69674796e-01 -7.49335468e-01 2.46969610e-02
1.77097484e-01 -5.74728847e-01 6.83100522e-01 -6.80727124e-01
-9.13270339e-02 -3.74364793e-01 1.76550522e-01 -9.62864399e-01
6.46516979e-01 -8.52901578e-01 2.79832095e-01 -1.08987045e+00
-4.54011291e-01 1.93375051e-01 3.33441347e-01 1.57336488e-01
1.23058736e-01 -6.46842597e-03 2.45519921e-01 -2.29693964e-01
2.98039764e-01 7.54322529e-01 1.34591711e+00 -1.35643840e-01
-2.72950679e-01 2.40510330e-01 -3.77694726e-01 4.87646371e-01
9.99560177e-01 -1.49210513e-01 -5.82873702e-01 -1.59409940e-01
-7.95473754e-01 6.31067276e-01 -1.98129773e-01 -9.04827356e-01
2.62630016e-01 -2.75504827e-01 -1.88969392e-02 -2.79556721e-01
8.49229276e-01 -6.64564788e-01 8.62637386e-02 2.63117880e-01
-4.55097914e-01 -2.47433826e-01 3.80542129e-01 -7.36680031e-02
-2.60259062e-01 -4.33716953e-01 1.05234659e+00 -9.33003351e-02
-4.63847667e-01 1.28739163e-01 -8.66354764e-01 -1.51805073e-01
1.28157604e+00 -2.79123753e-01 -5.93044832e-02 -8.89524817e-01
-5.46905577e-01 -1.03959389e-01 3.95515449e-02 8.04980218e-01
1.01385522e+00 -1.48021603e+00 -9.24643934e-01 8.73083055e-01
-1.18603006e-01 -7.05588698e-01 5.17252445e-01 1.24478355e-01
-5.42771518e-01 5.74727535e-01 -5.09185910e-01 -2.84596831e-01
-1.62170744e+00 4.08705980e-01 3.99602324e-01 7.85628915e-01
-1.77696809e-01 9.47612524e-01 -1.47181407e-01 -4.73685950e-01
4.48335379e-01 1.68786272e-01 -2.55714178e-01 -5.86494133e-02
6.52694702e-01 3.72543216e-01 -2.49064282e-01 -9.64486539e-01
-1.74228251e-01 4.29618269e-01 2.80970961e-01 -8.26768219e-01
7.82300711e-01 -3.51238638e-01 7.16105103e-02 5.71046531e-01
1.37181056e+00 4.80232358e-01 -9.22906578e-01 1.10793822e-01
-5.58053970e-01 -5.15571773e-01 -9.81631572e-04 -4.45672631e-01
-8.32532406e-01 1.06354570e+00 6.96126819e-01 -1.60500079e-01
8.49215925e-01 -2.40187034e-01 9.82613146e-01 1.87181696e-01
1.47529006e-01 -1.36312377e+00 8.88173506e-02 4.27626461e-01
1.10498512e+00 -1.08305168e+00 -7.73548007e-01 -3.46734792e-01
-1.11073458e+00 1.24639416e+00 5.88262558e-01 4.18205202e-01
7.61201382e-01 4.22339261e-01 5.34422398e-01 6.62963390e-01
-6.78025067e-01 6.83628619e-02 1.73826247e-01 9.58863080e-01
5.94817102e-01 -4.81977174e-03 -1.72629684e-01 5.24027944e-01
-6.31438732e-01 3.17010224e-01 5.72812736e-01 4.02735054e-01
-5.99332273e-01 -1.05143380e+00 -6.31488144e-01 1.01745337e-01
-7.42570341e-01 -2.76806235e-01 -1.63805127e-01 2.44386032e-01
-6.74194470e-02 1.47628653e+00 1.24417588e-01 -5.02997875e-01
1.12875476e-01 7.05055714e-01 2.54948795e-01 -6.32757843e-01
-3.67603600e-01 3.95167083e-01 1.00511856e-01 -7.04406574e-02
-1.37919754e-01 -4.40036267e-01 -1.14403665e+00 -3.71067107e-01
-3.40209782e-01 1.98281854e-01 9.22868192e-01 7.50693619e-01
3.32563907e-01 6.23522103e-01 1.09942544e+00 -7.04399526e-01
-4.32081968e-01 -1.28152931e+00 -5.78177273e-01 -1.44773021e-01
4.51775521e-01 -1.39929831e-01 -2.72796512e-01 4.56468910e-01] | [14.812342643737793, 6.516161918640137] |
7e86cef6-90b4-4392-9593-b64d2a91fe84 | challenging-mitosis-detection-algorithms | 2211.16852 | null | https://arxiv.org/abs/2211.16852v1 | https://arxiv.org/pdf/2211.16852v1.pdf | Challenging mitosis detection algorithms: Global labels allow centroid localization | Mitotic activity is a crucial proliferation biomarker for the diagnosis and prognosis of different types of cancers. Nevertheless, mitosis counting is a cumbersome process for pathologists, prone to low reproducibility, due to the large size of augmented biopsy slides, the low density of mitotic cells, and pattern heterogeneity. To improve reproducibility, deep learning methods have been proposed in the last years using convolutional neural networks. However, these methods have been hindered by the process of data labelling, which usually solely consist of the mitosis centroids. Therefore, current literature proposes complex algorithms with multiple stages to refine the labels at pixel level, and to reduce the number of false positives. In this work, we propose to avoid complex scenarios, and we perform the localization task in a weakly supervised manner, using only image-level labels on patches. The results obtained on the publicly available TUPAC16 dataset are competitive with state-of-the-art methods, using only one training phase. Our method achieves an F1-score of 0.729 and challenges the efficiency of previous methods, which required multiple stages and strong mitosis location information. | ['Valery Naranjo', 'Emiel Janssen', 'Sandra Morales', 'Julio Silva-Rodríguez', 'Umay Kiraz', 'Claudio Fernandez-Martín'] | 2022-11-30 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 3.09949249e-01 1.10730194e-01 -3.58629704e-01 3.94381806e-02
-9.24657822e-01 -5.71067572e-01 5.16307592e-01 6.84589863e-01
-8.22426260e-01 8.48125935e-01 -2.15696529e-01 -2.17160821e-01
1.07728645e-01 -6.46538019e-01 -4.47341681e-01 -1.29999268e+00
2.49101475e-01 5.11591256e-01 4.72849309e-01 2.97396064e-01
2.44767129e-01 4.43402559e-01 -1.33632159e+00 1.84983537e-01
6.24271274e-01 9.04356539e-01 2.04946935e-01 6.69463038e-01
-1.10281110e-01 4.36166108e-01 -4.83524591e-01 -2.52554059e-01
-6.38025701e-02 -3.55368316e-01 -6.10595167e-01 7.41517097e-02
2.57170230e-01 -1.26644775e-01 -1.85332015e-01 1.10857022e+00
4.97397482e-01 -5.84109187e-01 7.47917414e-01 -7.73359478e-01
-3.23700786e-01 4.04317409e-01 -8.43262017e-01 3.13623756e-01
-2.73190379e-01 2.64676839e-01 7.19605386e-01 -5.81071496e-01
7.13703513e-01 3.53377402e-01 7.99074054e-01 5.54953992e-01
-1.32962108e+00 -4.79867965e-01 -3.51889044e-01 1.09984152e-01
-1.67175019e+00 -3.57237756e-01 5.44465899e-01 -5.51279545e-01
6.65272951e-01 1.38086095e-01 7.10169375e-01 8.16964686e-01
3.21542650e-01 5.59948504e-01 1.37087846e+00 -4.16656584e-01
3.37834120e-01 1.85150981e-01 -8.49689692e-02 7.40660965e-01
5.63809574e-01 -2.83972591e-01 -1.79425299e-01 2.48023719e-01
8.90145123e-01 1.81659162e-01 -3.26227874e-01 -3.32360178e-01
-1.38806546e+00 5.94801426e-01 4.58077192e-01 9.08702791e-01
-2.45979372e-02 -1.66705810e-02 4.40176249e-01 -2.72356153e-01
3.92497718e-01 3.78364325e-01 -2.60068208e-01 -2.92611327e-02
-1.20875514e+00 -1.64796531e-01 5.43256044e-01 3.06961328e-01
6.18097842e-01 -6.77673638e-01 -3.32284212e-01 4.43208218e-01
2.10014418e-01 2.32654348e-01 5.95881999e-01 -5.45759857e-01
-6.14026487e-02 9.74523723e-01 -8.55880007e-02 -8.72487009e-01
-9.63741958e-01 -9.01113689e-01 -1.38221776e+00 1.64663211e-01
1.11100662e+00 2.44745106e-01 -1.10146725e+00 1.42695820e+00
4.20243323e-01 1.47142008e-01 -2.20139131e-01 9.24840629e-01
5.77327847e-01 1.54377684e-01 1.14417389e-01 -2.98047066e-01
1.41721940e+00 -1.02334750e+00 -6.88467920e-01 -9.18916240e-02
1.15209579e+00 -6.72039330e-01 9.92908835e-01 3.19132388e-01
-8.71409893e-01 -2.87152112e-01 -1.06771815e+00 -1.63927332e-01
-5.36492229e-01 6.53331101e-01 6.52020991e-01 7.38097847e-01
-1.29142833e+00 4.16749150e-01 -1.01175392e+00 -5.89501500e-01
8.66644740e-01 5.35130918e-01 -4.21269327e-01 7.81040117e-02
-5.94321311e-01 7.06706941e-01 3.50566506e-01 2.48029158e-01
-5.93137681e-01 -7.89559126e-01 -5.48422277e-01 1.12505928e-01
1.36418164e-01 -6.22876525e-01 9.88778412e-01 -5.98152459e-01
-1.40104127e+00 1.19853616e+00 -1.38199285e-01 -2.60106981e-01
6.19982243e-01 4.57413107e-01 1.75011501e-01 2.04892665e-01
-2.93933284e-02 8.61588538e-01 2.28642821e-01 -8.57395053e-01
-8.88316810e-01 -4.11684394e-01 -1.97262511e-01 -1.72912464e-01
-3.53935987e-01 -5.87921023e-01 -6.56429231e-01 -4.25528765e-01
1.16283014e-01 -9.14885938e-01 -4.36619848e-01 3.62948835e-01
-1.33747131e-01 -1.38587609e-01 4.38396186e-01 -5.31895638e-01
1.09358704e+00 -2.32189393e+00 -5.98566756e-02 4.01504934e-02
2.87056804e-01 3.12525272e-01 1.78290904e-01 -1.83375031e-01
2.20612705e-01 2.35413074e-01 -2.77465850e-01 -4.94874358e-01
-1.31180659e-01 -1.43779248e-01 4.30151522e-01 9.89633799e-01
1.84277385e-01 9.77389634e-01 -9.50399876e-01 -8.20335627e-01
1.69699326e-01 5.31318069e-01 -3.51946592e-01 4.94700810e-03
-7.47716501e-02 4.98910129e-01 6.10238910e-02 1.00691092e+00
6.44003928e-01 -7.77716756e-01 2.26024300e-01 -2.66646236e-01
-6.49810433e-02 -5.63272797e-02 -8.50021839e-01 1.77368522e+00
-1.82537735e-01 5.72342932e-01 1.59389392e-01 -9.08204854e-01
4.97237742e-01 3.58823389e-01 6.03702128e-01 -5.35007298e-01
3.91480505e-01 5.18343747e-01 1.08899318e-01 -3.80996287e-01
1.42299503e-01 -1.32162362e-01 1.34310141e-01 2.90967703e-01
-9.17032454e-03 1.95962582e-02 5.51852405e-01 -2.10005373e-01
1.30141950e+00 -1.62161663e-01 4.48991418e-01 -2.79626548e-01
6.34973645e-01 4.98561189e-02 6.17526531e-01 5.07655799e-01
-5.00958443e-01 9.09583688e-01 6.91881418e-01 -4.43175882e-01
-1.03995049e+00 -6.93031609e-01 -3.44007343e-01 3.50322157e-01
1.53952390e-01 -3.13039869e-02 -6.95269585e-01 -7.14145005e-01
-2.60799736e-01 -8.17447454e-02 -9.30201352e-01 7.82808736e-02
-4.77634728e-01 -1.21683860e+00 6.97749913e-01 4.78145868e-01
4.35691595e-01 -7.10120261e-01 -5.43422699e-01 3.73881668e-01
-1.82087392e-01 -1.12039351e+00 -2.49347880e-01 3.75003904e-01
-8.22192430e-01 -1.28492033e+00 -1.06340468e+00 -1.01739538e+00
1.22448456e+00 5.01595400e-02 8.80372882e-01 3.95398855e-01
-6.47167206e-01 -3.57713223e-01 -1.09286591e-01 -1.84937432e-01
-1.89799488e-01 4.66376126e-01 -4.47683394e-01 -2.55040005e-02
4.89467472e-01 -3.48251134e-01 -9.63959694e-01 2.08915740e-01
-1.05075622e+00 3.32115948e-01 1.22969317e+00 1.13238418e+00
1.01416135e+00 1.95032015e-01 4.17562366e-01 -1.02992141e+00
-9.72547978e-02 -1.98233590e-01 -6.02748632e-01 9.30440575e-02
-5.32346904e-01 -2.63492018e-01 5.04544020e-01 -4.38654512e-01
-7.41374910e-01 3.96342397e-01 -2.36704081e-01 1.18327424e-01
-4.26048130e-01 5.38001418e-01 1.14597820e-01 -3.25127780e-01
6.38663530e-01 1.57670334e-01 1.63554087e-01 -7.05003813e-02
-2.83211380e-01 5.57146907e-01 5.12685001e-01 5.40613048e-02
4.81756568e-01 7.56351531e-01 4.11973447e-01 -6.38207674e-01
-8.67778122e-01 -7.80507386e-01 -6.00671172e-01 -2.30088219e-01
7.89133906e-01 -7.99118161e-01 -5.36313713e-01 8.13052177e-01
-9.75502372e-01 -4.72898453e-01 -1.54897287e-01 4.66853678e-01
-3.62143099e-01 4.23573643e-01 -1.02083552e+00 -3.45586628e-01
-3.76622587e-01 -1.12363386e+00 1.31428134e+00 5.42192817e-01
-1.97352767e-01 -8.67946744e-01 2.50211298e-01 5.50087392e-01
5.31824648e-01 4.11233544e-01 8.90231967e-01 -3.61924887e-01
-7.06588864e-01 -3.83169860e-01 -3.56479764e-01 -1.41859144e-01
3.14624041e-01 1.67516783e-01 -1.14679790e+00 -3.89393806e-01
-4.21462625e-01 -1.34026870e-01 9.73159432e-01 5.51664352e-01
1.35579586e+00 8.43401849e-02 -8.75307560e-01 6.96771979e-01
1.46568978e+00 -1.35040238e-01 6.17279887e-01 5.00576317e-01
3.44429195e-01 5.39029300e-01 4.25331950e-01 1.24783941e-01
2.89804459e-01 2.70953685e-01 3.34173024e-01 -6.16395533e-01
-3.93199325e-01 1.16962023e-01 -2.22893521e-01 5.91545641e-01
-1.27496812e-02 -2.54100740e-01 -1.00749922e+00 9.44393396e-01
-1.66731846e+00 -5.68579018e-01 -8.01366419e-02 1.92475736e+00
1.04335320e+00 3.31791908e-01 -4.82596830e-02 5.72808146e-01
5.69079936e-01 -1.68382555e-01 -3.11720163e-01 1.72259942e-01
-1.74142510e-01 1.29058570e-01 5.90472043e-01 2.48344660e-01
-1.26002622e+00 5.62823951e-01 6.10602951e+00 9.06725168e-01
-1.33143926e+00 1.21143587e-01 1.12228775e+00 -1.11325808e-01
1.06914572e-01 -2.37776175e-01 -8.59766006e-01 6.47306800e-01
6.04658723e-01 2.62430876e-01 -7.87408873e-02 3.47798854e-01
5.50316647e-02 -4.29611057e-01 -1.08096492e+00 9.00908887e-01
5.38332909e-02 -1.49733686e+00 -3.90927881e-01 3.41923952e-01
8.11065078e-01 -9.86970812e-02 5.33158705e-03 1.48038656e-01
-2.66741574e-01 -1.07345569e+00 1.49348285e-02 5.51374733e-01
8.37890327e-01 -5.13723969e-01 1.44871342e+00 3.94748658e-01
-9.35025394e-01 1.57900915e-01 -2.09782928e-01 1.87748000e-01
-9.12050977e-02 1.05406940e+00 -1.12163270e+00 1.37133986e-01
6.44718587e-01 4.65120584e-01 -8.87457907e-01 1.43621469e+00
5.91140427e-02 5.45046806e-01 -5.32423019e-01 -2.84558654e-01
4.28425819e-02 1.43917471e-01 -2.66505461e-02 1.38812912e+00
5.52562237e-01 -2.61752248e-01 5.20525910e-02 6.78424239e-01
-5.46394810e-02 1.07132934e-01 -1.31002456e-01 -1.20634258e-01
1.13536477e-01 1.76599503e+00 -1.54928160e+00 -2.53886450e-02
-2.93914884e-01 5.28961360e-01 3.85076106e-01 3.96573991e-02
-7.38506377e-01 -2.69243360e-01 4.24258672e-02 4.36237723e-01
1.33753434e-01 -1.14333361e-01 -5.43233097e-01 -8.39483440e-01
1.66752432e-02 -5.03636599e-01 3.53957683e-01 -1.46363497e-01
-1.12605250e+00 2.90282786e-01 -5.44296265e-01 -1.09138405e+00
2.26879567e-01 -6.45511031e-01 -4.81314182e-01 4.80388552e-01
-1.86057842e+00 -1.20105255e+00 -5.10119200e-01 1.37029812e-01
3.85709673e-01 3.54821593e-01 7.94961691e-01 5.48122346e-01
-5.85776031e-01 7.27138817e-01 -1.84993315e-02 4.81582526e-03
8.85032892e-01 -1.37667739e+00 -1.69611275e-01 6.61933005e-01
-2.43457317e-01 3.07021946e-01 4.90555763e-01 -3.52355897e-01
-1.09717500e+00 -9.94814336e-01 1.08181202e+00 -2.32658431e-01
5.95607698e-01 -3.03144336e-01 -8.24826241e-01 1.69524089e-01
-4.12119962e-02 3.44227403e-01 9.44420874e-01 -1.34361312e-01
2.57602274e-01 -2.42734939e-01 -1.14764130e+00 6.57146633e-01
5.49981236e-01 -3.04898649e-01 1.40140772e-01 3.20265114e-01
1.66248545e-01 -5.78685880e-01 -9.05214846e-01 5.59897602e-01
4.72702265e-01 -8.43753040e-01 4.80185390e-01 4.42749918e-01
2.32131511e-01 -7.97328055e-01 4.27946061e-01 -9.27862048e-01
-5.49403667e-01 -5.91035224e-02 1.00626824e-02 1.21873116e+00
6.34764135e-01 -4.31141526e-01 1.32323289e+00 7.10019618e-02
-1.42528579e-01 -9.85060275e-01 -1.19577146e+00 -2.57595599e-01
-4.67471667e-02 2.58474946e-01 2.57449687e-01 6.30132675e-01
1.58374012e-01 1.48704067e-01 1.62383080e-01 1.42765895e-01
4.78214920e-01 -5.60648143e-02 5.98950028e-01 -1.15243161e+00
-2.47604504e-01 -8.74520898e-01 -6.82676256e-01 -7.02769160e-01
-1.32897690e-01 -7.55333543e-01 1.59189358e-01 -1.45613158e+00
5.00740826e-01 -4.21283990e-01 -4.98623282e-01 5.01017272e-01
-3.30717891e-01 8.57029021e-01 -4.19424504e-01 2.47151047e-01
-6.92987800e-01 1.01385430e-01 1.39996147e+00 -5.52396894e-01
5.83817214e-02 -2.39002392e-01 -6.04382575e-01 7.24201202e-01
8.66967261e-01 -4.50272560e-01 -6.13007657e-02 -1.39267653e-01
3.34637016e-01 -3.48578274e-01 3.06605875e-01 -1.33492291e+00
6.64945722e-01 1.23733483e-01 8.44581783e-01 -7.95678794e-01
2.24780917e-01 -7.83688247e-01 2.55203515e-01 7.04842627e-01
-1.40979275e-01 -4.48350579e-01 1.73911095e-01 5.08371830e-01
-2.18977824e-01 -1.54660538e-01 8.21480632e-01 -3.75450891e-03
-7.55577758e-02 3.07169288e-01 -5.15183866e-01 -4.19885844e-01
1.26911116e+00 -4.28808600e-01 -6.30205154e-01 3.12854022e-01
-3.86136234e-01 1.02936603e-01 7.77445316e-01 -1.46876514e-01
9.32013318e-02 -9.91209447e-01 -4.86482590e-01 1.04031809e-01
2.26438940e-01 5.23060322e-01 4.01224107e-01 1.44892645e+00
-9.32892382e-01 5.74509919e-01 -1.36866763e-01 -1.05694401e+00
-1.16819417e+00 5.29284596e-01 4.10766572e-01 -8.96798849e-01
-4.64452326e-01 7.81903207e-01 1.97469085e-01 -2.00900927e-01
3.41060072e-01 -4.97428179e-01 -5.27101457e-01 1.00713447e-02
4.26341921e-01 1.26189947e-01 4.35342014e-01 -3.06244314e-01
-3.63434225e-01 5.99448800e-01 -3.71701509e-01 3.39899302e-01
1.06978905e+00 -4.01706398e-02 -2.35138983e-01 2.47526228e-01
9.03716207e-01 -2.22281724e-01 -1.38935471e+00 -9.24601853e-02
1.21467300e-01 -2.55968243e-01 1.60108387e-01 -7.90594816e-01
-1.19099593e+00 7.47960389e-01 8.62684369e-01 8.66752192e-02
1.20094335e+00 -2.40384638e-02 6.59804404e-01 1.39339000e-01
3.00007999e-01 -1.07856417e+00 -1.15852661e-01 1.88653037e-01
1.47196472e-01 -1.37118876e+00 -4.08840962e-02 -4.99343246e-01
6.80337101e-02 1.10241306e+00 6.55761182e-01 1.82374850e-01
5.12030542e-01 5.55072367e-01 2.80963510e-01 -5.58170788e-02
-6.25916779e-01 -1.19530320e-01 -1.00433059e-01 4.05962110e-01
7.01390862e-01 -5.35946004e-02 -6.21707797e-01 6.60360098e-01
1.05058506e-01 2.80954242e-01 4.58350360e-01 9.80576396e-01
-2.98516899e-01 -9.91225302e-01 -2.05346927e-01 5.17830253e-01
-9.00381744e-01 2.19427049e-01 -3.27194214e-01 8.38604987e-01
2.84058839e-01 6.80410504e-01 1.08148009e-01 6.48379922e-02
-6.19967431e-02 -3.99548650e-01 4.65140045e-01 -3.78574252e-01
-5.24800241e-01 2.72442758e-01 -3.03429753e-01 -2.12667957e-01
-6.35282695e-01 -5.22616386e-01 -1.16012204e+00 -1.07649766e-01
-4.43036020e-01 1.41793452e-02 5.29357076e-01 8.93197358e-01
2.12391779e-01 6.72282517e-01 4.10134614e-01 -6.96883380e-01
-1.27602875e-01 -9.80960131e-01 -7.10144222e-01 1.90513819e-01
3.42804044e-01 -5.19238174e-01 -5.38486421e-01 2.85801798e-01] | [14.960721015930176, -3.107600212097168] |
8ae1c731-6abb-456f-b164-b2e705ea0a64 | hierarchically-supervised-latent-dirichlet | null | null | http://papers.nips.cc/paper/4313-hierarchically-supervised-latent-dirichlet-allocation | http://papers.nips.cc/paper/4313-hierarchically-supervised-latent-dirichlet-allocation.pdf | Hierarchically Supervised Latent Dirichlet Allocation | We introduce hierarchically supervised latent Dirichlet allocation (HSLDA), a model for hierarchically and multiply labeled bag-of-word data. Examples of such data include web pages and their placement in directories, product descriptions and associated categories from product hierarchies, and free-text clinical records and their assigned diagnosis codes. Out-of-sample label prediction is the primary goal of this work, but improved lower-dimensional representations of the bag-of-word data are also of interest. We demonstrate HSLDA on large-scale data from clinical document labeling and retail product categorization tasks. We show that leveraging the structure from hierarchical labels improves out-of-sample label prediction substantially when compared to models that do not. | ['Adler J. Perotte', 'Frank Wood', 'Noemie Elhadad', 'Nicholas Bartlett'] | 2011-12-01 | null | null | null | neurips-2011-12 | ['product-categorization'] | ['miscellaneous'] | [-9.10821036e-02 2.19459206e-01 -8.34406257e-01 -8.55098724e-01
-9.47195351e-01 -5.28358102e-01 3.66161674e-01 8.96021366e-01
-1.03788495e-01 3.32916737e-01 7.99031734e-01 -2.47106835e-01
-2.69503981e-01 -6.96242034e-01 -8.18733796e-02 -7.23184824e-01
-2.92982697e-01 1.15326428e+00 -2.42945462e-01 3.27058643e-01
2.33279858e-02 1.94837779e-01 -1.03465176e+00 7.01494515e-01
2.05519468e-01 8.54081333e-01 -2.51996905e-01 2.19455689e-01
-3.16212475e-01 8.59989643e-01 -1.65379737e-02 -2.10931748e-01
1.08047903e-01 -1.03758134e-01 -6.12468362e-01 6.33619964e-01
2.32718259e-01 -4.10375148e-01 -8.62017944e-02 7.80369401e-01
2.99953520e-01 1.18162828e-02 1.35476625e+00 -1.26532185e+00
-7.55593777e-01 6.45615041e-01 -7.51574278e-01 -3.13017182e-02
2.41143003e-01 -2.68351436e-01 1.58816862e+00 -8.66909802e-01
6.46005571e-01 1.33075476e+00 6.94659650e-01 1.86744079e-01
-1.63501060e+00 -6.33660913e-01 7.53053576e-02 -2.12512076e-01
-1.38419795e+00 -2.32719898e-01 6.01314127e-01 -1.25959551e+00
8.92428815e-01 -1.53219402e-01 3.35836232e-01 9.71270144e-01
2.67058015e-01 8.02623987e-01 9.59727943e-01 -4.13798988e-01
4.82730836e-01 2.82384723e-01 8.11297059e-01 7.81247973e-01
2.98884898e-01 -3.37975174e-01 -2.68966824e-01 -1.01250982e+00
4.97381181e-01 5.42058825e-01 3.83317530e-01 -9.47423697e-01
-7.87680924e-01 1.46341527e+00 9.03334916e-02 5.76237366e-02
-5.86779833e-01 -8.72457251e-02 4.84499186e-01 -2.89331466e-01
1.03847361e+00 1.22791529e-01 -5.36505520e-01 2.70648450e-02
-8.75512362e-01 1.75441235e-01 8.63634348e-01 1.23795450e+00
6.26519799e-01 -4.12883431e-01 -2.17616752e-01 1.21625495e+00
7.06966877e-01 2.80009136e-02 7.21488953e-01 -6.63303256e-01
2.41739020e-01 8.04212451e-01 1.37693688e-01 -8.72467041e-01
-8.38282883e-01 -2.82298654e-01 -6.41182899e-01 -4.41780239e-01
1.34150624e-01 1.09132476e-01 -1.11979115e+00 1.33848393e+00
3.46198976e-01 -4.33093518e-01 -1.27182484e-01 6.00746632e-01
7.90163875e-01 4.02255833e-01 7.17149615e-01 -4.78055120e-01
1.77810454e+00 -7.12136090e-01 -9.09095824e-01 -6.15323558e-02
1.19635236e+00 -4.14771944e-01 7.25639939e-01 3.00022960e-01
-7.26853788e-01 -2.15978429e-01 -6.17122054e-01 -3.09682757e-01
-4.20941412e-01 -1.59705997e-01 9.75308955e-01 6.23531818e-01
-6.61050975e-01 2.31703296e-01 -9.70185041e-01 -4.42874402e-01
7.64214277e-01 1.53850585e-01 -2.01290786e-01 -4.04859841e-01
-9.89948571e-01 3.89094323e-01 2.15568408e-01 -4.93548006e-01
-7.49910355e-01 -6.26503050e-01 -1.14489555e+00 2.00886503e-01
3.94495279e-02 -4.99602318e-01 1.39071941e+00 -1.79519579e-01
-6.54125333e-01 1.10022187e+00 -1.03052549e-01 -2.88672864e-01
-1.14090100e-01 2.53236473e-01 -3.76723975e-01 -4.57320958e-02
3.95159453e-01 8.04651380e-01 6.83294237e-01 -1.02135074e+00
-6.98957205e-01 -7.56532192e-01 -3.85497779e-01 8.69256482e-02
-5.93194246e-01 -3.68216671e-02 -8.85936245e-02 -5.94174385e-01
3.31641048e-01 -1.09784067e+00 -4.10380214e-01 -3.50066692e-01
-4.86637622e-01 -8.75957251e-01 2.94253320e-01 -6.86729968e-01
1.31068659e+00 -2.23585796e+00 -1.99244037e-01 8.32406506e-02
6.79670095e-01 -6.49650931e-01 7.00171292e-02 4.68491226e-01
-2.01427668e-01 1.94198713e-01 2.19160363e-01 -6.27000749e-01
3.47025096e-01 1.57429263e-01 -1.69792652e-01 5.46043575e-01
-2.24641159e-01 7.42597997e-01 -8.69453788e-01 -8.18877280e-01
-3.31105143e-01 1.52090460e-01 -7.47273743e-01 -6.65125996e-02
-2.41788715e-01 5.88663444e-02 -5.30659318e-01 7.04604685e-01
1.52163148e-01 -7.28769422e-01 3.82888019e-01 -3.39771241e-01
3.17165554e-01 4.74532068e-01 -5.91971695e-01 1.49756598e+00
-1.25186548e-01 1.90958530e-01 -4.12349939e-01 -6.66455805e-01
7.24836707e-01 2.80325264e-01 1.20213926e+00 -3.18930566e-01
9.79991332e-02 -2.49269679e-01 -1.67933583e-01 -3.51425320e-01
2.80564606e-01 -4.68075603e-01 -6.05634272e-01 8.14728022e-01
-9.41476822e-02 1.54452071e-01 9.61392522e-02 4.66800660e-01
1.05213201e+00 -3.59324843e-01 7.15839326e-01 -2.03121528e-01
-2.77978420e-01 3.99398416e-01 5.48586071e-01 5.14099061e-01
-1.92065015e-01 4.45017308e-01 7.64084518e-01 -5.19884884e-01
-1.28080380e+00 -9.04714763e-01 -6.55356109e-01 1.63550401e+00
-4.99220133e-01 -6.20483398e-01 -3.67498100e-01 -9.61197138e-01
5.75831771e-01 7.83324063e-01 -6.66146874e-01 9.44241360e-02
4.95095663e-02 -9.06870544e-01 2.54520737e-02 6.04057074e-01
-3.34120840e-01 -7.93513775e-01 -9.27338153e-02 3.32121551e-01
9.18804929e-02 -9.60537434e-01 -5.21310031e-01 5.79299688e-01
-8.48916173e-01 -1.08007491e+00 -6.29344344e-01 -1.06371927e+00
8.53381395e-01 -4.94555905e-02 1.33208466e+00 -5.52939355e-01
-4.14062887e-01 4.75016415e-01 -4.59766895e-01 -3.30256402e-01
-5.08962512e-01 8.75325203e-02 2.80381799e-01 -2.11586095e-02
1.08886421e+00 -2.21709043e-01 -5.27143955e-01 3.77107859e-01
-8.90614569e-01 -5.37935272e-02 2.58143604e-01 9.06814575e-01
7.80564845e-01 9.79816020e-02 6.03744447e-01 -1.42055964e+00
6.98342085e-01 -8.41717601e-01 -4.62043911e-01 4.70129438e-02
-9.41299856e-01 3.16128740e-03 -5.56496717e-02 -4.07678515e-01
-6.54068291e-01 3.11752647e-01 1.11424319e-01 -9.85919461e-02
-3.37680012e-01 5.74880123e-01 1.02145076e-01 6.95939183e-01
6.47372305e-01 -2.98092693e-01 -1.84196010e-02 -9.04096961e-01
7.18642890e-01 1.47174621e+00 -1.49177730e-01 -1.34717003e-01
5.41378669e-02 4.15918529e-01 -2.14541882e-01 -4.27842468e-01
-1.19418335e+00 -1.19720483e+00 -8.37648213e-01 4.53961968e-01
1.07159495e+00 -1.23228228e+00 -6.68927312e-01 -8.38095769e-02
-7.13695705e-01 -2.44629696e-01 -3.98408622e-01 5.97203135e-01
-5.60796142e-01 2.68552959e-01 -1.22995436e+00 -6.74954534e-01
-1.99507430e-01 -1.01715004e+00 1.31715953e+00 -5.22188783e-01
-7.21616209e-01 -1.19289851e+00 3.13941061e-01 7.21382678e-01
-4.41038758e-02 -8.52400362e-02 1.72865713e+00 -1.35666311e+00
4.50884998e-02 -6.14301860e-01 -2.99410298e-02 1.59776956e-01
5.52599549e-01 -6.96891129e-01 -7.25129247e-01 -4.92484629e-01
1.45432234e-01 -4.42617267e-01 8.13476682e-01 7.86485255e-01
1.00371099e+00 -3.89871597e-01 -7.43169844e-01 1.15336388e-01
1.24194801e+00 2.92134821e-01 1.02895446e-01 -3.39191630e-02
8.98512006e-01 1.05807447e+00 4.76983756e-01 9.16557789e-01
5.40757358e-01 9.52092648e-01 -4.88257818e-02 -1.74280167e-01
1.82485029e-01 -2.28455275e-01 3.62276323e-02 8.21606100e-01
9.87075388e-01 -6.87655509e-02 -1.32892156e+00 6.95836365e-01
-1.78997302e+00 -4.83344138e-01 -3.40863824e-01 1.75893426e+00
1.00909400e+00 3.64682674e-02 5.75218916e-01 -1.33605540e-01
8.45513940e-01 -1.69107720e-01 -7.85352111e-01 -2.86052406e-01
3.81500542e-01 -4.58973736e-01 7.53143847e-01 2.97889501e-01
-1.50358820e+00 6.68843269e-01 7.05336905e+00 8.06516349e-01
-2.36952767e-01 5.10817647e-01 1.00779235e+00 -1.40672028e-01
-3.65691595e-02 -2.53517687e-01 -1.24629223e+00 5.47721624e-01
1.04017580e+00 1.01157874e-01 -9.14128199e-02 1.26769555e+00
2.52077878e-01 -5.17239980e-02 -1.53458142e+00 1.10488749e+00
3.26587796e-01 -9.53604043e-01 3.03624440e-02 8.19342792e-01
9.38715041e-01 -1.64216325e-01 3.31514120e-01 4.28206027e-01
9.70777869e-01 -7.93149829e-01 1.42929822e-01 -3.15682031e-02
9.18814301e-01 -6.13535881e-01 7.22880721e-01 3.41416240e-01
-6.59618795e-01 -3.49108726e-01 -4.44518894e-01 1.43358842e-01
3.56395602e-01 9.67313170e-01 -1.27094901e+00 -2.60066122e-01
4.37195599e-01 8.85868669e-01 -5.03363907e-01 9.73321438e-01
1.97141096e-01 7.77807355e-01 -5.67533039e-02 1.27611220e-01
3.02270323e-01 -1.56504996e-02 -2.29472145e-01 1.17691624e+00
4.40362804e-02 2.00700939e-01 6.34641051e-01 5.08633137e-01
-1.38361543e-01 3.90665293e-01 -7.01957881e-01 -4.11534488e-01
3.34126115e-01 1.16569686e+00 -8.77030969e-01 -5.96886694e-01
-5.53335965e-01 5.36362529e-01 8.29304010e-02 1.09116018e-01
-5.09827495e-01 6.98997527e-02 6.72393799e-01 4.84797925e-01
3.86829406e-01 -2.34949946e-01 -3.70939672e-01 -8.77440929e-01
-6.69553876e-01 -6.00966573e-01 7.92480230e-01 -6.49691463e-01
-2.00917959e+00 2.42517486e-01 2.36251384e-01 -1.24606276e+00
-5.23504436e-01 -4.15194154e-01 5.89134037e-01 5.13023615e-01
-1.03479707e+00 -1.21730113e+00 1.97757222e-02 2.82303184e-01
6.03166461e-01 -3.92612129e-01 1.08664775e+00 3.83194983e-01
-3.50506455e-01 3.43445837e-01 6.99517071e-01 4.42214340e-01
8.57082188e-01 -1.44683552e+00 4.17995781e-01 -1.71093330e-01
1.97411090e-01 5.61792850e-01 3.98298174e-01 -1.10509706e+00
-8.82738173e-01 -1.09683919e+00 9.59873199e-01 -7.90513992e-01
9.15174484e-01 -7.20549047e-01 -7.27501094e-01 8.44107926e-01
-1.35123789e-01 -2.16518462e-01 1.63838458e+00 6.81725204e-01
-5.70928216e-01 1.92058891e-01 -1.15261340e+00 1.77236468e-01
7.33374953e-01 -6.98027492e-01 -5.20519674e-01 1.27267110e+00
8.81820142e-01 1.07285334e-02 -1.05183053e+00 -1.17936283e-01
4.63246405e-01 -3.41865182e-01 8.57459426e-01 -1.11377811e+00
4.38676506e-01 3.24595451e-01 -2.56786942e-01 -1.30206168e+00
-9.01946604e-01 6.04070723e-02 -6.70544291e-03 1.11281574e+00
3.83515894e-01 -2.04406664e-01 1.11366642e+00 1.06049049e+00
7.55788609e-02 -7.12817609e-01 -4.39755648e-01 -6.06320083e-01
1.45278201e-01 -3.63878936e-01 2.53068328e-01 1.16892886e+00
5.83035588e-01 7.33380437e-01 -2.11669475e-01 -2.17323810e-01
9.62851465e-01 2.21514419e-01 2.30490535e-01 -1.55414653e+00
-2.38474295e-01 -7.43539780e-02 -5.09676695e-01 -9.68211412e-01
3.21423978e-01 -1.16503417e+00 3.33496965e-02 -1.82280838e+00
6.98150873e-01 -7.96696842e-01 -2.35713720e-01 9.35306191e-01
2.64607400e-01 3.46909672e-01 -1.26095161e-01 7.14137018e-01
-9.08649921e-01 2.01992631e-01 6.59884095e-01 -3.58907491e-01
-3.14978868e-01 -7.72102326e-02 -7.39424706e-01 6.27715766e-01
4.82570767e-01 -9.75737572e-01 -3.67497146e-01 -6.36778399e-02
7.84699768e-02 9.57852975e-02 -2.66330600e-01 -2.14760303e-01
1.36377901e-01 5.73411472e-02 3.40413779e-01 -7.18512177e-01
3.02179784e-01 -8.95174742e-01 1.96698844e-01 2.78107911e-01
-1.07447410e+00 -1.13932982e-01 -2.64008731e-01 9.47939336e-01
9.54516989e-04 -1.77102268e-01 5.16856253e-01 -1.51004612e-01
-2.62264609e-01 1.97391093e-01 -6.54271007e-01 -1.16570637e-01
9.31187391e-01 3.03784329e-02 -1.87735960e-01 -3.73569369e-01
-1.44501698e+00 2.07896605e-01 3.23023617e-01 4.27876383e-01
2.58144617e-01 -1.62858081e+00 -5.84410489e-01 2.35278741e-01
5.38876951e-01 -1.01562366e-01 8.42477456e-02 6.11181796e-01
-1.26018956e-01 9.38287437e-01 2.19020903e-01 -6.38652205e-01
-1.22345734e+00 1.00617039e+00 -2.54282683e-01 -6.09474659e-01
-5.39798081e-01 4.88806903e-01 7.50288844e-01 -3.01602989e-01
5.43771029e-01 -2.92110264e-01 -5.23519695e-01 7.64379621e-01
3.34887266e-01 1.36837766e-01 1.31168306e-01 -5.88324189e-01
-2.43056118e-01 1.82518795e-01 -6.19232535e-01 -4.07307083e-03
1.53628731e+00 -2.71608174e-01 -1.81429639e-01 7.07338214e-01
1.57311547e+00 -4.88441825e-01 -1.08393741e+00 -5.38162649e-01
3.54733020e-01 -2.81401664e-01 1.19163617e-01 -6.62800968e-01
-5.41874051e-01 7.57442236e-01 7.04555035e-01 4.87249553e-01
5.57359576e-01 5.25884688e-01 6.32280886e-01 1.19875863e-01
5.36157489e-01 -8.87113273e-01 1.65592089e-01 9.54576433e-02
4.05352861e-01 -1.28044724e+00 2.23450571e-01 -4.81291473e-01
-9.32075024e-01 5.32836020e-01 5.42061636e-03 2.98769951e-01
1.14654255e+00 1.50313750e-01 -1.24644101e-01 -6.30414665e-01
-8.04861248e-01 -1.53037876e-01 2.32505560e-01 6.07543051e-01
6.53008103e-01 3.17467511e-01 -1.99398950e-01 9.59590435e-01
1.50600180e-01 -2.77208328e-01 3.58042002e-01 9.32763696e-01
-4.15349275e-01 -1.05268037e+00 -3.69732648e-01 1.16191053e+00
-5.73723495e-01 -3.31620932e-01 -1.61169142e-01 1.92363963e-01
-9.09529999e-02 1.08474851e+00 3.02688330e-01 -1.65921837e-01
4.14826050e-02 4.86013800e-01 1.49621079e-02 -1.37514889e+00
-3.09294432e-01 6.12620115e-01 2.31177598e-01 -2.43522808e-01
-1.87050983e-01 -9.30217445e-01 -1.14613307e+00 6.98884651e-02
-2.30439529e-01 3.02483201e-01 7.82103658e-01 7.59443164e-01
5.34063637e-01 1.59655124e-01 6.21036172e-01 -5.92984200e-01
-6.57546997e-01 -1.06578076e+00 -1.24394655e+00 8.02375913e-01
2.51906395e-01 -5.67013741e-01 -4.75081176e-01 4.77674276e-01] | [9.52619743347168, 4.393645286560059] |
0028681e-51b0-4bd9-b838-7729aa03e471 | enforcing-reasoning-in-visual-commonsense | 1910.11124 | null | https://arxiv.org/abs/1910.11124v2 | https://arxiv.org/pdf/1910.11124v2.pdf | Enforcing Reasoning in Visual Commonsense Reasoning | The task of Visual Commonsense Reasoning is extremely challenging in the sense that the model has to not only be able to answer a question given an image, but also be able to learn to reason. The baselines introduced in this task are quite limiting because two networks are trained for predicting answers and rationales separately. Question and image is used as input to train answer prediction network while question, image and correct answer are used as input in the rationale prediction network. As rationale is conditioned on the correct answer, it is based on the assumption that we can solve Visual Question Answering task without any error - which is over ambitious. Moreover, such an approach makes both answer and rationale prediction two completely independent VQA tasks rendering cognition task meaningless. In this paper, we seek to address these issues by proposing an end-to-end trainable model which considers both answers and their reasons jointly. Specifically, we first predict the answer for the question and then use the chosen answer to predict the rationale. However, a trivial design of such a model becomes non-differentiable which makes it difficult to train. We solve this issue by proposing four approaches - softmax, gumbel-softmax, reinforcement learning based sampling and direct cross entropy against all pairs of answers and rationales. We demonstrate through experiments that our model performs competitively against current state-of-the-art. We conclude with an analysis of presented approaches and discuss avenues for further work. | ['Hammad A. Ayyubi', 'Md. Mehrab Tanjim', 'David J. Kriegman'] | 2019-10-21 | null | null | null | null | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 3.68905395e-01 4.83622342e-01 1.32138461e-01 -5.97216487e-01
-9.31416392e-01 -7.47568369e-01 7.45998859e-01 4.49179821e-02
-4.57512885e-01 6.82909429e-01 2.20581442e-01 -5.45397699e-01
1.06297277e-01 -6.77054584e-01 -7.11165369e-01 -3.60085964e-01
5.55752754e-01 4.07671332e-01 2.44029179e-01 -1.17950805e-01
5.25709212e-01 1.91738844e-01 -1.52805793e+00 7.14931130e-01
7.75097311e-01 1.18080544e+00 4.77730811e-01 8.62080932e-01
-3.82472813e-01 1.68742621e+00 -6.63784146e-01 -8.20072711e-01
1.00013457e-01 -8.46949220e-01 -1.23304021e+00 -1.48685202e-01
6.65527225e-01 -4.14916307e-01 -1.26236295e-02 1.03189898e+00
3.14343810e-01 9.84057784e-02 1.06327379e+00 -1.23569572e+00
-1.11445880e+00 3.06705177e-01 -2.47991577e-01 2.26117000e-01
4.74692762e-01 3.95502716e-01 1.36856294e+00 -8.37872148e-01
6.90683663e-01 1.20353627e+00 4.18799877e-01 9.48768437e-01
-1.12559092e+00 -2.21329600e-01 3.08997840e-01 7.01321423e-01
-8.27661991e-01 -3.94142479e-01 9.99273002e-01 -4.40023482e-01
9.01312232e-01 3.30148816e-01 5.67972481e-01 1.17892158e+00
-5.80191463e-02 1.14984989e+00 1.21862018e+00 -4.34972614e-01
3.99984866e-01 2.70375699e-01 6.25691339e-02 7.30996490e-01
-3.84039432e-01 -1.53606564e-01 -2.08600968e-01 6.06835745e-02
6.07549608e-01 -2.52644449e-01 -3.46368432e-01 -4.81547564e-01
-9.79406714e-01 1.14309013e+00 8.33193481e-01 2.69346267e-01
-6.46745086e-01 4.73808587e-01 1.90114453e-01 4.49108154e-01
-2.45288312e-02 4.39913720e-01 -3.57130170e-01 2.56125107e-02
-9.44473326e-01 3.45161855e-01 7.75717735e-01 4.05731797e-01
5.95919490e-01 8.67317989e-03 -5.91669858e-01 6.78599894e-01
4.68845010e-01 2.54546672e-01 3.39033723e-01 -1.12405109e+00
4.19598192e-01 5.46177447e-01 7.44325295e-02 -1.02518904e+00
-3.37941259e-01 -2.12945700e-01 -6.09755039e-01 5.03702581e-01
7.70700157e-01 3.50635387e-02 -9.94326055e-01 1.92853165e+00
2.20940292e-01 -2.11183995e-01 2.12135732e-01 1.21385717e+00
1.31968343e+00 5.56663156e-01 3.64548087e-01 -1.44072741e-01
1.57693267e+00 -1.04516816e+00 -7.46810853e-01 -5.20867765e-01
2.27312490e-01 -6.41541779e-01 1.51002741e+00 1.33803084e-01
-1.29726028e+00 -6.00119114e-01 -8.44736099e-01 -5.33224046e-01
-3.50106984e-01 1.84050158e-01 3.75042200e-01 3.14275473e-01
-9.98038471e-01 2.81276256e-01 -1.88818678e-01 -1.39856800e-01
5.02549529e-01 9.10961106e-02 -1.05438612e-01 3.96659002e-02
-1.24521923e+00 1.33498228e+00 3.37860525e-01 1.98002532e-03
-6.99076951e-01 -4.52911586e-01 -8.09774935e-01 3.53464276e-01
4.53673363e-01 -1.04585958e+00 1.45166838e+00 -1.42445314e+00
-1.50492418e+00 1.18443739e+00 -2.01500967e-01 -6.46331370e-01
6.95665658e-01 -4.00825543e-03 -1.74653128e-01 4.92298037e-01
1.44915029e-01 9.31410968e-01 1.00689554e+00 -1.52496064e+00
-4.52240944e-01 -2.54083425e-01 5.80580235e-01 1.31336197e-01
1.12610243e-01 -1.98055759e-01 -4.02133971e-01 -4.27477270e-01
-3.67572624e-03 -6.05843186e-01 -1.45063298e-02 1.33174405e-01
-3.70677441e-01 -3.27953607e-01 6.59818470e-01 -8.44102681e-01
9.20954466e-01 -2.04094887e+00 1.20792024e-01 -1.06303349e-01
2.47442216e-01 1.38783216e-01 -1.53179809e-01 1.90203980e-01
-1.10445812e-01 3.92404422e-02 -2.26129800e-01 -2.10478216e-01
3.36487859e-01 1.90582767e-01 -5.36967278e-01 2.94583768e-01
4.38086867e-01 1.29367661e+00 -7.70405173e-01 -6.90625608e-01
2.27684960e-01 3.75456512e-01 -4.82392728e-01 5.48325658e-01
-6.82722628e-01 2.41062522e-01 -3.68429035e-01 2.64117509e-01
5.01403153e-01 -5.48917592e-01 -2.78642820e-03 -2.32087761e-01
1.68959081e-01 2.59772539e-01 -9.54643250e-01 1.38930237e+00
-4.76475656e-01 6.97770357e-01 -9.08664465e-02 -1.18904936e+00
8.51681411e-01 3.34900796e-01 -1.88539013e-01 -1.06934178e+00
3.02742213e-01 1.64832085e-01 -1.28854007e-01 -8.89550686e-01
2.14567408e-01 -6.63451135e-01 -1.90828852e-02 3.76465768e-01
6.55063475e-03 -4.22561169e-01 1.37852669e-01 1.27385542e-01
7.97261477e-01 3.54099274e-01 4.66941178e-01 9.22112837e-02
9.58873451e-01 2.11580899e-02 2.20163569e-01 7.80912817e-01
-4.44531351e-01 7.29714751e-01 7.95217991e-01 -3.73540074e-01
-8.14398646e-01 -1.21246314e+00 4.27582800e-01 9.27243710e-01
1.71895444e-01 1.14742108e-01 -7.29654849e-01 -1.10565507e+00
-1.87101841e-01 1.28373420e+00 -9.50717807e-01 3.60659063e-02
-5.58146119e-01 -8.82462636e-02 1.10689953e-01 4.89578664e-01
6.83428884e-01 -1.51023924e+00 -1.03253424e+00 -6.41371310e-02
-3.99398595e-01 -1.00775433e+00 -1.50985718e-01 2.33476073e-01
-5.59897602e-01 -1.24223578e+00 -6.64868057e-01 -7.86899388e-01
4.10813361e-01 -2.16932371e-02 1.42077851e+00 2.10928008e-01
6.66518062e-02 6.90829992e-01 -9.93893966e-02 -2.85786033e-01
-4.04951572e-01 -3.20802569e-01 -7.59069502e-01 1.43813401e-01
4.24335897e-01 -3.15955758e-01 -8.78220916e-01 -3.08413506e-01
-8.12596023e-01 1.46682933e-01 5.63769042e-01 8.16846192e-01
5.30171990e-01 -3.96668851e-01 7.61948228e-01 -7.42949903e-01
8.44145775e-01 -3.88429165e-01 -3.02350849e-01 6.59075797e-01
-2.36444846e-01 2.88275033e-01 8.01932812e-01 -1.75752088e-01
-1.36106133e+00 1.82430699e-01 -4.15069014e-01 -3.80538583e-01
-2.77675807e-01 3.25533807e-01 -7.56267309e-02 3.16126943e-01
6.76109731e-01 1.65737614e-01 -4.50535398e-03 -1.72856838e-01
6.90153360e-01 4.22407061e-01 7.42225111e-01 -2.68749326e-01
7.19668150e-01 3.57844740e-01 -1.37862355e-01 -6.31236017e-01
-1.24466980e+00 -3.54288936e-01 -3.58831257e-01 -3.96299899e-01
1.34137046e+00 -6.92905307e-01 -1.05992341e+00 -2.35614002e-01
-1.56696200e+00 -2.43485183e-01 -4.30964291e-01 2.30989024e-01
-9.67310965e-01 2.91767031e-01 -3.18405300e-01 -1.05431914e+00
-3.79145801e-01 -1.11302650e+00 7.70746827e-01 3.23919743e-01
-4.18420583e-01 -1.13590133e+00 -1.47049010e-01 8.02353263e-01
3.49082500e-01 3.42455924e-01 1.17733049e+00 -7.37483501e-01
-6.18377566e-01 5.41319102e-02 -7.04628766e-01 3.96599263e-01
-2.48361126e-01 -1.81856155e-01 -1.20589614e+00 1.90344274e-01
4.51361328e-01 -8.14307690e-01 1.05063462e+00 3.94317836e-01
1.10850251e+00 -2.08622366e-01 1.85857400e-01 1.48773044e-01
1.58925879e+00 -7.45143835e-03 6.03847861e-01 4.28582668e-01
4.80500937e-01 8.83838356e-01 4.60022330e-01 1.79277554e-01
8.34778130e-01 4.85688716e-01 9.36736107e-01 -6.58215284e-02
-3.80698025e-01 -1.40252486e-01 8.47504735e-02 1.38533294e-01
3.72607335e-02 -2.20784247e-01 -8.03621173e-01 5.96189320e-01
-1.99870384e+00 -1.42145765e+00 -4.02869940e-01 1.90836048e+00
8.92582536e-01 -6.57943636e-02 7.37761632e-02 9.53025818e-02
3.46733809e-01 5.31027876e-02 -5.00807285e-01 -6.24947011e-01
-2.20986437e-02 1.33676514e-01 -1.18293971e-01 6.19863868e-01
-1.06566036e+00 8.89397800e-01 5.84854984e+00 5.39884090e-01
-1.03699172e+00 1.69985339e-01 7.25331962e-01 4.69145924e-02
-6.50098264e-01 1.06084846e-01 -1.97669506e-01 1.23947352e-01
6.61485910e-01 3.61150503e-01 4.35842931e-01 7.69238472e-01
6.79247873e-03 -3.58040661e-01 -1.10335243e+00 1.09283948e+00
2.61154681e-01 -1.34976244e+00 1.78123321e-02 -5.58079064e-01
4.28633124e-01 -2.84943461e-01 1.45035923e-01 5.84135175e-01
2.50433028e-01 -1.13354862e+00 1.12093186e+00 6.38941526e-01
4.28199679e-01 -5.95423460e-01 5.19280016e-01 4.15433735e-01
-7.70329058e-01 -2.74460614e-01 -2.17138916e-01 -2.57624745e-01
1.08303659e-01 2.48721883e-01 -7.60050118e-01 3.10897410e-01
5.72578847e-01 1.59573808e-01 -7.58594751e-01 8.00624847e-01
-8.55721831e-01 3.31320971e-01 4.53481711e-02 -2.85717010e-01
4.86922652e-01 2.59871215e-01 1.19907185e-01 1.03493941e+00
5.21849422e-03 1.23539269e-01 1.77744199e-02 1.28305233e+00
-4.29796316e-02 1.63994670e-01 -2.68431306e-01 1.80125624e-01
3.48035730e-02 1.08490264e+00 -7.24808991e-01 -3.55459303e-01
-5.26890576e-01 1.14427674e+00 7.83595741e-01 5.37192881e-01
-9.64325905e-01 -1.71555921e-01 1.82697281e-01 -1.52374536e-01
4.44058031e-01 1.82229578e-01 -4.50728804e-01 -1.21643412e+00
1.28319919e-01 -8.09566140e-01 6.62907541e-01 -1.22961271e+00
-1.34421694e+00 6.29039764e-01 -1.69782266e-01 -8.63097847e-01
-6.03050590e-01 -7.80750930e-01 -6.79794610e-01 9.58483100e-01
-2.01753354e+00 -1.23858857e+00 -2.95380205e-01 6.47716224e-01
5.62422335e-01 8.69582817e-02 6.91492200e-01 -1.68561354e-01
-7.42755607e-02 2.27040291e-01 -2.79633939e-01 1.85754709e-02
5.77235997e-01 -1.54649734e+00 -2.11373921e-02 6.37642622e-01
3.41050148e-01 3.77673447e-01 1.01086664e+00 -3.01911056e-01
-1.03374636e+00 -5.80786765e-01 1.17792845e+00 -6.13394737e-01
6.07978821e-01 -1.40165433e-01 -9.12765443e-01 5.21351516e-01
6.17284596e-01 4.03162651e-02 6.31227970e-01 -2.44323518e-02
-5.32328784e-01 1.35951042e-01 -1.25711262e+00 8.71375024e-01
4.71621275e-01 -6.88122451e-01 -1.11393297e+00 2.95406729e-01
5.11538148e-01 -1.31268427e-01 -2.99607813e-01 1.07922442e-01
4.49278533e-01 -1.45781505e+00 1.06066942e+00 -6.04517460e-01
8.53485763e-01 -1.84760839e-01 -2.61381596e-01 -9.82422531e-01
-1.45718500e-01 -2.28785902e-01 -2.63585925e-01 1.09687614e+00
4.57994610e-01 -3.30543786e-01 7.50993311e-01 8.23390663e-01
3.32320899e-01 -8.19343925e-01 -1.02053094e+00 -4.63786215e-01
1.73944697e-01 -4.65137452e-01 3.77659291e-01 7.68778563e-01
-5.10021932e-02 9.27003443e-01 -3.07697505e-01 -7.12668672e-02
4.28339750e-01 5.51263094e-01 6.18930757e-01 -9.55010772e-01
-4.67808157e-01 -6.82639658e-01 4.81956527e-02 -1.05357325e+00
3.33012938e-01 -8.21522713e-01 2.09856585e-01 -2.12391305e+00
3.09162170e-01 1.09202862e-02 -1.26502365e-01 4.39749479e-01
-4.60114092e-01 2.58446932e-01 6.11673295e-01 5.47067448e-02
-6.21622860e-01 5.99945188e-01 1.52021503e+00 -1.85374409e-01
-1.96673293e-02 -1.18075654e-01 -7.63765872e-01 8.27608585e-01
8.24921250e-01 -1.86292276e-01 -7.26942062e-01 -4.56034511e-01
4.87071127e-01 4.61073726e-01 9.52065468e-01 -7.00621903e-01
1.55114353e-01 -2.03426972e-01 5.97283542e-01 -6.41101658e-01
5.64797997e-01 -9.41070855e-01 -4.33097392e-01 3.36896569e-01
-6.61459565e-01 -9.53629315e-02 5.31932153e-02 4.64341193e-01
-2.26976946e-01 -7.60613143e-01 7.91939378e-01 -4.51750934e-01
-8.24636579e-01 -1.64924234e-01 -1.64868549e-01 2.62921482e-01
1.01285386e+00 -3.38521212e-01 -3.25816900e-01 -7.94411242e-01
-8.44590068e-01 2.71902591e-01 1.57036215e-01 2.52426922e-01
8.24117064e-01 -1.21227264e+00 -6.48355305e-01 -2.97086179e-01
1.78662300e-01 -4.23379302e-01 4.67566371e-01 6.62280560e-01
-3.03094953e-01 3.78044903e-01 -1.08126149e-01 -3.38520110e-01
-1.24907851e+00 9.54096735e-01 2.44908854e-01 -3.08167070e-01
-3.79810125e-01 9.80928302e-01 4.73436236e-01 -5.29707909e-01
2.33371302e-01 -1.83008298e-01 -6.37478828e-01 1.09903537e-01
4.83833432e-01 -2.34242499e-01 -3.84486198e-01 -5.25914788e-01
-2.52306193e-01 4.00047481e-01 8.14603195e-02 -1.52375206e-01
1.19900572e+00 -1.79206029e-01 2.74400730e-02 4.77645189e-01
1.15671015e+00 -2.61850089e-01 -1.33175111e+00 -4.63631898e-02
-3.45622897e-02 -2.47479632e-01 -1.32184803e-01 -1.17190731e+00
-8.28955173e-01 1.27950203e+00 4.22594249e-01 4.04570162e-01
1.17523444e+00 2.21997395e-01 4.50787038e-01 3.20201218e-01
-2.26113126e-01 -1.06475663e+00 6.36093795e-01 4.27430183e-01
1.23420799e+00 -1.49818003e+00 -2.07059801e-01 -9.36160237e-02
-9.28230405e-01 1.35478246e+00 6.78774059e-01 -1.52636215e-01
1.63105860e-01 -2.58765727e-01 2.51393735e-01 -4.54284340e-01
-8.15112233e-01 -4.63752627e-01 5.69720507e-01 4.47689414e-01
5.08913577e-01 -2.38579452e-01 -3.04021001e-01 5.42438149e-01
-1.86546251e-01 9.32560768e-03 3.33054543e-01 6.83173299e-01
-5.17464221e-01 -7.80749321e-01 -3.95655483e-01 2.61601359e-01
-4.59967703e-01 -1.85117170e-01 -6.03017986e-01 7.52158761e-01
2.70489324e-02 1.04624140e+00 -1.89760193e-01 -8.13041255e-02
2.41080418e-01 2.89043218e-01 6.86439157e-01 -3.76924098e-01
-6.76827133e-01 -3.30947995e-01 1.55886590e-01 -3.81869763e-01
-4.92845893e-01 -4.54428643e-01 -1.39226723e+00 -4.07070406e-02
-1.41070157e-01 -3.93294245e-02 5.32489657e-01 1.23295164e+00
4.02498171e-02 6.59667194e-01 1.15924746e-01 -3.66876721e-01
-8.42641771e-01 -6.93181634e-01 -2.48995759e-02 6.23251796e-01
4.87935543e-01 -4.10476744e-01 -2.71544337e-01 2.36186370e-01] | [10.77768611907959, 1.7963685989379883] |
1946372b-b1dc-46cf-8d86-f6c0a2889c4a | 3d-convolution-neural-network-based-person | 2106.03136 | null | https://arxiv.org/abs/2106.03136v1 | https://arxiv.org/pdf/2106.03136v1.pdf | 3D Convolution Neural Network based Person Identification using Gait cycles | Human identification plays a prominent role in terms of security. In modern times security is becoming the key term for an individual or a country, especially for countries which are facing internal or external threats. Gait analysis is interpreted as the systematic study of the locomotive in humans. It can be used to extract the exact walking features of individuals. Walking features depends on biological as well as the physical feature of the object; hence, it is unique to every individual. In this work, gait features are used to identify an individual. The steps involve object detection, background subtraction, silhouettes extraction, skeletonization, and training 3D Convolution Neural Network on these gait features. The model is trained and evaluated on the dataset acquired by CASIA B Gait, which consists of 15000 videos of 124 subjects walking pattern captured from 11 different angles carrying objects such as bag and coat. The proposed method focuses more on the lower body part to extract features such as the angle between knee and thighs, hip angle, angle of contact, and many other features. The experimental results are compared with amongst accuracies of silhouettes as datasets for training and skeletonized image as training data. The results show that extracting the information from skeletonized data yields improved accuracy. | ['Rijo Jackson Tom', 'Supraja P', 'Ravi Shekhar Tiwari'] | 2021-06-06 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 1.38181940e-01 -4.91280019e-01 -9.72434804e-02 -1.64259989e-02
2.62266934e-01 -3.05653382e-02 1.78062782e-01 2.91203763e-02
-9.91293013e-01 6.27060175e-01 3.06839347e-02 2.72397012e-01
1.28268406e-01 -8.67562056e-01 -1.68852359e-01 -7.83011615e-01
-4.58887845e-01 1.72880486e-01 4.50879425e-01 -2.67154396e-01
3.17911118e-01 6.83329403e-01 -1.49922633e+00 -2.30955333e-01
2.73987800e-01 9.94325340e-01 -5.80890551e-02 8.62061143e-01
4.22290921e-01 3.44958365e-01 -6.78438187e-01 -4.28225607e-01
2.28399560e-01 -3.15720826e-01 -4.73178506e-01 2.69527227e-01
-4.53075729e-02 -3.97998691e-01 -2.75115043e-01 9.16298270e-01
7.18232751e-01 9.79902819e-02 7.85073876e-01 -1.23749447e+00
-1.78355724e-01 -1.28313258e-01 -9.51066256e-01 5.65862119e-01
3.59862775e-01 2.64994413e-01 3.68341565e-01 -5.70716143e-01
3.66591394e-01 1.11493206e+00 8.76393676e-01 4.90542501e-01
-7.51878560e-01 -6.71876073e-01 -5.21457314e-01 7.02569306e-01
-1.41415036e+00 -3.16748917e-01 8.01268518e-01 -4.55199540e-01
6.64009750e-01 -5.01476713e-02 9.82145071e-01 8.79965603e-01
6.80331588e-01 5.50593317e-01 6.74188793e-01 -5.09065449e-01
-3.59744951e-02 -2.32120052e-01 4.43220258e-01 1.01276827e+00
7.61905551e-01 -9.02083586e-04 -5.49026132e-01 1.82586417e-01
6.71910942e-01 -1.22775987e-01 -1.60950929e-01 -9.75978747e-02
-1.03766096e+00 6.75637364e-01 1.00065149e-01 6.20988943e-02
-3.28786105e-01 1.33474991e-02 7.26615429e-01 -1.21077612e-01
-1.13300063e-01 -3.33526194e-01 -2.23280326e-01 -3.42378378e-01
-6.81603491e-01 2.23538697e-01 4.77308095e-01 3.24027598e-01
3.73798490e-01 2.45599747e-01 3.74084175e-01 6.03090465e-01
3.61105800e-01 7.18467772e-01 9.00570452e-01 -6.27861381e-01
3.65241081e-01 7.72700608e-01 1.87848266e-02 -1.52458704e+00
-7.21811473e-01 1.20907743e-02 -8.95955145e-01 2.86520958e-01
5.32131433e-01 -1.81747496e-01 -8.56841207e-01 1.40499020e+00
5.82470179e-01 -5.77882156e-02 -5.05983345e-02 1.01110518e+00
9.35379744e-01 3.91462743e-01 2.60577172e-01 1.11908108e-01
1.93490279e+00 -4.75426286e-01 -5.70631146e-01 -2.73701072e-01
2.22077191e-01 -6.57595754e-01 6.34406567e-01 4.76341188e-01
-5.45313537e-01 -8.78084064e-01 -1.33411360e+00 2.56297618e-01
-6.52692497e-01 5.21257162e-01 2.42168233e-01 1.08596110e+00
-3.59862477e-01 6.17397785e-01 -9.67129052e-01 -8.38714063e-01
2.44623274e-01 5.51546454e-01 -6.90470517e-01 3.50038022e-01
-1.29667640e+00 9.58821356e-01 6.56141937e-01 3.16843957e-01
-3.87417257e-01 1.20847575e-01 -9.90164578e-01 -2.05510542e-01
3.06132846e-02 -6.24523282e-01 5.38998902e-01 -5.64571023e-01
-1.17146289e+00 1.04439795e+00 2.37875566e-01 -5.55496395e-01
5.31753302e-01 -2.25472882e-01 -6.35949850e-01 3.96453440e-01
1.66423798e-01 2.87211031e-01 1.03201604e+00 -4.60302770e-01
-7.36154735e-01 -6.59854054e-01 -4.29095715e-01 1.89677700e-01
-3.91314179e-01 1.92382067e-01 -2.30203554e-01 -5.39209068e-01
-2.82006115e-02 -7.70387709e-01 2.28846505e-01 1.12491228e-01
-3.74061257e-01 1.07698515e-02 1.22308218e+00 -1.40659726e+00
1.10734880e+00 -2.13094020e+00 3.01489066e-02 2.01318428e-01
1.47722010e-02 5.02334416e-01 4.26877797e-01 2.01891482e-01
-9.41779912e-02 -2.00759247e-01 -2.20300168e-01 5.56959724e-03
-1.83179706e-01 1.52915359e-01 3.90217245e-01 8.23968589e-01
2.59987921e-01 3.41202557e-01 -4.17972833e-01 -1.04160106e+00
3.47171277e-01 3.86494488e-01 -7.29112625e-02 1.47198783e-02
6.36107266e-01 2.43303999e-01 -5.93088567e-01 8.32443774e-01
6.12918794e-01 2.66936153e-01 -1.34136498e-01 -4.09166634e-01
1.45715371e-01 -5.83322465e-01 -1.40094435e+00 1.15331292e+00
-2.55186632e-02 5.29716671e-01 -4.18666657e-03 -1.24053860e+00
1.01465034e+00 3.05536896e-01 5.77590466e-01 -4.14493620e-01
5.34547031e-01 -4.80051339e-02 1.12969927e-01 -9.30823684e-01
3.18669498e-01 -5.88864554e-03 -2.56698161e-01 1.99300930e-01
1.72414593e-02 3.88618588e-01 5.95745385e-01 -1.74724266e-01
8.26026201e-01 2.33301863e-01 8.02296221e-01 -8.07522833e-02
9.22529876e-01 -1.13369435e-01 6.63959503e-01 1.05751202e-01
-6.79367900e-01 2.44003341e-01 4.32689637e-02 -8.11806321e-01
-1.00654376e+00 -1.23427701e+00 -4.25214469e-02 3.97249252e-01
3.21564138e-01 -1.51316583e-01 -8.02479565e-01 -3.84456456e-01
-1.29850931e-03 -7.39096329e-02 -6.06814325e-01 -5.24263203e-01
-8.43403876e-01 -1.03339112e+00 7.33626485e-01 6.30581021e-01
1.30319810e+00 -1.24765217e+00 -1.55270720e+00 1.95946440e-01
-2.75191933e-01 -1.19681299e+00 -2.76942551e-01 -1.65925890e-01
-8.10495794e-01 -1.37105179e+00 -7.18655050e-01 -6.96853042e-01
4.99337226e-01 -1.40858276e-04 6.27975106e-01 3.57389182e-01
-9.45442379e-01 2.26371765e-01 -2.50325650e-01 -4.55459028e-01
-7.79499263e-02 -6.33045733e-02 3.80866259e-01 1.91830415e-02
5.79836965e-01 -3.81415457e-01 -7.17259169e-01 3.17422360e-01
-3.72365236e-01 -1.70464143e-01 4.88747746e-01 8.40546548e-01
-6.35480881e-02 5.98420560e-01 2.16840178e-01 -7.73287937e-02
4.79775459e-01 -1.00157358e-01 -2.46875584e-01 1.35348484e-01
-9.04348716e-02 -3.92985046e-01 5.68329751e-01 -2.70333052e-01
-7.79128015e-01 1.59418434e-01 -5.45773096e-02 6.22470565e-02
-4.91215378e-01 1.38462484e-01 -2.50837713e-01 -1.80695802e-02
4.58641827e-01 4.03085262e-01 2.21406564e-01 -2.71423310e-01
-3.36810112e-01 7.65868306e-01 8.33102643e-01 -5.65582156e-01
8.49401474e-01 5.23063779e-01 4.63321477e-01 -1.49375916e+00
-9.78437066e-02 -6.43685162e-01 -1.09846497e+00 -7.72607148e-01
1.16356981e+00 -5.24168134e-01 -1.02909589e+00 9.28065479e-01
-8.64589930e-01 3.49494219e-01 1.24240905e-01 6.56461596e-01
-4.66201276e-01 8.68974745e-01 -6.08940780e-01 -9.93234754e-01
-6.96277499e-01 -8.79220903e-01 8.48966599e-01 6.45351171e-01
-4.45524275e-01 -7.78073132e-01 -5.68751507e-02 3.83320183e-01
3.30530666e-02 7.24470317e-01 6.21459663e-01 -3.91343653e-01
1.28994837e-01 -6.25289083e-01 -3.28315236e-02 2.64724910e-01
3.91251028e-01 2.47318625e-01 -5.58576167e-01 -2.63269365e-01
-2.54624691e-02 -4.18682359e-02 6.31784141e-01 3.92323524e-01
4.94589299e-01 -2.23758727e-01 -4.29496020e-01 3.65975708e-01
1.27462280e+00 4.49821472e-01 6.60894930e-01 6.27274692e-01
5.37378073e-01 6.58148885e-01 7.42171407e-01 5.86386144e-01
1.46932779e-02 5.13371766e-01 2.97236770e-01 -2.21667513e-02
-9.00814310e-02 1.54783383e-01 4.99929190e-01 5.78494489e-01
-5.17223597e-01 1.82803333e-01 -8.99391711e-01 5.12377799e-01
-1.45098639e+00 -1.18003929e+00 -1.36375085e-01 2.24088502e+00
2.87908614e-01 3.19066316e-01 8.07208300e-01 8.27858388e-01
1.17446959e+00 -1.98196203e-01 -3.56432825e-01 -2.00108618e-01
1.62722141e-01 2.94035375e-01 5.37941456e-01 5.54727279e-02
-1.36994004e+00 4.89397168e-01 5.54024982e+00 5.77848375e-01
-1.15385556e+00 -2.63308734e-01 2.70142168e-01 2.90473193e-01
1.09069526e+00 -5.15188992e-01 -7.49295413e-01 7.85066664e-01
6.62107170e-01 -9.74996164e-02 -5.84294728e-04 8.07404995e-01
3.03022385e-01 -4.22804505e-01 -4.08542961e-01 8.92716050e-01
1.57228261e-01 -5.31480968e-01 -1.88816398e-01 -2.60471180e-02
-1.72893964e-02 -6.67876184e-01 -5.34596816e-02 -1.41157523e-01
-1.51721820e-01 -7.63742864e-01 5.35723388e-01 3.77350569e-01
5.02197027e-01 -9.36581910e-01 1.04538584e+00 3.05652499e-01
-1.65841901e+00 -1.76393330e-01 -2.45576054e-01 -8.21304321e-02
2.24747330e-01 6.52809441e-02 -7.38123298e-01 6.03011429e-01
9.92726862e-01 6.96946204e-01 -6.47294939e-01 1.19813812e+00
-1.12803653e-01 4.55137134e-01 -4.80285048e-01 -1.62475154e-01
-1.41812623e-01 -3.18706751e-01 5.15072048e-01 1.28075290e+00
4.22539949e-01 -4.51595373e-02 1.66020915e-02 2.65099764e-01
5.39417267e-01 2.41785824e-01 -6.10845745e-01 -9.90876276e-03
3.87191139e-02 1.26518083e+00 -1.20604837e+00 -2.85710424e-01
-2.34088063e-01 9.12981093e-01 -2.80018836e-01 1.66607630e-02
-9.87135828e-01 -7.96760678e-01 5.62724292e-01 3.09484184e-01
2.33597755e-01 -3.90379190e-01 -3.58347557e-02 -8.93480182e-01
-9.54732522e-02 -5.64606845e-01 7.76782453e-01 -4.69086111e-01
-9.23677027e-01 8.30761716e-02 1.13935016e-01 -1.24393857e+00
-5.48129119e-02 -1.01581168e+00 -7.71479309e-01 6.91724479e-01
-6.71049833e-01 -1.16773319e+00 -6.58423901e-01 5.92971325e-01
5.82814276e-01 -3.19025606e-01 6.17170155e-01 4.89232749e-01
-8.84220958e-01 3.12803179e-01 -3.04598778e-01 7.30846524e-01
4.71912295e-01 -9.47582960e-01 3.11314672e-01 1.15867496e+00
-2.55307645e-01 5.87723494e-01 8.47413242e-01 -9.42185760e-01
-1.00617397e+00 -5.71982622e-01 6.52610362e-01 -1.02012247e-01
3.41557175e-01 3.23598385e-02 -3.60559225e-01 4.89276648e-01
-1.53003588e-01 -1.05456367e-01 5.96162617e-01 -5.24622083e-01
4.01235908e-01 -1.19131215e-01 -1.39786971e+00 6.00726783e-01
6.72516942e-01 8.16984847e-02 -7.18235373e-01 -1.18870474e-01
-9.50705633e-02 -1.75865650e-01 -7.60550559e-01 3.07853520e-01
1.07153976e+00 -9.90341425e-01 1.32097399e+00 -5.46107173e-01
1.94694236e-01 -4.90956366e-01 3.86787616e-02 -8.00710022e-01
-2.24168628e-01 -4.19840887e-02 -7.09110871e-02 9.97110307e-01
-2.98397571e-01 -4.69141692e-01 1.13310492e+00 2.89452970e-01
3.02482098e-01 -3.99004579e-01 -1.02374220e+00 -7.79155374e-01
-4.56265032e-01 -3.31459567e-02 3.32456857e-01 6.67595446e-01
5.52324541e-02 1.71405852e-01 -5.93902230e-01 1.82290614e-01
1.09479463e+00 -3.32672149e-01 1.02997518e+00 -1.38344622e+00
-1.04008995e-01 -9.92543399e-02 -1.33313227e+00 -2.42350146e-01
-1.99677438e-01 -3.28834176e-01 -3.24137658e-01 -1.31928408e+00
4.44580428e-02 2.61684358e-01 -5.79046682e-02 3.45161498e-01
-8.30709115e-02 5.05588710e-01 4.55376245e-02 -3.26365642e-02
-5.26504479e-02 2.15313017e-01 1.01213419e+00 -2.07210541e-01
4.30977866e-02 1.80113882e-01 7.60054141e-02 9.99401808e-01
1.00332367e+00 -2.42414564e-01 -1.19873956e-01 2.36021221e-01
-4.02821451e-01 -5.27138337e-02 5.64995289e-01 -1.59781039e+00
4.51119244e-02 -3.67070846e-02 1.07115984e+00 -6.93366528e-01
5.17762959e-01 -9.52507257e-01 2.46019751e-01 1.35487366e+00
3.86756092e-01 3.44577909e-01 8.38183016e-02 2.82241017e-01
-3.76550071e-02 -4.10715699e-01 9.63537872e-01 -3.28827679e-01
-1.14874101e+00 3.34034748e-02 -6.77123368e-01 -2.36182824e-01
1.32574630e+00 -9.76984680e-01 -1.05622455e-01 -1.29855990e-01
-9.15305734e-01 -1.08321182e-01 3.23947132e-01 1.18326299e-01
7.79283702e-01 -1.25118124e+00 -5.25685847e-01 3.91565204e-01
-3.28726023e-02 -4.74537164e-01 2.49774873e-01 8.16388547e-01
-1.06955218e+00 1.81446418e-01 -1.11571085e+00 -4.97673154e-01
-1.99601257e+00 4.44343328e-01 2.53102481e-01 -1.85050488e-01
-7.03211367e-01 2.74953842e-01 -3.36597860e-01 1.42929748e-01
-6.89989105e-02 -7.81583637e-02 -6.71992660e-01 5.94921745e-02
3.85525346e-01 8.58623743e-01 -2.27635428e-01 -1.13661540e+00
-6.00056827e-01 1.13460565e+00 2.27385730e-01 8.17150772e-02
1.00584710e+00 -9.54712331e-02 9.61261392e-02 1.53239682e-01
9.19154584e-01 -2.32502908e-01 -8.02487016e-01 2.55273700e-01
-1.06006652e-01 -5.39889812e-01 -4.90204215e-01 -2.62513310e-01
-1.01575482e+00 1.04098535e+00 1.07649004e+00 7.73071451e-03
1.07152247e+00 -5.91456592e-01 9.25820768e-01 1.96919903e-01
5.43118358e-01 -1.29730892e+00 2.37370744e-01 7.55907921e-03
4.37752753e-01 -1.08355486e+00 3.44025373e-01 -3.23002696e-01
-4.52542603e-01 1.36398244e+00 6.97540224e-01 -3.30532759e-01
5.90247273e-01 1.81582302e-01 -7.12471157e-02 -3.69778983e-02
1.99965984e-02 -1.39402017e-01 1.33121446e-01 9.33048368e-01
1.52715936e-01 -2.82716844e-02 -5.96983135e-01 4.47796911e-01
-2.89534241e-01 1.11157589e-01 3.31853122e-01 1.31695580e+00
-7.25518525e-01 -6.37524366e-01 -8.57948899e-01 4.02242362e-01
-6.13951206e-01 4.65897322e-01 -1.95653021e-01 1.27048123e+00
5.06350636e-01 7.33836949e-01 -1.20605692e-01 -5.78649461e-01
3.32148105e-01 1.16263740e-01 5.28424859e-01 -1.00521363e-01
-2.58518398e-01 -3.82205307e-01 2.01607332e-01 -1.91230461e-01
-3.79885286e-01 -7.44552553e-01 -1.28489614e+00 -4.64425474e-01
-2.17816401e-02 1.06244117e-01 6.30447805e-01 7.79212296e-01
-3.72304380e-01 2.76895761e-01 3.18149060e-01 -8.90855134e-01
-3.73199195e-01 -8.10759366e-01 -8.47203493e-01 6.47867322e-01
8.28871429e-02 -1.16486669e+00 -3.08210373e-01 5.42385340e-01] | [14.157527923583984, 1.460263729095459] |
ab0fca47-5c78-464c-a118-5e62955cd790 | exact-recovery-in-the-general-hypergraph | 2105.04770 | null | https://arxiv.org/abs/2105.04770v2 | https://arxiv.org/pdf/2105.04770v2.pdf | Exact Recovery in the General Hypergraph Stochastic Block Model | This paper investigates fundamental limits of exact recovery in the general d-uniform hypergraph stochastic block model (d-HSBM), wherein n nodes are partitioned into k disjoint communities with relative sizes (p1,..., pk). Each subset of nodes with cardinality d is generated independently as an order-d hyperedge with a certain probability that depends on the ground-truth communities that the d nodes belong to. The goal is to exactly recover the k hidden communities based on the observed hypergraph. We show that there exists a sharp threshold such that exact recovery is achievable above the threshold and impossible below the threshold (apart from a small regime of parameters that will be specified precisely). This threshold is represented in terms of a quantity which we term as the generalized Chernoff-Hellinger divergence between communities. Our result for this general model recovers prior results for the standard SBM and d-HSBM with two symmetric communities as special cases. En route to proving our achievability results, we develop a polynomial-time two-stage algorithm that meets the threshold. The first stage adopts a certain hypergraph spectral clustering method to obtain a coarse estimate of communities, and the second stage refines each node individually via local refinement steps to ensure exact recovery. | ['Vincent Y. F. Tan', 'Qiaosheng Zhang'] | 2021-05-11 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.43008757e-01 6.02798700e-01 -5.42653263e-01 2.38802209e-01
-9.31263924e-01 -8.25264871e-01 3.07240874e-01 1.96439832e-01
-2.69845445e-02 5.78681886e-01 5.02060987e-02 -3.80736589e-01
-3.10972214e-01 -8.56283128e-01 -7.35122263e-01 -1.26350045e+00
-7.25748479e-01 8.61071289e-01 3.75126183e-01 3.02431025e-02
1.79013059e-01 2.75418252e-01 -7.77594388e-01 -3.85096595e-02
6.11389637e-01 3.09077650e-01 -2.21581236e-02 1.03741705e+00
2.78374821e-01 4.10336673e-01 -3.40159863e-01 -2.79075205e-01
6.15860701e-01 -5.27038574e-01 -9.73069370e-01 5.70713460e-01
-2.14517996e-01 -1.48313835e-01 -8.85258436e-01 1.46308005e+00
2.51709908e-01 -5.20911098e-01 7.95521379e-01 -1.56570256e+00
-5.26168287e-01 1.37945461e+00 -1.22067797e+00 1.21669337e-01
3.55482340e-01 -1.87459096e-01 1.19010961e+00 -2.71013319e-01
7.68797338e-01 9.83968377e-01 6.69396639e-01 3.57575357e-01
-1.67826617e+00 -8.97427917e-01 -3.79052550e-01 3.95209230e-02
-2.23029780e+00 -3.62846583e-01 5.40614963e-01 -5.08946896e-01
3.08422565e-01 1.69992149e-01 5.33774436e-01 5.50892413e-01
-8.73302966e-02 5.06946266e-01 9.87915277e-01 -6.73466086e-01
3.39014322e-01 1.44171655e-01 1.22804999e-01 6.46295309e-01
1.10586226e+00 -1.01441368e-01 -2.93901473e-01 -6.68205142e-01
7.15600252e-01 -1.22011222e-01 -7.64729977e-01 -7.76697814e-01
-1.08997655e+00 9.48106706e-01 1.74167126e-01 4.10647333e-01
-1.54179111e-01 3.74338746e-01 4.36948426e-02 5.69027185e-01
1.12521395e-01 -4.12467510e-01 5.36173731e-02 4.48942214e-01
-1.21003532e+00 5.36241522e-03 1.16905499e+00 1.55549991e+00
8.78800094e-01 -4.26417679e-01 1.54897675e-01 -1.26860989e-02
3.57145637e-01 7.15835989e-01 -7.22508073e-01 -9.56159234e-01
7.14887083e-01 1.12142012e-01 3.61497253e-01 -1.14048052e+00
-3.70388627e-02 -4.55260456e-01 -1.21448684e+00 -2.15104252e-01
4.08477783e-01 4.25935350e-02 -6.02818012e-01 2.01633668e+00
3.37085813e-01 3.61421168e-01 1.64822593e-01 9.81120110e-01
1.45621717e-01 5.68860829e-01 -4.73637074e-01 -5.63100755e-01
9.78491783e-01 -2.72850811e-01 -4.34035331e-01 2.55494326e-01
4.93517399e-01 -2.78694421e-01 1.68087706e-01 3.01260591e-01
-1.20318031e+00 3.24316502e-01 -1.12481391e+00 4.00883168e-01
3.22767317e-01 -4.16102856e-01 1.96009144e-01 1.01204598e+00
-1.49533880e+00 3.01177114e-01 -6.95953965e-01 -3.85822117e-01
2.10862666e-01 2.82207012e-01 -3.27258676e-01 -3.47287476e-01
-9.86542940e-01 2.04281017e-01 5.36962390e-01 -2.43167039e-02
-1.26105607e+00 -1.77850887e-01 -4.54418659e-01 1.50750503e-01
3.96068901e-01 -7.59743392e-01 7.09798455e-01 -3.84040505e-01
-6.93998992e-01 1.01730216e+00 -2.36934230e-01 -5.87904215e-01
3.89063269e-01 7.98941791e-01 -1.30905151e-01 9.52929378e-01
2.83102363e-01 -5.19145131e-02 5.96655250e-01 -1.64142704e+00
-5.47125578e-01 -4.22489882e-01 4.36830446e-02 -9.71463695e-02
-1.45110577e-01 -4.99757789e-02 -6.44393265e-01 -5.83584793e-02
6.35622501e-01 -1.27914751e+00 -4.71640080e-01 -1.76771402e-01
-9.04013216e-01 1.27423748e-01 3.64619493e-01 -4.37328726e-01
1.21178544e+00 -1.93409538e+00 4.19217646e-01 1.02557015e+00
9.96743739e-01 -2.78796911e-01 7.58011490e-02 9.44797158e-01
2.69437671e-01 4.42570239e-01 -5.03877759e-01 -3.10281485e-01
-5.08377850e-02 1.02973349e-01 2.60318648e-02 1.31308198e+00
-5.92656612e-01 2.48704791e-01 -9.34031069e-01 -4.59584028e-01
-3.69409353e-01 9.28114653e-02 -6.73920035e-01 -3.62913549e-01
9.44789350e-02 2.62293130e-01 -5.88879645e-01 3.24876368e-01
1.26008976e+00 -9.06137168e-01 9.13176954e-01 2.07010210e-01
2.39251018e-01 -1.36231855e-01 -1.65035844e+00 1.31058609e+00
1.46969885e-01 4.81775731e-01 9.25380111e-01 -9.25167680e-01
2.91276932e-01 7.33243227e-01 5.10083795e-01 3.50800335e-01
1.04637474e-01 2.73636758e-01 -1.38491482e-01 -2.41506100e-01
3.53595763e-01 -3.63877416e-01 -2.23595098e-01 9.28911150e-01
-3.20397347e-01 3.66538376e-01 -7.81944022e-02 1.17352712e+00
1.59040284e+00 -8.59495759e-01 3.08448046e-01 -5.80901623e-01
1.32820442e-01 8.29153657e-02 4.60446149e-01 1.13680840e+00
-2.40842655e-01 5.27891219e-01 7.82800198e-01 4.98625278e-01
-1.38129663e+00 -1.11353016e+00 2.70623490e-02 1.95179909e-01
7.40467668e-01 -5.37539303e-01 -7.65690744e-01 -1.82771057e-01
6.70210421e-02 2.51506060e-01 -6.89972520e-01 -2.94302441e-02
1.51277885e-01 -8.88324559e-01 5.41952074e-01 5.35925701e-02
4.26054686e-01 -1.94558620e-01 5.89316525e-02 4.73119318e-02
-5.06226301e-01 -1.31612110e+00 -6.51525378e-01 4.01286222e-02
-7.84902334e-01 -1.41494334e+00 -6.52736604e-01 -8.02145958e-01
9.28545713e-01 9.97008801e-01 7.99912155e-01 3.10406119e-01
1.99896455e-01 5.10440528e-01 -4.35782433e-01 3.69672120e-01
-6.05771065e-01 1.01406559e-01 2.07143952e-03 -1.59046575e-02
1.70960993e-01 -7.38645256e-01 -7.62041509e-01 1.13142721e-01
-8.61024320e-01 -1.25945061e-02 4.43888038e-01 3.22365016e-01
4.41154778e-01 6.32282913e-01 7.47052655e-02 -9.35418010e-01
3.09434593e-01 -1.13019240e+00 -6.90483570e-01 1.80578589e-01
-4.09105480e-01 -1.54400513e-01 1.80409536e-01 -1.44935310e-01
-4.89423037e-01 -7.29777664e-03 5.53655505e-01 -3.44253719e-01
2.50297695e-01 5.45216262e-01 -1.97335511e-01 -2.25663096e-01
3.09455842e-01 1.56281739e-01 -1.11135639e-01 -1.38557106e-01
4.17209148e-01 8.46939266e-01 4.02278006e-01 -3.70214731e-01
1.31487489e+00 9.27349985e-01 3.44865859e-01 -9.38103080e-01
-3.80290896e-01 -7.90734470e-01 -5.93846619e-01 -2.60242503e-02
4.75193113e-01 -1.13049221e+00 -9.39089894e-01 2.47560307e-01
-9.85517323e-01 -4.64081645e-01 5.55102490e-02 3.47339243e-01
-4.54336822e-01 7.58072078e-01 -9.97729003e-01 -1.34586501e+00
2.95850914e-02 -6.93699837e-01 9.17556763e-01 -3.10681313e-01
2.38956451e-01 -9.25782323e-01 8.82638767e-02 1.43964320e-01
4.49578501e-02 2.88650334e-01 8.22179914e-01 -4.41291004e-01
-1.16561973e+00 -2.33739674e-01 -4.24588770e-01 -1.83500156e-01
-2.19790563e-01 -1.85699046e-01 -4.55202609e-01 -7.18524158e-01
-1.10757776e-01 1.81782059e-02 8.82929802e-01 4.86579150e-01
8.75813186e-01 -5.83852410e-01 -9.10073459e-01 3.70674878e-01
1.80048239e+00 -2.96932638e-01 7.17528999e-01 -3.38702723e-02
6.72055483e-01 2.13303417e-01 5.86162396e-02 5.98048389e-01
5.75383306e-01 3.06116343e-01 3.27122927e-01 2.26754576e-01
2.53711641e-01 -1.96363896e-01 1.58437237e-01 9.70066428e-01
9.95300412e-02 -6.74725175e-01 -8.31591189e-01 9.56825554e-01
-1.71477962e+00 -1.11933446e+00 -6.07746422e-01 2.67233419e+00
1.06154358e+00 5.91957681e-02 4.05408680e-01 3.78991783e-01
1.30802763e+00 -1.18421815e-01 -6.06985278e-02 2.30521470e-01
-3.39804053e-01 -1.00911058e-01 1.12812424e+00 8.62564683e-01
-5.98548830e-01 5.14982343e-01 6.13703394e+00 9.12898898e-01
-2.65997022e-01 3.04064810e-01 4.32081908e-01 1.14219800e-01
-7.78187752e-01 4.55986857e-01 -6.51721418e-01 3.23730856e-01
1.06249595e+00 -5.56162834e-01 5.11683941e-01 4.88048226e-01
8.71454179e-02 -1.71232432e-01 -8.24997962e-01 7.08867252e-01
-1.08467549e-01 -1.33088303e+00 -1.78090587e-01 7.95649409e-01
1.18052304e+00 -9.36342105e-02 -2.79370338e-01 -3.80024165e-01
8.98373604e-01 -5.61826587e-01 5.34081221e-01 9.94201526e-02
1.10696256e+00 -9.10473764e-01 4.01970059e-01 6.79969251e-01
-1.37335348e+00 -1.18440144e-01 -5.50267935e-01 3.07336837e-01
2.57243931e-01 9.43757176e-01 -8.72210741e-01 6.79491580e-01
4.31834966e-01 2.68714100e-01 1.79615155e-01 1.28893089e+00
1.68720543e-01 9.38479424e-01 -6.34851813e-01 2.75441021e-01
9.50170308e-02 -5.05540848e-01 9.50302780e-01 1.13291287e+00
5.20065844e-01 8.15772295e-01 2.98546970e-01 6.81944728e-01
-4.29315060e-01 -1.71301335e-01 -6.40755117e-01 -1.92660671e-02
9.63578165e-01 9.31558430e-01 -1.04974604e+00 -3.73749852e-01
-2.12617397e-01 1.05891538e+00 2.09674016e-02 5.27162135e-01
-6.40264034e-01 -4.32499737e-01 3.41717750e-01 5.27401686e-01
5.81162512e-01 -4.47929025e-01 -2.41904020e-01 -1.13729763e+00
-1.98398486e-01 -6.63347960e-01 4.42267835e-01 -3.41912389e-01
-1.16037107e+00 1.90258726e-01 -1.67326890e-02 -1.01918840e+00
2.33613864e-01 1.87868878e-01 -4.44177181e-01 7.61513710e-01
-1.07209420e+00 -8.24106634e-01 -3.94392274e-02 5.78345895e-01
-5.46830475e-01 3.10017049e-01 4.38355058e-01 2.38422155e-01
-2.89637983e-01 4.83409792e-01 5.88626802e-01 1.57625809e-01
1.33659214e-01 -1.24638700e+00 1.68957427e-01 1.06622815e+00
-1.04695313e-01 6.15102351e-01 9.98923838e-01 -8.72695506e-01
-1.37513280e+00 -1.00639582e+00 1.08191061e+00 2.16756854e-02
8.20495844e-01 -5.56779087e-01 -6.36436164e-01 8.65287006e-01
1.62505910e-01 -1.92957535e-01 6.77534044e-01 -1.68928891e-01
-3.91133577e-01 1.70479730e-01 -1.49863064e+00 4.29772079e-01
1.18559885e+00 -6.39757693e-01 3.62228714e-02 5.62474728e-01
5.77785194e-01 -1.42220678e-02 -8.28419864e-01 4.58062114e-03
2.26874650e-01 -8.01963925e-01 6.55352056e-01 -2.95641184e-01
1.06571518e-01 -5.02738416e-01 -4.46207434e-01 -8.68823111e-01
-5.40684283e-01 -1.21778953e+00 -1.89666301e-01 9.64408696e-01
4.52364385e-01 -5.32072723e-01 1.09441018e+00 2.50235975e-01
4.78156477e-01 -2.75712132e-01 -1.00461996e+00 -7.90376306e-01
4.21195813e-02 -1.84555247e-01 5.22837996e-01 1.08262658e+00
4.23970848e-01 1.86822727e-01 -5.23432255e-01 9.19626355e-01
1.51995492e+00 3.18496257e-01 6.64610207e-01 -1.32970738e+00
-5.44497728e-01 -2.07330123e-01 -3.13166857e-01 -1.30524874e+00
-2.24996004e-02 -1.08695257e+00 4.28157784e-02 -1.46389747e+00
1.13213599e+00 -8.37317705e-01 9.18730274e-02 -6.78408220e-02
5.67227527e-02 3.33268821e-01 -4.61956188e-02 5.48834264e-01
-8.99935246e-01 1.56794146e-01 1.03305352e+00 -3.75640541e-02
-1.25516117e-01 3.53633136e-01 -9.46111679e-01 1.96368590e-01
6.79420769e-01 -7.02560961e-01 -4.62605089e-01 -2.29534544e-02
3.41796368e-01 9.20293808e-01 4.22886372e-01 -6.83324695e-01
5.88049293e-01 -8.40583369e-02 -1.75073415e-01 -7.68888354e-01
4.33600321e-02 -8.59289765e-01 6.11401081e-01 6.62809730e-01
-4.33250815e-01 -4.40331578e-01 -4.15005744e-01 1.39292431e+00
3.28753561e-01 -2.36017674e-01 7.04106987e-01 1.33920163e-01
-2.43135709e-02 5.84905863e-01 -5.94646931e-01 4.79136519e-02
1.17798007e+00 -4.20532227e-01 -2.55835950e-01 -1.07438338e+00
-8.09762061e-01 4.46110904e-01 8.21041703e-01 -4.59558517e-01
4.43291813e-01 -1.20172095e+00 -1.15320015e+00 -1.31845281e-01
1.61637276e-01 -4.37651612e-02 3.36931318e-01 1.10950923e+00
-3.33487988e-01 2.88418829e-01 5.05194783e-01 -6.51456952e-01
-1.39203620e+00 7.87151039e-01 2.08133399e-01 -1.64375603e-01
-7.05122113e-01 7.17176080e-01 6.48298711e-02 2.03916356e-01
1.20994359e-01 2.42510810e-01 5.34030437e-01 -4.06576961e-01
3.94609243e-01 4.84372556e-01 -3.27806473e-01 -8.36006224e-01
-1.99982762e-01 2.35839933e-01 1.46953529e-02 -5.78869283e-01
1.10203803e+00 -7.81447411e-01 -3.64896744e-01 7.98714161e-02
1.26623726e+00 4.59561020e-01 -8.92923892e-01 -4.88006085e-01
-9.46814194e-02 -6.60982311e-01 -5.01842424e-02 -6.76510558e-02
-1.29251349e+00 4.10094827e-01 8.16217139e-02 6.97492182e-01
9.37659502e-01 5.87121367e-01 6.97323382e-01 1.05311751e-01
1.03715920e+00 -5.78429818e-01 -4.06926155e-01 -8.22584890e-03
2.47265890e-01 -6.44124031e-01 2.54427344e-02 -7.66089559e-01
-2.51978517e-01 7.92657495e-01 -2.63804793e-01 -1.04827270e-01
8.09434116e-01 3.59389603e-01 -7.64039278e-01 -3.21709663e-01
-6.43124163e-01 -2.58875400e-01 -2.91460633e-01 6.72558308e-01
-1.07144989e-01 3.98705721e-01 -2.00042903e-01 3.41661721e-01
7.88236707e-02 -2.02247560e-01 9.60183680e-01 5.80923855e-01
-8.72240484e-01 -1.04202521e+00 -5.20131528e-01 3.78361106e-01
-4.83072013e-01 -1.39573589e-01 -3.03702265e-01 7.28402495e-01
-3.30516487e-01 1.21276760e+00 -1.01813026e-01 -4.64138061e-01
-3.07068586e-01 -6.34709239e-01 6.17377758e-01 -6.30208611e-01
9.79513749e-02 3.15480739e-01 5.27129434e-02 -9.60258171e-02
-4.13116962e-01 -7.79105902e-01 -1.02861118e+00 -1.18016696e+00
-6.62057579e-01 7.32599735e-01 1.21212125e-01 6.79925561e-01
9.26009715e-02 -2.27972299e-01 1.07422364e+00 -4.23959881e-01
-5.44722080e-01 -6.73661888e-01 -1.36537683e+00 -2.74559762e-02
5.88936925e-01 -1.45435765e-01 -9.59316254e-01 1.40757894e-03] | [6.895145893096924, 5.1207475662231445] |
3c2e9a52-31c8-4ac7-a760-95010ab4815c | on-practical-robust-reinforcement-learning | 2305.06657 | null | https://arxiv.org/abs/2305.06657v2 | https://arxiv.org/pdf/2305.06657v2.pdf | On Practical Robust Reinforcement Learning: Practical Uncertainty Set and Double-Agent Algorithm | We study a robust reinforcement learning (RL) with model uncertainty. Given nominal Markov decision process (N-MDP) that generate samples for training, an uncertainty set is defined, which contains some perturbed MDPs from N-MDP for the purpose of reflecting potential mismatched between training (i.e., N-MDP) and testing environments. The objective of robust RL is to learn a robust policy that optimizes the worst-case performance over an uncertainty set. In this paper, we propose a new uncertainty set containing more realistic MDPs than the existing ones. For this uncertainty set, we present a robust RL algorithm (named ARQ-Learning) for tabular case and characterize its finite-time error bound. Also, it is proved that ARQ-Learning converges as fast as Q-Learning and the state-of-the-art robust Q-Learning while ensuring better robustness to real-world applications. Next, we propose {\em pessimistic} agent that efficiently tackles the key bottleneck for the extension of ARQ-Learning into the case with larger or continuous state spaces. Incorporating the idea of pessimistic agents into the famous RL algorithms such as Q-Learning, deep-Q network (DQN), and deep deterministic policy gradient (DDPG), we present PRQ-Learning, PR-DQN, and PR-DDPG, respectively. Noticeably, the proposed idea can be immediately applied to other model-free RL algorithms (e.g., soft actor critic). Via experiments, we demonstrate the superiority of our algorithms on various RL applications with model uncertainty. | ['SongNam Hong', 'Ukjo Hwang'] | 2023-05-11 | null | null | null | null | ['q-learning'] | ['methodology'] | [-3.29479694e-01 2.98298031e-01 -3.29430461e-01 4.48689200e-02
-1.26501441e+00 -4.34759885e-01 2.41091624e-01 -2.64356993e-02
-4.71490353e-01 1.32451558e+00 -1.39739245e-01 -4.58721101e-01
-5.45555711e-01 -7.70840168e-01 -9.16446924e-01 -1.02322185e+00
-3.04874361e-01 5.99312127e-01 2.09928658e-02 -2.12510914e-01
2.81920910e-01 2.45966941e-01 -1.13192081e+00 -4.53060448e-01
1.05622411e+00 1.20200408e+00 9.37521756e-02 5.06565630e-01
2.49129921e-01 8.49947393e-01 -7.91932642e-01 -1.90946683e-01
4.56600338e-01 -2.51270086e-01 -3.65050763e-01 -7.63941705e-02
-3.39468539e-01 -4.93845522e-01 -2.11684674e-01 1.12772465e+00
7.84981310e-01 6.60007417e-01 5.54529846e-01 -1.51103544e+00
-2.51725018e-01 8.17029476e-01 -5.82093358e-01 -1.54676987e-02
5.43369129e-02 7.08005846e-01 5.44309795e-01 -3.35716933e-01
2.41500095e-01 1.81791139e+00 1.60812303e-01 8.59706283e-01
-9.46702778e-01 -6.80849135e-01 4.91792053e-01 5.38078742e-03
-1.14899635e+00 -4.21089754e-02 5.54236412e-01 -1.61768988e-01
6.27875566e-01 -1.57127023e-01 4.58222002e-01 1.27166724e+00
6.98612690e-01 1.06910586e+00 1.29134822e+00 -5.42938858e-02
9.54093397e-01 -1.61737114e-01 -4.65390474e-01 5.11145651e-01
2.34392986e-01 7.10994661e-01 2.76948586e-02 -1.96357980e-01
8.46514881e-01 -5.40560856e-02 -2.39976216e-02 -4.44095731e-01
-9.52144206e-01 9.17015851e-01 1.20239235e-01 -3.20075214e-01
-4.96232361e-01 4.55631793e-01 5.12182415e-01 5.54373443e-01
2.00169653e-01 6.00652933e-01 -5.37586749e-01 -2.92983770e-01
-4.74304616e-01 6.45978928e-01 7.16727078e-01 9.56404984e-01
4.14845318e-01 7.78309584e-01 -6.58955157e-01 3.91909570e-01
5.02718151e-01 8.14675093e-01 5.81535101e-01 -1.40085864e+00
6.37317479e-01 3.90622541e-02 8.77786696e-01 -5.75962126e-01
-4.53851074e-01 -5.04797935e-01 -8.13786268e-01 5.90229869e-01
3.07136118e-01 -7.92867124e-01 -6.30477488e-01 1.92999041e+00
3.95400167e-01 3.80053610e-01 6.01728439e-01 8.16235602e-01
-1.43776312e-02 8.49029720e-01 -1.68264925e-01 -8.76842082e-01
7.94187605e-01 -7.97753096e-01 -8.37857068e-01 1.37905283e-02
9.68850553e-02 -1.18638217e-01 1.02113235e+00 7.05681801e-01
-1.19345653e+00 -4.51502204e-01 -1.17891240e+00 9.89906073e-01
9.26219672e-02 -4.19495970e-01 8.72253999e-02 5.95921516e-01
-7.35223472e-01 8.02778304e-01 -7.02796221e-01 1.33310616e-01
1.92644820e-01 3.48550797e-01 3.32663715e-01 2.53915250e-01
-1.53426898e+00 9.51451600e-01 7.96810627e-01 1.63503170e-01
-1.86196220e+00 -4.14747477e-01 -6.95781291e-01 -1.88828185e-01
1.31444871e+00 -4.90506262e-01 1.74309003e+00 -7.14261532e-01
-2.34134030e+00 -1.78085789e-01 4.58931148e-01 -6.77846849e-01
8.99872482e-01 -2.70932615e-01 -3.43215972e-01 7.88318589e-02
-2.76283026e-01 2.80451238e-01 1.45660233e+00 -1.36030960e+00
-7.13132560e-01 -1.41519159e-01 1.88858077e-01 4.00215328e-01
2.03027591e-01 -2.42772534e-01 9.23099369e-02 -4.64975059e-01
-3.61184090e-01 -9.12205637e-01 -7.37009227e-01 -4.40114349e-01
-4.35122788e-01 -4.39750999e-01 5.40984213e-01 -2.41535395e-01
1.15873730e+00 -1.81890166e+00 1.57967880e-01 3.62854421e-01
-1.70861706e-01 3.71641397e-01 -1.99876085e-01 5.52893937e-01
3.53042394e-01 1.46389738e-01 -2.83776522e-01 -7.20588677e-03
5.16910732e-01 5.60696483e-01 -5.95708311e-01 4.16314781e-01
2.43254155e-01 8.90646279e-01 -1.31048572e+00 -5.47624528e-01
1.83408797e-01 -1.10874318e-01 -2.45233670e-01 4.69600618e-01
-7.89513767e-01 5.98551273e-01 -8.79936159e-01 5.97166598e-01
6.13179386e-01 2.35052675e-01 3.24003883e-02 2.86545068e-01
-1.54374227e-01 -4.25193369e-01 -1.70932806e+00 1.26424181e+00
-5.83275139e-01 -2.11355641e-01 1.49318591e-01 -1.09545171e+00
1.06811690e+00 4.12319332e-01 4.84783560e-01 -7.47868299e-01
2.55721182e-01 2.16337532e-01 -1.25204578e-01 -3.70404691e-01
3.74032736e-01 -3.32323164e-01 -2.62725651e-01 4.96119082e-01
-5.09231277e-02 -4.49436814e-01 1.24640167e-01 -2.07856804e-01
7.89553642e-01 4.30178732e-01 4.88378853e-01 -2.20187053e-01
6.74865127e-01 -5.17087460e-01 1.04297459e+00 1.04169583e+00
-7.49975979e-01 -7.33469427e-02 8.49960804e-01 -1.85261145e-01
-8.20972621e-01 -1.13670635e+00 1.52427793e-01 8.54247749e-01
2.72601455e-01 1.45310417e-01 -5.06288528e-01 -9.03820455e-01
1.96262568e-01 9.76551116e-01 -4.01215464e-01 -4.60385323e-01
-4.91691083e-01 -6.76801920e-01 3.88292819e-01 2.46481791e-01
5.38837135e-01 -1.20139599e+00 -7.62995958e-01 6.26665890e-01
2.78670996e-01 -7.54966915e-01 -2.93321103e-01 1.82900280e-01
-7.89932132e-01 -9.14468110e-01 -6.04444027e-01 -3.12388122e-01
1.82122171e-01 -4.39198405e-01 9.27767873e-01 -7.47282207e-01
4.40409392e-01 4.40862209e-01 -1.74456716e-01 -6.17041528e-01
-6.79028690e-01 -2.32548550e-01 5.29363990e-01 -3.86142097e-02
-3.99046868e-01 -2.93946326e-01 -6.29951119e-01 4.69154537e-01
-1.09222889e+00 -5.30175388e-01 7.03384042e-01 9.24477160e-01
9.83640850e-01 4.15813923e-01 1.15900695e+00 -5.75826108e-01
1.10307193e+00 -5.38457990e-01 -1.08999157e+00 3.94976616e-01
-9.52557623e-01 6.49032116e-01 1.04941034e+00 -6.04156911e-01
-1.21563470e+00 -2.59223431e-01 -1.15299430e-02 -6.97158515e-01
2.22690925e-01 2.90445834e-01 -3.19983482e-01 1.22964799e-01
5.36591709e-01 2.29656547e-01 1.12952158e-01 -5.38535379e-02
4.67340171e-01 5.98987639e-01 3.76364052e-01 -1.04546392e+00
8.54241014e-01 1.06105067e-01 2.72377014e-01 -1.58966199e-01
-8.01120281e-01 1.04065865e-01 -1.06257118e-01 -2.74685204e-01
3.68512511e-01 -7.04345584e-01 -1.22327054e+00 3.25456381e-01
-7.07044303e-01 -6.13620222e-01 -7.00208127e-01 5.53901613e-01
-1.39559793e+00 1.93840876e-01 -4.41954553e-01 -1.52652228e+00
-3.99957001e-01 -1.24004614e+00 6.65811121e-01 4.16531742e-01
5.27151942e-01 -9.21259642e-01 5.50156832e-01 -1.11686960e-01
1.28452688e-01 6.45963728e-01 6.27547145e-01 -5.79898834e-01
-3.49721491e-01 2.03087449e-01 3.45584810e-01 5.29633880e-01
-1.94305077e-01 3.87854315e-02 -6.09078586e-01 -7.66692817e-01
1.96691066e-01 -8.20619345e-01 5.24525940e-01 4.69778150e-01
1.03397381e+00 -7.84270942e-01 -1.28845070e-02 3.77596259e-01
1.59716761e+00 8.15067708e-01 3.71104538e-01 5.05444229e-01
2.69570768e-01 3.25604647e-01 1.24012303e+00 1.00364554e+00
2.42761880e-01 3.52023005e-01 1.02011096e+00 5.63205898e-01
8.56280863e-01 -4.70427126e-01 9.16345596e-01 4.33194906e-01
1.90018669e-01 -3.35720032e-01 -6.17989361e-01 1.59126878e-01
-2.33232522e+00 -8.79884481e-01 6.95550978e-01 2.53837466e+00
8.76461506e-01 4.12867010e-01 2.96403468e-01 -2.04833373e-02
6.78067029e-01 1.78889275e-01 -1.35890007e+00 -7.85009205e-01
7.90629070e-03 6.84590861e-02 5.14510810e-01 4.14321423e-01
-1.04312968e+00 7.66687214e-01 5.49808598e+00 1.20912659e+00
-8.59586596e-01 -1.38350606e-01 6.95634067e-01 -4.56389412e-02
-8.12853575e-02 -3.81666631e-01 -6.85930490e-01 5.60750902e-01
1.26384699e+00 -4.76555616e-01 7.31022894e-01 9.88474131e-01
7.09688842e-01 2.39670575e-02 -9.73395526e-01 8.59317362e-01
-4.51869994e-01 -7.45308459e-01 -2.74532467e-01 -2.48801857e-01
1.12632155e+00 -2.45360896e-01 2.87064433e-01 1.04984486e+00
8.93309593e-01 -1.00577426e+00 7.00772583e-01 6.29085958e-01
6.05170429e-01 -1.37891388e+00 9.28788245e-01 6.45975947e-01
-8.26235235e-01 -7.40019739e-01 -6.38306916e-01 1.94228366e-01
4.51575927e-02 2.74169266e-01 -5.54464221e-01 7.97391832e-01
2.53198475e-01 4.83879864e-01 -1.64734200e-02 9.42675412e-01
-4.43337083e-01 4.30481970e-01 -1.34807199e-01 -2.50435203e-01
7.26672530e-01 -3.36326748e-01 7.28348970e-01 7.00383067e-01
3.67013216e-01 -1.25243351e-01 5.73473573e-01 8.39199662e-01
1.53897822e-01 1.86013915e-02 -4.55180049e-01 1.08565383e-01
6.34117663e-01 1.08023071e+00 -3.58629555e-01 -2.30571792e-01
5.83396554e-02 5.06834209e-01 2.70771086e-01 4.34367150e-01
-9.67916310e-01 -2.44775623e-01 6.41058326e-01 -6.08499348e-01
2.00335741e-01 -3.73068973e-02 3.36442351e-01 -8.67493153e-01
-1.75132915e-01 -1.10929036e+00 5.41055679e-01 -4.46231484e-01
-1.46076119e+00 3.33246827e-01 7.79421404e-02 -1.44113386e+00
-9.29505944e-01 -3.39903057e-01 -4.57516104e-01 6.13273203e-01
-1.70296371e+00 -5.79982221e-01 3.41377348e-01 7.71710634e-01
5.44489861e-01 -2.07955420e-01 4.37296748e-01 -4.31874931e-01
-9.20643270e-01 4.69286144e-01 8.77101302e-01 -3.31469923e-01
6.91355824e-01 -1.58190918e+00 2.89713591e-02 6.85318828e-01
-4.58101124e-01 9.40849707e-02 8.50532830e-01 -4.72503006e-01
-1.58036959e+00 -1.50695539e+00 -4.04675007e-01 -1.98240206e-01
8.67286086e-01 -8.34294129e-03 -6.40490055e-01 4.83979791e-01
2.87472755e-02 8.86735544e-02 -1.58768028e-01 -5.63481510e-01
-2.53116675e-02 -4.54591513e-01 -1.45684934e+00 7.18426228e-01
6.19256020e-01 -9.59367305e-03 -5.47608018e-01 2.46531770e-01
1.15517879e+00 -6.80832565e-01 -1.00068200e+00 3.81016046e-01
3.21188778e-01 -6.94096208e-01 7.24485219e-01 -7.67606676e-01
2.41209604e-02 -2.59948462e-01 -1.07094288e-01 -1.96687198e+00
-5.28185181e-02 -1.43768156e+00 -7.42645085e-01 8.86300206e-01
1.85758583e-02 -7.45608449e-01 5.39818048e-01 2.37293541e-01
-1.43677562e-01 -1.01782537e+00 -1.35968542e+00 -1.34304774e+00
4.82920408e-01 -3.94553363e-01 7.78236091e-01 3.07690203e-01
8.02646577e-03 -1.18914671e-01 -6.65985584e-01 2.89316505e-01
9.19935167e-01 9.17161927e-02 4.86811399e-01 -7.69607842e-01
-5.39209187e-01 -3.70400876e-01 7.36944005e-02 -7.33816445e-01
5.06179333e-01 -2.19088033e-01 5.05652487e-01 -1.27795708e+00
-2.19160646e-01 -4.42083359e-01 -8.37994218e-01 1.42219082e-01
-3.17372680e-01 -6.46091700e-01 4.34004158e-01 2.60300115e-02
-1.11467171e+00 1.11327565e+00 1.71000791e+00 -2.13406831e-01
-5.63748360e-01 5.57799757e-01 -4.74446058e-01 4.41595495e-01
1.18748081e+00 -3.80463839e-01 -9.13426876e-01 1.10473491e-01
2.10942686e-01 7.70972073e-01 -6.25997037e-02 -1.06201601e+00
-1.54191837e-01 -8.23750734e-01 -3.76265384e-02 -4.68600959e-01
1.52611556e-02 -6.83196187e-01 -2.00746104e-01 8.34046900e-01
-5.09240925e-01 1.50117591e-01 9.95258689e-02 9.31559622e-01
-6.86275735e-02 -3.52852345e-01 9.32296038e-01 -3.36341709e-01
-6.55746639e-01 6.69651449e-01 -3.40571702e-01 5.65808475e-01
1.24064052e+00 3.55362356e-01 -3.53325546e-01 -5.92466295e-01
-4.26717758e-01 8.89435768e-01 -1.14961959e-01 1.59620777e-01
7.93338478e-01 -1.17293513e+00 -6.77743733e-01 -1.45089030e-01
-1.38747573e-01 1.53553456e-01 1.58642367e-01 5.51896095e-01
3.23324092e-02 4.51447189e-01 -2.21252531e-01 -2.73793280e-01
-3.53844523e-01 1.04694986e+00 6.96030378e-01 -7.65379906e-01
-2.53759116e-01 3.18352103e-01 -1.58450589e-01 -5.37326038e-01
4.59475696e-01 -5.03773570e-01 -1.39248863e-01 -6.50488064e-02
4.87201661e-01 7.44985759e-01 -1.58595622e-01 3.96207981e-02
-5.66661730e-02 2.49058157e-01 2.45178133e-01 -5.15470028e-01
1.09722507e+00 -1.83456719e-01 3.49411726e-01 6.68153465e-01
6.37005150e-01 -5.11843622e-01 -2.18273687e+00 -2.04297379e-01
1.01469480e-01 -3.85676064e-02 -1.12126082e-01 -8.53261948e-01
-9.83905375e-01 6.84281766e-01 8.48155618e-01 8.89119208e-02
9.49156404e-01 -4.96974677e-01 5.40222526e-01 7.21641600e-01
6.83861554e-01 -1.46151829e+00 3.20906907e-01 5.90890706e-01
8.96829128e-01 -1.22518826e+00 -9.33370739e-02 6.43401206e-01
-1.10991096e+00 1.02674329e+00 9.44167793e-01 -2.96088308e-01
4.12942231e-01 1.41018420e-01 -9.14966613e-02 4.03203011e-01
-1.13007641e+00 -2.35632673e-01 -2.57221997e-01 7.82043815e-01
-4.39869732e-01 2.92038620e-01 -2.50328451e-01 6.28778100e-01
1.40036747e-01 1.26676172e-01 6.33517563e-01 9.48448122e-01
-6.50905788e-01 -8.33687365e-01 -5.85505366e-01 2.25049779e-01
-3.03120792e-01 4.12851423e-01 2.91944295e-01 9.31910098e-01
-4.98299003e-02 1.10300446e+00 -1.41661584e-01 -1.53578803e-01
4.40310031e-01 -1.62177786e-01 3.47239435e-01 -2.40452111e-01
-3.97713810e-01 1.25517070e-01 -2.24536806e-01 -8.82761538e-01
-1.92457289e-01 -4.98199612e-01 -1.44898033e+00 -2.40691807e-02
3.11808381e-02 3.71562988e-01 4.58669633e-01 9.75870609e-01
-2.34738383e-02 5.65534294e-01 1.12565386e+00 -5.10041952e-01
-1.80857170e+00 -8.78687799e-01 -9.37472820e-01 -3.48164812e-02
5.39181054e-01 -7.73800015e-01 -3.23732972e-01 -6.32146955e-01] | [4.312176704406738, 2.366952657699585] |
3d530d18-0439-47a0-9249-7c08f21a6c03 | lamner-code-comment-generation-using | 2204.09654 | null | https://arxiv.org/abs/2204.09654v1 | https://arxiv.org/pdf/2204.09654v1.pdf | LAMNER: Code Comment Generation Using Character Language Model and Named Entity Recognition | Code comment generation is the task of generating a high-level natural language description for a given code method or function. Although researchers have been studying multiple ways to generate code comments automatically, previous work mainly considers representing a code token in its entirety semantics form only (e.g., a language model is used to learn the semantics of a code token), and additional code properties such as the tree structure of a code are included as an auxiliary input to the model. There are two limitations: 1) Learning the code token in its entirety form may not be able to capture information succinctly in source code, and 2) The code token does not contain additional syntactic information, inherently important in programming languages. In this paper, we present LAnguage Model and Named Entity Recognition (LAMNER), a code comment generator capable of encoding code constructs effectively and capturing the structural property of a code token. A character-level language model is used to learn the semantic representation to encode a code token. For the structural property of a token, a Named Entity Recognition model is trained to learn the different types of code tokens. These representations are then fed into an encoder-decoder architecture to generate code comments. We evaluate the generated comments from LAMNER and other baselines on a popular Java dataset with four commonly used metrics. Our results show that LAMNER is effective and improves over the best baseline model in BLEU-1, BLEU-2, BLEU-3, BLEU-4, ROUGE-L, METEOR, and CIDEr by 14.34%, 18.98%, 21.55%, 23.00%, 10.52%, 1.44%, and 25.86%, respectively. Additionally, we fused LAMNER's code representation with the baseline models, and the fused models consistently showed improvement over the non-fused models. The human evaluation further shows that LAMNER produces high-quality code comments. | ['Fatemeh Fard', 'Fuxiang Chen', 'Rishab Sharma'] | 2022-04-05 | null | null | null | null | ['code-comment-generation', 'comment-generation'] | ['computer-code', 'natural-language-processing'] | [ 9.36447158e-02 1.71943992e-01 -3.59158903e-01 -4.24879789e-01
-9.61690664e-01 -6.47276819e-01 5.45165122e-01 4.59004402e-01
-7.73243681e-02 4.84150618e-01 4.23050702e-01 -3.47268909e-01
7.34310091e-01 -7.89055347e-01 -7.45692611e-01 -1.00977384e-01
3.09931044e-03 -8.37909430e-02 1.97689429e-01 -1.05735406e-01
6.29755437e-01 -2.47648478e-01 -1.49592543e+00 7.31947124e-01
1.35314071e+00 5.06021619e-01 4.42163438e-01 8.02172482e-01
-7.68358827e-01 1.17304230e+00 -8.45097125e-01 -6.10899687e-01
-1.86163008e-01 -7.90216088e-01 -9.48532522e-01 -1.36923566e-01
9.47640389e-02 -1.98455110e-01 2.55063176e-01 1.10861874e+00
1.84479982e-01 -3.30849916e-01 8.41981053e-01 -1.23570704e+00
-8.49507511e-01 1.05746007e+00 -5.23003936e-01 -5.11155427e-01
7.40726411e-01 -1.26908496e-01 1.16190815e+00 -1.03388524e+00
7.42712438e-01 9.69152331e-01 5.74072361e-01 8.41441870e-01
-9.26532447e-01 -6.08717680e-01 -1.37469262e-01 -3.61822218e-01
-1.15628815e+00 -9.08925310e-02 4.38952506e-01 -8.45248640e-01
1.23250139e+00 4.43003736e-02 1.85298935e-01 8.40235651e-01
5.04428148e-01 7.85290956e-01 6.93406880e-01 -5.74218929e-01
1.85293809e-01 2.58680373e-01 1.92051694e-01 9.12995219e-01
3.24766427e-01 -2.72980809e-01 -1.29871204e-01 -3.40039670e-01
2.26202965e-01 5.67461066e-02 -2.15756781e-02 -4.59557995e-02
-1.17081773e+00 9.88831937e-01 4.05035496e-01 1.26530290e-01
-9.01707187e-02 3.46300185e-01 8.96210968e-01 1.86883152e-01
2.08577916e-01 5.19311011e-01 -4.73045200e-01 -5.83836377e-01
-9.28292871e-01 3.93450856e-01 1.01828647e+00 1.53748643e+00
1.10405838e+00 1.29241288e-01 -3.89475346e-01 8.89765143e-01
5.69154978e-01 6.27434492e-01 7.19970942e-01 -6.57733023e-01
7.73990750e-01 1.06471300e+00 1.46503523e-01 -7.40131259e-01
5.49296476e-02 -3.46895941e-02 -5.34315765e-01 5.77446967e-02
-1.14553727e-01 -2.62111217e-01 -9.18662667e-01 1.53834128e+00
-2.49388292e-01 -1.70413747e-01 4.03380334e-01 4.33505446e-01
1.17942226e+00 8.81835103e-01 1.24642096e-01 1.76226407e-01
1.28227782e+00 -1.22115660e+00 -4.81739730e-01 -3.51381361e-01
1.08196211e+00 -1.00555182e+00 1.13106298e+00 -1.16998106e-01
-8.02407146e-01 -5.45855820e-01 -8.78947377e-01 -3.09640989e-02
-2.22401038e-01 5.25892437e-01 4.39198732e-01 6.46591246e-01
-9.55762029e-01 3.67638506e-02 -5.96550882e-01 -2.66135931e-01
3.49851549e-02 -9.55671817e-02 -2.44180188e-01 -7.14067221e-02
-9.78365064e-01 5.40652394e-01 6.26797616e-01 -6.34494722e-01
-1.11110699e+00 -7.04316020e-01 -1.28143346e+00 2.90577829e-01
5.35430759e-02 -3.02195668e-01 1.60501266e+00 -1.24336958e+00
-1.17161942e+00 8.10017884e-01 -4.22741205e-01 -3.31461817e-01
7.17236847e-02 8.19515660e-02 -5.75829089e-01 -1.94471970e-01
3.71495396e-01 7.68188477e-01 3.63109171e-01 -1.37976599e+00
-6.65992320e-01 2.56093502e-01 3.67841184e-01 -1.68725342e-01
-1.76900074e-01 4.43968862e-01 -3.65291864e-01 -8.36125314e-01
-4.19555902e-01 -9.51636791e-01 -1.20292157e-01 -3.04053843e-01
-3.48664969e-01 -2.31384218e-01 5.44508398e-01 -8.05336237e-01
1.63977289e+00 -2.26493287e+00 -3.70821416e-01 -6.28366917e-02
-3.90191227e-02 3.27990383e-01 -3.46139222e-01 9.43616331e-01
-1.22692354e-01 5.53448260e-01 -6.92983508e-01 -7.82823861e-02
1.22676887e-01 2.65010614e-02 -3.78473490e-01 -1.64972380e-01
3.30053091e-01 8.72798026e-01 -1.04390872e+00 -4.00875717e-01
-3.34494799e-01 3.21491450e-01 -1.00501204e+00 4.82874930e-01
-4.99244213e-01 -2.13875651e-01 -4.15308088e-01 4.77652162e-01
5.46260834e-01 -3.28316063e-01 1.96662173e-02 2.93601871e-01
-1.73519447e-01 6.87870741e-01 -1.05631912e+00 1.85690594e+00
-9.61344004e-01 5.68954051e-01 -4.12538946e-01 -4.05650437e-01
1.27109456e+00 4.59228396e-01 1.87470042e-03 -4.51885045e-01
-2.56011695e-01 4.34986562e-01 -1.59423560e-01 -7.67084539e-01
7.67286479e-01 1.22510448e-01 -6.52326882e-01 7.30196953e-01
-2.64450740e-02 -7.78090116e-03 4.73030329e-01 5.13456404e-01
1.26160038e+00 1.20130382e-01 5.89420676e-01 -1.85346097e-01
6.54745996e-01 8.30255970e-02 5.45211494e-01 5.74055076e-01
1.88082591e-01 7.95470238e-01 6.62099600e-01 -1.68317840e-01
-1.12213480e+00 -8.60659242e-01 1.33708999e-01 1.03979957e+00
1.60926324e-03 -1.06916893e+00 -9.38098133e-01 -1.09906840e+00
-1.86584979e-01 9.16670144e-01 -4.76405025e-01 -2.34380573e-01
-5.65020740e-01 -3.52088124e-01 8.74243796e-01 5.51496446e-01
5.42538524e-01 -1.25398338e+00 -6.64926410e-01 1.94552511e-01
-4.93510693e-01 -7.90437758e-01 -7.98746109e-01 -5.18676750e-02
-5.87549329e-01 -9.97956932e-01 -5.15134692e-01 -1.07417965e+00
1.08164310e+00 3.03977728e-02 1.30278718e+00 5.12323916e-01
-1.50150023e-02 -1.00729480e-01 -9.04808760e-01 -3.45954090e-01
-1.16117144e+00 7.13413209e-02 -6.64184630e-01 -3.33832294e-01
3.94332677e-01 3.65342796e-02 -2.27827847e-01 -4.86111529e-02
-1.15673840e+00 1.77241713e-01 6.20559394e-01 8.38148415e-01
3.98438983e-02 -2.37166524e-01 5.23551881e-01 -1.24667382e+00
7.80049205e-01 -7.13421047e-01 -3.71257603e-01 2.99412936e-01
-6.56582296e-01 4.16243702e-01 9.96235490e-01 -1.23185486e-01
-1.32846141e+00 1.27866015e-01 -3.32279116e-01 3.27929974e-01
-1.41375557e-01 9.22681212e-01 -1.50484145e-01 3.79355431e-01
8.65416110e-01 3.00017953e-01 -2.07283393e-01 -3.22295487e-01
2.42997542e-01 1.22611237e+00 2.58819669e-01 -8.86602104e-01
7.51935959e-01 -1.76973809e-02 -6.57495856e-01 -2.17498034e-01
-4.52620029e-01 -5.00472367e-01 -3.18135262e-01 1.04178675e-01
7.61709988e-01 -1.15551972e+00 -1.70428991e-01 4.40160006e-01
-1.49248254e+00 -7.44099766e-02 -3.31337079e-02 2.80092418e-01
-4.24295664e-01 5.37878692e-01 -7.55011559e-01 -5.55964947e-01
-4.59497809e-01 -1.49180639e+00 1.12881899e+00 2.17045441e-01
-4.27244067e-01 -8.32562387e-01 1.76437527e-01 1.37216300e-01
4.94529158e-01 3.31761748e-01 1.16869700e+00 -6.10012889e-01
-3.91802132e-01 -1.96220979e-01 -3.22268218e-01 4.17561471e-01
3.26585352e-01 3.93152565e-01 -7.18944669e-01 -2.24510446e-01
-3.33438784e-01 -4.19662356e-01 5.25304973e-01 -3.26136142e-01
1.08134151e+00 -7.30161726e-01 -1.74239844e-01 3.92589688e-01
1.70684826e+00 4.37470257e-01 7.18703747e-01 5.68268374e-02
6.07290208e-01 3.70627612e-01 4.36967701e-01 6.11392081e-01
7.29182303e-01 4.33033317e-01 5.03219008e-01 2.01030433e-01
-1.07511975e-01 -6.11854076e-01 9.17261779e-01 1.29431725e+00
3.75580758e-01 -2.68092722e-01 -1.16739583e+00 7.80977130e-01
-1.61966085e+00 -9.75880265e-01 -4.25813556e-01 2.02738762e+00
1.23128140e+00 -7.78387189e-02 -2.54591964e-02 -1.41586900e-01
7.60221660e-01 -1.22950569e-01 -2.01522529e-01 -9.15833592e-01
2.05647260e-01 9.67178639e-05 4.09987152e-01 3.20903152e-01
-7.40357935e-01 7.74369776e-01 5.77450657e+00 5.72377503e-01
-1.08878541e+00 4.57678549e-02 1.67490482e-01 6.56761169e-01
-7.04454780e-01 4.65109020e-01 -7.88653672e-01 8.39403689e-01
1.27511811e+00 -6.77404642e-01 3.37505072e-01 1.18150568e+00
2.38208137e-02 9.32322163e-03 -1.20515287e+00 7.56005049e-01
3.96078944e-01 -1.26941669e+00 3.50263447e-01 -2.30861098e-01
1.11757088e+00 1.55543052e-02 -3.43914777e-01 6.01122737e-01
5.93799651e-01 -9.77631629e-01 9.08786118e-01 3.55494022e-01
9.72586453e-01 -6.88805819e-01 9.07994926e-01 4.41730380e-01
-1.30769968e+00 -8.85447580e-03 -3.47688019e-01 6.07002759e-03
-1.21225253e-01 5.71312666e-01 -1.18469024e+00 6.24598086e-01
3.34322095e-01 7.67481625e-01 -7.76242614e-01 1.15203142e+00
-5.00999331e-01 7.38818169e-01 3.05699229e-01 -4.13803071e-01
3.52435857e-01 4.04098094e-01 4.77754660e-02 1.81271029e+00
5.53207874e-01 -3.61540049e-01 3.38385403e-01 1.21274018e+00
-4.27322626e-01 2.95551509e-01 -4.77435201e-01 -2.11791277e-01
5.31883895e-01 1.06589413e+00 -4.63076979e-01 -5.70971251e-01
-7.81867981e-01 8.90667081e-01 1.24452874e-01 2.71629333e-01
-9.79703724e-01 -1.06382954e+00 4.91228104e-01 -6.11638948e-02
3.29003960e-01 1.39868809e-02 -1.52058125e-01 -1.33591974e+00
3.32603395e-01 -1.18212593e+00 1.92718774e-01 -8.65172267e-01
-8.78344536e-01 7.07689106e-01 -9.42484513e-02 -1.48671520e+00
-6.22492075e-01 -3.99201304e-01 -8.32766235e-01 1.01064527e+00
-1.46775174e+00 -9.55112517e-01 -3.90276998e-01 5.19927330e-02
7.45996714e-01 -3.40107143e-01 1.04486859e+00 3.45886588e-01
-2.48650193e-01 9.11852479e-01 1.09976627e-01 6.74084485e-01
6.39186442e-01 -1.37186658e+00 7.12223291e-01 1.19447434e+00
7.50002963e-03 1.06846738e+00 4.80492830e-01 -7.40788043e-01
-1.25762057e+00 -1.45438361e+00 1.32818651e+00 -3.30655634e-01
4.05423671e-01 -5.41002095e-01 -9.49102879e-01 6.04178727e-01
3.06335241e-01 -1.21031180e-01 7.81747758e-01 -3.42940778e-01
-7.29535878e-01 2.34175116e-01 -1.05857122e+00 4.22895968e-01
5.48255861e-01 -7.91119456e-01 -6.68272316e-01 2.44610701e-02
9.40651178e-01 -5.13665438e-01 -8.90712917e-01 4.12330255e-02
4.20356810e-01 -8.98356140e-01 3.65224004e-01 -4.53088343e-01
1.06978643e+00 -5.73970675e-01 -1.65548831e-01 -1.23253381e+00
-3.05243731e-02 -3.28652173e-01 5.04641645e-02 1.65296578e+00
7.99919128e-01 -2.34186232e-01 5.06059170e-01 4.80408967e-01
-3.46796125e-01 -5.07533312e-01 -3.81294698e-01 -7.49973893e-01
2.06960991e-01 -3.45953524e-01 1.00017524e+00 1.03512204e+00
3.81644607e-01 2.19717562e-01 -1.32664397e-01 -7.42901191e-02
1.32769227e-01 2.23606765e-01 8.38861108e-01 -9.05142367e-01
-2.63652861e-01 -3.06841999e-01 -2.14452744e-01 -1.05009210e+00
4.95677054e-01 -1.36216414e+00 3.23125660e-01 -1.80082417e+00
5.55643618e-01 -4.15746182e-01 1.57207865e-02 6.51368916e-01
-3.64414662e-01 -3.57634068e-01 1.76555991e-01 2.25778967e-01
-4.14373428e-01 3.13626289e-01 8.72660100e-01 -3.62170905e-01
1.26736701e-01 -1.39912531e-01 -7.68381238e-01 4.70678419e-01
8.62795353e-01 -8.16669405e-01 -5.24477839e-01 -6.91338241e-01
5.26828349e-01 2.09035203e-01 7.54115433e-02 -9.30100858e-01
1.48575068e-01 -2.53334761e-01 -1.92325175e-01 -2.08812028e-01
-3.40018988e-01 -5.49001873e-01 3.44294049e-02 6.52463675e-01
-6.13104761e-01 3.00278753e-01 -3.58417146e-02 1.50771856e-01
-4.99223202e-01 -9.00384545e-01 6.79490745e-01 -2.84190416e-01
-9.25030529e-01 3.02705765e-02 -4.25705254e-01 4.99254674e-01
9.44933534e-01 -1.04448855e-01 -4.59775746e-01 -1.50117233e-01
8.11986029e-02 1.70856267e-02 8.87317896e-01 8.90327036e-01
6.84020221e-01 -1.34020090e+00 -8.70205224e-01 2.79197991e-01
8.57959986e-01 -1.90293446e-01 -1.91074356e-01 1.30672574e-01
-8.71339858e-01 2.94744849e-01 -1.61739036e-01 -3.24972183e-01
-1.13521588e+00 4.26033944e-01 6.54993430e-02 -2.81852305e-01
-2.61505812e-01 5.66181123e-01 3.05497888e-02 -7.33142018e-01
-3.02577838e-02 -6.68694198e-01 -2.68343426e-02 -3.77517819e-01
6.08227909e-01 7.72332400e-02 5.38478605e-02 -7.32418299e-01
-4.44114476e-01 3.91071588e-01 -7.81897530e-02 2.52151877e-01
1.03661323e+00 2.49985680e-01 -4.63369012e-01 3.42644215e-01
1.58190310e+00 3.39583069e-01 -8.91322315e-01 -7.49231279e-02
3.72525424e-01 -3.47715020e-01 -5.71676373e-01 -9.37916756e-01
-8.75752091e-01 8.52068305e-01 8.50701854e-02 6.56988248e-02
8.45268250e-01 -1.04989126e-01 8.74055743e-01 2.92748094e-01
5.44228137e-01 -8.25665116e-01 6.74106777e-02 9.57796931e-01
7.14137912e-01 -1.21503341e+00 -4.69826818e-01 -2.96086520e-01
-7.61840045e-01 1.28372025e+00 8.36912572e-01 3.02033145e-02
3.36736232e-01 6.54960632e-01 2.22534999e-01 3.19201499e-02
-9.88861680e-01 1.51020318e-01 1.90533400e-01 4.12103862e-01
1.17172122e+00 2.96679419e-02 -4.96982723e-01 5.79365492e-01
-1.74800023e-01 -7.25005567e-02 1.07867348e+00 1.12468040e+00
-5.96152127e-01 -1.46898139e+00 -1.16971627e-01 4.86957490e-01
-6.75043106e-01 -4.15537596e-01 -3.47833931e-01 4.43187326e-01
1.54494867e-01 1.05399168e+00 -2.49977767e-01 -4.29713130e-01
2.44625002e-01 1.48305863e-01 -1.79864213e-01 -1.30068374e+00
-1.00156522e+00 -4.39360023e-01 8.09540227e-02 -2.34481856e-01
-3.04315031e-01 -3.48789424e-01 -1.91145897e+00 -1.88554406e-01
-2.88381487e-01 7.52323389e-01 7.09754169e-01 5.59215605e-01
5.55752993e-01 4.82841730e-01 6.38094664e-01 -9.66815725e-02
-4.65631187e-01 -8.35543513e-01 -1.32071048e-01 6.71675742e-01
3.72414440e-01 -1.07635885e-01 -3.02668095e-01 6.29230797e-01] | [7.678467750549316, 7.909031391143799] |
952c8673-c36d-497c-a66f-6b31e73b4c3e | financial-sentiment-analysis-using-finbert | 2306.02136 | null | https://arxiv.org/abs/2306.02136v1 | https://arxiv.org/pdf/2306.02136v1.pdf | Financial sentiment analysis using FinBERT with application in predicting stock movement | We apply sentiment analysis in financial context using FinBERT, and build a deep neural network model based on LSTM to predict the movement of financial market movement. We apply this model on stock news dataset, and compare its effectiveness to BERT, LSTM and classical ARIMA model. We find that sentiment is an effective factor in predicting market movement. We also propose several method to improve the model. | ['Andy Zeng', 'Tingsong Jiang'] | 2023-06-03 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-9.33077633e-01 -6.61782861e-01 -3.46761614e-01 -3.70378822e-01
1.62048429e-01 -5.86486340e-01 8.90054226e-01 -4.59373802e-01
-6.73615158e-01 7.35609412e-01 5.74749768e-01 -7.41842568e-01
1.39772296e-01 -1.34942794e+00 -5.88573813e-01 -1.41135484e-01
-3.72162133e-01 1.04270592e-01 3.38328481e-01 -8.28878999e-01
8.81609917e-01 -2.01450177e-02 -6.96179867e-01 6.61107659e-01
9.58226696e-02 1.47944450e+00 -9.84700099e-02 8.02625358e-01
-7.77643383e-01 2.06129313e+00 -1.05812812e+00 -4.35002185e-02
4.29946601e-01 -4.81004804e-01 -6.14547670e-01 -4.59881127e-01
-5.22230446e-01 -4.82817411e-01 -3.39080304e-01 9.27339196e-01
1.32233933e-01 4.30383295e-01 6.38269782e-01 -5.92985809e-01
-1.43607545e+00 1.44148695e+00 -7.22846746e-01 1.32621825e+00
-4.79625445e-03 -4.99563396e-01 6.21468067e-01 -7.09434807e-01
3.80866587e-01 1.17747450e+00 9.18396294e-01 -1.29516991e-02
-2.85527706e-01 -8.33942294e-01 6.46653533e-01 1.92585781e-01
-3.51102203e-01 2.36807819e-02 7.64094949e-01 -5.99282861e-01
1.18119180e+00 -1.32128760e-01 1.09654641e+00 9.91507411e-01
1.01979363e+00 9.64990914e-01 1.00707412e+00 -4.40227032e-01
8.54346603e-02 9.56043229e-02 5.38004339e-01 1.18176363e-01
-1.33441150e-01 6.83805496e-02 -6.77959979e-01 4.47484143e-02
8.39539826e-01 6.10191345e-01 2.67443031e-01 8.43131542e-01
-9.88728881e-01 1.28210258e+00 7.06823230e-01 8.03604722e-01
-8.02280068e-01 4.40285027e-01 6.50093317e-01 7.36383557e-01
1.12603819e+00 3.39256138e-01 -1.04723954e+00 -3.58721912e-01
-9.56666172e-01 4.45465088e-01 9.65907097e-01 3.07513088e-01
-4.48190682e-02 7.29831219e-01 2.10285589e-01 5.01718819e-01
4.53611761e-01 5.66541314e-01 1.55393898e+00 -2.50267327e-01
5.99172488e-02 4.74272549e-01 6.52206913e-02 -1.12989962e+00
-6.49158657e-01 -3.40024412e-01 -7.23367870e-01 1.68207332e-01
-9.40264612e-02 -8.54043305e-01 -8.02918732e-01 7.36356378e-01
-3.96210462e-01 5.21033585e-01 2.78575242e-01 2.94482887e-01
6.62498832e-01 1.16989183e+00 -1.84188709e-01 -3.41816872e-01
1.08477724e+00 -1.17459941e+00 -1.19481790e+00 -1.14439562e-01
3.84693563e-01 -6.30615354e-01 3.67150456e-01 5.58749557e-01
-9.06707585e-01 -5.11412501e-01 -8.42469037e-01 5.01562595e-01
-1.11907446e+00 -2.07146734e-01 8.79142284e-01 3.12047720e-01
-8.61011505e-01 7.87048042e-01 -9.53864694e-01 2.99507141e-01
1.01119315e-03 3.17228585e-01 4.33005899e-01 1.08943141e+00
-1.76132655e+00 1.00882077e+00 6.18460298e-01 5.46995163e-01
-1.47025511e-01 -1.96092561e-01 -6.66781962e-01 -2.72788286e-01
-1.28448531e-01 -2.19761923e-01 1.67010188e+00 -1.30068207e+00
-1.83006477e+00 -1.01425193e-01 8.69798809e-02 -1.21340013e+00
3.76039177e-01 -4.70065445e-01 -1.03064036e+00 -4.57999617e-01
-1.99730799e-01 -6.01909943e-02 5.56167126e-01 -2.49587938e-01
-1.10187066e+00 -7.72210956e-03 4.23082039e-02 -1.86262637e-01
-4.63632971e-01 5.38357854e-01 3.65188658e-01 -1.07870889e+00
-1.92646638e-01 -5.03083467e-01 -3.08880001e-01 -1.15100181e+00
-2.17350181e-02 -5.45424283e-01 8.80699754e-01 -8.58439803e-01
1.48609650e+00 -1.70451355e+00 -6.08041406e-01 2.63392746e-01
-2.40027979e-01 8.06738287e-02 3.02312136e-01 4.37217653e-01
8.16278718e-03 3.20679843e-01 2.31483892e-01 1.02325991e-01
-4.18472737e-02 1.20324098e-01 -1.06241620e+00 8.50049406e-02
-1.46317691e-01 1.24909377e+00 -4.14055258e-01 1.16242386e-01
6.39563426e-02 2.02438638e-01 -1.37720391e-01 -2.46587992e-01
-4.39239562e-01 1.67771205e-02 -7.03492284e-01 4.38960731e-01
4.99095529e-01 -3.42499882e-01 -2.79625714e-01 3.11187834e-01
-6.28553689e-01 3.69120508e-01 -8.78280461e-01 8.16495538e-01
-1.72087207e-01 1.08631492e+00 -8.95548463e-01 -1.07915223e+00
1.27565265e+00 2.86872000e-01 2.13894129e-01 -8.77992630e-01
4.55066115e-01 3.98420811e-01 1.82187960e-01 -5.94632745e-01
7.62366652e-01 -3.94531071e-01 -6.18141703e-02 5.95200419e-01
-2.73219287e-01 4.73738372e-01 9.01214629e-02 -3.31061453e-01
6.42063260e-01 1.16541006e-01 1.55187637e-01 -4.33955580e-01
5.25193632e-01 6.73611835e-02 5.49613118e-01 7.34230161e-01
-3.60935479e-02 -2.09817246e-01 6.43016636e-01 -1.29700875e+00
-5.35684705e-01 -3.34393591e-01 -1.62500143e-01 1.26725996e+00
-1.20044366e-01 2.38280624e-01 -2.94230223e-01 -6.28089070e-01
9.20401663e-02 8.96271884e-01 -1.07014024e+00 2.74650574e-01
-5.87651968e-01 -1.21139920e+00 1.80541664e-01 9.51738358e-01
8.43961895e-01 -1.69400561e+00 -6.28385186e-01 5.62020898e-01
3.85685802e-01 -6.35910273e-01 -1.22965917e-01 2.04424858e-01
-7.68728554e-01 -7.57299840e-01 -6.40564740e-01 -7.17041373e-01
-1.79132566e-01 -3.11730236e-01 9.32642937e-01 -3.13113779e-01
5.43970585e-01 -1.21584728e-01 -5.48659325e-01 -1.20897102e+00
-8.24735388e-02 9.92521644e-02 7.82675669e-02 -2.27803081e-01
7.62729645e-01 -2.00355858e-01 -6.91283584e-01 -4.68627065e-02
-9.17452037e-01 -5.02239347e-01 7.40597397e-02 5.85279942e-01
8.79708901e-02 3.88490349e-01 1.02521098e+00 -1.00339472e+00
1.31332672e+00 -8.28565240e-01 -7.56838202e-01 -5.24766371e-02
-8.01384568e-01 -6.27719387e-02 6.12594903e-01 -5.38199186e-01
-1.06383157e+00 -5.69418609e-01 -8.93302783e-02 -4.32417803e-02
5.62766612e-01 1.43729353e+00 1.12543750e+00 1.23452112e-01
1.31907195e-01 3.83592188e-01 -1.34731978e-01 -5.65110505e-01
1.87077578e-02 5.16264617e-01 1.53225273e-01 3.05768818e-01
1.30968899e-01 3.78908813e-01 -5.65868974e-01 -5.45337319e-01
-9.11647022e-01 -6.17025327e-03 -4.86213505e-01 -3.38204950e-01
8.61661553e-01 -8.11421096e-01 -9.11055744e-01 9.85884964e-01
-9.71524596e-01 -3.88266087e-01 -2.35558441e-03 9.60582733e-01
-1.29230365e-01 -2.71655113e-01 -1.53039229e+00 -1.31023347e+00
-4.93107319e-01 -6.19777322e-01 2.79029220e-01 2.84038752e-01
-4.52080369e-02 -1.82681239e+00 6.33411705e-01 -2.87975609e-01
9.08099353e-01 4.49463546e-01 2.00062037e-01 -1.28602040e+00
-1.69068307e-01 -5.68164468e-01 1.25630692e-01 5.43804526e-01
9.27498639e-02 3.25498521e-01 -6.84008777e-01 1.66583031e-01
4.96141046e-01 5.51193133e-02 1.32759786e+00 9.98293400e-01
6.33342445e-01 -5.14236808e-01 -7.16090351e-02 3.79894078e-01
1.16787648e+00 1.14394498e+00 4.07035202e-01 1.24806023e+00
4.37281996e-01 2.75502861e-01 5.37240744e-01 4.82052505e-01
6.42444849e-01 -2.40514949e-01 1.11805927e-02 5.69415092e-02
9.39442515e-01 -2.25382019e-02 8.94580483e-01 1.55003607e+00
-3.34740013e-01 -2.32618719e-01 -8.52940738e-01 1.07811414e-01
-2.23347163e+00 -1.25384688e+00 -3.30547154e-01 1.14843869e+00
3.28588784e-01 7.26500571e-01 5.42359650e-01 -8.89643133e-02
5.11122584e-01 3.65010679e-01 -4.45365369e-01 -6.19511008e-01
-3.99976343e-01 1.91188395e-01 9.99522388e-01 4.57849681e-01
-1.44516230e+00 1.18007696e+00 8.27596760e+00 3.32309008e-01
-1.69543004e+00 -6.73633516e-02 8.68646801e-01 1.34607498e-02
-2.44661853e-01 -3.52444470e-01 -8.22056592e-01 8.81317019e-01
1.37063825e+00 -4.81484711e-01 -1.02316633e-01 8.77801299e-01
3.05937707e-01 1.01226019e-02 -4.36033428e-01 3.90224367e-01
7.58626163e-02 -1.70598471e+00 1.07296124e-01 1.22708991e-01
8.13668847e-01 2.58903414e-01 2.85974026e-01 5.58547020e-01
6.64581597e-01 -8.26674640e-01 8.03775787e-01 1.30262768e+00
-5.31414509e-01 -1.04836690e+00 1.38069987e+00 2.70601451e-01
-1.12589240e+00 -3.18696648e-01 -6.61807835e-01 -9.36398327e-01
3.30049396e-01 2.15671897e-01 -4.00053203e-01 3.32840979e-01
9.27226841e-01 1.29773450e+00 -2.00102642e-01 5.33517838e-01
-4.25303951e-02 8.06666553e-01 -2.07428023e-01 -6.46004498e-01
9.00683045e-01 -2.58478910e-01 -6.67997915e-03 1.05076003e+00
5.42843640e-01 -1.22992764e-03 6.15641363e-02 4.74455059e-01
3.69148880e-01 3.15650284e-01 -6.78788185e-01 -1.66322172e-01
1.34521633e-01 6.16303205e-01 -1.04656398e+00 -7.44898975e-01
-7.71332085e-01 5.89246213e-01 -3.76885712e-01 3.60899061e-01
-8.04789841e-01 -4.75232422e-01 2.33902171e-01 -4.50166613e-01
6.15173340e-01 -1.37371272e-01 -3.37807149e-01 -1.17460918e+00
-1.05543949e-01 -4.23885822e-01 6.41439259e-01 -7.71267593e-01
-1.39981151e+00 8.98257077e-01 -3.13662201e-01 -1.16506636e+00
-8.18692803e-01 -1.00343585e+00 -1.17758060e+00 7.50864804e-01
-1.62051547e+00 -6.81226552e-01 6.16638362e-01 6.10926986e-01
7.23893523e-01 -9.41252112e-01 5.02186894e-01 5.69174178e-02
-7.83219993e-01 -1.23843119e-01 5.83437741e-01 5.69608986e-01
1.93413302e-01 -1.36844099e+00 1.09060955e+00 7.25757062e-01
1.82154663e-02 5.32345533e-01 5.62096000e-01 -7.88088977e-01
-7.85865486e-01 -8.14968467e-01 6.68071151e-01 -5.25579989e-01
1.53551555e+00 1.45015806e-01 -7.59831429e-01 1.30846667e+00
7.31044292e-01 -5.37312925e-01 6.96631074e-01 5.33599630e-02
1.63945913e-01 -1.54113127e-02 -6.16093159e-01 3.00831616e-01
1.89879891e-02 -1.99189126e-01 -1.31131172e+00 6.73666447e-02
1.16835558e+00 -1.14713646e-01 -1.04558504e+00 2.93304652e-01
7.44400144e-01 -9.95095611e-01 6.64873958e-01 -7.33962774e-01
6.12145782e-01 9.31665450e-02 1.71505257e-01 -1.50419450e+00
-4.28742468e-01 -3.57606262e-01 -1.42729014e-01 9.82024550e-01
8.95454049e-01 -1.25326896e+00 6.13592982e-01 5.97045422e-01
2.30063111e-01 -4.08269107e-01 -4.87096608e-01 -7.27808714e-01
4.12018687e-01 -4.44326788e-01 9.05358434e-01 1.38340926e+00
4.17809099e-01 2.21209973e-01 -7.02876687e-01 -3.28526914e-01
-2.82520145e-01 7.83015490e-02 3.60099107e-01 -1.14151275e+00
-1.67596444e-01 -9.09505665e-01 -1.75619721e-01 -9.60770428e-01
2.23579958e-01 -2.48321623e-01 -3.59217435e-01 -1.44130933e+00
-4.93323177e-01 3.82191837e-01 -1.12849069e+00 -1.16787627e-02
2.71657467e-01 -1.00465395e-01 1.92109235e-02 3.60124379e-01
-2.53703743e-01 3.14422041e-01 1.06731665e+00 -4.44976687e-02
-4.76526320e-01 3.23766321e-01 -5.59190094e-01 9.92030799e-01
1.04471684e+00 -1.11936115e-01 9.38619394e-03 -1.82751268e-01
8.23053181e-01 1.92050964e-01 -2.02322289e-01 -6.29318118e-01
4.98463154e-01 -2.91377872e-01 9.20269191e-01 -1.04433846e+00
-1.24964453e-01 -4.33188826e-01 -2.51938403e-01 9.33842361e-01
-6.34975076e-01 9.46261764e-01 4.49150473e-01 3.40833575e-01
-8.73422265e-01 -2.92782605e-01 5.21843508e-02 -4.05324548e-01
-9.80268300e-01 2.63974041e-01 -8.23174655e-01 -2.59294033e-01
7.19126999e-01 -9.33024101e-03 -3.25903505e-01 -6.54908299e-01
-6.57656372e-01 4.52645391e-01 -5.01922965e-01 8.26529324e-01
3.86604786e-01 -1.33542061e+00 -5.45474350e-01 2.14232117e-01
-6.08142972e-01 -5.18419147e-01 4.33106732e-04 5.52875400e-01
-1.03919435e+00 5.76520085e-01 -2.38863185e-01 1.23854233e-02
-5.54367125e-01 5.60550570e-01 6.71232522e-01 -1.73943073e-01
-5.69772422e-01 9.93181705e-01 -3.57943699e-02 -1.32275984e-01
4.29823361e-02 -9.12615299e-01 -1.11936688e+00 6.41004503e-01
9.68746841e-01 4.38702345e-01 -4.15918589e-01 -2.36838847e-01
-1.87998846e-01 5.29522598e-01 -2.53014296e-01 -3.22290033e-01
1.85754275e+00 -1.82921924e-02 -4.83946085e-01 1.48797667e+00
8.17953587e-01 -1.48214310e-01 -8.31912220e-01 -6.82232007e-02
7.63794839e-01 -8.10434893e-02 2.12348655e-01 -5.00948548e-01
-1.27495468e+00 3.76889408e-01 4.35700357e-01 1.29898393e+00
8.30593407e-01 -5.09932160e-01 1.14014971e+00 6.40015006e-01
-1.11351058e-01 -1.60070646e+00 -1.23004451e-01 1.20446801e+00
7.09969878e-01 -1.11912513e+00 -2.71107763e-01 6.30438745e-01
-7.78697670e-01 1.65435827e+00 3.34774613e-01 -1.11669767e+00
1.64634013e+00 5.64176321e-01 7.57095635e-01 -3.12381804e-01
-1.11733508e+00 1.40778586e-01 -4.61091176e-02 -1.04452208e-01
8.32596183e-01 1.19810365e-01 -2.64259636e-01 8.16564381e-01
-9.06860530e-01 3.60783935e-01 6.90058827e-01 9.10770059e-01
-6.93777204e-01 -6.82853341e-01 -2.38806516e-01 6.51620388e-01
-1.20494676e+00 -2.31495664e-01 -3.80089432e-01 9.55807507e-01
-3.33367407e-01 6.48505092e-01 9.57103372e-01 -5.90567410e-01
4.09083992e-01 2.09271669e-01 -4.27689165e-01 -1.34484366e-01
-1.04474771e+00 5.21485507e-01 -1.67382210e-01 -1.39790297e-01
-7.39624143e-01 -6.44042671e-01 -1.14110589e+00 -6.33210659e-01
-2.37304881e-01 2.37199396e-01 4.64785278e-01 1.16212296e+00
-1.92195088e-01 9.83347535e-01 9.78370309e-01 -4.84570414e-01
-5.01655936e-01 -1.46789432e+00 -1.12508893e+00 -1.71446800e-01
7.17723966e-01 -3.49908113e-01 -4.37742800e-01 -7.08737150e-02] | [4.448798656463623, 4.246963977813721] |
abd71044-e2a9-4f9d-a4aa-92c318640d4a | the-invertible-u-net-for-optical-flow-free | 2103.09576 | null | https://arxiv.org/abs/2103.09576v3 | https://arxiv.org/pdf/2103.09576v3.pdf | The U-Net based GLOW for Optical-Flow-free Video Interframe Generation | Video frame interpolation is the task of creating an interframe between two adjacent frames along the time axis. So, instead of simply averaging two adjacent frames to create an intermediate image, this operation should maintain semantic continuity with the adjacent frames. Most conventional methods use optical flow, and various tools such as occlusion handling and object smoothing are indispensable. Since the use of these various tools leads to complex problems, we tried to tackle the video interframe generation problem without using problematic optical flow . To enable this , we have tried to use a deep neural network with an invertible structure, and developed an U-Net based Generative Flow which is a modified normalizing flow. In addition, we propose a learning method with a new consistency loss in the latent space to maintain semantic temporal consistency between frames. The resolution of the generated image is guaranteed to be identical to that of the original images by using an invertible network. Furthermore, as it is not a random image like the ones by generative models, our network guarantees stable outputs without flicker. Through experiments, we \sam {confirmed the feasibility of the proposed algorithm and would like to suggest the U-Net based Generative Flow as a new possibility for baseline in video frame interpolation. This paper is meaningful in that it is the world's first attempt to use invertible networks instead of optical flows for video interpolation. | ['Donghoon Han', 'Nojun Kwak', 'Saem Park'] | 2021-03-17 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 1.47713795e-01 -6.74858093e-02 1.44246835e-02 -1.39121577e-01
1.38916567e-01 -1.99351966e-01 5.85324347e-01 -5.83557904e-01
-3.52948487e-01 1.08824646e+00 -2.86098979e-02 -1.59374654e-01
1.76736280e-01 -1.02700281e+00 -8.25420260e-01 -7.38118112e-01
4.12366800e-02 -7.72370547e-02 5.06786346e-01 -1.56887785e-01
3.61873627e-01 5.42123556e-01 -1.56668472e+00 6.49083257e-02
9.92506564e-01 8.23226988e-01 3.01832050e-01 5.75926185e-01
-4.51241732e-01 1.06590700e+00 -6.13731027e-01 -3.41908514e-01
4.86350566e-01 -9.76888299e-01 -7.47570217e-01 1.67266577e-01
5.63211501e-01 -6.64124191e-01 -5.16718984e-01 1.23166466e+00
2.15460181e-01 4.17147756e-01 5.05528808e-01 -1.44677556e+00
-9.40496206e-01 3.49035412e-01 -4.10575658e-01 2.19929442e-01
4.46111083e-01 1.84916317e-01 4.53567535e-01 -5.42430401e-01
8.86822045e-01 1.22799814e+00 7.37874210e-01 6.57569766e-01
-1.23148477e+00 -7.52962828e-01 -2.06569005e-02 3.93011510e-01
-1.26205349e+00 -3.79798025e-01 9.32146549e-01 -3.90536934e-01
4.30665761e-01 2.82598615e-01 8.86764228e-01 1.06928575e+00
3.67299318e-01 4.00879383e-01 9.62535560e-01 -5.98478258e-01
-1.16781987e-01 7.57427290e-02 -1.53632835e-01 6.76461279e-01
5.08725643e-02 -1.23128071e-02 -2.00889841e-01 2.59136260e-01
1.39508832e+00 1.32079795e-01 -6.17191613e-01 -1.29080385e-01
-1.32083428e+00 7.11220145e-01 5.40793657e-01 6.00929558e-01
-8.73456299e-02 3.11203837e-01 8.66552740e-02 1.29116535e-01
4.24401641e-01 2.14716122e-01 1.01654939e-01 6.65157065e-02
-1.21685743e+00 2.26350039e-01 7.02621818e-01 8.92278135e-01
8.65597069e-01 4.05073702e-01 -1.61641255e-01 4.65178430e-01
2.82733560e-01 2.99687952e-01 5.74801445e-01 -1.43537879e+00
2.60038495e-01 1.90303206e-01 2.06033498e-01 -1.35135317e+00
-3.26617435e-02 1.56524999e-03 -1.23020566e+00 6.22639000e-01
7.82646537e-01 -1.78520173e-01 -7.14783132e-01 1.82105863e+00
1.01861872e-01 5.96072316e-01 -2.81623900e-01 1.00175738e+00
6.36867762e-01 9.50213850e-01 -1.46592721e-01 -5.56772292e-01
8.33402812e-01 -7.78677166e-01 -1.24280238e+00 4.42932218e-01
1.90738946e-01 -1.05435383e+00 9.73113418e-01 3.67873341e-01
-1.31543934e+00 -9.83290493e-01 -1.01368284e+00 -3.00359458e-01
-1.20566070e-01 -1.71624839e-01 6.05537057e-01 5.64068139e-01
-1.28994405e+00 9.09191072e-01 -6.86465144e-01 -1.75785378e-01
-3.32278647e-02 1.70953840e-01 -4.02387828e-01 4.52845991e-01
-1.32199514e+00 7.45161355e-01 5.45308828e-01 4.36620682e-01
-4.49644804e-01 -3.81704450e-01 -7.54964888e-01 -5.35184406e-02
2.29941532e-02 -9.19947386e-01 7.06691563e-01 -1.51876175e+00
-1.66580248e+00 3.21514457e-01 -5.04393160e-01 -3.69575053e-01
9.76147234e-01 -1.01630256e-01 -3.74388397e-01 3.28263529e-02
9.60652083e-02 7.21036196e-01 1.06135011e+00 -1.23900962e+00
-6.64090991e-01 2.96083719e-01 2.29741856e-01 -1.60751030e-01
-2.77152240e-01 -1.08340792e-01 -2.85862058e-01 -9.66254592e-01
2.05598533e-01 -8.07580292e-01 -3.98050360e-02 2.47632205e-01
-1.46705002e-01 -4.12531337e-03 1.12035084e+00 -8.82773459e-01
1.37738788e+00 -2.03278995e+00 7.85345361e-02 1.28386959e-01
1.49713069e-01 3.27471346e-01 4.83579785e-02 9.22658443e-02
-3.55585426e-01 1.87900394e-01 -2.41192713e-01 -3.14820498e-01
-3.99607688e-01 2.23802298e-01 -4.05913234e-01 3.98289174e-01
1.04119718e-01 5.31233847e-01 -9.38238144e-01 -6.64324820e-01
4.04220134e-01 8.69969785e-01 -7.01935232e-01 1.68375686e-01
-2.22142830e-01 7.96396077e-01 8.99979845e-03 7.89018199e-02
8.34359884e-01 9.85579938e-02 -2.12935343e-01 -4.30879921e-01
-3.36763084e-01 -3.72900479e-02 -1.46634138e+00 1.63370550e+00
-3.58873367e-01 8.26163352e-01 -2.06554100e-01 -8.42187226e-01
1.02231765e+00 4.64028209e-01 6.13809586e-01 -5.34510851e-01
1.62770569e-01 1.69052511e-01 -1.47995502e-01 -6.71033859e-01
4.70094711e-01 -1.22478515e-01 5.85697830e-01 1.83820948e-01
1.27249556e-02 -1.22505486e-01 3.82364631e-01 2.77664550e-02
5.38611233e-01 5.87332129e-01 -3.23547050e-02 -1.66063964e-01
8.85195553e-01 -4.11412984e-01 7.81084299e-01 5.36687076e-01
-8.36990550e-02 9.20563638e-01 4.33894783e-01 -8.46012056e-01
-1.25897539e+00 -1.07977462e+00 -1.29317254e-01 2.41879642e-01
3.44298571e-01 -1.38594076e-01 -8.51882279e-01 -4.14616019e-01
-4.93717134e-01 5.08817315e-01 -4.09533411e-01 1.38468444e-01
-9.11125660e-01 -5.35622656e-01 3.74318987e-01 3.15425754e-01
9.33394372e-01 -1.12329209e+00 -1.74701944e-01 3.61505926e-01
-5.89882791e-01 -1.04945529e+00 -7.94866085e-01 -1.97107598e-01
-9.31145489e-01 -9.86983895e-01 -9.00115430e-01 -9.76970375e-01
7.05255508e-01 2.57361352e-01 9.23388898e-01 2.03251943e-01
1.42567549e-02 -3.61958683e-01 -2.66016543e-01 1.73541859e-01
-5.80904186e-01 -2.52622366e-01 2.72762142e-02 3.40790093e-01
-8.98086354e-02 -7.97646463e-01 -6.43574238e-01 2.59947181e-01
-1.24989259e+00 2.83688396e-01 -8.00590366e-02 8.51604283e-01
4.49577242e-01 1.68918446e-01 2.80170768e-01 -5.33217132e-01
5.89635551e-01 -9.62938517e-02 -6.34088159e-01 1.03573933e-01
-1.50033653e-01 1.43243879e-01 9.47594047e-01 -4.61884588e-01
-1.31182778e+00 5.90738207e-02 -3.49662393e-01 -6.17126107e-01
-1.07643558e-02 -1.26555651e-01 -7.12499842e-02 -1.85843915e-01
5.55909693e-01 2.07958937e-01 1.64321586e-01 -2.43841767e-01
4.10009265e-01 4.49112535e-01 6.05654240e-01 -2.87991703e-01
9.08221543e-01 5.72990537e-01 2.17858553e-01 -9.85940158e-01
-4.19421285e-01 5.74691817e-02 -6.89840555e-01 -3.64934713e-01
1.22787106e+00 -6.17871702e-01 -9.27579880e-01 5.45513272e-01
-1.66139758e+00 -1.16535425e-01 -3.34395945e-01 6.46660268e-01
-5.50725102e-01 7.90498078e-01 -9.00760174e-01 -7.09642172e-01
1.97864510e-02 -1.35505927e+00 5.02243161e-01 4.50513870e-01
-8.16867277e-02 -1.15458632e+00 3.45006473e-02 -6.63348520e-03
4.15417045e-01 3.81027848e-01 4.43121791e-01 2.22641110e-01
-9.84519362e-01 1.07307754e-01 -2.93508261e-01 6.54635429e-01
5.61987042e-01 5.83596647e-01 -8.89296591e-01 -1.36508986e-01
3.57062846e-01 3.29203665e-01 8.28251600e-01 5.37565410e-01
1.33341205e+00 -4.12830174e-01 8.62053111e-02 1.05125833e+00
1.42866492e+00 5.93591273e-01 1.14765751e+00 3.17076415e-01
8.93868327e-01 3.59849215e-01 2.35333979e-01 1.94059521e-01
8.64567980e-02 7.51137614e-01 3.59626263e-01 -1.95405081e-01
-4.30216998e-01 -2.26962984e-01 3.09489876e-01 7.94874668e-01
-7.52223492e-01 -2.16773197e-01 -3.17060262e-01 2.19483048e-01
-1.85306036e+00 -1.55396807e+00 -4.40118253e-01 2.33806562e+00
7.53801346e-01 1.31833732e-01 2.01662257e-02 1.89529777e-01
8.58777404e-01 2.05877006e-01 -3.18948994e-03 -5.00177085e-01
-2.06386372e-01 1.94305107e-01 4.51325029e-01 8.79628122e-01
-1.09558344e+00 7.58034766e-01 6.18390799e+00 9.21983719e-01
-1.41283774e+00 1.21528171e-01 6.74473882e-01 2.20587254e-01
-4.57243562e-01 9.33108106e-02 -8.03985536e-01 8.82822275e-01
5.55197001e-01 6.00515492e-02 4.93047923e-01 4.05921638e-01
6.43537700e-01 -2.07120478e-01 -7.36001968e-01 9.52570915e-01
3.12032271e-03 -1.44555533e+00 2.59890318e-01 -2.35928506e-01
8.00575733e-01 -6.96072161e-01 -1.55219808e-01 -2.04302773e-01
-6.75181765e-03 -8.79379034e-01 6.67600632e-01 8.53630006e-01
6.77510798e-01 -6.40155494e-01 6.50816917e-01 1.79045290e-01
-1.21380579e+00 2.86224186e-01 -3.80612433e-01 -2.28695825e-01
4.48840976e-01 5.96300066e-01 -4.73911583e-01 5.65808594e-01
6.13886952e-01 6.99411511e-01 -2.69242972e-01 1.15434325e+00
-1.75431609e-01 1.94760934e-01 -1.67414948e-01 2.74113894e-01
8.05461630e-02 -7.01452553e-01 4.91431892e-01 1.12010825e+00
6.03132188e-01 -1.53383061e-01 -7.08042702e-04 1.05089080e+00
6.85584247e-02 5.63293695e-02 -8.78394783e-01 3.85258824e-01
2.69818883e-02 9.42723036e-01 -7.75939286e-01 -4.28796798e-01
-4.71185714e-01 1.24001932e+00 -2.24857986e-01 5.75336695e-01
-1.19413137e+00 -4.94160146e-01 5.14887452e-01 2.77077168e-01
-1.03806369e-01 -4.06709164e-01 -1.59142882e-01 -1.41617572e+00
1.43579677e-01 -5.56818783e-01 -1.12453595e-01 -8.75478625e-01
-1.08886743e+00 8.50681365e-01 -4.28314880e-02 -1.70335603e+00
-4.49238151e-01 -2.93639630e-01 -6.41152024e-01 1.12316120e+00
-1.46500289e+00 -7.70606339e-01 -4.50586587e-01 8.75375211e-01
6.07788086e-01 3.18975598e-02 4.04161781e-01 5.56937039e-01
-3.01760644e-01 3.13948750e-01 -1.15842139e-02 1.80178002e-01
9.47067142e-01 -1.02833605e+00 1.80572987e-01 1.29613984e+00
3.44236642e-02 6.24498427e-01 8.46782267e-01 -6.42336011e-01
-8.93896997e-01 -9.13324952e-01 1.01973414e+00 -9.17635113e-02
5.75489998e-01 -1.00294603e-02 -1.04998899e+00 7.49443948e-01
5.00664294e-01 3.44407968e-02 3.13027836e-02 -7.13662088e-01
1.67209759e-01 -4.01467919e-01 -1.07359719e+00 7.45015979e-01
1.07709289e+00 -3.20964426e-01 -4.79429454e-01 2.27230087e-01
9.01156843e-01 -3.61730933e-01 -7.03077316e-01 3.66239965e-01
4.93248433e-01 -1.45756435e+00 1.12294805e+00 -1.16846040e-01
5.80078840e-01 -7.38842666e-01 1.84051484e-01 -1.07751334e+00
-4.29989100e-01 -7.89413929e-01 1.87404811e-01 1.34256840e+00
-7.45023936e-02 -6.89758897e-01 7.67845273e-01 5.52960992e-01
-3.05542741e-02 -1.60808817e-01 -7.43102849e-01 -7.69898593e-01
-1.25473723e-01 -1.80630878e-01 3.26530129e-01 1.11816132e+00
-3.81193280e-01 -8.94552767e-02 -7.58214951e-01 2.40097363e-02
7.29393184e-01 -1.23393729e-01 8.01890612e-01 -1.06775451e+00
-3.00023377e-01 -4.18669760e-01 -3.42625737e-01 -1.11384881e+00
2.20085278e-01 -4.72605199e-01 -2.18495000e-02 -1.37392628e+00
-2.85370022e-01 -4.54824984e-01 -7.23208941e-05 -4.09296155e-02
-7.72598237e-02 4.39255893e-01 3.98281634e-01 2.60528088e-01
1.56230435e-01 4.86332506e-01 1.84103870e+00 2.44577564e-02
-3.91398162e-01 8.50813538e-02 -6.67402893e-02 8.54557216e-01
6.43949687e-01 -1.11035280e-01 -5.29761016e-01 -3.70086282e-01
-1.26440585e-01 3.50070447e-01 3.70937675e-01 -1.26127744e+00
1.43262014e-01 -2.27547631e-01 5.96818805e-01 -3.83071274e-01
3.47259820e-01 -1.08215141e+00 7.08903015e-01 5.37329793e-01
-7.57110491e-02 1.57291219e-01 -1.12458289e-01 1.66929439e-01
-5.08233964e-01 -4.78107631e-01 9.17488277e-01 -3.54880720e-01
-7.59175420e-01 2.56546229e-01 -2.27797791e-01 -3.37932914e-01
9.63357747e-01 -6.12755597e-01 -1.96565166e-01 -4.92521942e-01
-8.44170332e-01 -2.99417645e-01 5.73213935e-01 2.79748797e-01
5.46826839e-01 -1.37456501e+00 -3.15531760e-01 4.08979863e-01
-6.56579971e-01 -2.41769087e-02 2.02171132e-01 6.68145239e-01
-1.22370923e+00 2.70067215e-01 -6.83559120e-01 -4.59450334e-01
-1.04006755e+00 6.87173665e-01 3.83994550e-01 6.14530146e-02
-6.61684692e-01 4.93211538e-01 3.05432945e-01 1.15594499e-01
2.05864176e-01 -5.33119857e-01 -2.97863215e-01 -1.84520837e-02
5.45606911e-01 3.51974159e-01 -3.06006700e-01 -6.01019561e-01
9.20774564e-02 6.69466496e-01 4.56232041e-01 -1.53684393e-01
1.08760583e+00 -3.15380841e-01 -3.97596657e-01 3.56746495e-01
1.15171111e+00 5.79810552e-02 -1.46022868e+00 2.74420381e-01
-4.99188662e-01 -8.46982062e-01 -1.61334112e-01 1.11367647e-02
-1.40122640e+00 7.48920619e-01 5.61841011e-01 5.58081031e-01
1.29893422e+00 -7.45373428e-01 8.71719241e-01 5.51489135e-03
4.12670523e-01 -7.85145938e-01 -6.14372939e-02 3.96809131e-01
8.37062776e-01 -1.07254434e+00 -2.71892428e-01 -5.61730564e-01
-1.71064943e-01 1.53302574e+00 6.46983624e-01 -3.94388169e-01
6.13946915e-01 2.14486256e-01 2.47125588e-02 5.47773540e-01
-2.76532769e-01 -1.02227889e-01 1.38892561e-01 5.83583891e-01
7.34560847e-01 -4.26255465e-01 -7.72239208e-01 -2.95127422e-01
-1.51819721e-01 5.69138348e-01 8.38175058e-01 5.17700911e-01
-3.84315372e-01 -1.37268198e+00 -7.21491992e-01 -6.76105097e-02
-4.70353097e-01 -8.57278332e-03 2.91320741e-01 8.07129085e-01
4.74140793e-01 7.39417076e-01 2.01952145e-01 -1.91080540e-01
1.28239691e-01 2.40939520e-02 5.48252106e-01 4.82634306e-02
-2.89756209e-01 1.97396860e-01 -3.17231536e-01 -6.51236594e-01
-8.57012033e-01 -3.56202096e-01 -1.04202366e+00 -6.15766287e-01
-3.92969176e-02 -2.05860916e-03 4.32989180e-01 7.73253202e-01
-1.29723847e-01 6.31624222e-01 5.91054261e-01 -1.14469624e+00
9.28681642e-02 -6.79842830e-01 -5.12483180e-01 8.36547256e-01
4.17238563e-01 -5.72274804e-01 -4.39285398e-01 6.43555164e-01] | [10.782602310180664, -1.490761160850525] |
7b1e7276-470c-4601-8cd0-9cf91baa544d | document-modeling-with-external-attention-for | null | null | https://aclanthology.org/P18-1188 | https://aclanthology.org/P18-1188.pdf | Document Modeling with External Attention for Sentence Extraction | Document modeling is essential to a variety of natural language understanding tasks. We propose to use external information to improve document modeling for problems that can be framed as sentence extraction. We develop a framework composed of a hierarchical document encoder and an attention-based extractor with attention over external information. We evaluate our model on extractive document summarization (where the external information is image captions and the title of the document) and answer selection (where the external information is a question). We show that our model consistently outperforms strong baselines, in terms of both informativeness and fluency (for CNN document summarization) and achieves state-of-the-art results for answer selection on WikiQA and NewsQA. | ['Yi Chang', 'Jiangsheng Yu', 'Nikos Papasarantopoulos', 'Mirella Lapata', 'Shay B. Cohen', 'Shashi Narayan', 'Ronald Cardenas'] | 2018-07-01 | null | null | null | acl-2018-7 | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 4.18832511e-01 4.69916672e-01 -1.45403966e-01 -5.21879733e-01
-1.32481539e+00 -6.62782550e-01 9.21724916e-01 3.59116167e-01
-6.57925308e-01 7.04544604e-01 1.09698081e+00 -1.29848659e-01
2.76714295e-01 -5.96627831e-01 -8.82292271e-01 -9.78629962e-02
4.84037519e-01 8.03169191e-01 -2.30291467e-02 -2.41684332e-01
6.59610271e-01 -1.74291745e-01 -9.49287653e-01 1.00004172e+00
1.15421605e+00 6.42787635e-01 2.45377004e-01 1.21946681e+00
-4.57108289e-01 1.31790411e+00 -1.18880343e+00 -7.97838807e-01
-3.73417825e-01 -7.06116378e-01 -1.57431638e+00 -1.55087328e-02
1.08593857e+00 -7.42690325e-01 -4.34109628e-01 7.69524217e-01
5.60076475e-01 1.08480535e-01 8.40316474e-01 -4.94912505e-01
-1.15682650e+00 9.57266986e-01 -3.15540820e-01 3.11554790e-01
4.39608723e-01 1.45554423e-01 1.45687282e+00 -8.17790806e-01
8.92402411e-01 1.47652090e+00 2.25799724e-01 7.22584367e-01
-1.19741118e+00 -5.73243909e-02 3.69805396e-01 2.25229144e-01
-5.75160325e-01 -7.02838302e-01 3.52097780e-01 -2.53864110e-01
1.62332654e+00 4.19600099e-01 2.52485871e-01 1.23648703e+00
4.17597860e-01 1.58881843e+00 3.42078626e-01 -3.91198456e-01
2.91630030e-02 -4.66303807e-03 6.33330107e-01 7.06236064e-01
1.09286331e-01 -8.71090710e-01 -6.82341635e-01 1.09344393e-01
3.59886810e-02 -4.98770416e-01 -3.12715352e-01 2.08550811e-01
-8.58757079e-01 9.86605108e-01 2.84170389e-01 -1.41301394e-01
-3.53105843e-01 5.91658890e-01 8.02438676e-01 1.85263857e-01
6.53475940e-01 9.87263620e-01 -3.59602630e-01 -1.66860536e-01
-1.18523884e+00 4.48868006e-01 1.13794434e+00 1.20352924e+00
2.20118314e-01 2.04075888e-01 -1.08095682e+00 8.77950549e-01
-2.48853788e-02 6.30279362e-01 6.11143768e-01 -1.13033724e+00
1.06234407e+00 4.02484566e-01 1.27929702e-01 -6.96177602e-01
-2.55961448e-01 -6.35656536e-01 -7.12220132e-01 -5.77807307e-01
-6.87482506e-02 -1.79637015e-01 -1.20532286e+00 1.42956817e+00
-3.82342398e-01 -6.89352870e-01 3.04353148e-01 5.14212966e-01
1.58064401e+00 1.01202643e+00 8.12490098e-03 -1.60487056e-01
1.40073073e+00 -1.67999041e+00 -1.08765757e+00 -6.54075980e-01
8.21492136e-01 -5.50033152e-01 1.05003142e+00 1.67588770e-01
-1.65547109e+00 -3.84353966e-01 -8.05074751e-01 -8.09335828e-01
-3.02815735e-02 3.53706241e-01 1.96204990e-01 -3.40297297e-02
-1.19513237e+00 3.53145778e-01 -6.01130188e-01 -3.95638317e-01
6.67424798e-01 2.64063060e-01 -2.57292598e-01 -1.96677312e-01
-9.00573075e-01 1.08191657e+00 3.51976275e-01 -1.41269550e-01
-1.01133060e+00 -5.08838117e-01 -1.06215894e+00 4.95021522e-01
2.62647897e-01 -1.38063371e+00 1.88847280e+00 -8.04613292e-01
-1.37551963e+00 8.68991792e-01 -4.03223544e-01 -8.78192008e-01
2.92564064e-01 -7.81232655e-01 -1.18295178e-02 6.80784166e-01
4.05470222e-01 9.16641653e-01 7.34863162e-01 -1.02282143e+00
-3.15077305e-01 -3.58366966e-01 2.41075471e-01 4.32262540e-01
-4.53769356e-01 2.26604849e-01 -6.79391980e-01 -5.82713783e-01
-3.22981566e-01 -5.06583452e-01 -4.00127321e-02 -7.84776688e-01
-9.64034736e-01 -3.10123563e-01 6.14907503e-01 -1.28356278e+00
1.30865288e+00 -1.39421725e+00 6.23846889e-01 -4.69634384e-01
1.74021378e-01 2.90280193e-01 -5.67020774e-01 6.49488807e-01
3.41469973e-01 4.29078609e-01 -2.99211860e-01 -7.69436419e-01
9.95803922e-02 -1.98860377e-01 -7.97984779e-01 -1.42044336e-01
5.70921659e-01 1.44038951e+00 -9.02669013e-01 -5.95648050e-01
-3.47244054e-01 1.60097793e-01 -5.90899646e-01 3.05449218e-01
-7.29472220e-01 1.03669958e-02 -6.70648158e-01 3.95268768e-01
2.43480235e-01 -3.02957326e-01 -3.21984589e-01 -1.01879805e-01
2.01320037e-01 8.91877055e-01 -3.24717492e-01 1.88850522e+00
-4.33970034e-01 9.82708931e-01 1.53164774e-01 -4.73630548e-01
5.96648037e-01 2.34960511e-01 -2.21910805e-01 -6.43013775e-01
6.89682513e-02 -9.25402269e-02 -2.13503718e-01 -7.14652240e-01
1.10217643e+00 2.76074529e-01 -1.32834420e-01 6.61790013e-01
7.38186657e-01 -6.36338711e-01 7.30605543e-01 1.02306819e+00
1.46643400e+00 1.17167048e-01 1.89957544e-01 -1.73971951e-01
3.47167224e-01 7.96094313e-02 -1.70537829e-01 1.31995809e+00
3.71287614e-01 7.84790576e-01 8.02044749e-01 -1.25038669e-01
-9.18487966e-01 -7.58150399e-01 2.37617388e-01 1.18928874e+00
-4.86226469e-01 -1.01338732e+00 -1.01429617e+00 -1.08156109e+00
-1.09678522e-01 1.10552752e+00 -5.97868264e-01 -2.82642961e-01
-6.96370900e-01 -3.13686281e-01 6.11729741e-01 9.08596814e-01
5.36477506e-01 -1.25229049e+00 -5.12141287e-01 1.81994855e-01
-5.47965944e-01 -1.16634643e+00 -7.38698542e-01 1.98908344e-01
-9.38228011e-01 -5.44775009e-01 -6.86109424e-01 -6.50059521e-01
5.33389151e-01 5.78692481e-02 1.67054427e+00 1.75384656e-01
5.30581325e-02 7.47430742e-01 -2.77558118e-01 -6.16753638e-01
-4.59297806e-01 6.75842106e-01 -4.99824882e-01 -3.75182956e-01
2.12547481e-01 -3.96619253e-02 -5.53359449e-01 -4.39980954e-01
-1.03012025e+00 3.58424008e-01 6.81896031e-01 8.54137301e-01
2.23226562e-01 -8.07978213e-01 8.78635347e-01 -1.19750905e+00
1.24465919e+00 -1.56914249e-01 -2.29937192e-02 4.13519591e-01
-4.43893671e-02 3.84287506e-01 6.19969547e-01 3.55230346e-02
-1.28755546e+00 -2.81996727e-01 -3.38037997e-01 1.69216543e-01
-6.74242973e-02 7.23866642e-01 4.34695072e-02 6.73927784e-01
5.81139624e-01 3.29049826e-01 -3.61893982e-01 -4.33677733e-01
7.16233552e-01 7.81549275e-01 7.79616356e-01 -5.12969911e-01
3.43206733e-01 3.20523679e-01 -3.86862934e-01 -9.17987585e-01
-1.62584329e+00 -5.02088070e-01 -5.61423779e-01 4.15348969e-02
1.08293641e+00 -9.50497985e-01 -3.23647648e-01 2.20587388e-01
-1.84298575e+00 -1.73637018e-01 -4.76597846e-01 1.11491449e-01
-7.15025723e-01 2.62768596e-01 -1.08819771e+00 -5.62030852e-01
-1.17337799e+00 -1.07572699e+00 1.42068064e+00 2.02054396e-01
-6.26043499e-01 -9.77163911e-01 1.44656867e-01 7.68277705e-01
4.45903689e-01 -1.41681731e-01 1.12491727e+00 -9.65724885e-01
-6.08110845e-01 -2.52242327e-01 -3.13318342e-01 5.55133879e-01
-2.72925496e-01 -1.55698642e-01 -1.09250319e+00 -9.33232158e-02
9.43768304e-03 -7.47229457e-01 1.79668009e+00 4.84843761e-01
1.12089217e+00 -7.87112653e-01 -1.65013701e-01 1.75231144e-01
9.44318891e-01 -2.16730997e-01 7.52106130e-01 2.23486856e-01
7.43744850e-01 7.35667944e-01 1.47792757e-01 1.93032384e-01
4.84473497e-01 4.06825483e-01 2.67053902e-01 1.06760040e-01
-4.17910993e-01 -2.25570053e-01 4.59136754e-01 1.15930319e+00
2.94599563e-01 -8.03858161e-01 -6.94992244e-01 8.39057326e-01
-1.96331739e+00 -1.18448722e+00 -2.04893887e-01 1.55006218e+00
1.21292686e+00 2.15260282e-01 -1.70764059e-01 -5.37465155e-01
3.27524781e-01 3.72034997e-01 -4.60563242e-01 -9.45210516e-01
-2.49882460e-01 1.90641180e-01 -3.36580724e-02 8.18296552e-01
-1.05475259e+00 1.20999479e+00 6.80676937e+00 5.55268705e-01
-5.45796871e-01 -4.43829224e-02 6.15035594e-01 -3.61898273e-01
-4.56374556e-01 -1.10392302e-01 -9.83491898e-01 1.31346837e-01
9.59065914e-01 -3.04811751e-03 -4.78780903e-02 5.32313168e-01
-5.63300028e-02 -1.87428385e-01 -1.39823055e+00 3.88633609e-01
9.43811476e-01 -1.71557701e+00 7.36073792e-01 -4.43232685e-01
9.13456440e-01 8.61153826e-02 -1.87231582e-02 6.80981934e-01
4.28956538e-01 -1.11666322e+00 6.80938482e-01 6.58373892e-01
4.24242020e-01 -5.16466618e-01 8.30064356e-01 3.36795777e-01
-5.29781163e-01 9.37981829e-02 -2.04801470e-01 -5.44548891e-02
3.42773885e-01 2.71117687e-01 -7.50412762e-01 4.30478603e-01
5.28284669e-01 7.10963607e-01 -8.28821301e-01 8.90520215e-01
-7.82532036e-01 7.45935500e-01 5.32005131e-02 -1.77904621e-01
4.81206954e-01 5.94916381e-02 6.34542525e-01 1.71354473e+00
-1.90640107e-01 1.07087210e-01 -9.69103873e-02 8.60715747e-01
-8.11046064e-01 2.26422876e-01 -5.27386367e-01 -2.81738430e-01
8.37784167e-03 1.13339055e+00 -3.30980599e-01 -7.24112332e-01
-3.53534400e-01 1.27254009e+00 6.07524693e-01 5.88136137e-01
-3.51703405e-01 -8.41251135e-01 -2.15014983e-02 -2.89913982e-01
5.09414613e-01 6.93077892e-02 -2.95139879e-01 -1.49698913e+00
2.09877744e-01 -1.09711874e+00 4.60649520e-01 -1.08045208e+00
-8.92278314e-01 5.96102536e-01 -1.54321507e-01 -4.80432332e-01
-5.38268030e-01 -4.30918247e-01 -8.46969903e-01 6.90816462e-01
-1.52463150e+00 -1.23352885e+00 -8.71443469e-03 9.91789475e-02
1.23049831e+00 -1.25196531e-01 8.79423201e-01 -2.18951359e-01
-4.61572647e-01 3.24382454e-01 9.58223268e-02 2.74336606e-01
7.57569611e-01 -1.54603219e+00 7.87710071e-01 9.10584331e-01
3.56496960e-01 6.82191312e-01 8.55675757e-01 -8.16396534e-01
-1.30339289e+00 -1.08310699e+00 1.13512337e+00 -9.97281849e-01
5.09685040e-01 -4.35646087e-01 -7.27821469e-01 1.11622322e+00
1.09088659e+00 -6.96586013e-01 6.83285236e-01 1.45297974e-01
-3.95812541e-01 2.87016928e-01 -6.45076513e-01 6.77767932e-01
6.09057188e-01 -5.84704399e-01 -1.10821009e+00 6.92200065e-01
1.22487736e+00 -4.51387823e-01 -4.22788382e-01 1.08708598e-01
2.77314067e-01 -3.43893737e-01 6.95652902e-01 -1.35068464e+00
1.26531959e+00 1.68065220e-01 4.16553952e-02 -1.47921801e+00
-1.85207546e-01 -3.54098201e-01 -7.97637939e-01 1.35274410e+00
6.17241263e-01 8.75846893e-02 5.15758932e-01 5.14455020e-01
-6.44780099e-01 -8.23593795e-01 -4.61875439e-01 -4.60384071e-01
2.89712667e-01 -3.12946998e-02 1.99721918e-01 3.51291358e-01
2.25577489e-01 1.28220904e+00 -2.49122590e-01 -4.16309893e-01
2.09892809e-01 2.41028354e-01 7.98355639e-01 -9.21635926e-01
-2.62030631e-01 -6.28250957e-01 1.56112462e-01 -1.43391979e+00
5.24464965e-01 -1.01250327e+00 2.02242851e-01 -2.66306496e+00
7.98328400e-01 7.86176801e-01 1.24368012e-01 3.53707045e-01
-2.61343151e-01 2.14958470e-02 2.89228708e-01 -6.20111004e-02
-1.48293924e+00 8.50498617e-01 1.18420696e+00 -7.51228511e-01
-2.12570429e-02 -1.39652893e-01 -1.07667530e+00 7.35788763e-01
4.85527098e-01 -3.83231550e-01 -1.72442868e-01 -1.22568274e+00
4.85006511e-01 2.04467401e-01 1.11777388e-01 -7.13038743e-01
4.67391759e-01 9.00881961e-02 5.31547129e-01 -9.77563083e-01
4.02282208e-01 2.06056740e-02 -1.06979513e+00 2.62028873e-01
-1.17655003e+00 1.03105038e-01 3.43556851e-01 3.96953702e-01
-4.09567147e-01 -7.76208222e-01 2.24137932e-01 -3.79669964e-01
-1.34132609e-01 -2.16725357e-02 -5.19897223e-01 6.62497282e-01
2.83647746e-01 2.71487623e-01 -1.10120666e+00 -1.01099491e+00
-4.40252304e-01 7.10963488e-01 1.02616340e-01 3.20960999e-01
8.65085304e-01 -9.56925571e-01 -1.26546323e+00 -4.26650107e-01
1.17181540e-01 2.86678839e-02 1.00154020e-01 6.12684608e-01
-4.55864102e-01 9.97258127e-01 1.77091211e-01 -2.66481727e-01
-1.20470679e+00 1.92316636e-01 1.29935592e-01 -6.88593507e-01
-4.86838728e-01 9.87462759e-01 3.00037563e-01 -2.29206607e-01
3.97938550e-01 -4.78897631e-01 -4.24226016e-01 2.93586969e-01
8.52448165e-01 1.75200179e-01 3.62553239e-01 -3.30245852e-01
8.93827120e-04 1.98158547e-01 -7.55033731e-01 -2.28410974e-01
1.35504222e+00 -7.38797933e-02 -4.11162317e-01 1.67235121e-01
1.16981196e+00 4.90473956e-02 -9.49071765e-01 -1.58737466e-01
2.10546210e-01 1.48884347e-03 9.78796277e-03 -1.21957266e+00
-4.81588483e-01 1.15730250e+00 -3.10188591e-01 8.55035484e-02
7.96860874e-01 2.63577461e-01 8.97335708e-01 1.14951086e+00
-3.95362735e-01 -1.46733034e+00 5.15890658e-01 1.06683934e+00
1.41289485e+00 -1.21212435e+00 1.62997812e-01 1.03859596e-01
-8.51238847e-01 1.10071695e+00 7.70885050e-01 -1.26494572e-01
-1.93769872e-01 1.08749218e-01 -3.54013592e-01 -5.04650295e-01
-1.27321494e+00 -1.14575356e-01 6.41914427e-01 2.05051064e-01
7.01954305e-01 -3.27103019e-01 -2.23120540e-01 9.12519872e-01
-3.07604462e-01 -3.82144928e-01 8.47257733e-01 8.80036294e-01
-6.52787626e-01 -6.97082400e-01 6.40704706e-02 8.91692817e-01
-7.05673456e-01 -4.96169031e-01 -8.83714139e-01 4.06868100e-01
-7.16419280e-01 9.20485735e-01 3.62627767e-02 1.03619702e-01
3.65235686e-01 3.38478416e-01 6.66111410e-01 -1.22484124e+00
-9.59388137e-01 -2.17163131e-01 7.40730047e-01 -3.75623256e-01
-4.32990715e-02 -4.52501804e-01 -1.15596104e+00 2.59466976e-01
-3.88979197e-01 2.93485940e-01 7.24623263e-01 1.08338034e+00
5.35686493e-01 8.77410769e-01 2.57314474e-04 -4.79219466e-01
-8.06925416e-01 -1.36624062e+00 1.13736674e-01 3.13607007e-01
5.46507657e-01 2.67904788e-01 -4.16717008e-02 2.63446391e-01] | [12.346600532531738, 9.352109909057617] |
bec05676-d6cc-47fd-8db1-31d5b2a5ac7e | active-visual-information-gathering-for | 2007.08037 | null | https://arxiv.org/abs/2007.08037v3 | https://arxiv.org/pdf/2007.08037v3.pdf | Active Visual Information Gathering for Vision-Language Navigation | Vision-language navigation (VLN) is the task of entailing an agent to carry out navigational instructions inside photo-realistic environments. One of the key challenges in VLN is how to conduct a robust navigation by mitigating the uncertainty caused by ambiguous instructions and insufficient observation of the environment. Agents trained by current approaches typically suffer from this and would consequently struggle to avoid random and inefficient actions at every step. In contrast, when humans face such a challenge, they can still maintain robust navigation by actively exploring the surroundings to gather more information and thus make more confident navigation decisions. This work draws inspiration from human navigation behavior and endows an agent with an active information gathering ability for a more intelligent vision-language navigation policy. To achieve this, we propose an end-to-end framework for learning an exploration policy that decides i) when and where to explore, ii) what information is worth gathering during exploration, and iii) how to adjust the navigation decision after the exploration. The experimental results show promising exploration strategies emerged from training, which leads to significant boost in navigation performance. On the R2R challenge leaderboard, our agent gets promising results all three VLN settings, i.e., single run, pre-exploration, and beam search. | ['Jianbing Shen', 'Wenguan Wang', 'Tianmin Shu', 'Hanqing Wang', 'Wei Liang'] | 2020-07-15 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4046_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123670307.pdf | eccv-2020-8 | ['vision-language-navigation'] | ['computer-vision'] | [ 1.52205423e-01 8.65658000e-02 -1.70882214e-02 -3.99941981e-01
-4.91088510e-01 -5.26998758e-01 6.19198143e-01 -9.98192877e-02
-9.95884001e-01 7.18686819e-01 1.48456007e-01 -5.65750420e-01
-1.96151912e-01 -6.79575801e-01 -6.20291114e-01 -8.14441919e-01
-4.42334637e-02 6.06741667e-01 2.48492718e-01 -5.46924233e-01
3.85839760e-01 2.28558749e-01 -1.52000391e+00 -4.10361677e-01
1.13420320e+00 8.92894268e-01 8.15223098e-01 6.57440662e-01
2.39817780e-02 8.28521967e-01 -2.58231461e-01 2.21897259e-01
5.14520645e-01 -3.27905715e-01 -6.71225667e-01 1.03639521e-01
9.18074623e-02 -4.52392817e-01 -2.71501213e-01 1.16246259e+00
5.59596837e-01 8.06414366e-01 2.96436608e-01 -1.09392321e+00
-4.56553012e-01 5.45592904e-01 -5.24877489e-01 1.68400675e-01
5.16503155e-01 7.77401030e-01 7.16351748e-01 -4.92911279e-01
6.46542788e-01 1.34348857e+00 1.83787793e-01 7.51289964e-01
-9.18323338e-01 -2.72366375e-01 7.56907344e-01 3.53322685e-01
-8.65292430e-01 -4.35942650e-01 4.96017158e-01 -1.41045913e-01
7.00949669e-01 4.16154265e-02 6.88609779e-01 1.40487051e+00
1.43644437e-01 7.53070116e-01 1.13039207e+00 -3.09037507e-01
7.10335851e-01 1.22978417e-02 -1.19589344e-02 7.56289482e-01
2.68976361e-01 3.98009598e-01 -5.96315622e-01 1.73357844e-01
6.10627055e-01 -1.53684467e-01 -5.28904378e-01 -6.37183905e-01
-1.25062799e+00 7.18000948e-01 8.81000042e-01 -3.26923020e-02
-8.54466498e-01 -9.39718354e-03 2.52780080e-01 2.38470778e-01
-5.47364831e-01 6.69292212e-01 -2.61930227e-01 -4.15210605e-01
-3.00378680e-01 2.32565835e-01 5.03282487e-01 7.76571631e-01
7.41390884e-01 1.08549133e-01 -1.54442132e-01 5.74606299e-01
3.84734899e-01 5.92931569e-01 6.45260692e-01 -1.40495110e+00
5.81521869e-01 4.28910971e-01 4.83199060e-01 -8.37636352e-01
-5.06611228e-01 -3.62989694e-01 -7.91659296e-01 9.27071869e-01
3.91587824e-01 -3.70229095e-01 -1.26885772e+00 1.98398876e+00
4.74034846e-01 -3.07980567e-01 4.86815751e-01 1.16731381e+00
4.23171401e-01 4.79772002e-01 1.17522795e-02 7.86996167e-03
1.39145446e+00 -1.43925214e+00 -7.27861166e-01 -8.88008237e-01
3.65283638e-01 -3.48410100e-01 1.30019820e+00 6.09007120e-01
-8.30801249e-01 -6.50040865e-01 -1.06391978e+00 5.62944822e-02
-2.62862146e-01 -1.53783008e-01 6.11081958e-01 3.31463367e-01
-1.20961833e+00 1.10289440e-01 -9.11813200e-01 -4.91473675e-01
2.11482182e-01 1.76788568e-01 -4.96807516e-01 -2.36651450e-01
-1.02445555e+00 9.60737407e-01 4.98241276e-01 4.70223635e-01
-1.17719245e+00 -7.77098164e-02 -8.32178533e-01 -1.45397782e-01
9.83909786e-01 -7.78931022e-01 1.38103116e+00 -6.30724192e-01
-1.61534369e+00 1.96470141e-01 -1.72643393e-01 -7.45803058e-01
7.52863228e-01 -3.33092511e-01 1.42284974e-01 -1.14463761e-01
3.25121760e-01 1.00585258e+00 4.94872242e-01 -1.42453909e+00
-1.04968882e+00 -5.61178923e-01 4.41592485e-01 8.50327551e-01
5.24892239e-03 -7.17387915e-01 -7.76891470e-01 -1.20859273e-01
4.19894606e-01 -1.22674835e+00 -8.81036699e-01 -6.03258274e-02
-4.64234293e-01 -5.93368262e-02 6.55084729e-01 -4.22823042e-01
6.71839893e-01 -1.69715369e+00 2.06000164e-01 1.86949253e-01
8.25833902e-02 2.07534522e-01 -2.31220797e-01 1.70894191e-01
6.27339661e-01 -1.85510993e-01 -2.77674496e-02 -3.87667954e-01
5.25401533e-02 4.76431161e-01 -3.28062445e-01 9.39186513e-02
-5.12856424e-01 8.31821918e-01 -1.07248807e+00 -1.08012222e-01
3.55277568e-01 3.41286749e-01 -4.57093447e-01 4.49886709e-01
-3.64789546e-01 9.85545397e-01 -7.99872458e-01 6.29033625e-01
5.34832478e-01 -2.48040259e-01 -2.06497479e-02 2.37193435e-01
-2.67665267e-01 1.54511109e-01 -1.22657120e+00 1.78362989e+00
-6.79146409e-01 5.82955122e-01 3.75718057e-01 -6.40515566e-01
7.63390958e-01 -2.35390693e-01 -4.17890772e-02 -1.38187790e+00
3.51733528e-02 2.16151457e-02 1.74057670e-02 -4.62551415e-01
5.63080013e-01 4.41427976e-01 1.11178182e-01 3.40083182e-01
-6.46406770e-01 5.74809946e-02 1.91103488e-01 1.25481382e-01
1.21295643e+00 2.70806164e-01 3.45599920e-01 1.09133907e-01
6.79622352e-01 1.53648078e-01 6.61844134e-01 1.42406130e+00
-4.91297871e-01 1.29646778e-01 8.99454653e-02 -6.76650286e-01
-5.69578826e-01 -9.43355083e-01 4.71279025e-01 1.23995066e+00
6.37703240e-01 -4.26819548e-02 -6.56071484e-01 -6.58009946e-01
-3.58190238e-01 1.04932964e+00 -6.72628224e-01 -1.31757319e-01
-6.86818182e-01 -4.93308604e-01 3.07869334e-02 5.17739356e-01
9.78723109e-01 -1.39290452e+00 -1.41600525e+00 1.73998356e-01
-4.89742219e-01 -9.78669047e-01 -4.34594423e-01 4.05211210e-01
-5.33184528e-01 -9.39453185e-01 -3.39991122e-01 -6.19557202e-01
8.78023148e-01 6.64778590e-01 8.27916265e-01 9.12958309e-02
5.54972216e-02 4.02018100e-01 -4.21702653e-01 -2.86090344e-01
-2.41930366e-01 -2.43853852e-02 1.66324258e-01 -4.90216762e-01
9.47390944e-02 -3.51509750e-01 -9.70585883e-01 6.00944459e-01
-5.46919167e-01 9.42906737e-02 9.14655387e-01 8.39901030e-01
7.29833305e-01 3.19424868e-01 2.07054228e-01 -4.61182505e-01
8.30376267e-01 -2.42128879e-01 -9.31469083e-01 2.78762549e-01
-9.08026159e-01 3.44381809e-01 7.22932637e-01 -3.95579517e-01
-1.24773526e+00 3.25467922e-02 -2.07736507e-01 1.63408443e-02
-2.24797159e-01 3.14055741e-01 -1.80820584e-01 -1.10150382e-01
6.73668027e-01 5.22723913e-01 -1.69134475e-02 -1.29088819e-01
5.41577995e-01 3.32932234e-01 7.14398563e-01 -5.53752542e-01
6.65166914e-01 6.14278972e-01 -1.05979085e-01 -7.30050087e-01
-6.10097647e-01 -3.99325132e-01 -3.44777733e-01 -1.60914361e-01
8.28065217e-01 -8.67339611e-01 -9.53261912e-01 2.91151255e-01
-6.86627865e-01 -7.07683206e-01 -1.33186102e-01 3.67595613e-01
-6.66137755e-01 4.46618289e-01 2.76685916e-02 -1.00448704e+00
-3.21176797e-01 -1.57886398e+00 6.89044476e-01 7.29005158e-01
-2.04139035e-02 -7.86827326e-01 1.45175606e-01 5.01781642e-01
6.06802881e-01 3.25647779e-02 4.52367961e-01 -4.06788290e-01
-1.09032702e+00 6.09936006e-02 -3.82630192e-02 -1.17508136e-01
5.51565401e-02 -5.85039556e-01 -5.28920233e-01 -5.65882981e-01
4.74102013e-02 -5.66516340e-01 9.48031843e-01 4.46056843e-01
7.05498636e-01 -3.53081614e-01 -3.70249808e-01 7.05419421e-01
1.16692102e+00 7.20104992e-01 4.61217105e-01 1.06055796e+00
4.61948246e-01 6.31068587e-01 9.83394444e-01 3.24715108e-01
8.00584316e-01 7.78448284e-01 1.01473045e+00 1.29410818e-01
1.55025452e-01 -5.06847441e-01 4.67691809e-01 9.73848775e-02
5.08807264e-02 -5.33510149e-01 -8.14217210e-01 3.93119961e-01
-2.15604973e+00 -7.75097609e-01 4.55070525e-01 2.19693232e+00
4.95691448e-01 4.05228853e-01 -1.72858089e-01 -3.95370215e-01
2.45628148e-01 2.64923126e-01 -1.03394616e+00 -1.58925042e-01
-4.83367033e-02 -5.38103640e-01 4.98888105e-01 8.17245662e-01
-8.57011795e-01 1.24861395e+00 5.46288729e+00 4.17301953e-01
-1.02027392e+00 -2.49755621e-01 4.99545664e-01 9.87745523e-02
-2.52274692e-01 1.46100312e-01 -8.81074429e-01 3.67138773e-01
5.96088886e-01 2.40071297e-01 7.84249544e-01 1.01257491e+00
4.60197270e-01 -7.26343095e-01 -8.65865529e-01 9.05475736e-01
-1.86423585e-01 -1.04577744e+00 -3.85514349e-02 1.04862958e-01
5.69464386e-01 1.98696598e-01 2.63217181e-01 5.31582654e-01
8.82338464e-01 -1.11280835e+00 7.23288536e-01 4.36167538e-01
-3.61819677e-02 -5.46720505e-01 6.80371344e-01 7.89812028e-01
-1.04130042e+00 -5.07899225e-01 -4.09738570e-01 -6.77090213e-02
4.63567555e-01 2.57740188e-02 -9.62893784e-01 4.15746629e-01
7.82109022e-01 1.51093423e-01 -5.16148150e-01 1.12733400e+00
-7.34426737e-01 -2.17872914e-02 -2.47225210e-01 -2.25547358e-01
7.43792951e-01 -5.57354629e-01 6.69482589e-01 6.16952896e-01
3.61479193e-01 1.93755597e-01 5.77377498e-01 6.94334269e-01
4.27102327e-01 -2.22642615e-01 -5.88378906e-01 2.51308829e-01
5.08104622e-01 1.02770603e+00 -8.26507807e-01 -3.87998819e-02
-6.17707148e-02 9.28016126e-01 5.63525438e-01 6.15476489e-01
-6.41705513e-01 -5.88168427e-02 6.12803698e-01 -3.57703120e-01
3.70060503e-01 -4.05159324e-01 -1.84704155e-01 -7.67703712e-01
1.64406598e-01 -8.07522237e-01 2.34131381e-01 -8.83700907e-01
-7.00087607e-01 9.95701373e-01 -3.04684550e-01 -9.05927658e-01
-5.85655689e-01 -4.27745551e-01 -6.11667335e-01 7.71297038e-01
-1.70663178e+00 -7.90226161e-01 -7.37468004e-01 3.87695223e-01
8.86364043e-01 -2.17265412e-01 4.54984844e-01 -3.02747577e-01
-3.90650570e-01 2.24426463e-01 -1.28092021e-01 -1.17447272e-01
4.49753881e-01 -1.34208512e+00 4.76257801e-01 1.03927076e+00
4.49897908e-02 6.74937069e-01 1.02259529e+00 -7.09757209e-01
-1.56908429e+00 -7.58054256e-01 2.42270067e-01 -4.48670864e-01
2.33600542e-01 -1.31300315e-01 -6.86049521e-01 6.30939126e-01
1.78124949e-01 -2.70742834e-01 1.49414212e-01 -3.73670235e-02
-4.47535925e-02 7.92592615e-02 -9.43051815e-01 1.20137787e+00
1.25156355e+00 -5.66219725e-02 -6.06514931e-01 1.22629531e-01
7.59110034e-01 -6.13324702e-01 7.93395787e-02 2.55228937e-01
5.40939629e-01 -1.29139316e+00 1.01964855e+00 -4.32858974e-01
-1.13087617e-01 -6.02703691e-01 -2.38577858e-01 -1.44388795e+00
-2.42432207e-01 -8.61608803e-01 1.04839697e-01 6.13228679e-01
5.10827422e-01 -5.80097318e-01 9.43894446e-01 6.96829796e-01
-3.82214189e-02 -7.90677428e-01 -9.25159931e-01 -5.97595870e-01
-3.24538887e-01 -3.40803921e-01 3.46151859e-01 2.70929903e-01
-3.49814892e-01 2.15441063e-01 -4.81905013e-01 4.59598511e-01
7.32820392e-01 6.29223362e-02 1.13941348e+00 -9.15169954e-01
3.39425057e-02 -3.94939542e-01 -2.26064138e-02 -1.63320768e+00
2.38405429e-02 -3.09175968e-01 7.09491134e-01 -2.03901291e+00
-9.04433616e-03 -5.21417201e-01 -1.59993961e-01 4.12350833e-01
-2.58063346e-01 -4.38143402e-01 2.56498545e-01 2.13542655e-01
-1.08227754e+00 7.61151254e-01 1.40781212e+00 -3.71383280e-02
-6.36361361e-01 2.16870010e-01 -9.57603216e-01 7.40187109e-01
6.79393888e-01 6.30207956e-02 -9.03360069e-01 -6.08557582e-01
2.51684129e-01 2.19934598e-01 2.67485738e-01 -1.03451991e+00
8.50061476e-01 -3.64538103e-01 2.49452144e-01 -7.80213714e-01
4.82913047e-01 -9.40087616e-01 -1.44525751e-01 6.66094005e-01
-4.35547471e-01 3.23267281e-01 2.72726510e-02 9.61400211e-01
2.74204724e-02 -8.67629498e-02 6.78425789e-01 -4.64958042e-01
-1.39362764e+00 2.23149940e-01 -5.16137540e-01 1.62495375e-01
9.37535107e-01 -3.74696970e-01 -4.31172550e-01 -7.40529537e-01
-6.99384689e-01 1.11562264e+00 5.97231269e-01 5.89411199e-01
8.30718875e-01 -1.12429011e+00 -3.68509054e-01 3.16217542e-01
5.43911234e-02 3.08106720e-01 3.40785146e-01 6.49273813e-01
-4.01863486e-01 4.82174903e-01 -3.47035348e-01 -6.77411497e-01
-9.54937339e-01 5.04148006e-01 3.96001816e-01 -4.22302783e-01
-7.12935328e-01 9.40516055e-01 4.07801002e-01 -5.45015693e-01
7.48393416e-01 -2.27707043e-01 -4.68248516e-01 -1.86919346e-01
6.85530663e-01 2.75281698e-01 -1.49300009e-01 -3.96028697e-01
-2.92962432e-01 5.79222798e-01 -3.46410036e-01 -4.20811743e-01
9.18592334e-01 -7.32547939e-01 2.86515206e-01 3.60162444e-02
4.90881562e-01 -1.32363319e-01 -1.94173968e+00 -2.33654797e-01
-5.70652336e-02 -5.53610563e-01 1.22345559e-01 -1.18796778e+00
-8.35236967e-01 5.91812313e-01 6.92737043e-01 3.47344726e-02
9.61294472e-01 -2.49662161e-01 6.50241733e-01 1.09987497e+00
7.67527223e-01 -1.22017646e+00 3.15046757e-01 7.61865914e-01
8.68779182e-01 -1.67374063e+00 -1.91761270e-01 1.55066848e-01
-9.75445032e-01 9.72393513e-01 1.07623839e+00 3.23764205e-01
1.84835106e-01 -8.07708874e-03 4.55708861e-01 -2.24204421e-01
-8.59455228e-01 -3.88743281e-01 4.03332189e-02 8.70665014e-01
-4.40380543e-01 -1.68129042e-01 2.22078353e-01 1.94507524e-01
-3.05154562e-01 -3.42584401e-01 5.50783932e-01 1.01343250e+00
-9.17174101e-01 -6.78806663e-01 -3.44205290e-01 -1.02197481e-02
2.58289218e-01 2.59812921e-01 -2.01418653e-01 8.10868204e-01
-4.79732640e-02 1.17642593e+00 -3.05493861e-01 -2.58053452e-01
5.06159782e-01 -4.21970159e-01 8.09451714e-02 -2.57750392e-01
-8.27105790e-02 -1.41218334e-01 -5.41998893e-02 -1.04225993e+00
-2.00277507e-01 -5.33372581e-01 -1.51178133e+00 -4.46036225e-03
-3.98689434e-02 2.58360833e-01 7.36892581e-01 9.00629103e-01
2.89107233e-01 5.64752042e-01 4.44979310e-01 -9.67966974e-01
-9.26629126e-01 -6.05455220e-01 2.61249617e-02 1.11759380e-01
6.72720492e-01 -8.07908773e-01 -4.92016047e-01 -4.97700959e-01] | [4.52008056640625, 0.5736924409866333] |
f760fc43-cf9d-48c1-a73b-40b89e2ac2ac | streaming-belief-propagation-for-community | 2106.04805 | null | https://arxiv.org/abs/2106.04805v2 | https://arxiv.org/pdf/2106.04805v2.pdf | Streaming Belief Propagation for Community Detection | The community detection problem requires to cluster the nodes of a network into a small number of well-connected "communities". There has been substantial recent progress in characterizing the fundamental statistical limits of community detection under simple stochastic block models. However, in real-world applications, the network structure is typically dynamic, with nodes that join over time. In this setting, we would like a detection algorithm to perform only a limited number of updates at each node arrival. While standard voting approaches satisfy this constraint, it is unclear whether they exploit the network information optimally. We introduce a simple model for networks growing over time which we refer to as streaming stochastic block model (StSBM). Within this model, we prove that voting algorithms have fundamental limitations. We also develop a streaming belief-propagation (StreamBP) approach, for which we prove optimality in certain regimes. We validate our theoretical findings on synthetic and real data. | ['Jakab Tardos', 'Ashkan Norouzi-Fard', 'Andrea Montanari', 'Filipe Miguel Goncalves de Almeida', 'Andre Linhares', 'Mohammadhossein Bateni', 'Yuchen Wu'] | 2021-06-09 | null | http://proceedings.neurips.cc/paper/2021/hash/e2a2dcc36a08a345332c751b2f2e476c-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/e2a2dcc36a08a345332c751b2f2e476c-Paper.pdf | neurips-2021-12 | ['stochastic-block-model'] | ['graphs'] | [ 3.49178761e-01 1.39587834e-01 -3.52138638e-01 -4.02441509e-02
-3.95098954e-01 -7.79997945e-01 5.12239873e-01 7.10959911e-01
-2.58492559e-01 3.46213400e-01 -1.47590384e-01 -3.24352711e-01
-1.79979384e-01 -1.02784681e+00 -5.44210970e-01 -7.67295003e-01
-6.51084840e-01 5.89213550e-01 7.64785409e-01 5.97043410e-02
3.05737764e-01 3.07729572e-01 -8.23575616e-01 -1.21432312e-01
5.35516560e-01 4.31139946e-01 6.68610781e-02 1.07063615e+00
1.30494669e-01 7.76544034e-01 -3.33276033e-01 -3.00940096e-01
3.53493422e-01 -3.74032855e-01 -7.43546128e-01 4.27582890e-01
-2.08434224e-01 -2.61690885e-01 -5.22126257e-01 1.22721326e+00
3.76802057e-01 -2.98460662e-01 5.18520713e-01 -1.53520620e+00
4.65388969e-02 1.22845781e+00 -1.10589421e+00 3.35028052e-01
3.36516082e-01 -3.14594150e-01 1.35283995e+00 -5.31327307e-01
7.73642659e-01 9.16928649e-01 8.84711623e-01 1.77432552e-01
-1.43210900e+00 -6.32323027e-01 2.50851005e-01 8.02295562e-03
-1.47091103e+00 -4.85716194e-01 7.77683318e-01 -5.18088639e-01
3.11871678e-01 2.65823781e-01 8.90865624e-01 6.17497265e-01
-4.47467752e-02 8.55899930e-01 9.04645562e-01 -5.06425202e-01
5.46415389e-01 -1.55678183e-01 2.34344229e-01 5.93875766e-01
8.90936553e-01 -3.91980559e-01 -5.96574366e-01 -7.44710565e-01
7.56968558e-01 3.48226689e-02 -2.67645776e-01 -8.10598671e-01
-1.13208914e+00 1.17485571e+00 7.25113377e-02 2.49654964e-01
-4.36200738e-01 5.69706440e-01 3.58561724e-01 4.40468162e-01
5.83675623e-01 -3.13424170e-01 3.06799207e-02 3.76160890e-02
-1.32170689e+00 1.62190780e-01 1.32417786e+00 1.10678625e+00
5.31478107e-01 -5.06339967e-01 2.54921883e-01 3.88621539e-01
5.65418899e-01 4.31728274e-01 -3.29617977e-01 -1.11588609e+00
3.93927693e-01 2.60526210e-01 2.33873978e-01 -1.14336658e+00
-2.12204814e-01 -5.20646274e-01 -1.26223958e+00 -1.54921398e-01
4.88943130e-01 -1.64191902e-01 -3.67852181e-01 1.70610225e+00
4.80834275e-01 2.36758962e-01 -1.91435948e-01 4.20035809e-01
-1.00963330e-03 5.90065718e-01 -5.32810450e-01 -8.40198040e-01
8.78076136e-01 -6.04958177e-01 -4.09106046e-01 1.57977089e-01
5.06465256e-01 -4.42982346e-01 1.14351213e-01 5.39137244e-01
-1.44796848e+00 3.47364396e-01 -9.25125182e-01 7.18882859e-01
4.71276551e-01 -6.85126960e-01 4.95637238e-01 9.09162402e-01
-1.34580958e+00 5.31581104e-01 -1.19450831e+00 -8.16543639e-01
2.98743367e-01 2.29962572e-01 2.39397176e-02 -2.18648687e-01
-7.12086618e-01 1.17584020e-01 8.99213254e-02 8.57877284e-02
-1.15498734e+00 -2.27141455e-01 -3.99556041e-01 2.57883724e-02
6.07771754e-01 -7.12041855e-01 1.43879080e+00 -8.25240254e-01
-7.94151068e-01 6.43597722e-01 -5.03448963e-01 -7.64579356e-01
8.43370020e-01 3.99988919e-01 1.97665289e-01 6.59025908e-01
8.29227865e-02 9.01169628e-02 6.07534409e-01 -1.43192267e+00
-7.26211190e-01 -8.02758932e-02 3.60650688e-01 -3.61772068e-02
-4.55802113e-01 5.24946451e-01 -3.29799235e-01 -4.90426451e-01
5.13398886e-01 -1.01792920e+00 -6.62617385e-01 1.84316322e-01
-3.97659093e-01 -2.26009458e-01 5.39356291e-01 3.18626463e-02
1.38800049e+00 -1.80846465e+00 3.36531661e-02 7.78494716e-01
7.86884189e-01 -1.79414064e-01 1.98509425e-01 1.05197036e+00
3.03379029e-01 5.50174594e-01 -3.44822139e-01 -5.29272377e-01
-2.23485500e-01 3.38056624e-01 -1.87295437e-01 1.02069449e+00
-3.45927894e-01 3.69474262e-01 -1.26217699e+00 -6.97301924e-01
-3.07818264e-01 -6.48766160e-02 -8.84167373e-01 -3.06958467e-01
3.40290815e-02 7.28157982e-02 -5.97994030e-01 4.72351193e-01
7.48942256e-01 -8.48641455e-01 6.25905693e-01 5.22000432e-01
-2.86581982e-02 -7.03098774e-02 -1.73962235e+00 1.12776423e+00
2.75429487e-01 6.66025460e-01 9.24095333e-01 -1.30340493e+00
3.61992091e-01 5.86013794e-01 8.65642309e-01 4.28336591e-01
-9.21782926e-02 1.82127625e-01 5.82531095e-02 -1.47107109e-01
3.85563433e-01 -2.45829359e-01 1.37483269e-01 9.87140715e-01
-4.59692895e-01 1.18630506e-01 4.54826653e-01 8.49407315e-01
1.79174531e+00 -8.94415617e-01 3.88543755e-01 -2.14112446e-01
2.42820904e-02 3.89621109e-02 7.22800434e-01 1.46977818e+00
-5.39847732e-01 5.91150880e-01 8.99370492e-01 2.26124585e-01
-1.15578306e+00 -1.12641752e+00 1.62105441e-01 7.51643240e-01
3.20931017e-01 -6.74746275e-01 -7.21434534e-01 -1.83898002e-01
3.84003646e-03 -7.09022582e-02 -4.56066340e-01 2.64133632e-01
-4.02588189e-01 -8.77153695e-01 4.46718901e-01 2.10435987e-01
2.16501430e-01 -5.72820127e-01 -3.37670922e-01 5.34818590e-01
-2.53894359e-01 -1.03127182e+00 -4.44268793e-01 5.81111163e-02
-1.11545920e+00 -1.31582475e+00 -6.33685827e-01 -7.69536853e-01
6.91319585e-01 9.45182979e-01 1.08728445e+00 4.11226034e-01
3.19069661e-02 7.79946625e-01 -5.65160334e-01 -1.07012227e-01
-6.55900478e-01 2.15593219e-01 1.10028557e-01 1.58976257e-01
3.28358859e-02 -7.71009564e-01 -8.19625258e-01 1.74182519e-01
-8.96532834e-01 -3.04233897e-02 3.08604866e-01 4.30400848e-01
2.16488913e-01 5.80588222e-01 3.60703260e-01 -8.14699292e-01
6.15633845e-01 -1.02367246e+00 -6.32940769e-01 5.20858876e-02
-5.08246541e-01 -4.09618169e-01 2.19433397e-01 -3.21806222e-01
-5.26877344e-01 -8.57012123e-02 4.27532613e-01 -2.06033409e-01
3.15108359e-01 9.79112089e-01 3.19552511e-01 -1.47133350e-01
3.30816120e-01 1.11681759e-01 3.01223304e-02 -1.35937139e-01
1.79983512e-01 6.85066164e-01 1.76663205e-01 -6.22379065e-01
9.51981068e-01 1.16412246e+00 1.25271797e-01 -1.05728555e+00
-4.38583195e-01 -8.87336314e-01 -5.99885941e-01 -4.57587093e-01
1.32421792e-01 -9.69029009e-01 -7.43271470e-01 5.37257552e-01
-1.18241191e+00 -4.57082212e-01 -8.09004456e-02 1.96980283e-01
-5.07775724e-01 8.71153116e-01 -9.95589375e-01 -1.37812304e+00
-7.87935629e-02 -5.10678113e-01 5.04289627e-01 -1.45061329e-01
-2.31553987e-01 -1.03566074e+00 3.62731725e-01 -3.81336212e-02
1.27095580e-01 2.91606933e-01 4.20352966e-01 -6.36898935e-01
-8.83877993e-01 -3.39311928e-01 -1.74659669e-01 -9.62665584e-03
-1.49213344e-01 5.83535135e-01 -3.81594867e-01 -6.90767884e-01
-9.36559401e-03 6.28662258e-02 9.46299791e-01 6.10118628e-01
8.70837271e-01 -3.32787216e-01 -7.35847473e-01 1.48030013e-01
1.41324127e+00 -2.73601502e-01 1.64044291e-01 -3.83921936e-02
3.52932006e-01 5.70045233e-01 2.11860538e-01 8.91543567e-01
4.85335976e-01 2.33366355e-01 4.80174392e-01 1.85374528e-01
3.62954706e-01 -1.90600932e-01 3.35359961e-01 1.34660125e+00
-4.88809757e-02 -5.75688064e-01 -1.00274575e+00 1.04632878e+00
-2.05776453e+00 -1.24893606e+00 -5.27030170e-01 2.21915984e+00
8.81026447e-01 5.52099884e-01 5.07726908e-01 4.92730856e-01
1.19093871e+00 2.23257631e-01 -3.58670503e-01 5.21179549e-02
-2.97769696e-01 -7.77341276e-02 7.95013011e-01 5.26410758e-01
-9.95677650e-01 3.49954218e-01 6.94509935e+00 7.46839106e-01
-5.86134195e-01 2.92137891e-01 4.51781422e-01 -8.66534859e-02
-4.08637613e-01 4.64099258e-01 -6.24883056e-01 3.14431638e-01
9.55486953e-01 -6.39888585e-01 1.90883622e-01 6.07434511e-01
6.17877901e-01 -3.22437614e-01 -1.03271520e+00 5.93430459e-01
-3.36929172e-01 -1.36020851e+00 -3.89826000e-01 4.03864592e-01
1.04847145e+00 2.96877533e-01 -4.31529045e-01 -3.75617027e-01
1.02141273e+00 -6.14249349e-01 6.40396893e-01 1.42804280e-01
6.11522913e-01 -5.60791790e-01 2.90322185e-01 7.72275984e-01
-1.43383253e+00 -2.52942324e-01 -2.97873020e-01 -2.51073092e-01
4.53156531e-01 9.95144844e-01 -7.25936532e-01 4.04609174e-01
4.38267857e-01 9.05485690e-01 -1.73843220e-01 1.53517330e+00
9.29961577e-02 1.25311685e+00 -9.59271669e-01 -8.86479020e-02
1.76335931e-01 -2.27541983e-01 9.51175153e-01 1.13398910e+00
3.12346309e-01 1.71519488e-01 2.95071810e-01 5.36738634e-01
-2.08890930e-01 -2.23993719e-01 -5.64410210e-01 -7.20213950e-02
9.36521530e-01 1.08096516e+00 -1.32552743e+00 -3.22981268e-01
-5.31951547e-01 6.52861893e-01 8.31726640e-02 4.15253609e-01
-4.79956686e-01 -2.26137996e-01 3.97939831e-01 5.01266658e-01
5.71783066e-01 -5.24750650e-01 -2.42240965e-01 -1.23691463e+00
1.48670569e-01 -6.51834786e-01 5.19654691e-01 -3.28425050e-01
-1.28472507e+00 -2.20838800e-01 -1.95012905e-03 -9.44687724e-01
-4.99225333e-02 1.21056870e-01 -8.78931224e-01 1.94891542e-01
-1.08880353e+00 -8.29160929e-01 -2.62000114e-02 4.94915694e-01
1.71957389e-01 3.36988896e-01 3.00086528e-01 1.71534032e-01
-5.39906323e-01 1.83371663e-01 7.24866033e-01 3.38903487e-01
3.07866752e-01 -1.30136943e+00 3.65872234e-01 1.10907984e+00
1.28463790e-01 5.77823281e-01 1.04157472e+00 -8.27831030e-01
-1.29256880e+00 -8.43233883e-01 1.00267136e+00 -3.50136869e-02
1.09234667e+00 -6.60175085e-01 -7.70859420e-01 7.07519531e-01
4.24237214e-02 2.18126491e-01 5.38441479e-01 -2.78187245e-02
-2.39579841e-01 8.56389478e-03 -1.05261326e+00 5.46376646e-01
1.20161712e+00 -3.93875092e-01 -1.07721269e-01 5.07205725e-01
5.06373584e-01 3.00397664e-01 -7.95557201e-01 2.04446524e-01
4.65827107e-01 -8.83825719e-01 7.10035741e-01 -4.10602033e-01
1.79393724e-01 -3.95354778e-01 -9.72765908e-02 -9.63595450e-01
-2.89252281e-01 -1.21015847e+00 -4.07742709e-01 1.05649865e+00
1.29010081e-01 -6.15675628e-01 1.29540932e+00 1.99286997e-01
6.37120306e-01 -5.02239764e-01 -1.04741669e+00 -7.90917337e-01
2.36665249e-01 -3.80373389e-01 -3.07559725e-02 9.92835581e-01
3.10498238e-01 2.69960836e-02 -1.08552851e-01 2.62080133e-01
1.25926459e+00 1.01933762e-01 6.86405659e-01 -1.60875893e+00
-3.62914294e-01 -5.75697958e-01 -2.09351525e-01 -1.35058391e+00
-2.23362803e-01 -7.39828765e-01 -8.30485765e-03 -1.40954387e+00
8.58116388e-01 -7.53453434e-01 -8.53110328e-02 -1.84139967e-01
6.49613813e-02 1.76498946e-02 3.55614610e-02 6.79210305e-01
-1.02674878e+00 1.06191255e-01 6.31912291e-01 -5.53760827e-02
-2.99067679e-03 5.36002278e-01 -5.45912206e-01 8.05948615e-01
6.54793978e-01 -8.14439476e-01 -3.35031450e-01 -4.33621369e-02
8.53277326e-01 4.89862174e-01 3.49066764e-01 -7.67543256e-01
9.05144989e-01 -1.82440415e-01 -3.18742871e-01 -9.38761950e-01
-2.85961702e-02 -6.33973181e-01 2.58620292e-01 8.48363161e-01
-3.55930299e-01 -7.31027266e-03 -6.70835376e-01 1.35992730e+00
3.98806520e-02 -4.30797786e-01 7.07625926e-01 -7.93181285e-02
4.20133397e-02 5.29467046e-01 -8.73146415e-01 1.76504746e-01
1.22864342e+00 -3.82771254e-01 -4.32150159e-03 -1.04162788e+00
-7.97609687e-01 5.30249834e-01 7.85033166e-01 -3.14931542e-01
3.21316391e-01 -9.11310375e-01 -1.12238812e+00 -3.11111867e-01
1.03358209e-01 1.36796549e-01 -5.96227609e-02 1.09762108e+00
-6.42508268e-01 5.08060455e-02 5.86068273e-01 -8.20102394e-01
-1.48924637e+00 6.41083241e-01 2.32261360e-01 -3.20722431e-01
-3.78578931e-01 6.42913997e-01 -6.08100221e-02 8.41408223e-03
2.90217310e-01 4.12453972e-02 1.81244463e-01 1.90127864e-01
3.26834470e-01 5.44816434e-01 -2.74534672e-01 -3.63433957e-01
-1.96671084e-01 7.37734437e-02 -8.80482718e-02 -5.93713224e-01
1.32910430e+00 -6.33510113e-01 -4.45587397e-01 5.60405612e-01
7.87040770e-01 1.95793733e-01 -1.01716709e+00 -6.31340921e-01
2.97385722e-01 -2.45280504e-01 -2.10733965e-01 2.96793938e-01
-9.53225791e-01 5.63111961e-01 -1.29857972e-01 1.04865456e+00
7.74723589e-01 2.25803956e-01 3.48415583e-01 4.74308223e-01
7.51533031e-01 -9.20972228e-01 -1.19249620e-01 4.24604982e-01
2.78744876e-01 -9.66010392e-01 6.70718178e-02 -6.68620169e-01
1.54391125e-01 9.90343630e-01 -1.76796883e-01 -2.51284063e-01
1.14024591e+00 4.16747302e-01 -6.20944500e-01 -1.97739229e-01
-1.11609828e+00 -1.70320030e-02 -7.29861915e-01 4.90697920e-01
1.23278871e-01 1.65836737e-01 -2.78626025e-01 1.88360795e-01
1.25461265e-01 -3.76981758e-02 1.21596622e+00 1.12851357e+00
-9.59782720e-01 -1.10362542e+00 -4.19663221e-01 5.70860028e-01
-6.96547568e-01 -1.19640477e-01 -3.95597279e-01 5.98955393e-01
-5.14875293e-01 1.34504271e+00 2.00184837e-01 -2.33244877e-02
-2.31941342e-01 -3.41865480e-01 3.48001093e-01 -8.58277857e-01
-3.44017833e-01 2.62594759e-01 5.26310988e-02 -1.43561587e-01
-7.69839048e-01 -1.10090172e+00 -7.27648079e-01 -1.00609887e+00
-6.73057854e-01 3.71379048e-01 3.39280069e-01 7.34012246e-01
-4.18837108e-02 -6.18929006e-02 9.84727323e-01 -3.78689528e-01
-7.55999982e-01 -8.42470706e-01 -9.91234064e-01 -9.55987945e-02
4.63948756e-01 -1.99925795e-01 -7.37287104e-01 1.65524542e-01] | [6.888237953186035, 5.121920585632324] |
62f2c76b-d470-4f79-bbc4-9032921a1526 | what-makes-us-laugh-investigations-into | null | null | https://aclanthology.org/W18-1101 | https://aclanthology.org/W18-1101.pdf | What makes us laugh? Investigations into Automatic Humor Classification | Most scholarly works in the field of computational detection of humour derive their inspiration from the incongruity theory. Incongruity is an indispensable facet in drawing a line between humorous and non-humorous occurrences but is immensely inadequate in shedding light on what actually made the particular occurrence a funny one. Classical theories like Script-based Semantic Theory of Humour and General Verbal Theory of Humour try and achieve this feat to an adequate extent. In this paper we adhere to a more holistic approach towards classification of humour based on these classical theories with a few improvements and revisions. Through experiments based on our linear approach and performed on large data-sets of jokes, we are able to demonstrate the adaptability and show componentizability of our model, and that a host of classification techniques can be used to overcome the challenging problem of distinguishing between various categories and sub-categories of jokes. | ['Vikram Ahuja', 'Navjyoti Singh', 'Taradheesh Bali'] | 2018-06-01 | null | null | null | ws-2018-6 | ['humor-detection'] | ['natural-language-processing'] | [-2.07824960e-01 -3.03825647e-01 -6.77204579e-02 -2.89719477e-02
1.63807109e-01 -5.70737302e-01 9.44212496e-01 2.07537800e-01
-2.25337848e-01 6.16004586e-01 3.65819663e-01 -3.87160331e-01
-2.50902504e-01 -7.17910111e-01 6.29006177e-02 -3.51304531e-01
3.30268443e-01 3.08778942e-01 2.48560756e-01 -7.98986793e-01
8.67150009e-01 3.19716036e-01 -1.33688080e+00 1.77634805e-01
6.34918332e-01 3.47332329e-01 -3.38041745e-02 4.91508961e-01
-3.03902596e-01 1.63184953e+00 -6.72556400e-01 -5.76392651e-01
-9.85484011e-03 -1.03648162e+00 -1.14424312e+00 4.39121388e-02
1.69381142e-01 2.52458304e-02 -1.30164504e-01 9.12988424e-01
2.77154326e-01 3.54766920e-02 7.73461163e-01 -1.08586884e+00
-6.03313029e-01 6.11433864e-01 -2.55919516e-01 2.07066432e-01
8.10798049e-01 7.07036257e-02 9.89739656e-01 -7.37905145e-01
5.89318037e-01 9.99869227e-01 8.95970345e-01 4.09274936e-01
-1.05311334e+00 -2.32860565e-01 -6.65609717e-01 6.78682923e-01
-1.04713774e+00 -2.67483830e-01 1.12328172e+00 -7.69487143e-01
7.47713804e-01 5.96220434e-01 7.56003916e-01 9.27315533e-01
1.23137757e-01 4.57660586e-01 1.65251291e+00 -6.97118402e-01
1.51930436e-01 5.22984862e-01 3.49937677e-01 4.73842531e-01
2.05550089e-01 -1.77916184e-01 -5.93199849e-01 -2.12401003e-01
5.00263512e-01 -1.93488330e-01 -3.13461334e-01 1.43150032e-01
-8.29470336e-01 1.06149089e+00 2.98123807e-01 9.55149829e-01
-1.34288341e-01 -1.29566938e-02 8.62031281e-01 5.20733297e-01
2.55733337e-02 7.04192460e-01 1.65613040e-01 -3.36964041e-01
-1.17379856e+00 4.27884936e-01 1.15537691e+00 4.31860268e-01
4.21610415e-01 1.34109274e-01 1.39677197e-01 7.16224849e-01
-3.12488098e-02 2.09125414e-01 7.18707979e-01 -8.22140753e-01
-1.61111102e-01 8.16795409e-01 1.06905788e-01 -1.53135169e+00
-4.75047290e-01 -3.56533319e-01 -5.01770079e-01 2.96717614e-01
5.40431559e-01 4.18296039e-01 -3.07393819e-01 1.29809296e+00
9.35572386e-02 -3.48874003e-01 -2.33200729e-01 1.04186845e+00
8.49169314e-01 1.84558883e-01 7.61135593e-02 -3.12896609e-01
1.46239340e+00 -6.62653506e-01 -7.45947123e-01 -7.14339390e-02
6.65434062e-01 -1.06925476e+00 1.42274404e+00 5.59201241e-01
-9.66613770e-01 -2.21136630e-01 -1.07895112e+00 -1.88247427e-01
-2.84431785e-01 -3.65783423e-01 4.32487994e-01 7.92375207e-01
-6.03076041e-01 7.54668355e-01 -1.31214097e-01 -4.07116354e-01
-1.78545415e-02 -2.44829044e-01 -2.47399598e-01 3.40024233e-01
-1.23004043e+00 1.46893990e+00 3.61500382e-01 -7.27345347e-02
-3.00313324e-01 -2.85883963e-01 -4.51985091e-01 -1.62338555e-01
5.42087615e-01 -5.24069309e-01 9.48547006e-01 -1.37820005e+00
-1.17939568e+00 1.33845472e+00 -5.28718010e-02 -5.07942975e-01
6.76614940e-01 4.20854725e-02 -2.79417425e-01 2.53380895e-01
7.98491240e-02 -3.26229423e-01 8.87275875e-01 -1.24925876e+00
4.71895337e-02 -2.91083366e-01 8.90755677e-04 -1.51807934e-01
-4.01846319e-01 4.37432438e-01 4.08988625e-01 -4.90971118e-01
9.31506045e-03 -7.42814302e-01 3.15872729e-01 -5.14912426e-01
-3.34629089e-01 -2.85946935e-01 6.12220943e-01 -6.48459435e-01
1.65976000e+00 -1.75634074e+00 9.05716270e-02 6.09414950e-02
6.73077166e-01 2.92218477e-01 5.10847807e-01 1.05958486e+00
1.28029168e-01 2.36670196e-01 -6.21124581e-02 1.64965332e-01
2.92159617e-01 2.49387890e-01 -3.57838303e-01 4.80626136e-01
-3.30477744e-01 9.33151603e-01 -9.98672664e-01 -6.92641556e-01
1.82495371e-01 1.68996558e-01 -1.59342274e-01 3.55560482e-02
1.81560386e-02 1.33322060e-01 -2.86656111e-01 4.31860775e-01
3.02503407e-01 -3.29802841e-01 4.02202338e-01 2.82815516e-01
-2.62103707e-01 5.96190274e-01 -7.29771674e-01 9.46787179e-01
-2.46680707e-01 8.75224829e-01 -2.52231896e-01 -8.85563135e-01
1.09704292e+00 3.24074209e-01 1.03804246e-01 -6.37763023e-01
4.31303144e-01 5.08547485e-01 1.93341464e-01 -7.20287859e-01
8.17492247e-01 -9.67609584e-01 -1.66293398e-01 4.85285521e-01
-1.95046917e-01 -2.63232499e-01 2.49545962e-01 2.48091444e-01
8.39932561e-01 1.96219608e-01 9.77941215e-01 -4.87276196e-01
8.82129431e-01 2.22056344e-01 1.21450596e-01 6.47031248e-01
-3.70640129e-01 2.46917874e-01 7.27765203e-01 -6.55576587e-01
-1.22227442e+00 -7.64811397e-01 -3.60384613e-01 1.04692268e+00
4.03118581e-02 -5.29795945e-01 -5.91786504e-01 -2.23622978e-01
-1.44889832e-01 9.94384348e-01 -6.16284788e-01 1.05260268e-01
-4.59484875e-01 -5.28430879e-01 7.24641740e-01 6.43223226e-02
4.37716335e-01 -1.20350587e+00 -1.23900640e+00 2.11052984e-01
-1.89782932e-01 -9.19950187e-01 2.64560521e-01 1.34350792e-01
-5.87362587e-01 -1.33468843e+00 -1.53916314e-01 -5.39402246e-01
9.88454744e-02 4.78853852e-01 1.26210213e+00 9.04113412e-01
-2.28966564e-01 2.33786240e-01 -7.07998097e-01 -3.97817492e-01
-9.73471224e-01 -3.33836526e-01 -5.78682050e-02 -5.30269086e-01
8.83128881e-01 -7.97946990e-01 -2.17415005e-01 2.51181573e-01
-8.47205460e-01 -2.68283840e-02 -1.16007909e-01 8.44236493e-01
-3.93982828e-01 -1.36274904e-01 6.68147564e-01 -1.09830546e+00
8.85630488e-01 -7.19322205e-01 6.49728924e-02 -1.01839550e-01
-8.15153658e-01 -2.02563897e-01 9.76997733e-01 -1.51310831e-01
-7.81607389e-01 -6.47097468e-01 -2.82995272e-02 -3.55468839e-02
-5.32998238e-03 5.02150834e-01 3.11496556e-01 -7.86644816e-02
1.21582985e+00 3.73685420e-01 4.19391170e-02 -3.30648720e-01
4.21481252e-01 7.89650381e-01 6.42418206e-01 -7.19880104e-01
7.55748987e-01 4.46151942e-01 1.36244312e-01 -8.93548727e-01
-8.50318134e-01 -6.82748497e-01 -6.78021491e-01 -5.57666183e-01
4.72273290e-01 -4.23436642e-01 -9.44029212e-01 2.60229617e-01
-1.12547934e+00 -7.18984695e-04 -1.36281922e-01 1.06682619e-02
-6.56853855e-01 8.95462573e-01 -6.80068910e-01 -1.19833803e+00
-2.08442643e-01 -4.75034416e-01 2.19555646e-01 7.95331374e-02
-9.35186923e-01 -1.23153031e+00 2.54194349e-01 5.46112597e-01
3.65142643e-01 4.50338691e-01 1.06649971e+00 -8.34209681e-01
4.68881391e-02 -2.98383027e-01 4.86867167e-02 2.25016356e-01
6.88178241e-02 -1.02010325e-01 -8.09088707e-01 1.62126780e-01
7.31378257e-01 -5.67094386e-01 6.20798171e-01 -2.75075018e-01
2.55609781e-01 -5.10161340e-01 1.29804909e-01 -1.92435402e-02
1.73772752e+00 -3.43658440e-02 9.09970284e-01 8.67867351e-01
4.37127948e-01 8.61111462e-01 3.64511609e-01 6.41838133e-01
2.33408108e-01 7.35308349e-01 4.11302030e-01 4.50811684e-01
-1.83682427e-01 -2.24983469e-01 2.27857992e-01 1.07242465e+00
-3.11885715e-01 6.95695430e-02 -1.10105586e+00 6.63820326e-01
-1.74438310e+00 -1.62979770e+00 -1.08823383e+00 2.07504678e+00
9.54784811e-01 3.00130934e-01 5.44818878e-01 6.28513455e-01
3.93118411e-01 3.00668478e-01 2.97798336e-01 -9.59632218e-01
-1.26513541e-01 -5.95679060e-02 -3.18652280e-02 5.40157199e-01
-4.93098795e-01 9.49557006e-01 6.60586357e+00 7.32303679e-01
-9.27176058e-01 7.40685835e-02 2.16028132e-02 1.76094472e-01
-4.47811812e-01 2.28980020e-01 -2.40884691e-01 6.84631765e-01
8.26576650e-01 -4.01237011e-01 5.97473502e-01 8.31217110e-01
4.16064292e-01 -5.42550981e-01 -9.37630892e-01 8.39807808e-01
6.02619708e-01 -1.08770800e+00 -2.15289131e-01 -2.05103680e-01
5.39869249e-01 -5.11505485e-01 -3.00537258e-01 1.29359275e-01
-4.30639051e-02 -1.34792888e+00 7.60740638e-01 4.54167485e-01
1.95809364e-01 -3.57698172e-01 8.76843452e-01 5.15033364e-01
-4.88591135e-01 -2.88066175e-02 -3.19773525e-01 -9.89987850e-01
1.14373200e-01 3.33594501e-01 -8.37188363e-01 4.61260736e-01
3.40148568e-01 4.79418635e-01 -5.10513723e-01 8.16621244e-01
-4.58787203e-01 6.54954374e-01 -2.96232644e-02 -4.71993506e-01
2.18748361e-01 -2.80001700e-01 7.14149654e-01 1.42586672e+00
-3.91086005e-02 -2.09901016e-02 -2.24224389e-01 1.08568525e+00
3.50805074e-01 4.38651234e-01 -6.81421638e-01 -1.71698034e-02
4.19063151e-01 1.31251609e+00 -8.11328113e-01 -3.35966617e-01
-3.09216112e-01 8.41156006e-01 3.02651137e-01 -2.68451184e-01
-7.21242428e-01 -1.88291594e-02 7.40002245e-02 2.86957115e-01
-1.28086254e-01 -3.09764653e-01 -9.93845642e-01 -9.42386806e-01
-1.08670242e-01 -9.68572319e-01 1.89907119e-01 -7.42815137e-01
-1.32008827e+00 3.62684280e-01 -7.06173182e-02 -1.06359982e+00
-3.05331260e-01 -7.02985048e-01 -7.96925664e-01 8.75120044e-01
-1.12340450e+00 -1.16290510e+00 -4.74625438e-01 3.32442045e-01
2.77122170e-01 8.49360600e-02 8.89964223e-01 -1.58940315e-01
1.00956544e-01 1.17143162e-01 -7.29314759e-02 2.50380784e-02
6.44656122e-01 -1.16921484e+00 -2.37663984e-01 6.79357290e-01
1.15102813e-01 7.30699003e-01 1.50412965e+00 -5.25568962e-01
-1.05916166e+00 -8.08090493e-02 1.37174678e+00 -9.06809092e-01
1.31354916e+00 1.82573318e-01 -1.18434238e+00 2.90029734e-01
4.08831447e-01 -7.74714470e-01 7.70468473e-01 3.03050965e-01
-7.81993806e-01 6.01455986e-01 -8.39570522e-01 5.87312162e-01
7.52925634e-01 -6.61461353e-01 -1.42822301e+00 3.58549237e-01
-2.14217766e-03 1.01790071e-01 -6.65457487e-01 -5.80688268e-02
5.22373617e-01 -1.59449935e+00 6.37670338e-01 -7.07663238e-01
1.19611728e+00 -1.48499027e-01 -3.06481179e-02 -8.55514884e-01
-3.70537281e-01 -6.23353302e-01 2.19140708e-01 1.06715465e+00
-2.58199304e-01 -4.09113020e-01 5.39697289e-01 2.91233361e-01
-3.14009450e-02 -5.09122968e-01 -8.61305177e-01 -8.66876543e-01
5.11116147e-01 -5.96781015e-01 5.05593494e-02 1.38296163e+00
9.15181637e-01 6.00294650e-01 -8.61684680e-01 -7.95002580e-01
4.35243577e-01 2.30420053e-01 9.33820367e-01 -1.27156460e+00
-3.76546174e-01 -8.83913517e-01 -7.03966796e-01 -4.72739667e-01
7.25426152e-02 -1.05601442e+00 -2.41290197e-01 -1.26227736e+00
7.29138017e-01 6.98525012e-02 5.09722764e-03 3.54344040e-01
-1.00493051e-01 5.52887321e-01 6.22177124e-01 6.48053288e-01
-4.22540963e-01 1.85474783e-01 9.79692936e-01 2.23950699e-01
-1.62540048e-01 -5.22809148e-01 -7.72333682e-01 9.26712871e-01
8.17168117e-01 -5.38102746e-01 -2.17514932e-01 3.22935790e-01
7.71836162e-01 6.71465769e-02 8.67457747e-01 -6.11108720e-01
1.38761491e-01 -4.53888357e-01 -1.30708441e-01 -7.42003620e-02
1.32277429e-01 -7.13009715e-01 2.31299937e-01 5.76780796e-01
-1.59213871e-01 -3.45511198e-01 -9.13174972e-02 2.65459597e-01
-8.36919174e-02 -9.29480970e-01 8.09919655e-01 -5.09083331e-01
-6.55208111e-01 -7.17974544e-01 -4.89775509e-01 1.81576878e-01
8.55910003e-01 -6.74344480e-01 -6.12363935e-01 -5.01215041e-01
-5.86174667e-01 -2.05373973e-01 9.04888153e-01 2.78898656e-01
3.58541280e-01 -1.21517515e+00 -7.91039288e-01 -3.20171773e-01
1.11502752e-01 -1.22584486e+00 1.97619379e-01 1.34964061e+00
-8.07810068e-01 3.94502014e-01 -5.13642192e-01 -1.35901690e-01
-1.38672745e+00 7.38545656e-01 3.05014133e-01 -1.25782967e-01
-8.45228136e-01 2.52607673e-01 -1.83097079e-01 -2.34819017e-02
-3.79872471e-01 6.62542582e-01 -4.12237287e-01 1.97067618e-01
4.83861804e-01 7.69582689e-01 -2.73983836e-01 -1.16067910e+00
-3.97819847e-01 2.75024384e-01 2.12925822e-01 -4.03918885e-03
1.04082501e+00 -1.64841235e-01 -6.46824658e-01 1.10685444e+00
8.28827441e-01 4.33165312e-01 -4.18281883e-01 9.43656638e-03
3.65091801e-01 -8.22545350e-01 -3.59128058e-01 -9.40926552e-01
-1.87323205e-02 7.52554178e-01 -8.00255761e-02 9.26887751e-01
8.62850130e-01 -9.27134603e-02 6.93320632e-01 1.20686047e-01
4.96734530e-01 -1.24830616e+00 1.92415103e-01 5.28337121e-01
9.43347394e-01 -9.77004588e-01 2.75692761e-01 -3.50307196e-01
-7.27133095e-01 1.60479093e+00 2.90206194e-01 -4.27296788e-01
1.18327685e-01 -2.02081371e-02 7.28115588e-02 -4.95676577e-01
-5.83899856e-01 -1.91750169e-01 2.65012294e-01 2.56099701e-01
9.43356991e-01 2.35141665e-01 -1.45597875e+00 6.01526856e-01
-6.36315823e-01 -1.30259087e-02 1.04633987e+00 6.28689766e-01
-9.69192326e-01 -8.85529220e-01 -6.92517936e-01 1.52331412e-01
-6.20793402e-01 -1.23471327e-01 -9.49943304e-01 1.13727784e+00
1.03252001e-01 1.07498336e+00 -5.32499492e-01 -6.36982381e-01
-1.00806236e-01 3.88656199e-01 6.33132339e-01 -2.67562002e-01
-1.01017654e+00 -1.71387076e-01 3.95080566e-01 -2.09760055e-01
-7.12062418e-01 -4.74276811e-01 -1.14660227e+00 -1.01596117e+00
-2.07653463e-01 2.72795647e-01 4.54001904e-01 1.30922210e+00
-4.08922732e-01 7.16616511e-02 4.67243969e-01 -4.53627110e-01
-8.10078621e-01 -1.01461399e+00 -7.28832483e-01 9.25979793e-01
-3.30926739e-02 -4.14417893e-01 -6.88484251e-01 9.55089256e-02] | [8.91905403137207, 10.96629810333252] |
14332219-6965-4bbf-b538-182ebe2e7e09 | cross-lingual-cross-modal-consolidation-for | null | null | https://aclanthology.org/2022.findings-naacl.142 | https://aclanthology.org/2022.findings-naacl.142.pdf | Cross-Lingual Cross-Modal Consolidation for Effective Multilingual Video Corpus Moment Retrieval | Existing multilingual video corpus moment retrieval (mVCMR) methods are mainly based on a two-stream structure. The visual stream utilizes the visual content in the video to estimate the query-visual similarity, and the subtitle stream exploits the query-subtitle similarity. The final query-video similarity ensembles similarities from two streams. In our work, we pro- pose a simple and effective strategy termed as Cross-lingual Cross-modal Consolidation (C3 ) to improve mVCMR accuracy. We adopt the ensemble similarity as the teacher to guide the training of each stream, leading to a more powerful ensemble similarity. Meanwhile, we use the teacher for a specific language to guide the student for another language to exploit the complementary knowledge across languages. Ex- tensive experiments on mTVR dataset demonstrate the effectiveness of our C3 method. | ['Ping Li', 'Mingming Sun', 'Hanyu Peng', 'Tan Yu', 'Jiaheng Liu'] | null | null | null | null | findings-naacl-2022-7 | ['moment-retrieval', 'video-similarity'] | ['computer-vision', 'computer-vision'] | [-2.28310928e-01 -8.85700941e-01 -4.20735240e-01 -8.25714394e-02
-1.22415483e+00 -6.39337420e-01 8.30938697e-01 8.96009579e-02
-4.32456136e-01 1.42491758e-01 4.00861770e-01 -1.13249190e-01
7.52164051e-02 -3.78098458e-01 -5.56961119e-01 -6.56308651e-01
5.10070145e-01 5.18745743e-02 5.14530957e-01 -2.37646848e-01
4.93485659e-01 2.24931568e-01 -1.48806942e+00 5.53095341e-01
1.11299598e+00 8.50506365e-01 6.43530071e-01 4.70063150e-01
-5.42696178e-01 8.97495210e-01 -3.68043512e-01 -4.56018090e-01
-5.82188852e-02 -4.98044729e-01 -4.19762641e-01 2.29929462e-02
4.79756445e-01 -2.24874288e-01 -6.13118291e-01 1.04316771e+00
7.14341521e-01 5.28545797e-01 6.69673920e-01 -1.11447573e+00
-6.37717605e-01 5.16988993e-01 -6.66221380e-01 3.10440361e-01
7.11489379e-01 -5.03169820e-02 9.94308054e-01 -1.38217068e+00
8.63453627e-01 1.34078789e+00 1.86608419e-01 2.81318516e-01
-7.83444941e-01 -8.20491433e-01 4.06283855e-01 7.57162511e-01
-1.63635254e+00 -4.79305863e-01 8.80756736e-01 -3.27213943e-01
5.31413674e-01 1.40332669e-01 7.16107547e-01 7.95637965e-01
-2.55349372e-02 1.40973258e+00 8.54008913e-01 -3.00765693e-01
-1.16765253e-01 2.21376210e-01 -1.05015941e-01 6.56379461e-01
-3.59443754e-01 -1.05067447e-01 -7.27731109e-01 -7.80862644e-02
5.17543018e-01 2.39591941e-01 -4.03003275e-01 -4.44722772e-01
-1.33217728e+00 7.75679231e-01 2.83879310e-01 6.00244224e-01
-2.06528619e-01 -1.69935390e-01 6.86117887e-01 3.85785460e-01
3.97706091e-01 3.65940779e-02 -9.08550154e-03 -1.78663447e-01
-1.18792069e+00 3.58231569e-04 2.81395912e-01 1.03932822e+00
8.42298031e-01 -1.63607020e-02 -2.77412176e-01 1.18189383e+00
5.34434199e-01 6.31869137e-01 9.05011714e-01 -8.65635574e-01
7.34786451e-01 4.94448036e-01 -3.30306292e-01 -9.66738045e-01
3.92637789e-01 -2.10963264e-01 -4.46801782e-01 -4.47479278e-01
-1.10941343e-01 2.43833125e-01 -8.66785586e-01 1.48973811e+00
4.77328151e-01 4.78097200e-01 3.13616961e-01 9.06674922e-01
1.06249499e+00 1.01831698e+00 6.96314573e-02 -3.90761852e-01
1.26232517e+00 -1.21440578e+00 -6.09406650e-01 1.20961286e-01
7.44717002e-01 -1.17580068e+00 9.77101624e-01 2.82191664e-01
-1.01596618e+00 -6.94540143e-01 -1.00650918e+00 -1.85738489e-01
-2.35955685e-01 2.37044081e-01 1.41764328e-01 -4.45520282e-02
-7.76987731e-01 2.10761443e-01 -5.38657129e-01 -3.30656707e-01
3.21626998e-02 -2.12364376e-01 -4.17392671e-01 -6.16150975e-01
-1.12313449e+00 5.04026651e-01 5.47865152e-01 -1.02928787e-01
-9.49075162e-01 -5.70767939e-01 -9.90270495e-01 -1.82709366e-01
3.52494836e-01 -3.24447930e-01 9.75540400e-01 -1.01983225e+00
-1.28434563e+00 7.56009042e-01 -4.14593846e-01 -4.79719490e-02
3.78359258e-01 -2.78976530e-01 -5.34639835e-01 5.87675512e-01
2.10634872e-01 6.52946413e-01 8.39488268e-01 -1.26696205e+00
-9.00633693e-01 -2.02976689e-01 -7.62409493e-02 6.61797047e-01
-3.78287703e-01 1.11033924e-01 -1.12594509e+00 -9.86775279e-01
2.56334811e-01 -8.25586140e-01 2.59306759e-01 -3.87610048e-01
7.80866221e-02 -4.91347969e-01 9.03328955e-01 -7.06130445e-01
1.63028216e+00 -2.30229497e+00 4.20498639e-01 1.05163850e-01
1.56986006e-02 2.39694849e-01 -4.32508498e-01 6.69706583e-01
-9.74866599e-02 -2.70762205e-01 2.58538127e-01 -2.29362026e-01
-3.34136695e-01 8.51482004e-02 -4.26672339e-01 2.94741184e-01
-1.00799248e-01 7.91776776e-01 -1.22781181e+00 -1.20356977e+00
1.98223159e-01 2.85692960e-01 -4.50968951e-01 5.10282993e-01
1.55460134e-01 1.13180876e-01 -6.78212404e-01 6.63804471e-01
3.36058557e-01 -1.49379581e-01 1.36180788e-01 -4.92114246e-01
-3.26772869e-01 4.50017862e-02 -1.00693810e+00 2.10619140e+00
-5.78311205e-01 5.48287094e-01 -2.06871614e-01 -8.57304454e-01
9.53705192e-01 4.48710024e-01 5.20811021e-01 -7.99541116e-01
-3.77736166e-02 3.14033061e-01 -1.87008351e-01 -6.93672121e-01
6.31411374e-01 9.97437239e-02 3.94144654e-02 4.33692932e-01
1.70214400e-01 5.28097861e-02 4.05700862e-01 7.49616206e-01
5.96026659e-01 5.27539790e-01 2.88842708e-01 3.79141644e-02
9.54769731e-01 -3.80220041e-02 5.39149404e-01 3.57988566e-01
-2.86432266e-01 5.87410450e-01 7.38761202e-02 -8.60121995e-02
-8.05294216e-01 -1.10655689e+00 2.25248232e-01 1.27090800e+00
4.62543279e-01 -7.49086857e-01 -3.77220809e-01 -7.30596721e-01
-2.01469153e-01 4.85492319e-01 -2.56117582e-01 -2.86719561e-01
-6.87600553e-01 5.77334687e-03 1.97330937e-01 5.60367763e-01
5.28818011e-01 -8.96500587e-01 -1.88248843e-01 1.02878079e-01
-6.31597042e-01 -1.03462911e+00 -1.10701191e+00 -4.37380284e-01
-6.27938628e-01 -8.24666977e-01 -1.09380305e+00 -9.96823132e-01
3.91972274e-01 9.24841702e-01 8.36589336e-01 1.98067203e-01
1.52631730e-01 6.36317968e-01 -7.65559256e-01 2.33379349e-01
-2.06111193e-01 -8.82612392e-02 1.30081668e-01 3.87476832e-02
4.45890725e-01 -4.99721497e-01 -6.61940038e-01 2.07526609e-01
-7.98219919e-01 9.46552232e-02 5.47225654e-01 8.02509785e-01
9.02403474e-01 -3.04224133e-01 3.99933249e-01 -4.65544641e-01
5.72009683e-01 -6.72022045e-01 -3.15055668e-01 7.36905932e-01
-4.79712427e-01 -2.52654199e-02 5.64608634e-01 -8.35671425e-01
-9.57911789e-01 3.64339980e-03 4.04201709e-02 -1.19474924e+00
1.70572594e-01 6.95966423e-01 -1.10042103e-01 8.25910047e-02
6.28680214e-02 7.20390975e-01 -3.84388715e-01 -3.62817615e-01
4.42169458e-01 7.51352966e-01 7.64454246e-01 -8.16260993e-01
7.41796613e-01 1.83194652e-01 -3.95073146e-01 -6.62384391e-01
-5.69385529e-01 -8.86685908e-01 -4.91692036e-01 -5.71793854e-01
9.71116960e-01 -1.33420956e+00 -3.54075193e-01 1.89130008e-01
-1.11929107e+00 1.59531787e-01 1.69043228e-01 7.48746216e-01
-2.33444214e-01 6.54615462e-01 -6.06859624e-01 -5.75525820e-01
-3.68074685e-01 -1.25308728e+00 1.12591040e+00 3.00742567e-01
1.82429701e-01 -7.97446072e-01 2.52392232e-01 3.73320788e-01
-2.57944316e-02 -3.61822009e-01 6.52711749e-01 -8.57515633e-01
-5.63397229e-01 -7.40665942e-02 -3.03625703e-01 1.51736304e-01
1.64881609e-02 2.71669894e-01 -6.08583331e-01 -4.76406246e-01
-1.57382697e-01 -4.21036601e-01 8.12937438e-01 2.99258232e-02
9.07285273e-01 -2.98728030e-02 -1.48312435e-01 5.04917979e-01
1.33654737e+00 3.94881457e-01 4.64505643e-01 3.24610978e-01
9.07295465e-01 4.07597184e-01 1.03719127e+00 1.95511505e-01
6.43975496e-01 7.59871900e-01 -2.48552665e-01 2.14414135e-01
-1.77869946e-01 -6.36749446e-01 7.45253205e-01 1.91303205e+00
-1.00221947e-01 2.10504085e-02 -6.68585539e-01 5.64132750e-01
-1.90181375e+00 -1.18594432e+00 2.47989282e-01 2.16559434e+00
8.37393761e-01 -2.13729456e-01 4.89610769e-02 -2.28690416e-01
7.41688013e-01 3.61158013e-01 -3.00259918e-01 -1.01015858e-01
-3.62956747e-02 -3.46985132e-01 -6.49411604e-02 3.25798035e-01
-7.54377306e-01 1.12356162e+00 5.27698660e+00 1.47001910e+00
-1.21093464e+00 1.93978697e-01 1.33117795e-01 -9.66629758e-02
-6.15016997e-01 1.58376694e-01 -6.35033429e-01 5.69401681e-01
5.45546710e-01 -4.01901692e-01 2.98973799e-01 7.20431507e-01
5.23281693e-02 -7.77583122e-02 -8.22558403e-01 1.31384146e+00
5.59871972e-01 -1.08895004e+00 4.74827915e-01 -1.64991572e-01
7.35328972e-01 -8.41567945e-03 2.26074662e-02 5.89680612e-01
1.46722108e-01 -3.36420417e-01 7.37747133e-01 5.69496810e-01
8.32330167e-01 -1.02041173e+00 3.96957695e-01 2.74166763e-01
-1.92506349e+00 2.11217955e-01 -3.44956100e-01 6.91692829e-01
2.35281542e-01 9.60306004e-02 -3.36069852e-01 8.76312196e-01
6.17802083e-01 1.14950883e+00 -8.15124333e-01 8.60979676e-01
-1.26770705e-01 1.78445294e-01 8.43839720e-02 1.36857480e-01
4.20340598e-01 -5.57541788e-01 6.40408516e-01 1.24090695e+00
4.60109055e-01 1.81686640e-01 4.76220280e-01 1.89199597e-01
-3.61934900e-02 4.90746081e-01 -5.17109096e-01 -1.22383408e-01
8.17275941e-01 1.32374704e+00 -4.71180260e-01 -7.11306930e-01
-7.04106271e-01 1.19312966e+00 4.20158833e-01 5.22488058e-01
-6.59992158e-01 -4.86579239e-01 2.92495579e-01 -3.51901621e-01
4.72976655e-01 -2.36288175e-01 4.01795447e-01 -1.60785341e+00
-1.28950626e-01 -1.11023176e+00 6.97288990e-01 -1.11349642e+00
-1.25969422e+00 5.33227086e-01 2.21810251e-01 -1.74268365e+00
-3.93321425e-01 -7.16799796e-02 -5.37369549e-01 7.14760244e-01
-1.52342308e+00 -1.26732647e+00 -9.60774571e-02 1.09725249e+00
8.97398472e-01 -3.82629514e-01 2.78975070e-01 5.30325413e-01
-7.11751521e-01 6.42883837e-01 9.09297466e-02 8.06579068e-02
1.04754353e+00 -1.03369284e+00 -2.10934877e-02 1.06812000e+00
4.89048660e-01 7.66187727e-01 4.44360018e-01 -8.33269417e-01
-1.54011655e+00 -9.56422329e-01 8.81670773e-01 -9.49547291e-02
8.81561458e-01 -4.33330005e-03 -1.11400425e+00 4.58317816e-01
3.62631917e-01 -1.78567633e-01 6.68974221e-01 -2.11105525e-01
-7.29018271e-01 -2.66153067e-01 -5.39131105e-01 8.34937930e-01
8.06889594e-01 -1.11381364e+00 -8.52736115e-01 5.11625297e-02
9.71894920e-01 -3.38654667e-01 -9.74144220e-01 2.86726892e-01
7.05018997e-01 -7.08604693e-01 8.70746911e-01 -2.99215257e-01
6.59687042e-01 -4.79571968e-01 -4.36155140e-01 -1.27870631e+00
-6.54399842e-02 -3.81039351e-01 -1.37245163e-01 1.57074046e+00
2.60136556e-02 -2.09529564e-01 2.92779863e-01 -5.10487817e-02
-3.09254080e-01 -7.34557867e-01 -5.32169878e-01 -6.79554462e-01
-4.45068963e-02 -4.75766987e-01 4.14499104e-01 1.13876379e+00
5.36852665e-02 4.03544337e-01 -3.95852685e-01 -5.06053492e-02
4.28894937e-01 5.34753740e-01 7.43296564e-01 -7.84822702e-01
-2.21760124e-01 -3.49176854e-01 -1.95364594e-01 -1.39161479e+00
3.36644322e-01 -1.06334102e+00 5.69918230e-02 -1.24971139e+00
5.44242620e-01 -1.07955001e-02 -5.68695664e-01 8.96320567e-02
-5.20671368e-01 6.62756562e-02 5.14231741e-01 6.14654422e-01
-1.14877415e+00 6.76993430e-01 1.38615072e+00 -8.25644135e-02
-1.69231772e-01 -2.88694471e-01 -2.59562492e-01 4.94007409e-01
3.59562963e-01 -3.25998545e-01 -6.61877036e-01 -4.86705452e-01
-1.70580037e-02 2.24714607e-01 -5.16208000e-02 -8.09746981e-01
5.19005537e-01 -2.53026247e-01 3.55124623e-01 -1.04972088e+00
3.21980298e-01 -6.66975617e-01 -5.91565520e-02 3.32818985e-01
-3.59279424e-01 5.67919970e-01 -1.09729141e-01 5.78637958e-01
-6.24630570e-01 -6.82523325e-02 5.11244237e-01 -2.39995290e-02
-1.02490246e+00 5.39400756e-01 -1.71456322e-01 1.88606218e-01
8.62539172e-01 -6.94744140e-02 -3.88886839e-01 -5.10508537e-01
-2.42575601e-01 6.07325256e-01 5.70515692e-01 7.04131365e-01
9.52012420e-01 -1.89358926e+00 -7.11037040e-01 -6.09097406e-02
4.89868134e-01 -4.58921164e-01 4.44106758e-01 8.47339988e-01
-2.66934097e-01 3.61058921e-01 -2.80664116e-03 -7.84363389e-01
-1.57603300e+00 8.10808420e-01 -8.43473822e-02 -1.52994901e-01
-5.49600422e-01 6.19431973e-01 4.60479558e-01 -3.68699394e-02
9.81227010e-02 3.48031521e-01 -4.66863662e-01 3.28817666e-01
7.41169631e-01 2.49454916e-01 -2.16916040e-01 -1.07945848e+00
-3.56188416e-01 8.78192008e-01 -4.18171227e-01 -3.86625111e-01
9.17693734e-01 -5.03175616e-01 5.63707389e-02 6.99676275e-01
1.55380869e+00 1.10440627e-01 -9.06871974e-01 -6.40363634e-01
-1.04978263e-01 -5.78353882e-01 -8.90563987e-03 -2.53667027e-01
-1.19901919e+00 8.84631574e-01 5.01348913e-01 -3.04955274e-01
1.38761604e+00 8.86614025e-02 1.00548506e+00 3.59128147e-01
2.81632304e-01 -1.11029458e+00 3.45556200e-01 4.65459049e-01
7.32666135e-01 -1.06156766e+00 1.89771906e-01 -1.64561629e-01
-9.53555107e-01 9.32110429e-01 5.98501205e-01 -2.01177616e-02
3.60041440e-01 -2.45182678e-01 1.59272939e-01 1.91277415e-02
-9.89382446e-01 -3.65988731e-01 6.99796557e-01 3.70663285e-01
4.16294038e-01 -2.36347005e-01 -5.47237635e-01 3.45442921e-01
2.11165130e-01 -1.56949773e-01 5.57951853e-02 9.77462351e-01
-4.18902308e-01 -1.15462530e+00 -2.63808280e-01 4.49213721e-02
-2.92954415e-01 -2.32125625e-01 -1.66037247e-01 7.04251587e-01
6.43197447e-02 8.75346780e-01 2.95475572e-02 -6.81599021e-01
2.33300552e-01 1.09915743e-02 4.83642399e-01 -4.62788194e-01
-5.69122672e-01 5.73418260e-01 -2.61363357e-01 -6.97942257e-01
-6.76264226e-01 -7.25043833e-01 -1.12286711e+00 -1.53123394e-01
-3.35127771e-01 4.25086409e-01 4.72652316e-01 1.08603644e+00
2.20666990e-01 2.60350376e-01 9.77879643e-01 -7.22561300e-01
-2.30796412e-01 -7.75126159e-01 -3.43724400e-01 4.89180177e-01
1.69455409e-01 -5.61306536e-01 -3.14146310e-01 1.45354807e-01] | [10.300559997558594, 0.9188964366912842] |
9dd33e77-6192-42e1-92ab-471a1f16898b | a-brief-survey-on-person-recognition-at-a | 2212.08969 | null | https://arxiv.org/abs/2212.08969v1 | https://arxiv.org/pdf/2212.08969v1.pdf | A Brief Survey on Person Recognition at a Distance | Person recognition at a distance entails recognizing the identity of an individual appearing in images or videos collected by long-range imaging systems such as drones or surveillance cameras. Despite recent advances in deep convolutional neural networks (DCNNs), this remains challenging. Images or videos collected by long-range cameras often suffer from atmospheric turbulence, blur, low-resolution, unconstrained poses, and poor illumination. In this paper, we provide a brief survey of recent advances in person recognition at a distance. In particular, we review recent work in multi-spectral face verification, person re-identification, and gait-based analysis techniques. Furthermore, we discuss the merits and drawbacks of existing approaches and identify important, yet under explored challenges for deploying remote person recognition systems in-the-wild. | ['Rama Chellappa', 'Thirimachos Bourlai', 'Shuowen Hu', 'Carlos D. Castillo', 'Joshua Gleason', 'Neehar Peri', 'Chrisopher B. Nalty'] | 2022-12-17 | null | null | null | null | ['person-re-identification', 'person-recognition'] | ['computer-vision', 'computer-vision'] | [ 3.84396404e-01 -1.01694441e+00 3.75817060e-01 -5.46682358e-01
-3.21122140e-01 -6.74363494e-01 4.33096617e-01 -3.90644282e-01
-6.87266350e-01 8.03028762e-01 -5.18720224e-02 4.40456212e-01
-2.07479000e-01 -5.83481610e-01 -4.03626740e-01 -8.16145301e-01
-2.87272573e-01 8.55729878e-02 -7.04313219e-01 7.37982839e-02
-2.02697128e-01 1.13156593e+00 -1.85497546e+00 -1.43735260e-01
3.60002369e-01 8.62288177e-01 -4.27674025e-01 1.06049633e+00
6.57094896e-01 3.47696692e-01 -1.04622579e+00 -6.70923054e-01
5.15977323e-01 -1.03887059e-01 -3.75261426e-01 3.47136259e-01
1.35282767e+00 -7.99337626e-01 -9.00379241e-01 1.15931916e+00
9.54966426e-01 3.79861683e-01 3.71055931e-01 -1.54768491e+00
-1.02562952e+00 -1.81815565e-01 -3.53358090e-01 5.18572092e-01
9.22762275e-01 3.33376646e-01 -1.54076489e-02 -7.21725106e-01
2.03374296e-01 1.20552611e+00 1.46488655e+00 8.23176265e-01
-8.70604873e-01 -6.86212420e-01 -4.02083583e-02 3.79925519e-01
-1.90160584e+00 -9.37099755e-01 4.41120803e-01 -2.82703727e-01
1.02392125e+00 6.23599254e-02 7.67954707e-01 1.56004465e+00
-2.53824115e-01 2.52263904e-01 6.28474772e-01 -2.53278852e-01
-3.72530818e-01 -4.43583846e-01 1.25970095e-01 6.33611739e-01
7.28513837e-01 4.21851486e-01 -7.87067413e-01 -2.83849180e-01
8.67537796e-01 3.32960188e-01 -6.37293816e-01 -9.89546478e-02
-1.14895070e+00 3.38574797e-01 2.17021629e-01 1.68518960e-01
-5.43102682e-01 1.51007980e-01 3.29370409e-01 2.72159785e-01
2.84215122e-01 3.83105651e-02 -3.56444754e-02 -1.20787784e-01
-1.03978360e+00 3.12040478e-01 5.41643858e-01 7.23538339e-01
4.32654113e-01 4.98854995e-01 1.15471251e-01 7.23594248e-01
1.52106673e-01 9.75602150e-01 1.75070316e-01 -6.72383428e-01
2.41538122e-01 1.02926195e-01 2.11066186e-01 -1.18272388e+00
-3.09330344e-01 -3.30498338e-01 -1.23105693e+00 4.29073237e-02
5.26564598e-01 -5.11701465e-01 -7.65405357e-01 1.35884750e+00
9.52631831e-02 6.85374141e-01 1.62106335e-01 1.16515005e+00
8.43181431e-01 2.82229424e-01 -1.11948319e-01 -3.46725173e-02
1.19932365e+00 -6.25022352e-01 -7.16724932e-01 -5.23653924e-01
-1.34674519e-01 -4.53575283e-01 2.89101988e-01 2.75016099e-01
-6.25219464e-01 -7.73492873e-01 -1.01064932e+00 2.86238104e-01
-4.73561078e-01 3.72141361e-01 2.99871922e-01 1.19788647e+00
-1.23580134e+00 6.75719619e-01 -5.03685534e-01 -8.64371896e-01
4.12198305e-01 5.23939490e-01 -7.37656057e-01 -3.93144488e-01
-1.10199428e+00 5.50868154e-01 -1.55858025e-01 7.93864429e-01
-6.64235592e-01 -4.76688057e-01 -1.07496178e+00 -2.71793753e-01
-6.75043045e-03 -7.56462872e-01 7.51684904e-01 -1.00695884e+00
-1.17764425e+00 1.02189159e+00 -1.94751918e-01 -2.65344948e-01
4.79030937e-01 -5.09648502e-01 -1.18242788e+00 1.58869848e-01
6.02943413e-02 2.03839108e-01 1.03928936e+00 -8.08155119e-01
-3.48942369e-01 -9.48208392e-01 -2.40667164e-01 2.07121491e-01
-3.11770111e-01 6.78206623e-01 -2.26773068e-01 -5.67160785e-01
-1.27423584e-01 -9.93571460e-01 1.69637695e-01 2.70280689e-01
-2.98296094e-01 7.89291933e-02 9.39998746e-01 -9.70238566e-01
5.28817236e-01 -2.26676416e+00 -2.15933487e-01 6.37177005e-02
-1.29088461e-01 5.35251677e-01 -1.77688822e-01 2.28724435e-01
-3.34225297e-01 -3.00802857e-01 6.44782698e-03 -4.39800441e-01
-8.07626843e-02 -4.86972891e-02 1.42725930e-01 9.86869454e-01
-1.12444825e-01 6.72285557e-01 -5.86829722e-01 4.86295968e-02
4.21274453e-01 1.02878261e+00 2.40245789e-01 9.71658677e-02
9.10467088e-01 4.25711602e-01 2.51448929e-01 1.19016576e+00
9.87818837e-01 9.21508372e-02 -5.91781959e-02 -3.87296200e-01
-4.76789149e-03 -4.82321560e-01 -1.53444147e+00 1.35072386e+00
1.71084926e-01 1.03527975e+00 4.90719050e-01 -9.61137116e-01
7.86134958e-01 5.04071176e-01 5.11031449e-01 -1.96379706e-01
1.63381472e-02 5.99608012e-02 -3.34294409e-01 -6.34168684e-01
5.22680104e-01 1.67904362e-01 2.43226647e-01 4.18826848e-01
1.57525185e-02 8.42788041e-01 7.65450448e-02 -3.53125215e-01
8.83078218e-01 1.44410012e-02 2.97038972e-01 8.84396359e-02
8.01358700e-01 -3.64682615e-01 4.67095256e-01 9.19217348e-01
-8.69453788e-01 7.00086772e-01 -6.13474131e-01 -1.06136990e+00
-9.30695653e-01 -9.33949649e-01 -5.29536940e-02 7.95184612e-01
-1.09050468e-01 9.42950174e-02 -7.03385472e-01 -2.99891204e-01
2.69901067e-01 -4.25897002e-01 -4.58309382e-01 -4.43870164e-02
-7.97420740e-01 -1.05843747e+00 1.26249242e+00 8.30801904e-01
1.01358187e+00 -6.82771206e-01 -4.25285667e-01 -6.60233293e-03
-3.08988631e-01 -1.55829716e+00 -5.87467313e-01 -5.12657583e-01
-3.70782077e-01 -1.30052006e+00 -1.22573233e+00 -8.62830579e-01
6.51626468e-01 8.07309091e-01 9.10503685e-01 2.30083928e-01
-5.69680512e-01 8.29684675e-01 3.93201113e-02 -1.09194398e-01
3.49706620e-01 -4.31199878e-01 1.05486810e+00 4.65543032e-01
9.57959712e-01 -3.28993201e-01 -5.98633230e-01 2.79978037e-01
-2.59366989e-01 -8.56375933e-01 7.09427372e-02 7.29811788e-01
4.98158969e-02 4.02556747e-01 4.00367022e-01 1.39490068e-02
6.43424511e-01 -8.82257300e-04 -4.56846833e-01 4.13720876e-01
-1.97873577e-01 -7.54621685e-01 3.92502934e-01 -5.20754457e-01
-9.05395508e-01 1.94983870e-01 -2.70989630e-03 -5.02766788e-01
-9.84337866e-01 1.37450084e-01 -2.82942623e-01 -6.24802113e-01
6.55543447e-01 3.76275212e-01 1.39249817e-01 -2.57610828e-01
-2.01076388e-01 8.98984373e-01 9.63251948e-01 -2.68227965e-01
1.10486650e+00 5.90734303e-01 -3.13728660e-01 -1.38244462e+00
-6.47512734e-01 -6.03212416e-01 -1.04594553e+00 -5.90622902e-01
1.01775861e+00 -1.34359777e+00 -1.08589149e+00 1.43335235e+00
-9.88456190e-01 8.02103151e-03 1.48287445e-01 5.23318708e-01
1.93629831e-01 5.95465302e-01 -7.01848090e-01 -1.18231297e+00
-5.03093362e-01 -6.19687498e-01 1.15711963e+00 6.83071792e-01
-1.04329279e-02 -8.61377895e-01 -4.65265126e-04 7.69622386e-01
4.53906596e-01 3.36349010e-01 -1.98271856e-01 -2.22388253e-01
-3.99859667e-01 -6.07099593e-01 -3.45402509e-01 2.36447036e-01
4.43304837e-01 7.68352821e-02 -1.27342045e+00 -8.50963414e-01
-4.44853365e-01 -1.01307094e-01 6.58984065e-01 5.88302493e-01
8.79398584e-01 -1.63400859e-01 -3.96631300e-01 1.15081656e+00
1.21513915e+00 2.34989390e-01 5.39855540e-01 3.84273887e-01
9.31712866e-01 5.97893178e-01 -3.91454510e-02 3.78409207e-01
1.94126353e-01 6.74541414e-01 -2.67278366e-02 7.33804330e-02
-7.60696009e-02 2.65169024e-01 4.61384147e-01 -6.02742694e-02
-6.05165839e-01 -3.34045142e-01 -9.00204778e-01 3.80099803e-01
-1.57899821e+00 -1.48915112e+00 -2.08619386e-02 2.16198254e+00
-1.32831350e-01 -6.14888549e-01 3.76339406e-01 2.31927663e-01
1.31243098e+00 8.73369947e-02 -5.72540998e-01 2.91922271e-01
-5.12281775e-01 -5.12554077e-03 5.31088769e-01 1.89802855e-01
-1.57242835e+00 4.33297604e-01 7.03088856e+00 -2.53636986e-02
-1.02263236e+00 -1.88035637e-01 1.43814281e-01 -2.99982995e-01
8.70564640e-01 -7.57842064e-01 -1.14612806e+00 2.65974432e-01
8.05462956e-01 1.50054365e-01 7.29495227e-01 6.50002182e-01
1.64482445e-02 4.04648855e-02 -1.14204168e+00 1.62436032e+00
9.91613686e-01 -9.91454482e-01 -3.27416331e-01 2.54403681e-01
5.65358520e-01 1.19932536e-02 2.22171947e-01 -1.36708971e-02
-1.40519828e-01 -1.15395463e+00 2.58853823e-01 5.50190628e-01
9.33194876e-01 -9.72087324e-01 1.03035247e+00 -9.21454281e-02
-1.51962578e+00 -3.05151433e-01 -4.69693124e-01 -2.41753891e-01
2.48415276e-01 2.64716625e-01 -3.32889289e-01 4.99088407e-01
1.35439289e+00 1.12581658e+00 -6.79331303e-01 9.16593134e-01
2.26930171e-01 4.27362397e-02 -2.52050906e-01 2.49228850e-01
-1.12695657e-01 -3.25691134e-01 5.78753412e-01 1.10411882e+00
6.34018123e-01 1.16588935e-01 2.13790119e-01 6.12993360e-01
-4.61016409e-02 -6.64111376e-01 -9.43999052e-01 6.05466627e-02
3.63775522e-01 1.05938911e+00 -1.92057684e-01 -2.16647536e-01
-4.07739252e-01 1.36086857e+00 5.15588373e-03 6.72299385e-01
-6.91938996e-01 -2.42596373e-01 1.33609223e+00 -4.86718044e-02
-6.53147232e-03 -5.85945010e-01 3.05208832e-01 -1.45178568e+00
2.18182996e-01 -8.15286040e-01 5.34637928e-01 -9.12528753e-01
-1.68495250e+00 5.97866416e-01 -1.76371843e-01 -1.14134431e+00
-6.95544556e-02 -9.38791096e-01 -5.34010589e-01 9.98174965e-01
-1.39884830e+00 -1.34578145e+00 -8.90133500e-01 1.11422467e+00
3.65554661e-01 -8.31122518e-01 1.03963459e+00 8.14491510e-01
-1.09266841e+00 8.28866363e-01 1.46313563e-01 8.47938299e-01
8.69686425e-01 -9.10487473e-01 6.25634551e-01 1.48382056e+00
-1.31380975e-01 7.24257171e-01 3.77645075e-01 -7.10150123e-01
-1.53235197e+00 -1.13174677e+00 8.08129489e-01 -3.27045798e-01
2.37918094e-01 -2.04964146e-01 -7.31047630e-01 8.96999896e-01
8.38604718e-02 3.00269932e-01 1.07808304e+00 2.08246559e-01
-3.52181703e-01 -2.50022173e-01 -1.46196282e+00 2.84149379e-01
1.22445178e+00 -8.35997581e-01 -2.00983286e-01 2.92365968e-01
-2.85340488e-01 -1.21168219e-01 -7.41958141e-01 8.05247575e-02
1.00638533e+00 -8.77222419e-01 1.72599411e+00 -4.46924210e-01
-4.26067561e-01 -2.59884000e-01 -3.14343214e-01 -1.22186995e+00
-5.98147988e-01 -4.61644471e-01 -9.22572762e-02 1.14926755e+00
-4.33833122e-01 -7.35491514e-01 1.02826107e+00 7.90297270e-01
3.15834463e-01 3.71427447e-01 -1.01735866e+00 -1.06112218e+00
-3.30771685e-01 -5.90650104e-02 7.85960793e-01 1.09509587e+00
-5.02557039e-01 -4.64018919e-02 -9.58019197e-01 9.09134209e-01
1.43604648e+00 -1.90144360e-01 8.15104902e-01 -1.54044843e+00
1.57527804e-01 1.57515854e-02 -8.61829817e-01 -5.89857578e-01
3.14548343e-01 -2.37598777e-01 -1.45114854e-01 -1.26146674e+00
1.66807085e-01 1.92569122e-01 1.90615598e-02 1.96749434e-01
-5.37850335e-02 7.29693294e-01 7.12903887e-02 1.73586115e-01
-3.00783813e-01 2.44666278e-01 5.43230176e-01 -3.90980065e-01
1.81547269e-01 -2.93073826e-03 -3.83056879e-01 6.65145874e-01
9.15091813e-01 -3.94727848e-03 2.68985957e-01 -8.01935375e-01
-9.63263214e-02 -1.66221902e-01 1.17007852e+00 -1.60097182e+00
6.01009011e-01 1.29106581e-01 1.17974186e+00 -5.97120464e-01
7.72494256e-01 -9.31588531e-01 4.82420653e-01 3.78029853e-01
1.82536393e-01 4.52053964e-01 8.76862854e-02 5.32685220e-01
-9.78680179e-02 1.75372213e-02 8.06941867e-01 -2.84297049e-01
-9.89064872e-01 5.67140043e-01 -6.93288445e-01 -4.67005312e-01
9.48072970e-01 -7.58191288e-01 -4.69657063e-01 -4.44665015e-01
-7.24061012e-01 -1.38821408e-01 4.36812818e-01 5.31109035e-01
8.10640872e-01 -1.55633187e+00 -9.00136650e-01 6.72590792e-01
6.11566976e-02 -5.51004410e-01 6.65953398e-01 5.04104018e-01
-5.89147270e-01 3.57166559e-01 -5.76575637e-01 -6.16117358e-01
-1.84270465e+00 2.24212036e-01 1.01582265e+00 4.53545332e-01
-5.52034438e-01 9.23034012e-01 -5.50674796e-01 -6.00656092e-01
1.89494610e-01 4.52506781e-01 -1.81467354e-01 -4.19097953e-03
1.04496133e+00 7.49349535e-01 9.06143710e-02 -1.36043453e+00
-8.32392693e-01 9.29300368e-01 2.55385190e-01 9.26371664e-02
1.29439902e+00 -3.17891508e-01 -7.32504427e-02 9.84760188e-03
1.00796640e+00 -4.59765583e-01 -1.19963360e+00 -1.78269610e-01
-5.47964215e-01 -7.87404418e-01 -2.46126279e-02 -4.83062685e-01
-1.11830902e+00 7.79160380e-01 1.20411634e+00 -2.61913836e-02
9.39639330e-01 -5.27590930e-01 7.97759116e-01 8.57955813e-01
4.13633466e-01 -1.08882868e+00 1.01457238e-02 6.11128867e-01
6.38433099e-01 -1.41842330e+00 7.03366101e-02 -1.58915296e-01
-5.44900633e-02 1.26568699e+00 6.84853911e-01 5.51899821e-02
6.83426559e-01 1.99111745e-01 2.89262623e-01 -2.30751606e-03
8.12438130e-02 -1.99133173e-01 1.26981094e-01 1.41699553e+00
2.75384337e-01 -1.18056014e-01 9.16339576e-01 -9.92475152e-02
-1.93606615e-01 -1.41652953e-02 3.91228735e-01 9.50974166e-01
-3.35211828e-02 -6.72706604e-01 -1.10598099e+00 2.25809619e-01
-3.69384766e-01 7.81656802e-02 -5.04587531e-01 3.97645414e-01
2.86440849e-01 1.16143131e+00 3.07510823e-01 -4.17384595e-01
2.89235413e-01 2.26666540e-01 5.95865905e-01 -1.91098273e-01
-5.30993938e-01 -3.00545901e-01 1.63808614e-01 -3.07249278e-01
-1.05271673e+00 -1.16077304e+00 -3.30716163e-01 -9.07225192e-01
-1.60564065e-01 -2.83749342e-01 4.76645947e-01 8.80816400e-01
2.97038406e-01 2.51204729e-01 3.30591738e-01 -1.21640849e+00
-2.02006653e-01 -8.41044843e-01 -8.53949666e-01 2.79148042e-01
7.89216816e-01 -6.03531539e-01 -4.14614499e-01 3.40651631e-01] | [14.328944206237793, 1.036224126815796] |
1be7a59a-8331-413a-9448-4c8b78267b54 | elfis-expert-learning-for-fine-grained-image | 2303.09269 | null | https://arxiv.org/abs/2303.09269v1 | https://arxiv.org/pdf/2303.09269v1.pdf | ELFIS: Expert Learning for Fine-grained Image Recognition Using Subsets | Fine-Grained Visual Recognition (FGVR) tackles the problem of distinguishing highly similar categories. One of the main approaches to FGVR, namely subset learning, tries to leverage information from existing class taxonomies to improve the performance of deep neural networks. However, these methods rely on the existence of handcrafted hierarchies that are not necessarily optimal for the models. In this paper, we propose ELFIS, an expert learning framework for FGVR that clusters categories of the dataset into meta-categories using both dataset-inherent lexical and model-specific information. A set of neural networks-based experts are trained focusing on the meta-categories and are integrated into a multi-task framework. Extensive experimentation shows improvements in the SoTA FGVR benchmarks of up to +1.3% of accuracy using both CNNs and transformer-based networks. Overall, the obtained results evidence that ELFIS can be applied on top of any classification model, enabling the obtention of SoTA results. The source code will be made public soon. | ['Petia Radeva', 'Bhalaji Nagarajan', 'Ignacio Sarasúa', 'Marc Bolaños', 'Jesús M. Rodríguez-de-Vera', 'Pablo Villacorta'] | 2023-03-16 | null | null | null | null | ['fine-grained-image-recognition', 'fine-grained-visual-recognition'] | ['computer-vision', 'computer-vision'] | [ 3.12696621e-02 1.98763292e-02 -3.33014280e-01 -6.16910934e-01
-6.24349475e-01 -6.56066775e-01 7.01005340e-01 6.84622005e-02
-2.78329670e-01 5.51313281e-01 1.67972237e-01 -1.49753243e-01
-3.94756198e-01 -7.99125254e-01 -6.28090918e-01 -4.28842843e-01
2.15191156e-01 6.30155683e-01 4.17634338e-01 -1.38850152e-01
2.71794260e-01 7.00601339e-01 -2.10389924e+00 1.09683859e+00
6.95431709e-01 1.66212451e+00 -3.74856256e-02 3.22946042e-01
-3.97946805e-01 1.23933303e+00 -4.61603671e-01 -4.03274208e-01
1.83270201e-01 -2.88402918e-03 -8.19800913e-01 -2.46288881e-01
9.75000441e-01 -7.39079490e-02 1.00227885e-01 7.64420152e-01
1.85155794e-01 -3.25064510e-02 8.06411326e-01 -1.20017993e+00
-4.21172947e-01 8.20199430e-01 -6.93283826e-02 1.38603583e-01
-1.30684227e-01 -7.05642775e-02 1.30840540e+00 -1.15334880e+00
5.35853446e-01 1.34038424e+00 8.63531649e-01 3.65071595e-01
-1.11098742e+00 -7.59289682e-01 2.76015699e-01 4.46154296e-01
-1.60532200e+00 -4.76804882e-01 6.03657365e-01 -8.50189805e-01
1.16761804e+00 3.90801132e-01 2.95354128e-01 9.69982564e-01
2.83559337e-02 5.97293556e-01 1.36580276e+00 -3.23957443e-01
2.20278889e-01 5.05355000e-01 7.49118209e-01 5.87197244e-01
1.87145576e-01 1.80685535e-01 -5.88399827e-01 -3.45794251e-03
3.46016526e-01 -8.20487291e-02 -8.20508972e-03 -6.18179917e-01
-8.08472872e-01 9.26681459e-01 7.55997300e-01 6.31921709e-01
-3.11931282e-01 -1.03787951e-01 6.09600723e-01 2.70058483e-01
5.93889117e-01 5.49372911e-01 -5.20086050e-01 4.13235784e-01
-1.05450118e+00 1.20062001e-01 5.91502726e-01 7.54327655e-01
9.22859609e-01 3.46727185e-02 -5.76921701e-01 9.08492744e-01
1.86385080e-01 1.01559214e-01 5.66338599e-01 -6.78212523e-01
2.51047909e-01 1.07126069e+00 -1.86496198e-01 -1.11925948e+00
-4.38252896e-01 -1.05211961e+00 -7.41535604e-01 2.53941864e-01
1.91057935e-01 3.38593900e-01 -1.22403300e+00 1.50696969e+00
1.39393181e-01 1.37989283e-01 -2.80891433e-02 7.47372270e-01
1.27576947e+00 3.45583081e-01 2.01097071e-01 1.65505633e-01
1.06069207e+00 -1.03395915e+00 -3.61753285e-01 -2.48537064e-01
3.29755783e-01 -3.00614655e-01 9.52336967e-01 5.34263611e-01
-7.26011992e-01 -1.04452574e+00 -1.03608155e+00 5.26140518e-02
-8.03842485e-01 4.29774165e-01 3.70008111e-01 7.03469932e-01
-1.09789288e+00 7.24657118e-01 -4.09033298e-01 -1.85153663e-01
7.15251267e-01 3.80286008e-01 -4.04904157e-01 -6.16023988e-02
-1.05533576e+00 1.04227579e+00 7.85551190e-01 1.84632868e-01
-1.12117434e+00 -8.45948517e-01 -5.02248883e-01 1.88774928e-01
2.90265113e-01 -4.70555335e-01 9.01924789e-01 -1.23620021e+00
-1.20616639e+00 1.02049756e+00 4.66435328e-02 -6.87342405e-01
5.22490680e-01 -1.67911157e-01 -3.48523915e-01 3.08924802e-02
-1.85311548e-02 5.91546953e-01 9.48749185e-01 -1.38298106e+00
-1.00963509e+00 -4.23009872e-01 1.37038857e-01 -1.10241577e-01
-5.97923160e-01 -4.86685894e-03 -1.43374771e-01 -3.87319684e-01
-2.65301108e-01 -6.33618474e-01 9.36614126e-02 -3.92544717e-01
-4.18538421e-01 -6.31432652e-01 7.53204346e-01 -6.06043577e-01
1.28586376e+00 -1.98840904e+00 4.30151165e-01 2.42962778e-01
4.90098625e-01 5.23069799e-01 -1.00547485e-01 2.03476414e-01
-7.59997517e-02 6.92249760e-02 9.41170081e-02 -2.45984584e-01
1.05285130e-01 -3.76473702e-02 -4.07460213e-01 2.02291667e-01
1.44782528e-01 7.46098757e-01 -5.22360802e-01 -3.46865296e-01
4.80819911e-01 2.90315509e-01 -2.20232889e-01 1.96328506e-01
-3.39032173e-01 2.60678530e-01 -2.01390490e-01 4.62691814e-01
6.34373903e-01 -1.66706413e-01 1.92484483e-01 -6.90124512e-01
-3.01090896e-01 2.73256838e-01 -1.05175078e+00 1.27098346e+00
-6.62544250e-01 4.19056684e-01 -4.43355411e-01 -1.05421185e+00
1.01562583e+00 6.08317107e-02 7.68195391e-02 -6.50204957e-01
3.60790938e-01 4.52171355e-01 -1.82180535e-02 -9.12923813e-02
4.98538822e-01 -7.94745311e-02 -9.69322398e-02 5.76337762e-02
5.87681532e-01 3.37808609e-01 3.21111679e-01 -1.40925243e-01
8.33385050e-01 1.56943083e-01 5.11224329e-01 -4.10391629e-01
7.46473372e-01 2.74736911e-01 4.45525914e-01 9.24345970e-01
6.07345738e-02 4.53238457e-01 4.98363711e-02 -8.49498391e-01
-9.08385992e-01 -1.00733924e+00 -3.18915337e-01 1.31738865e+00
-1.85374603e-01 -3.64497036e-01 -5.82058549e-01 -9.09160316e-01
2.23683059e-01 7.98379004e-01 -8.51161122e-01 -2.60814250e-01
-2.69765913e-01 -4.64118451e-01 5.91668487e-01 6.93299353e-01
4.66839135e-01 -9.42343473e-01 -7.26899028e-01 1.90654531e-01
2.19368618e-02 -1.11746287e+00 2.61014163e-01 4.81724381e-01
-6.98443174e-01 -1.08735394e+00 -4.69574809e-01 -4.41444248e-01
2.23488405e-01 -2.79254187e-02 1.43693459e+00 5.38600497e-02
-3.92368361e-02 1.32427603e-01 -5.37102461e-01 -3.38597476e-01
-4.10083026e-01 3.71545970e-01 -1.54901296e-01 3.47527295e-01
5.86438060e-01 -3.82176250e-01 -2.73811430e-01 3.67515057e-01
-4.75587249e-01 1.17572859e-01 6.85486376e-01 7.25105166e-01
6.28165543e-01 2.01052636e-01 5.90515018e-01 -1.03998947e+00
2.06055343e-01 -3.83647621e-01 -5.21380007e-01 5.00767767e-01
-5.49080968e-01 1.51419222e-01 8.24838638e-01 -4.02440488e-01
-1.00966263e+00 8.74719247e-02 -1.50080770e-01 -7.29904294e-01
-4.54168618e-01 4.93021548e-01 -2.36551344e-01 -3.09917390e-01
8.55962217e-01 9.40904617e-02 -6.19292319e-01 -5.35786211e-01
4.71387208e-01 8.71657729e-01 4.22706246e-01 -3.91382217e-01
5.25938928e-01 3.14899296e-01 -2.14941427e-01 -5.95667779e-01
-1.31756234e+00 -6.25755906e-01 -8.08799028e-01 -3.22183669e-01
8.20673704e-01 -1.05037987e+00 -3.09323430e-01 4.05516028e-01
-9.38892305e-01 -3.83374840e-01 -1.87627390e-01 6.53729141e-02
-3.41153115e-01 9.04445164e-03 -2.05056638e-01 -6.70025766e-01
-3.62958759e-01 -9.80349779e-01 9.20057416e-01 2.17992768e-01
-2.56088674e-01 -8.56782496e-01 -1.42111763e-01 3.98821086e-01
5.53255439e-01 1.88017711e-01 1.00964022e+00 -1.06282437e+00
-5.27182758e-01 -1.37171686e-01 -3.77700061e-01 6.04303300e-01
9.45616141e-02 -9.78370104e-03 -1.40413654e+00 -2.69591331e-01
-3.34722191e-01 -5.49515963e-01 1.21768832e+00 2.66968846e-01
1.16651428e+00 -1.43111572e-01 -4.19808269e-01 3.99570346e-01
1.68317974e+00 1.67099327e-01 4.37828839e-01 4.36403573e-01
9.54244733e-01 5.47711849e-01 4.74252731e-01 3.62859875e-01
4.15839851e-01 8.75981569e-01 6.16726339e-01 1.41578361e-01
-4.80314910e-01 5.04654646e-03 -6.36398271e-02 4.77730304e-01
-1.18975401e-01 -2.19303980e-01 -1.10524476e+00 6.88457489e-01
-1.78021550e+00 -7.90178716e-01 -2.05487713e-01 2.12297225e+00
5.87385297e-01 2.39123642e-01 1.24712691e-01 3.03594559e-01
6.83933675e-01 4.90649138e-03 -4.97817665e-01 -4.25586641e-01
-4.91160341e-02 3.62770736e-01 3.68725985e-01 2.04820186e-01
-1.37889099e+00 1.06956804e+00 5.62308502e+00 1.04157341e+00
-1.22275364e+00 2.44649485e-01 4.51140136e-01 2.60671318e-01
-6.73129186e-02 -2.60004103e-01 -1.16628671e+00 2.39532098e-01
1.10147810e+00 1.29871860e-01 1.24053605e-01 1.02877259e+00
-3.02351266e-01 2.56176084e-01 -1.22971463e+00 8.04782033e-01
2.04647541e-01 -1.49029422e+00 3.94271642e-01 -1.36318237e-01
5.84743440e-01 1.04561463e-01 4.99405004e-02 5.75151324e-01
5.61506867e-01 -1.11976182e+00 1.11522496e+00 7.21649051e-01
7.09766388e-01 -9.20524299e-01 9.74435568e-01 3.45190406e-01
-1.30063081e+00 -4.25776035e-01 -5.66476941e-01 6.53738156e-02
-4.50921476e-01 5.88611066e-01 -7.99218237e-01 9.01274383e-01
1.00533760e+00 8.25950325e-01 -9.13798869e-01 1.11615920e+00
-7.29050720e-03 5.16754866e-01 -1.40699089e-01 1.49055570e-01
2.65653640e-01 2.93766588e-01 4.97186854e-02 1.28980911e+00
1.91804186e-01 -3.39799374e-01 2.71832973e-01 7.64297366e-01
-1.94773637e-02 1.31840065e-01 -4.76094902e-01 -4.37605232e-02
3.59215736e-01 1.43164134e+00 -6.25044286e-01 -4.79373634e-01
-2.76834458e-01 5.92861116e-01 7.53809929e-01 -6.08029887e-02
-7.29150534e-01 -2.93792397e-01 4.40292656e-01 -1.29760895e-02
6.32331192e-01 2.14475811e-01 -4.59289193e-01 -1.02211726e+00
-2.84206420e-01 -1.10447907e+00 6.93237007e-01 -6.23282075e-01
-1.39172041e+00 1.12617004e+00 6.56601489e-02 -1.13011658e+00
-2.31504515e-01 -9.38243985e-01 -3.46124053e-01 7.01422453e-01
-1.53425813e+00 -1.59506059e+00 -5.74433923e-01 6.60044372e-01
4.76948738e-01 -4.98918176e-01 9.61242378e-01 2.80852109e-01
-4.58310932e-01 6.93238974e-01 3.84480432e-02 1.13361448e-01
4.33727294e-01 -1.33296132e+00 4.27132770e-02 8.82818282e-01
4.19441730e-01 4.12410289e-01 5.62729836e-01 -5.19659460e-01
-8.07323456e-01 -1.54855466e+00 8.98869574e-01 -5.80914259e-01
5.99655449e-01 -3.90399933e-01 -9.23279107e-01 6.18713915e-01
1.69104263e-01 1.27468377e-01 7.51586318e-01 3.49596977e-01
-9.20933723e-01 -4.28801537e-01 -1.04522598e+00 1.19033113e-01
1.01915050e+00 -6.38946414e-01 -7.79879689e-01 1.17125914e-01
4.52803761e-01 -2.20843166e-01 -9.74776626e-01 6.66869879e-01
6.33370101e-01 -1.20465899e+00 9.85539377e-01 -7.01141775e-01
4.65321392e-01 -3.39210033e-01 -5.01945376e-01 -1.45363379e+00
-5.55291414e-01 3.23622257e-01 -1.16919465e-01 1.27587509e+00
3.67909282e-01 -4.28897977e-01 5.93060076e-01 5.68357073e-02
-2.93756843e-01 -4.09209609e-01 -8.83836746e-01 -8.49717796e-01
1.74014494e-01 -5.40211380e-01 5.40804505e-01 7.72778749e-01
-5.38108468e-01 4.86472875e-01 -3.63762647e-01 2.13423356e-01
6.89864874e-01 4.31524932e-01 6.16684318e-01 -1.73261559e+00
-2.28348225e-01 -5.39859653e-01 -5.39706469e-01 -2.85873353e-01
3.89492691e-01 -1.13370693e+00 2.67810747e-03 -1.61933327e+00
9.53766927e-02 -5.84514976e-01 -7.33359993e-01 8.61812234e-01
1.15545683e-01 5.48310280e-01 3.37385923e-01 1.37539938e-01
-7.50384271e-01 2.96745896e-01 6.84018731e-01 -3.54195446e-01
6.09178878e-02 -5.39988093e-02 -6.45275116e-01 6.93277717e-01
8.32004428e-01 -4.19892013e-01 -4.48278010e-01 -3.48703831e-01
8.26118290e-02 -4.01343465e-01 5.50571978e-01 -1.37558866e+00
1.31040752e-01 5.15192971e-02 3.87431800e-01 -7.10466385e-01
2.02978119e-01 -9.83860373e-01 3.98727208e-01 3.03947419e-01
-4.88820761e-01 -3.68949443e-01 3.70307863e-01 2.95090824e-01
-4.12611812e-01 -1.33094177e-01 9.47591186e-01 -1.42849430e-01
-1.18042552e+00 2.43043661e-01 -1.56587437e-01 -1.25426248e-01
8.56967688e-01 -1.30952045e-01 -4.42894876e-01 1.63475260e-01
-8.24400961e-01 -1.27418160e-01 4.90921512e-02 4.69627380e-01
4.82844323e-01 -1.36032891e+00 -7.19942093e-01 -3.73667777e-02
4.71066475e-01 -3.22554320e-01 4.50430423e-01 5.91008902e-01
-4.36180979e-02 7.60611296e-01 -6.06805563e-01 -7.27139115e-01
-1.44554567e+00 7.48162389e-01 6.63153946e-01 -5.28330445e-01
-3.90008599e-01 8.26319814e-01 2.66395837e-01 -5.87964177e-01
2.98017293e-01 -2.62185603e-01 -8.31134498e-01 4.41949487e-01
5.36742449e-01 2.04290405e-01 4.75145161e-01 -8.31859827e-01
-5.97678781e-01 5.38697004e-01 -3.08364898e-01 4.35218871e-01
1.46881509e+00 4.98543344e-02 -1.71494000e-02 5.85986555e-01
1.11452305e+00 -2.00572491e-01 -1.02154994e+00 -3.78195643e-01
3.67411673e-01 -2.64369965e-01 3.10611188e-01 -1.28358924e+00
-1.12702382e+00 9.68605042e-01 8.03943634e-01 1.89832792e-01
1.26816821e+00 -3.36542800e-02 2.28097185e-01 3.60955656e-01
5.43030620e-01 -8.11351895e-01 -3.67054529e-02 5.40089369e-01
9.35987175e-01 -1.05246580e+00 -2.33529344e-01 -3.88537616e-01
-5.05590796e-01 1.23008728e+00 6.17703497e-01 -3.16126198e-01
5.71602225e-01 1.18091971e-01 3.33299823e-02 -1.02889724e-01
-7.97120452e-01 -4.86967415e-01 7.89851248e-01 6.20102227e-01
2.73360997e-01 5.70285171e-02 -4.35870886e-02 9.20032203e-01
-1.76644102e-02 1.73483163e-01 1.43375501e-01 4.52294648e-01
-5.45938730e-01 -1.00452101e+00 -1.34755790e-01 5.90370715e-01
-2.55814254e-01 -1.10645451e-01 -5.17345726e-01 7.73485065e-01
6.53225362e-01 9.22974169e-01 -2.89027505e-02 -7.87878454e-01
5.09079695e-01 2.15818584e-01 4.12494957e-01 -7.04531372e-01
-9.48164999e-01 -4.19379026e-01 3.37996930e-01 -5.66103756e-01
-6.12442434e-01 -2.68703461e-01 -7.50998139e-01 -5.79000711e-02
-1.39265746e-01 5.18338941e-02 3.97245705e-01 1.15856612e+00
3.43606502e-01 7.39944518e-01 5.43814600e-01 -8.08845222e-01
-3.61884654e-01 -9.41218913e-01 -3.14198285e-01 2.44708598e-01
2.36344919e-01 -1.03298903e+00 -3.55217278e-01 -2.22381681e-01] | [9.584187507629395, 2.172175645828247] |
24014c27-eb55-4908-a37b-11c48ecad400 | systemic-risk-of-optioned-portfolios | 2209.04685 | null | https://arxiv.org/abs/2209.04685v1 | https://arxiv.org/pdf/2209.04685v1.pdf | Systemic Risk of Optioned Portfolios: Controllability and Optimization | We investigate the portfolio selection problem against the systemic risk which is measured by CoVaR. We first demonstrate that the systemic risk of pure stock portfolios is essentially uncontrollable due to the contagion effect and the seesaw effect. Next, we prove that it is necessary and sufficient to introduce options to make the systemic risk controllable by the correlation hedging and the extreme loss hedging. In addition to systemic risk control, we show that using options can also enhance return-risk performance. Then, with a reasonable approximation of the conditional distribution of optioned portfolios, we show that the portfolio optimization problem can be formulated as a second-order cone program (SOCP) that allows for efficient computation. Finally, we carry out comprehensive simulations and empirical tests to illustrate the theoretical findings and the performance of our method. | ['Jiali Ma', 'Xueting Cui', 'Shushang Zhu', 'Xiaochuan Pang'] | 2022-09-10 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-5.15757442e-01 2.21247934e-02 -1.83396563e-01 3.41364145e-01
-2.55032390e-01 -1.03225231e+00 2.54196763e-01 -4.05312240e-01
-4.04424500e-03 7.26771832e-01 1.49363205e-01 -4.77319658e-01
-6.38327599e-01 -1.01736856e+00 -3.25007915e-01 -8.02081764e-01
-1.72877312e-01 -4.43850271e-02 -9.92631465e-02 -1.57431185e-01
2.38680288e-01 4.70744818e-01 -8.27855945e-01 -3.85681897e-01
8.60991955e-01 1.27451301e+00 -2.89685100e-01 1.71295837e-01
2.36090541e-01 1.04613554e+00 -4.29187864e-01 -6.80404544e-01
1.00546563e+00 -2.99897701e-01 -1.28441425e-02 -1.01399675e-01
-4.65103090e-01 -5.79460919e-01 -3.53846252e-02 1.36813617e+00
3.20033133e-01 1.64553344e-01 6.40272319e-01 -1.26084816e+00
-4.80122209e-01 6.45020068e-01 -6.03712797e-01 8.95733982e-02
-2.23289073e-01 2.00312644e-01 1.28125811e+00 -6.60677671e-01
3.26990068e-01 1.03554142e+00 6.19687915e-01 2.74283022e-01
-9.14574206e-01 -6.15389407e-01 2.10006639e-01 -6.81113482e-01
-8.83885086e-01 9.78166088e-02 7.16174483e-01 -5.78971207e-01
6.04030252e-01 6.79729581e-01 8.82820487e-01 4.64929521e-01
7.96401501e-01 4.33883399e-01 1.06424725e+00 -5.67078702e-02
3.63844275e-01 1.89524680e-01 6.88471571e-02 1.02648325e-01
9.59230721e-01 8.07165384e-01 -2.34240126e-02 -4.67357546e-01
9.91662920e-01 4.13352221e-01 -3.12958747e-01 -2.22484216e-01
-6.05516076e-01 1.03727067e+00 6.37400299e-02 -1.43713117e-01
-5.77929258e-01 2.43439004e-01 -3.70824635e-02 7.22532213e-01
2.82726020e-01 6.65790439e-01 -2.09142029e-01 1.18084840e-01
-4.76192206e-01 2.09169507e-01 1.19289887e+00 8.54806662e-01
-2.29610000e-02 3.72215629e-01 -1.98172167e-01 1.67735219e-01
5.46962857e-01 8.28364670e-01 -6.07723854e-02 -1.18031168e+00
7.51433551e-01 2.28850022e-01 6.61611438e-01 -9.55199897e-01
-1.47673905e-01 -7.39492357e-01 -5.94081342e-01 5.12217462e-01
3.74535918e-01 -7.24181354e-01 -9.18264221e-03 1.74914694e+00
-1.77508399e-01 -5.47005832e-02 4.37091514e-02 5.11706173e-01
-4.77082700e-01 5.63590527e-01 -4.05217230e-01 -1.00747442e+00
9.81408179e-01 -6.63084984e-01 -1.01787996e+00 9.09443349e-02
2.95642674e-01 -4.30869251e-01 5.68045378e-01 2.93159485e-01
-1.20381296e+00 4.49334770e-01 -9.79004085e-01 9.64937806e-01
2.87699819e-01 -3.32448959e-01 5.91718137e-01 9.05014038e-01
-7.44272470e-01 6.64047539e-01 -6.62731051e-01 3.45622152e-01
2.46709496e-01 1.54082671e-01 4.68531728e-01 6.71122551e-01
-1.12275183e+00 7.81981111e-01 -4.98017967e-02 1.12523168e-01
-9.69623208e-01 -1.08630478e+00 -2.07569048e-01 4.37578946e-01
8.02081943e-01 -6.12824142e-01 1.04252326e+00 -7.98986137e-01
-1.73046768e+00 -2.17651930e-02 5.84873199e-01 -5.69755495e-01
1.00934112e+00 -5.21576941e-01 2.62067541e-02 2.04159454e-01
-8.09174255e-02 -7.01071739e-01 4.55950290e-01 -8.70569944e-01
-3.02863866e-01 -1.60642087e-01 9.04455855e-02 9.10365433e-02
-4.96170849e-01 5.17795265e-01 3.60111028e-01 -1.09411597e+00
-2.30003312e-01 -9.61169899e-01 -6.39050961e-01 -2.82086492e-01
-5.88739336e-01 6.48576498e-01 -2.77696885e-02 -5.91970563e-01
1.45169544e+00 -1.93898344e+00 -2.18868658e-01 6.88271284e-01
-7.92314336e-02 -3.61475170e-01 3.34534347e-01 7.56634712e-01
-1.52527928e-01 6.12925112e-01 -2.44899839e-01 6.34949878e-02
3.86425763e-01 -3.73048037e-01 -1.07123661e+00 5.06343544e-01
-9.03784335e-02 9.00790811e-01 -4.46293861e-01 1.79857314e-01
-1.77849144e-01 -2.45052382e-01 -6.30876482e-01 2.34912992e-01
2.12242901e-02 -4.74055707e-02 -8.16654682e-01 5.37593663e-01
7.37473726e-01 -1.51925400e-01 3.70035321e-01 4.58981484e-01
-2.70611048e-01 7.21032219e-03 -1.32369506e+00 2.02461377e-01
-3.32392424e-01 1.92448005e-01 4.42964584e-02 -4.83250022e-01
6.89642727e-01 3.36319655e-01 4.55335766e-01 -2.02191383e-01
2.01121613e-01 3.88669282e-01 -7.63254389e-02 -1.01976886e-01
2.17217267e-01 -1.01176620e+00 -2.31801242e-01 1.00708401e+00
-5.80580950e-01 5.04652113e-02 -3.48170310e-01 4.21643071e-02
1.01779938e+00 -5.30547261e-01 4.52244133e-01 -5.91939211e-01
6.90655038e-02 -3.94691199e-01 8.50527346e-01 5.95004737e-01
-1.58983335e-01 -8.88388511e-03 1.27178693e+00 2.63677239e-01
-7.28873312e-01 -1.20512557e+00 7.35562369e-02 4.31490213e-01
-2.11343616e-02 7.20341355e-02 -4.94910449e-01 -4.91061240e-01
6.05091453e-01 9.88336146e-01 -6.87904239e-01 -2.17375830e-01
-2.53686577e-01 -1.21781838e+00 1.37389421e-01 6.27422452e-01
3.72437358e-01 -6.58598006e-01 -6.33368254e-01 1.37585714e-01
5.40550590e-01 -3.32272381e-01 -7.54637301e-01 1.98097527e-01
-8.36481988e-01 -9.51181531e-01 -7.95634687e-01 4.14899252e-02
4.64154959e-01 1.65722027e-01 8.03430140e-01 -2.24081829e-01
2.95906305e-01 5.24777651e-01 1.74612224e-01 -5.40238440e-01
1.55916940e-02 -6.37482047e-01 8.55322331e-02 2.54122019e-01
-2.98095942e-01 -5.66849232e-01 -7.41866946e-01 4.48973060e-01
-6.26051188e-01 -5.51728725e-01 1.05366126e-01 6.79652691e-01
5.28145194e-01 3.50568950e-01 9.19942915e-01 -5.44932365e-01
9.24087644e-01 -5.28208315e-01 -1.42164171e+00 5.23252368e-01
-8.95445168e-01 -1.93732500e-01 3.06452394e-01 -4.08763409e-01
-1.50576115e+00 -3.68270606e-01 5.27735353e-01 -3.57355803e-01
8.09825897e-01 6.58037603e-01 -4.54438359e-01 -1.71341926e-01
-1.07623801e-01 -1.54825404e-01 1.03043035e-01 -4.36791211e-01
2.65405953e-01 1.79412812e-01 2.65825558e-02 -4.81632411e-01
1.06250787e+00 7.28242457e-01 2.86559880e-01 -2.07870856e-01
-7.68368959e-01 2.32319742e-01 -1.09587483e-01 -3.35515976e-01
7.91304708e-01 -8.08003485e-01 -9.77211177e-01 4.32088912e-01
-6.55008435e-01 -2.78122842e-01 -5.00239968e-01 5.60747564e-01
-8.26421440e-01 -6.68239594e-03 -8.18120062e-01 -1.66115737e+00
-2.19261780e-01 -7.23977804e-01 3.97161096e-02 1.58292279e-01
2.17930347e-01 -1.24704635e+00 4.04733837e-01 -4.37360965e-02
5.94840765e-01 6.79136097e-01 6.40603900e-01 -8.28569651e-01
-1.09478354e+00 -4.31547225e-01 1.32284656e-01 3.13955873e-01
8.23988542e-02 3.84193063e-02 -4.12781656e-01 -2.80031383e-01
1.02942574e+00 1.61150083e-01 7.35868692e-01 4.42890525e-01
4.62296009e-01 -8.87917876e-01 -1.39585406e-01 7.39592314e-01
1.77275276e+00 6.75544024e-01 4.56160843e-01 7.90684223e-01
2.50500768e-01 9.69756544e-01 8.76588285e-01 9.44374621e-01
-1.46211371e-01 3.59729856e-01 3.66916955e-01 5.32077312e-01
9.23848689e-01 -2.57737219e-01 6.19223177e-01 7.25761116e-01
2.97703547e-03 -1.51464194e-01 -6.79025948e-01 2.95880914e-01
-1.85798538e+00 -1.31700182e+00 -1.45768166e-01 2.60496378e+00
6.38722599e-01 1.66359648e-01 2.91419744e-01 -3.16356808e-01
6.88672304e-01 -4.01246138e-02 -5.25090873e-01 -3.63015562e-01
-4.09911543e-01 -2.30030388e-01 9.87876534e-01 4.38338965e-01
-8.29693496e-01 3.77833158e-01 7.58530951e+00 6.00091398e-01
-6.35157824e-01 -4.47324887e-02 9.17549491e-01 -5.47064424e-01
-9.74579811e-01 1.75848991e-01 -7.06391215e-01 7.96744645e-01
7.34903932e-01 -1.11895967e+00 2.30676115e-01 9.02150869e-01
4.36336577e-01 2.24435732e-01 -6.25479460e-01 2.51637191e-01
-4.99018401e-01 -1.17892325e+00 -3.10338289e-01 4.82605666e-01
8.98787737e-01 -5.35584986e-01 4.88034308e-01 -3.93532030e-02
6.46889329e-01 -8.63721728e-01 8.36645365e-01 8.99164796e-01
2.67034501e-01 -1.25804400e+00 7.79509842e-01 1.57194361e-01
-9.93442059e-01 -7.23299801e-01 -4.34556276e-01 -2.04095751e-01
5.76703966e-01 7.38199413e-01 1.70927092e-01 4.43077087e-01
1.34983182e-01 2.92225957e-01 1.28000425e-02 1.31699741e+00
-3.71062338e-01 4.87004340e-01 -3.26074868e-01 2.13416032e-02
1.12499006e-01 -9.43857133e-01 8.21947813e-01 5.17611206e-01
5.80422878e-01 3.51568937e-01 -3.40559691e-01 1.25863111e+00
1.13142185e-01 -6.13401122e-02 -6.64482832e-01 -2.37035930e-01
4.78703380e-01 8.08294237e-01 -5.22724450e-01 6.63512349e-02
-4.16442871e-01 4.78832155e-01 -2.87777692e-01 5.58959663e-01
-7.82695115e-01 -5.79082370e-01 9.36853826e-01 -1.52635947e-01
3.40293527e-01 -8.18671957e-02 -8.97297382e-01 -1.24655879e+00
3.27413291e-01 -6.28948390e-01 4.51295793e-01 -1.98940784e-01
-1.52760315e+00 1.18425801e-01 -1.71946019e-01 -1.39284706e+00
-2.83944339e-01 -4.53607380e-01 -1.30786824e+00 8.71552646e-01
-1.31157839e+00 -2.16833770e-01 5.46638966e-01 1.96217194e-01
-2.73339987e-01 -4.31292981e-01 1.88790396e-01 -2.00235203e-01
-9.56286728e-01 5.11277556e-01 8.57187808e-01 -1.99095696e-01
1.69778481e-01 -1.33996701e+00 2.07161933e-01 9.59278286e-01
-5.00912070e-01 7.61156976e-01 5.64466059e-01 -1.14006662e+00
-1.47935629e+00 -1.08719194e+00 4.89629686e-01 -3.92618746e-01
1.54523885e+00 -3.96597385e-03 -6.85220599e-01 7.88948238e-01
1.97255969e-01 -4.29270685e-01 6.96635187e-01 -3.81160259e-01
-1.03454903e-01 -2.02772692e-01 -1.08474433e+00 7.93847442e-01
7.41662025e-01 -3.57365459e-01 -5.93842864e-01 2.48784766e-01
1.07466650e+00 1.63982794e-01 -9.60139751e-01 1.11057349e-01
4.28974330e-01 -9.17347431e-01 9.88419712e-01 -7.00978994e-01
3.23187202e-01 9.89103690e-02 -3.19230229e-01 -1.07922459e+00
-3.52348268e-01 -1.32722080e+00 -5.61719462e-02 1.50344408e+00
4.25607681e-01 -1.32600367e+00 5.17941475e-01 1.04085076e+00
4.55870003e-01 -6.88251972e-01 -9.99399126e-01 -1.62009656e+00
5.96617043e-01 -1.96163580e-01 7.60394394e-01 7.21785247e-01
3.03044200e-01 -4.48750645e-01 -6.03761375e-01 1.72463238e-01
1.04722393e+00 5.53796172e-01 1.77729562e-01 -8.23084414e-01
-6.45558059e-01 -6.61386967e-01 2.74117738e-01 -5.49114525e-01
5.32216206e-02 -5.83833098e-01 -1.42920151e-01 -8.41743648e-01
4.44022864e-01 -2.62029469e-01 -8.06246042e-01 -4.57194932e-02
-2.76543766e-01 -4.63955671e-01 7.68381119e-01 3.69450927e-01
-1.62128329e-01 9.22802985e-01 1.14718032e+00 2.01918676e-01
-3.81033033e-01 5.73452473e-01 -1.05481780e+00 7.98299670e-01
8.49562883e-01 -5.07755160e-01 -5.14086008e-01 2.01051801e-01
7.34653771e-01 7.36824572e-01 3.43994200e-01 -1.77356109e-01
-9.54275653e-02 -9.24825728e-01 -2.85813063e-01 -6.03512287e-01
-7.71834180e-02 -7.50802636e-01 4.23185110e-01 8.41059804e-01
-3.10481548e-01 1.24513008e-01 -1.10735729e-01 8.46588671e-01
-3.46016930e-03 -5.01044691e-01 7.04003513e-01 1.07546248e-01
4.09860462e-01 2.96244740e-01 -5.69911838e-01 1.83942035e-01
1.31613982e+00 4.87952292e-01 -4.81421024e-01 -6.89657390e-01
-4.58712608e-01 5.22431791e-01 5.04772902e-01 -1.29044652e-01
4.73147988e-01 -1.56986558e+00 -3.83930445e-01 -2.72643328e-01
-5.80467045e-01 -8.43540668e-01 5.73112331e-02 9.01426196e-01
-3.24786872e-01 5.84668815e-01 -1.40647486e-01 2.74773955e-01
-7.03279912e-01 8.27507317e-01 7.48214424e-01 -4.27152127e-01
-3.85074764e-01 6.51344836e-01 6.50192797e-01 2.36071259e-01
8.66522789e-02 -1.89503118e-01 -7.13807940e-02 2.57961094e-01
6.87730670e-01 8.01795006e-01 -6.03304029e-01 1.42312810e-01
-3.69420558e-01 4.72670436e-01 3.71968001e-01 -6.33493721e-01
1.41772354e+00 -3.44758391e-01 -3.53775084e-01 3.31582457e-01
8.12205434e-01 3.97400916e-01 -1.49153113e+00 2.79030204e-01
4.00911927e-01 -6.84358537e-01 -2.25748852e-01 -5.33014238e-01
-1.40449059e+00 7.01733768e-01 2.65903696e-02 6.34367108e-01
1.07698095e+00 -4.41321582e-01 4.40187395e-01 4.04206544e-01
4.73110020e-01 -9.97061968e-01 8.43144581e-02 3.36866796e-01
1.32082391e+00 -4.90140110e-01 -2.23236997e-02 -5.17163634e-01
-9.60351348e-01 9.07243848e-01 2.44913191e-01 -6.27909780e-01
1.25687838e+00 5.45584679e-01 -2.51528114e-01 2.70996913e-02
-1.05737543e+00 2.03777894e-01 2.11483333e-02 2.96930134e-01
-1.72369659e-01 3.61770749e-01 -4.52899575e-01 1.59159124e+00
-2.40236282e-01 -2.48109028e-01 1.01807272e+00 6.99553311e-01
-4.39565599e-01 -7.89293587e-01 -5.67231417e-01 5.08689582e-01
-9.31881785e-01 -1.44185573e-01 -4.26173329e-01 7.48236418e-01
-7.51153409e-01 8.02012265e-01 1.12765290e-01 -1.39591515e-01
3.52831811e-01 -1.55316532e-01 -1.00684606e-01 -2.92602688e-01
-6.91052794e-01 4.33243543e-01 -1.72838196e-02 -5.17133296e-01
2.41885260e-01 -9.14331198e-01 -8.74587834e-01 -4.04544264e-01
-5.67095280e-01 -5.23416400e-02 -3.19435559e-02 7.45918691e-01
-6.47249585e-03 4.52307016e-01 1.30299449e+00 -4.12982106e-02
-1.65042436e+00 -3.55388880e-01 -1.35212266e+00 -2.41261393e-01
1.03715129e-01 -5.86017728e-01 -1.06064510e+00 -4.83522981e-01] | [4.935173988342285, 3.954319953918457] |
bb3bc4aa-e04f-478b-9b69-bdb6b73cce97 | dynamic-decision-boundary-for-one-class | 2004.02273 | null | https://arxiv.org/abs/2004.02273v1 | https://arxiv.org/pdf/2004.02273v1.pdf | Dynamic Decision Boundary for One-class Classifiers applied to non-uniformly Sampled Data | A typical issue in Pattern Recognition is the non-uniformly sampled data, which modifies the general performance and capability of machine learning algorithms to make accurate predictions. Generally, the data is considered non-uniformly sampled when in a specific area of data space, they are not enough, leading us to misclassification problems. This issue cut down the goal of the one-class classifiers decreasing their performance. In this paper, we propose a one-class classifier based on the minimum spanning tree with a dynamic decision boundary (OCdmst) to make good prediction also in the case we have non-uniformly sampled data. To prove the effectiveness and robustness of our approach we compare with the most recent one-class classifier reaching the state-of-the-art in most of them. | ['Riccardo La Grassa', 'Nicola Landro', 'Ignazio Gallo'] | 2020-04-05 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 3.97327214e-01 3.13361794e-01 -3.38349223e-01 -3.88572991e-01
-2.78087646e-01 -2.39141747e-01 3.71965528e-01 4.87654418e-01
-3.48005444e-01 1.04857731e+00 -5.94536960e-01 -3.25916857e-01
-4.45051730e-01 -1.08897114e+00 -5.76217532e-01 -8.83392334e-01
1.67618110e-03 7.40000665e-01 6.85102284e-01 9.62915421e-02
3.67215455e-01 5.54204047e-01 -2.13656163e+00 4.99185920e-01
1.02550435e+00 1.15490699e+00 -5.62847555e-02 4.71667767e-01
-2.68230081e-01 3.41044366e-01 -6.64201379e-01 -6.81082606e-02
3.88762385e-01 -3.86663169e-01 -6.74355865e-01 9.66146812e-02
3.45535368e-01 1.60473168e-01 5.40732741e-01 9.58405793e-01
-5.13438880e-02 -1.26529276e-01 1.12940252e+00 -1.40809643e+00
-1.27550662e-01 4.15239185e-01 -6.42148077e-01 2.05859944e-01
4.67794463e-02 -2.19012052e-01 6.73456192e-01 -5.35350502e-01
6.42025292e-01 9.54312444e-01 7.18360245e-01 5.18912196e-01
-1.10711944e+00 -4.54759866e-01 2.54500002e-01 4.70823318e-01
-1.33489227e+00 3.60464156e-02 5.00642955e-01 -5.60385644e-01
4.88469094e-01 5.67559242e-01 7.03265965e-01 8.02061915e-01
2.61311114e-01 3.21588308e-01 1.31735563e+00 -5.73782921e-01
5.53977489e-01 4.21092361e-01 3.23110014e-01 2.68891245e-01
5.80614030e-01 2.15675011e-01 6.01130016e-02 -1.36198997e-01
1.27763346e-01 1.72868058e-01 -1.09385781e-01 -5.26275516e-01
-4.57362324e-01 7.51828134e-01 1.94582194e-01 7.43542492e-01
-3.39871645e-01 -5.59731960e-01 1.35925859e-01 4.47245240e-01
4.16930258e-01 3.68093967e-01 -4.57630187e-01 -6.62004277e-02
-1.07623661e+00 2.38375410e-01 9.62832749e-01 5.39042830e-01
6.19811296e-01 -3.79968077e-01 -1.36061758e-01 6.00838959e-01
-1.93390563e-01 2.52628550e-02 6.46909773e-01 -2.64200084e-02
4.21460569e-02 1.09205496e+00 1.23512678e-01 -9.73829091e-01
-4.90268737e-01 -8.01053286e-01 -7.31169879e-01 5.25694132e-01
7.48859048e-01 1.24370746e-01 -9.18976426e-01 1.34413600e+00
4.88958240e-01 1.88505635e-01 -1.08567975e-03 8.49430203e-01
1.61729082e-01 5.11259496e-01 -1.03275754e-01 -5.52090168e-01
7.81534195e-01 -8.27885568e-01 -4.86470759e-01 1.18029907e-01
6.81882143e-01 -5.64933479e-01 8.21162343e-01 1.04665256e+00
-2.45870605e-01 -6.11987770e-01 -1.20836461e+00 6.48704231e-01
-7.24933147e-01 1.38765901e-01 2.00562388e-01 8.28170896e-01
-4.99197692e-01 8.36782217e-01 -5.65042555e-01 -4.94281411e-01
4.97435540e-01 2.97544330e-01 -3.47240329e-01 -1.53316796e-01
-8.20004940e-01 1.13772285e+00 6.85087860e-01 -1.30992696e-01
-5.80491483e-01 -5.11428416e-01 -7.38559142e-02 -7.35386088e-02
4.94694918e-01 -1.27376691e-01 9.63070393e-01 -1.19531846e+00
-1.09888422e+00 7.78218746e-01 5.10487743e-02 -7.76558876e-01
1.08049428e+00 -3.88427787e-02 -4.64380175e-01 -1.90366730e-01
-2.30134681e-01 3.86015266e-01 8.03195119e-01 -1.26854908e+00
-9.37800527e-01 -5.65476835e-01 -2.02648178e-01 -1.48774371e-01
-2.84221143e-01 -5.43989241e-01 3.72604251e-01 -3.06228876e-01
2.19929129e-01 -7.90989459e-01 -2.20542789e-01 -8.19188431e-02
-4.26781237e-01 -5.22668481e-01 1.15507877e+00 -2.51371533e-01
1.24088180e+00 -2.11377382e+00 -1.63671091e-01 1.49268553e-01
-1.57311678e-01 4.08371329e-01 2.72714227e-01 3.69066954e-01
-2.74534315e-01 2.50001878e-01 -4.89791185e-01 -4.06385101e-02
-4.63632494e-01 3.71836722e-01 -2.11672992e-01 5.79563498e-01
3.08437973e-01 4.05523740e-02 -6.94178879e-01 -4.71141487e-01
3.69768322e-01 1.07215203e-01 -3.14684153e-01 1.26319885e-01
-3.50240737e-01 3.78023863e-01 -5.68143904e-01 3.14132065e-01
9.08829629e-01 1.24027006e-01 1.21587507e-01 1.86705232e-01
-2.77366996e-01 2.52020434e-02 -1.52457488e+00 8.96282971e-01
-2.77806103e-01 2.47513846e-01 -4.46703792e-01 -1.58885527e+00
1.30753100e+00 -1.04964254e-02 3.57602477e-01 -3.08838338e-01
1.51800334e-01 4.91084844e-01 2.53210902e-01 -3.61709774e-01
1.32806346e-01 -2.40593255e-01 3.54110241e-01 9.99581721e-03
-4.19693589e-01 8.15807804e-02 1.78225026e-01 -5.13728201e-01
9.25551116e-01 2.52423603e-02 6.28160238e-01 -6.42781973e-01
6.33960009e-01 1.95783973e-01 5.82507610e-01 7.22894371e-01
-2.25874092e-02 7.38589942e-01 6.33130074e-01 -5.49349904e-01
-1.04853904e+00 -6.53456390e-01 -7.67713666e-01 6.25184953e-01
1.92980375e-02 8.34708363e-02 -9.44584787e-01 -1.00942910e+00
2.09426284e-01 8.23722124e-01 -8.51280630e-01 -3.06396127e-01
-3.04637730e-01 -8.40481162e-01 1.53895348e-01 9.22407359e-02
2.88632840e-01 -7.80397594e-01 -1.17534280e+00 2.56999314e-01
2.61392355e-01 -8.84499013e-01 3.88990939e-01 4.60745573e-01
-1.04749024e+00 -1.52053523e+00 -4.72807616e-01 -5.13903022e-01
5.71427643e-01 -5.23851961e-02 8.68966997e-01 1.66712359e-01
-3.13712895e-01 -3.93507570e-01 -8.70965779e-01 -6.50017202e-01
-7.11723328e-01 4.01316047e-01 2.99574509e-02 3.62495512e-01
4.89416361e-01 -5.15845001e-01 -3.43847424e-01 6.15820587e-01
-8.13034534e-01 -1.84567630e-01 6.79690301e-01 8.11900198e-01
5.32522738e-01 4.83671039e-01 8.74140561e-01 -1.13447392e+00
3.24769318e-01 -5.46340764e-01 -5.91551363e-01 3.75751287e-01
-1.00585914e+00 5.77449799e-02 1.10787117e+00 -5.88958025e-01
-4.50506300e-01 1.19524738e-02 7.12607428e-02 -3.26133281e-01
-5.15306473e-01 1.61750197e-01 -1.33755326e-01 2.64003482e-02
8.84645820e-01 4.90803383e-02 2.06648614e-02 -6.85809970e-01
-2.15439677e-01 1.06241870e+00 -3.26342881e-02 -1.43250629e-01
4.52910334e-01 3.11562777e-01 3.43903869e-01 -7.09742665e-01
-7.55657494e-01 -3.72661561e-01 -7.39899337e-01 -1.99551642e-01
6.18532121e-01 -8.56617540e-02 -3.49519014e-01 4.28598702e-01
-8.42479706e-01 -4.07608822e-02 -3.88643980e-01 4.13976908e-01
-4.11301702e-01 2.80934393e-01 -1.43406773e-03 -1.23792386e+00
-8.67744163e-02 -1.08472097e+00 6.42188311e-01 2.65554219e-01
-9.26145762e-02 -6.46290362e-01 -8.17223117e-02 -1.06934898e-01
1.35701060e-01 5.88471711e-01 9.65098917e-01 -1.20122898e+00
-1.82250544e-01 -5.95966041e-01 1.40902594e-01 4.77256268e-01
1.14662170e-01 1.98666215e-01 -9.69868720e-01 -2.02120036e-01
1.89339202e-02 -2.58404464e-02 8.85041535e-01 3.20905685e-01
1.42654455e+00 -2.16263041e-01 -7.45406389e-01 1.41263500e-01
1.57688546e+00 4.77632791e-01 4.90312934e-01 3.21938992e-01
4.67336876e-03 8.18019271e-01 1.14930129e+00 5.28475165e-01
-1.62839875e-01 7.97459960e-01 7.11683750e-01 1.45391434e-01
1.62269145e-01 -1.48380265e-01 -1.58379540e-01 2.03592539e-01
1.54521301e-01 -2.79468119e-01 -8.96693647e-01 5.89769363e-01
-1.80563569e+00 -7.18917966e-01 -3.65436882e-01 2.60871911e+00
6.19398594e-01 4.05939817e-01 2.90364295e-01 9.35955822e-01
9.62026954e-01 -1.21847361e-01 -4.63156223e-01 -6.71035588e-01
8.42099488e-02 2.22270787e-01 5.10280311e-01 3.32316637e-01
-1.00509131e+00 4.04111207e-01 5.93087292e+00 1.15978181e+00
-1.31759965e+00 -7.32675754e-03 7.89302588e-01 1.70692220e-01
1.04373343e-01 -5.58303706e-02 -9.97668087e-01 7.15480089e-01
8.89071286e-01 1.20143496e-01 3.16111296e-02 1.18520391e+00
6.34425730e-02 -5.43690026e-01 -1.13721526e+00 7.26348639e-01
2.65206713e-02 -8.07810307e-01 -2.36663334e-02 1.60406440e-01
4.60987300e-01 -5.25141358e-01 -2.74699837e-01 1.69441879e-01
-1.79751724e-01 -1.19285858e+00 5.72352111e-01 5.31035900e-01
5.49977183e-01 -8.93011689e-01 8.36363196e-01 1.02119851e+00
-6.82647943e-01 -3.86622608e-01 -5.82606375e-01 -2.26035058e-01
-2.02687606e-01 1.18075168e+00 -1.13577545e+00 7.27242470e-01
5.05257607e-01 4.13442791e-01 -6.03151798e-01 1.36212409e+00
1.60439670e-01 6.19196117e-01 -5.84902287e-01 -5.68874836e-01
2.22386360e-01 -1.77786678e-01 5.65527320e-01 9.75124061e-01
4.28062141e-01 -2.31492203e-02 2.92939961e-01 6.75206661e-01
5.08968055e-01 3.45932633e-01 -7.18221188e-01 3.58967572e-01
4.39979315e-01 9.34602678e-01 -9.16941464e-01 -2.42921472e-01
-6.30906150e-02 5.14035165e-01 2.55394578e-01 -1.95516214e-01
-6.00582182e-01 -1.56655982e-01 2.18849108e-01 4.98977602e-01
3.47518802e-01 1.55344948e-01 -5.77137649e-01 -7.72470713e-01
2.60419577e-01 -6.84058130e-01 6.23306215e-01 -1.92733169e-01
-1.41797984e+00 8.36216509e-01 1.23822846e-01 -1.50634539e+00
-2.14411452e-01 -8.97364736e-01 -3.24114770e-01 5.62178135e-01
-1.31538665e+00 -7.09919810e-01 -2.71381021e-01 2.94776678e-01
5.25477946e-01 -5.30842319e-02 8.25062633e-01 1.98737383e-01
-3.81578445e-01 5.36535919e-01 1.17287338e-01 -3.51878673e-01
5.62353194e-01 -1.07921457e+00 -1.64690897e-01 7.95729339e-01
-3.82977687e-02 5.79309016e-02 1.00924826e+00 -5.43767035e-01
-8.64304483e-01 -8.96610558e-01 6.28818512e-01 -2.17827037e-01
2.53714234e-01 -3.24520409e-01 -1.12293470e+00 2.74300128e-01
-3.79201293e-01 1.69974253e-01 3.96274537e-01 1.17232881e-01
-1.18512243e-01 -3.25273186e-01 -1.60129547e+00 2.06964359e-01
7.40042329e-01 1.64464638e-01 -6.39239132e-01 2.54692286e-01
1.36755824e-01 -1.92176357e-01 -6.30152285e-01 8.07936847e-01
5.82438290e-01 -1.33089221e+00 4.46232557e-01 -7.09507763e-01
3.34899336e-01 -3.05517286e-01 -1.81133062e-01 -1.32566404e+00
6.92707300e-02 -7.38633797e-02 7.57381991e-02 9.45841312e-01
5.31400084e-01 -9.60120916e-01 9.96054947e-01 2.38052785e-01
1.46011472e-01 -1.09829509e+00 -1.28735411e+00 -9.59509075e-01
2.38568634e-01 -3.06562722e-01 6.75498366e-01 8.81334484e-01
-8.25661793e-02 -1.70302000e-02 -1.93340197e-01 -8.37327391e-02
4.69126463e-01 4.03124273e-01 5.80304146e-01 -1.71072364e+00
-3.79025310e-01 -3.89144599e-01 -8.73683691e-01 -6.16578162e-01
-1.88782483e-01 -6.92535698e-01 -1.06468573e-01 -1.18314981e+00
1.71484761e-02 -8.43029380e-01 -5.24737500e-02 2.88358092e-01
-1.97109282e-02 6.88538374e-03 1.20170645e-01 1.91409793e-02
-1.48513272e-01 7.22839311e-02 1.02771509e+00 2.05311462e-01
-1.77003533e-01 5.58575809e-01 -2.87548333e-01 5.72874486e-01
8.13929856e-01 -8.05608273e-01 -3.86047989e-01 3.46021295e-01
-1.60481289e-01 -7.45241567e-02 7.83288181e-02 -1.36800420e+00
-7.32805654e-02 -3.17625731e-01 3.88633490e-01 -7.02272892e-01
3.01740468e-02 -1.25479937e+00 3.36007297e-01 9.17833209e-01
-2.03929558e-01 -1.74312964e-01 -3.07400394e-02 5.40600717e-01
-2.05101460e-01 -8.19294572e-01 1.24769497e+00 1.93391331e-02
-4.39037830e-01 1.44738480e-01 -2.31123820e-01 -1.89377323e-01
1.62653983e+00 -6.65566862e-01 -2.93769747e-01 1.55253671e-02
-7.49876142e-01 -8.76960233e-02 5.41021287e-01 3.34729880e-01
2.81802446e-01 -8.46625924e-01 -6.15181327e-01 3.16038549e-01
2.12446302e-01 3.59517671e-02 1.16069265e-01 7.87105143e-01
-4.51433033e-01 3.32140267e-01 -2.76915580e-01 -7.80312598e-01
-1.33847654e+00 7.36266315e-01 6.63529456e-01 -3.08591336e-01
-5.79750359e-01 3.54938567e-01 -2.62675941e-01 -3.44248801e-01
2.22057365e-02 -5.36921322e-01 -4.98952806e-01 8.41061994e-02
3.90268177e-01 4.68840808e-01 4.76573825e-01 -1.59781247e-01
-3.34900916e-01 4.58532035e-01 -1.45461084e-02 1.23737998e-01
1.05233967e+00 3.71342897e-01 1.72897838e-02 9.30818439e-01
1.10403430e+00 -1.66074798e-01 -9.41176891e-01 2.88957506e-01
2.44312748e-01 -6.97146595e-01 -2.01421633e-01 -8.97533000e-01
-6.40794039e-01 8.69152844e-01 9.81066883e-01 8.13601315e-01
1.10316873e+00 -3.62347275e-01 2.33054131e-01 2.95085430e-01
6.88097417e-01 -1.08100927e+00 -4.22930062e-01 3.02365303e-01
7.00229049e-01 -1.05040312e+00 -6.71746060e-02 -6.76339209e-01
-5.03505886e-01 1.40761435e+00 6.33523643e-01 -3.03031474e-01
9.55170095e-01 4.25320029e-01 -1.36211932e-01 3.97054821e-01
-8.32573652e-01 9.94244590e-03 2.11151719e-01 6.53051198e-01
1.78818092e-01 3.16217095e-01 -9.52597916e-01 5.99417567e-01
-2.09651187e-01 3.23668659e-01 4.80167568e-01 8.16314340e-01
-7.57493854e-01 -1.21714592e+00 -2.85047323e-01 7.96596587e-01
-4.46697116e-01 3.44786525e-01 -4.49628830e-01 9.89187300e-01
5.39505720e-01 1.09540892e+00 2.27462232e-01 -4.85810846e-01
4.01622266e-01 2.00575307e-01 4.00732815e-01 -4.84259486e-01
-4.74923760e-01 -4.40863639e-01 4.20950167e-02 -1.92782298e-01
-3.92961890e-01 -7.81372547e-01 -1.01829815e+00 -1.65327430e-01
-4.66725171e-01 2.96846062e-01 6.35140479e-01 9.75199759e-01
1.96807593e-01 3.07739943e-01 9.00892913e-01 -1.90038592e-01
-9.38546479e-01 -1.22016478e+00 -7.02595055e-01 4.03298616e-01
2.68221051e-01 -9.94715929e-01 -6.20666325e-01 -3.41370523e-01] | [8.499464988708496, 4.211613655090332] |
d7c32477-e2f5-4f2f-aa9b-687c35b6a18f | fusing-rgbd-tracking-and-segmentation-tree | 2104.00205 | null | https://arxiv.org/abs/2104.00205v1 | https://arxiv.org/pdf/2104.00205v1.pdf | Fusing RGBD Tracking and Segmentation Tree Sampling for Multi-Hypothesis Volumetric Segmentation | Despite rapid progress in scene segmentation in recent years, 3D segmentation methods are still limited when there is severe occlusion. The key challenge is estimating the segment boundaries of (partially) occluded objects, which are inherently ambiguous when considering only a single frame. In this work, we propose Multihypothesis Segmentation Tracking (MST), a novel method for volumetric segmentation in changing scenes, which allows scene ambiguity to be tracked and our estimates to be adjusted over time as we interact with the scene. Two main innovations allow us to tackle this difficult problem: 1) A novel way to sample possible segmentations from a segmentation tree; and 2) A novel approach to fusing tracking results with multiple segmentation estimates. These methods allow MST to track the segmentation state over time and incorporate new information, such as new objects being revealed. We evaluate our method on several cluttered tabletop environments in simulation and reality. Our results show that MST outperforms baselines in all tested scenes. | ['Dmitry Berenson', 'Kun Huang', 'Andrew Price'] | 2021-04-01 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 5.22156417e-01 -1.77931786e-01 -8.44955519e-02 -2.58036792e-01
-5.78754663e-01 -9.24963355e-01 2.74780005e-01 1.44276381e-01
-3.38182747e-01 6.91212177e-01 -9.87458006e-02 -6.84923232e-02
1.79410547e-01 -3.97413343e-01 -6.60383523e-01 -3.14713836e-01
-2.86065377e-02 7.95729578e-01 1.13296533e+00 1.42730817e-01
2.60280669e-01 8.15339029e-01 -1.47478187e+00 1.03847042e-01
9.94265378e-01 5.36618710e-01 6.25392079e-01 9.76041496e-01
-3.44282627e-01 2.42140919e-01 -5.54860115e-01 1.77785054e-01
4.81806308e-01 -4.82943878e-02 -8.84137392e-01 8.19905281e-01
5.23313284e-01 -4.50969070e-01 1.63016841e-01 9.96652842e-01
2.02661380e-01 1.77831024e-01 2.15120181e-01 -9.77234304e-01
4.69679654e-01 3.83748263e-01 -8.90159905e-01 3.58700126e-01
8.32794428e-01 1.91323519e-01 3.29411924e-01 -5.45649648e-01
1.01252913e+00 1.40658939e+00 5.40275156e-01 3.22498381e-01
-1.39207673e+00 -2.14765996e-01 9.01563883e-01 -2.91382581e-01
-1.22875023e+00 -4.58737075e-01 6.07215345e-01 -6.80116415e-01
6.87388659e-01 4.11463529e-01 9.32870507e-01 5.91877937e-01
1.36522427e-01 1.18610942e+00 1.17284560e+00 -3.25690508e-01
2.46282294e-01 1.75069615e-01 2.31734678e-01 5.61289072e-01
2.65885264e-01 -1.60037920e-01 -2.74141699e-01 -2.13732123e-02
1.06525576e+00 -9.12007764e-02 -3.48186642e-01 -7.51464665e-01
-1.47541070e+00 2.73263872e-01 1.15075499e-01 1.06974795e-01
-3.42714101e-01 1.37230828e-01 6.09947219e-02 -1.25101358e-01
4.68459159e-01 3.77566367e-01 -6.32986903e-01 -2.28001684e-01
-1.19023609e+00 5.32558799e-01 8.08999658e-01 1.14430976e+00
5.64688027e-01 -1.92273900e-01 -2.30075538e-01 4.44175452e-01
4.11208302e-01 3.83978724e-01 7.89188296e-02 -1.11822951e+00
4.27738160e-01 3.65804076e-01 4.74772006e-01 -7.23296225e-01
-5.63389063e-01 -3.33831310e-01 -2.26090457e-02 3.49984527e-01
6.29277349e-01 -2.15856582e-01 -1.40773535e+00 1.36039662e+00
1.03772831e+00 3.66143107e-01 -3.17855388e-01 7.94179678e-01
4.46248353e-01 3.90216857e-01 -1.19251765e-01 -4.74162996e-01
1.21435952e+00 -1.00785100e+00 -8.71291399e-01 -4.39808518e-01
3.45440716e-01 -9.44502413e-01 6.80168271e-01 6.79755270e-01
-1.37161601e+00 -7.72245586e-01 -9.32574809e-01 1.59616426e-01
-1.84567362e-01 -5.17534375e-01 7.95662224e-01 6.28079534e-01
-1.04321539e+00 6.08673275e-01 -1.24816942e+00 -5.14013946e-01
4.71015006e-01 6.12622678e-01 9.23973247e-02 -1.39285550e-01
-5.08269966e-01 6.59466505e-01 6.22800589e-01 7.08595738e-02
-6.10649705e-01 -5.80989003e-01 -8.16651344e-01 -1.83670774e-01
9.55944896e-01 -6.66760147e-01 1.51916182e+00 -8.08552027e-01
-1.46370971e+00 6.03661180e-01 -4.30496126e-01 -2.66829014e-01
8.35507274e-01 -6.08875155e-01 2.02156790e-02 1.40137225e-01
2.43022814e-01 8.88540089e-01 6.88488722e-01 -1.74320138e+00
-9.29045141e-01 -4.62498128e-01 4.28973399e-02 6.59472346e-01
5.06135762e-01 -1.02260806e-01 -1.01497865e+00 -3.93380135e-01
6.65033519e-01 -9.44948673e-01 -7.09445775e-01 4.15560231e-02
-6.78839326e-01 2.42290154e-01 9.99643683e-01 -4.51028138e-01
1.05311275e+00 -1.86894107e+00 2.28371263e-01 1.38267562e-01
1.43137291e-01 3.08879435e-01 1.80762440e-01 9.08415914e-02
3.98269027e-01 -6.34668246e-02 -2.46490166e-01 -6.36980295e-01
-2.15314165e-01 2.93011904e-01 -2.83014148e-01 1.92477837e-01
-1.92134202e-01 8.17765772e-01 -1.06603897e+00 -8.91823232e-01
6.56037867e-01 1.80140957e-01 -4.57393914e-01 1.09541781e-01
-6.72400594e-01 9.09680068e-01 -6.51230812e-01 9.08264160e-01
1.07431674e+00 -1.77772224e-01 1.95658922e-01 -3.11693866e-02
-4.49879557e-01 -5.56864403e-02 -1.74557602e+00 2.04440427e+00
5.26181459e-02 3.40960056e-01 2.10215792e-01 -5.05290091e-01
4.40124124e-01 5.76062463e-02 7.05461264e-01 -2.26327434e-01
1.35556459e-01 -7.06276149e-02 -2.48305559e-01 -5.62434912e-01
7.06293821e-01 1.50454029e-01 -5.53695904e-03 1.50263444e-01
-4.04566556e-01 -6.82203352e-01 4.34121400e-01 1.62882373e-01
9.21219349e-01 6.34550989e-01 2.77211905e-01 -4.30322774e-02
1.16929941e-01 3.82289439e-01 7.87903488e-01 9.71407473e-01
-5.12700379e-01 7.19269454e-01 9.08400416e-02 -4.29454893e-01
-4.55828816e-01 -1.13821316e+00 -2.35898152e-01 4.53206539e-01
8.24288189e-01 -1.63362712e-01 -8.43245327e-01 -6.23499155e-01
-1.09077036e-01 5.42298317e-01 -2.33287632e-01 6.14909649e-01
-7.71838725e-01 -5.43733895e-01 -2.75370866e-01 3.87766033e-01
4.61052686e-01 -8.88849437e-01 -1.14333749e+00 6.40120089e-01
-2.29082063e-01 -1.59671175e+00 -5.43036282e-01 1.38734221e-01
-1.26020312e+00 -1.04625916e+00 -5.00260293e-01 -5.24668097e-01
5.83870947e-01 5.45866668e-01 9.35825586e-01 -5.86566739e-02
-3.98201436e-01 5.96940696e-01 -1.22304410e-01 -4.51146036e-01
-4.02083009e-01 -1.45065039e-01 -2.25949407e-01 -2.42241636e-01
-1.30269855e-01 -2.99298614e-01 -5.50029516e-01 4.74525928e-01
-8.76309514e-01 5.45101166e-01 2.55563080e-01 7.31677264e-02
1.09607446e+00 1.80634528e-01 3.60161327e-02 -1.20011055e+00
3.08883160e-01 -1.59017473e-01 -9.68884230e-01 3.20518821e-01
-1.40181363e-01 -1.76723942e-01 4.74762432e-02 -6.14516973e-01
-1.28042173e+00 6.26025319e-01 1.65471807e-01 -4.04817402e-01
-4.86780196e-01 2.33113125e-01 -4.89415526e-02 -4.68349569e-02
3.08342308e-01 -2.32145399e-01 -3.96468490e-01 -4.92919058e-01
4.32238907e-01 2.06058800e-01 5.73744535e-01 -5.90427399e-01
5.48082709e-01 7.34014153e-01 -5.76716512e-02 -8.08587492e-01
-8.42951655e-01 -7.68200278e-01 -1.16517210e+00 -4.75390226e-01
9.41986263e-01 -7.97522664e-01 -4.54128712e-01 3.74087214e-01
-1.27576017e+00 -5.89574277e-01 -4.07563597e-01 4.47958440e-01
-3.84910434e-01 6.14388347e-01 -4.12799746e-01 -9.57406104e-01
2.18502730e-01 -1.38593221e+00 1.31146491e+00 5.47225654e-01
-3.08768541e-01 -9.31874335e-01 3.90584022e-02 3.51803720e-01
-4.51344103e-02 7.69619644e-01 2.17950851e-01 -1.80847511e-01
-1.23545372e+00 -6.52903169e-02 1.76214408e-02 -3.59631091e-01
2.79594928e-01 2.03022078e-01 -9.22031462e-01 -3.90192896e-01
5.22498488e-02 2.42292657e-01 6.29583955e-01 9.32885945e-01
1.09360170e+00 1.79767594e-01 -1.06774187e+00 5.09411156e-01
1.18808043e+00 5.45457721e-01 3.34988922e-01 2.99687386e-01
8.16450119e-01 4.89864767e-01 1.12568355e+00 2.63691157e-01
4.13742393e-01 1.03050828e+00 3.42579752e-01 -2.99160898e-01
-2.31601715e-01 -7.32236505e-02 -1.00196213e-01 6.41654432e-01
3.88753340e-02 -4.46469635e-01 -8.42237413e-01 4.67493355e-01
-2.09409904e+00 -6.42080724e-01 -3.56272519e-01 2.26100159e+00
5.75057983e-01 5.90081036e-01 9.97355357e-02 -9.03231278e-02
8.18966389e-01 2.51841117e-02 -8.75293970e-01 -1.73469484e-01
2.73625493e-01 4.68248017e-02 4.84036773e-01 8.86261642e-01
-1.40586102e+00 1.23520350e+00 6.83643198e+00 2.80991554e-01
-7.85393894e-01 -1.06988817e-01 3.64021063e-01 1.94084551e-02
-1.54417023e-01 2.63663352e-01 -9.45939004e-01 3.61425757e-01
2.69082844e-01 1.21346287e-01 2.17207476e-01 6.42503262e-01
2.27611855e-01 -9.80662823e-01 -1.16456831e+00 8.87918830e-01
1.83460880e-02 -1.03981149e+00 -3.66791725e-01 4.96419519e-02
8.69617283e-01 8.01875349e-03 -1.38180748e-01 -8.25985987e-03
4.88775343e-01 -5.50747275e-01 7.61525750e-01 4.73651141e-01
3.38530391e-01 -2.74241835e-01 2.86221683e-01 5.79383552e-01
-1.36973643e+00 2.67009318e-01 5.12413718e-02 -2.24959143e-02
7.36787021e-01 6.32745922e-01 -1.14563072e+00 4.08031076e-01
5.77574670e-01 5.58537066e-01 -5.29394686e-01 1.53732598e+00
-1.61325932e-01 9.92241874e-02 -5.04550219e-01 3.08150232e-01
2.41013244e-02 -1.87680662e-01 8.26871634e-01 1.12511957e+00
7.41989091e-02 2.64042705e-01 8.91574025e-01 8.91773283e-01
4.61445153e-01 -3.37500930e-01 -4.63142425e-01 2.56714672e-01
5.08079052e-01 1.11848366e+00 -1.67957520e+00 -4.81763542e-01
-1.75553039e-01 1.22120214e+00 -1.71882808e-02 6.02568090e-01
-7.79105067e-01 1.59936309e-01 4.51327235e-01 2.15074733e-01
4.91837263e-01 -6.18351221e-01 -2.21925274e-01 -1.26897240e+00
-1.80513505e-03 -4.21028972e-01 2.48628646e-01 -8.66294563e-01
-5.93265653e-01 4.65225011e-01 3.54784071e-01 -1.09368026e+00
-7.85433426e-02 -1.67225346e-01 -7.59690627e-02 5.02056539e-01
-1.61076295e+00 -8.67587328e-01 -3.91835570e-01 3.46306175e-01
1.19012904e+00 6.71170354e-01 4.08824623e-01 7.73223788e-02
-3.98735970e-01 -4.15664241e-02 -7.78883100e-02 -3.88236493e-01
4.44430232e-01 -1.38329411e+00 5.51973283e-01 1.07757711e+00
2.08746955e-01 3.73006076e-01 8.25658858e-01 -1.19183123e+00
-1.10183585e+00 -7.31835067e-01 2.65042484e-01 -7.53955185e-01
1.66445643e-01 -6.82502389e-01 -9.79458809e-01 8.09962928e-01
-1.93054065e-01 1.04191437e-01 2.20040113e-01 -1.81031138e-01
3.38594198e-01 2.80222535e-01 -1.25619316e+00 6.71002626e-01
1.14589512e+00 1.56412557e-01 -3.48726749e-01 3.47105086e-01
1.09439027e+00 -1.38274324e+00 -6.02970719e-01 5.84625840e-01
5.61687171e-01 -1.03041875e+00 1.04404366e+00 -1.67488500e-01
-3.06036919e-01 -8.09068918e-01 5.61019517e-02 -9.53266919e-01
3.00334841e-02 -8.42612803e-01 -2.99943626e-01 7.94669807e-01
1.17863826e-01 -3.65770966e-01 1.03007734e+00 9.70774114e-01
-2.84716308e-01 -5.36921680e-01 -7.54513741e-01 -6.52666330e-01
-5.46504200e-01 -5.30030549e-01 4.80483621e-01 8.69065523e-01
-4.05560166e-01 7.36708567e-02 -1.15268439e-01 6.30735278e-01
7.11153805e-01 4.20850128e-01 9.97848928e-01 -1.38173807e+00
-2.34039664e-01 -1.24039583e-01 -3.45003903e-01 -1.48375058e+00
-2.46491432e-01 -3.19499582e-01 4.00599808e-01 -1.82156193e+00
1.57092005e-01 -6.41200185e-01 1.61298946e-01 1.98473901e-01
-3.51875275e-01 -1.64066538e-01 3.83518428e-01 1.83318987e-01
-8.91793549e-01 5.95504679e-02 1.63497913e+00 6.85298592e-02
-8.24022532e-01 4.29452300e-01 -1.91633508e-01 1.05363834e+00
5.22127032e-01 -4.85411257e-01 -5.69317341e-01 -5.76423168e-01
-1.46417931e-01 4.43491340e-01 1.13834739e-01 -1.15492237e+00
2.36067161e-01 -3.46019357e-01 5.16669214e-01 -1.19101012e+00
6.15401566e-01 -9.69664574e-01 3.76057178e-01 5.39227486e-01
1.04105331e-01 -5.91810159e-02 6.37569070e-01 7.31357098e-01
1.55403584e-01 -2.68230975e-01 5.33682764e-01 -4.22838092e-01
-1.00632012e+00 3.42058241e-01 -3.53724569e-01 -1.90028891e-01
1.31243968e+00 -9.24430430e-01 2.06029445e-01 -4.48767841e-02
-1.19468999e+00 5.97380817e-01 7.17552960e-01 4.27465737e-01
5.36162794e-01 -8.90844643e-01 -2.14510784e-01 1.89949110e-01
-2.46490791e-01 7.00019658e-01 2.18716934e-01 7.47758806e-01
-6.39965177e-01 2.61291116e-01 1.43921047e-01 -1.20213234e+00
-1.55515015e+00 6.19423926e-01 2.03312457e-01 -2.83363789e-01
-7.66480923e-01 7.58816361e-01 4.14278746e-01 -2.16899887e-01
4.18727666e-01 -8.46364737e-01 -1.32528514e-01 -9.00047049e-02
3.42076272e-01 3.16923589e-01 -1.46794483e-01 -4.44738746e-01
-2.65501499e-01 8.70378852e-01 -3.48692387e-01 -3.58661681e-01
1.04893279e+00 -6.40852153e-01 2.10317016e-01 8.05767655e-01
5.29256940e-01 -2.21960410e-03 -1.78310442e+00 -2.62290865e-01
6.78039193e-02 -9.56476092e-01 -4.03755680e-02 -8.16626966e-01
-6.81911349e-01 5.56080461e-01 6.75048649e-01 3.33070695e-01
8.84526849e-01 5.88823929e-02 8.02700162e-01 9.30269659e-02
6.51860118e-01 -1.11309314e+00 1.23903766e-01 3.98081183e-01
4.69865978e-01 -1.14797461e+00 3.44890028e-01 -9.87688303e-01
-4.95153785e-01 1.05658674e+00 8.09389353e-01 3.29697162e-01
4.10186768e-01 5.18584549e-01 7.23367929e-02 -4.34967160e-01
-2.97982395e-01 -3.93646389e-01 1.35566965e-01 6.51339829e-01
-1.64250806e-02 -4.27187681e-02 6.83657899e-02 -9.75311026e-02
1.51280895e-01 -2.13821437e-02 7.18640268e-01 1.35463202e+00
-6.47392094e-01 -1.09549451e+00 -7.74151266e-01 3.07454526e-01
-2.60751307e-01 4.00916100e-01 -1.00593232e-01 9.92180526e-01
1.83533981e-01 8.65739584e-01 4.27229591e-02 1.98586240e-01
4.22438473e-01 -2.70902097e-01 8.05904090e-01 -1.06365180e+00
-2.45700791e-01 7.15582967e-01 -5.27612399e-03 -8.21510434e-01
-7.57808566e-01 -1.07284319e+00 -1.47682500e+00 3.20684135e-01
-7.95580924e-01 -1.17784411e-01 7.63329566e-01 9.42754030e-01
2.23614469e-01 8.08173120e-01 2.92907596e-01 -1.35683918e+00
2.79266566e-01 -3.77295941e-01 -4.21446294e-01 2.05713168e-01
4.51025546e-01 -8.34079504e-01 -1.69398114e-02 2.70611554e-01] | [9.136979103088379, -0.48007291555404663] |
7f73422c-56e4-4615-b9e7-0aeae6782344 | metric-scale-truncation-robust-heatmaps-for | 2003.02953 | null | https://arxiv.org/abs/2003.02953v1 | https://arxiv.org/pdf/2003.02953v1.pdf | Metric-Scale Truncation-Robust Heatmaps for 3D Human Pose Estimation | Heatmap representations have formed the basis of 2D human pose estimation systems for many years, but their generalizations for 3D pose have only recently been considered. This includes 2.5D volumetric heatmaps, whose X and Y axes correspond to image space and the Z axis to metric depth around the subject. To obtain metric-scale predictions, these methods must include a separate, explicit post-processing step to resolve scale ambiguity. Further, they cannot encode body joint positions outside of the image boundaries, leading to incomplete pose estimates in case of image truncation. We address these limitations by proposing metric-scale truncation-robust (MeTRo) volumetric heatmaps, whose dimensions are defined in metric 3D space near the subject, instead of being aligned with image space. We train a fully-convolutional network to estimate such heatmaps from monocular RGB in an end-to-end manner. This reinterpretation of the heatmap dimensions allows us to estimate complete metric-scale poses without test-time knowledge of the focal length or person distance and without relying on anthropometric heuristics in post-processing. Furthermore, as the image space is decoupled from the heatmap space, the network can learn to reason about joints beyond the image boundary. Using ResNet-50 without any additional learned layers, we obtain state-of-the-art results on the Human3.6M and MPI-INF-3DHP benchmarks. As our method is simple and fast, it can become a useful component for real-time top-down multi-person pose estimation systems. We make our code publicly available to facilitate further research (see https://vision.rwth-aachen.de/metro-pose3d). | ['Bastian Leibe', 'István Sárándi', 'Timm Linder', 'Kai O. Arras'] | 2020-03-05 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-1.51537850e-01 2.74063230e-01 -1.99111439e-02 -5.83342552e-01
-7.00049460e-01 -5.59064806e-01 2.07289129e-01 -6.89847916e-02
-5.72137773e-01 4.53786284e-01 4.39636767e-01 1.57846063e-01
8.27616900e-02 -6.56345069e-01 -8.18618715e-01 -2.59773463e-01
-2.19368964e-01 9.43048537e-01 -1.17916465e-02 -1.59159452e-01
-2.32124552e-01 3.01360637e-01 -1.31111455e+00 -9.91939455e-02
4.66175437e-01 9.73592699e-01 -2.12942913e-01 6.81586981e-01
4.59534436e-01 -6.06252067e-02 -4.42288816e-01 -4.96881723e-01
4.55633968e-01 -2.36140087e-01 -7.90019512e-01 1.77957073e-01
8.47547114e-01 -4.87993598e-01 -4.93022054e-01 7.50082135e-01
7.76297152e-01 9.47665889e-03 5.23660183e-01 -1.22762656e+00
-3.67528349e-01 2.14979380e-01 -6.73284411e-01 -2.91304380e-01
7.16306388e-01 1.78813502e-01 6.72050059e-01 -8.31482470e-01
7.58650780e-01 1.36398172e+00 9.69848633e-01 6.29334867e-01
-1.11260974e+00 -5.55383623e-01 4.08322774e-02 7.44099217e-03
-1.49297261e+00 -1.17398836e-01 8.75273347e-01 -5.09003997e-01
7.83706129e-01 3.30385178e-01 1.09481013e+00 1.05245435e+00
3.03091854e-01 6.78162813e-01 1.00042844e+00 -1.64739057e-01
6.49587214e-02 -4.60359663e-01 -2.32821956e-01 1.00444436e+00
3.03271562e-01 1.15408897e-04 -5.33043385e-01 4.76783328e-02
1.10758698e+00 -3.29929553e-02 -1.27960920e-01 -9.69218552e-01
-1.36271048e+00 7.30108798e-01 8.80324900e-01 -1.67982370e-01
-2.27082387e-01 5.13795555e-01 3.30739737e-01 -9.74506512e-02
5.03349781e-01 3.79812896e-01 -5.95123291e-01 -2.16381580e-01
-9.01640773e-01 7.59263277e-01 5.34198105e-01 8.53512645e-01
5.27625918e-01 -3.88852030e-01 -6.85390159e-02 7.01589644e-01
2.80892760e-01 6.24017239e-01 2.84746915e-01 -1.20516288e+00
7.01726198e-01 7.03834474e-01 5.60307354e-02 -1.00121951e+00
-1.10646665e+00 -4.92537707e-01 -7.70109534e-01 1.11144833e-01
7.93803215e-01 -2.76753455e-01 -1.14297020e+00 1.82051265e+00
6.00617230e-01 -4.01018143e-01 -5.37504971e-01 1.62192154e+00
6.74146771e-01 1.26410514e-01 -1.92177162e-01 4.22780365e-01
1.64415324e+00 -7.83159614e-01 -1.84843853e-01 -3.67113650e-01
6.38970673e-01 -4.71334219e-01 1.09420538e+00 3.72441500e-01
-1.28853869e+00 -5.13365746e-01 -1.07856774e+00 -4.94974136e-01
-3.20555478e-01 1.77153796e-01 7.35555828e-01 6.49730682e-01
-9.08621132e-01 5.74634433e-01 -1.14634275e+00 -5.40989041e-01
1.90200150e-01 4.85253632e-01 -5.81052065e-01 8.75589773e-02
-1.16477621e+00 9.39315856e-01 8.06064904e-02 4.89224941e-01
-5.57773888e-01 -6.21444762e-01 -1.23011601e+00 -3.42702061e-01
3.54462147e-01 -1.10116768e+00 1.13431835e+00 -3.62980098e-01
-1.36061001e+00 1.10604024e+00 9.54966098e-02 -2.36364663e-01
1.12095511e+00 -6.70866728e-01 3.01338006e-02 2.57582724e-01
1.66996837e-01 1.02788377e+00 5.94070494e-01 -9.83393192e-01
-1.20981477e-01 -9.10070181e-01 1.09265618e-01 5.60312748e-01
-4.49694544e-02 -3.76888782e-01 -9.04342115e-01 -6.52952552e-01
5.00649452e-01 -1.32746458e+00 -1.94963887e-01 6.33636713e-01
-6.70672774e-01 6.85637817e-02 2.90117413e-01 -8.55013072e-01
7.92060733e-01 -1.70261478e+00 6.35442138e-01 2.76915669e-01
2.65904546e-01 -2.26077899e-01 4.05358560e-02 9.16071832e-02
1.01434238e-01 -1.18506059e-01 -2.67617345e-01 -5.65418541e-01
1.07992768e-01 -9.36740078e-03 3.16783071e-01 9.67753470e-01
-5.45900390e-02 1.06297815e+00 -7.43991613e-01 -5.35276473e-01
4.09773856e-01 7.22084045e-01 -7.88293362e-01 1.02385432e-02
-3.16508710e-02 5.58269680e-01 -2.29397103e-01 6.93256676e-01
5.69129705e-01 -1.29577205e-01 6.00160100e-02 -4.95825380e-01
1.22412220e-01 1.14505090e-01 -1.25228953e+00 2.44414854e+00
-4.10356611e-01 3.39594960e-01 3.93809117e-02 -5.96233189e-01
6.64928794e-01 3.85794826e-02 8.13811541e-01 -5.06644964e-01
2.23904029e-01 -4.21957783e-02 -1.80839866e-01 -1.43291935e-01
4.60849255e-01 -2.20200773e-02 -4.25284088e-01 2.16087714e-01
-2.34848894e-02 -2.54038215e-01 9.36424881e-02 -1.27216443e-01
7.62732387e-01 5.63690782e-01 4.02773730e-02 -5.44864126e-02
2.06718385e-01 -9.13315639e-02 5.15437126e-01 2.82144815e-01
-2.30157495e-01 1.15209949e+00 3.66542667e-01 -6.72214806e-01
-1.20149362e+00 -1.45830488e+00 -1.26123041e-01 8.19178283e-01
2.40484968e-01 -6.28106236e-01 -9.24878776e-01 -5.71634591e-01
3.69027108e-01 1.16795436e-01 -9.00503457e-01 -1.99315876e-01
-8.65415037e-01 -5.86856425e-01 6.23061359e-01 8.22004914e-01
5.48525810e-01 -5.89515269e-01 -9.03646052e-01 1.16272777e-01
-4.23766822e-01 -1.08379936e+00 -6.25612557e-01 -7.77196810e-02
-7.03823209e-01 -1.25084031e+00 -1.25489688e+00 -4.73473072e-01
7.02822387e-01 -2.16346219e-01 9.85876620e-01 -2.21951053e-01
-5.43572128e-01 4.66785669e-01 -6.29578531e-02 -7.17784017e-02
1.72794119e-01 1.94627553e-01 3.41818810e-01 -4.15340096e-01
6.61085621e-02 -4.81023341e-01 -1.12430191e+00 5.19947469e-01
-2.04142958e-01 3.07450116e-01 3.87823403e-01 5.08133352e-01
7.36376941e-01 -4.08817798e-01 -2.64949370e-02 -2.96933383e-01
2.72056699e-01 4.88371775e-03 -4.43826288e-01 5.56466356e-02
-1.63705483e-01 5.93452603e-02 2.31261075e-01 -2.69663841e-01
-5.96449256e-01 4.49494183e-01 -1.93277627e-01 -3.44086081e-01
-6.80392310e-02 3.17156255e-01 -9.62758064e-02 -1.49000278e-02
7.00727284e-01 -7.18882978e-02 1.48207501e-01 -6.02453113e-01
5.25787115e-01 3.53335798e-01 7.78324664e-01 -7.46399343e-01
7.58910060e-01 6.01988614e-01 2.65407681e-01 -5.57816446e-01
-6.75735652e-01 -3.11814517e-01 -1.05314612e+00 -3.80419821e-01
1.23971367e+00 -1.11283302e+00 -9.12092984e-01 5.19893467e-01
-1.00739229e+00 -3.99249732e-01 -9.09755453e-02 5.82987905e-01
-8.93199623e-01 3.15539062e-01 -6.97511435e-01 -4.73504126e-01
-4.86112714e-01 -1.09042048e+00 1.55468738e+00 -4.26509492e-02
-6.03653133e-01 -5.70682704e-01 6.31214604e-02 6.04169548e-01
-1.50405420e-02 6.61406517e-01 4.83084053e-01 -9.92355403e-04
-2.14299366e-01 -3.75289738e-01 -1.44465163e-01 -7.54474662e-03
-7.24715889e-02 -4.97553349e-01 -8.15512955e-01 -5.06909549e-01
-5.04704475e-01 -4.96708900e-01 6.73270702e-01 6.15195632e-01
1.24577832e+00 -9.57418084e-02 -2.18110815e-01 1.01864660e+00
9.86762345e-01 -6.21974647e-01 4.13558394e-01 3.83653134e-01
9.67184722e-01 6.29513383e-01 5.44239998e-01 4.65958327e-01
8.80918503e-01 1.14342785e+00 3.72598052e-01 -2.39518613e-01
-3.81524920e-01 -4.47565526e-01 1.60301790e-01 5.13081312e-01
-4.00590152e-01 2.80920148e-01 -1.03993571e+00 2.46138901e-01
-1.58190846e+00 -5.03571987e-01 7.68709034e-02 2.37785649e+00
9.09371138e-01 2.67431349e-01 5.26179731e-01 -2.31149178e-02
4.92296845e-01 1.24215215e-01 -7.89514184e-01 -8.97486359e-02
1.77770317e-01 6.19159779e-03 7.47380435e-01 5.49720466e-01
-1.29908395e+00 8.08938146e-01 5.47637892e+00 2.89398938e-01
-1.10390556e+00 5.39818965e-02 4.08276528e-01 -5.91040432e-01
1.41885012e-01 -3.47074389e-01 -7.37990558e-01 3.08396310e-01
5.64025462e-01 3.94108236e-01 3.31660718e-01 8.79982233e-01
1.95856281e-02 -1.01670772e-01 -1.26062953e+00 1.31883383e+00
3.96423452e-02 -7.96953917e-01 -2.96657443e-01 1.60325214e-01
4.50987011e-01 4.83409762e-02 -6.84113279e-02 2.32412547e-01
-2.66431365e-02 -1.00048828e+00 1.03361201e+00 4.98345941e-01
1.06179583e+00 -7.83930421e-01 4.32511777e-01 1.85043320e-01
-1.42189884e+00 1.99520305e-01 -4.42341059e-01 -1.79480895e-01
3.06433588e-01 5.02451420e-01 -7.10835218e-01 4.77289200e-01
8.81286800e-01 4.45324004e-01 -7.42837429e-01 9.57738757e-01
-2.48899162e-01 -2.27183960e-02 -6.96791172e-01 1.34309560e-01
-1.21457458e-01 9.75312367e-02 3.55392456e-01 9.66721892e-01
3.19805264e-01 -1.47363693e-01 2.84109324e-01 7.98706591e-01
-4.06653099e-02 -1.07873447e-01 -2.26814419e-01 4.18822110e-01
1.66553229e-01 1.08579016e+00 -7.73238957e-01 -2.78238114e-02
-1.41653225e-01 1.23029387e+00 3.21850240e-01 2.53855705e-01
-1.02368999e+00 -3.18555117e-01 8.63349617e-01 4.68350828e-01
-2.39663804e-03 -6.22627616e-01 -5.84813833e-01 -1.28052628e+00
3.36463720e-01 -5.89582384e-01 3.48585904e-01 -7.38737941e-01
-1.01005018e+00 3.70294154e-01 3.31614196e-01 -1.13091540e+00
-5.13980389e-01 -6.92281067e-01 -2.56943237e-03 7.06984460e-01
-8.27350438e-01 -1.36992908e+00 -4.44991589e-01 5.52010119e-01
2.27988034e-01 4.48889256e-01 8.05717409e-01 1.79214463e-01
-2.34760180e-01 8.31052423e-01 -4.47036386e-01 5.34199297e-01
8.38569939e-01 -1.33691943e+00 7.28007376e-01 5.67895770e-01
6.78260401e-02 5.36703289e-01 7.48375237e-01 -7.70195901e-01
-1.49924827e+00 -8.67892206e-01 7.57880867e-01 -8.03655326e-01
1.70203164e-01 -9.13906515e-01 -3.74740869e-01 7.16380596e-01
-4.14275825e-01 2.14558586e-01 4.36165899e-01 4.83395040e-01
-4.34438229e-01 -1.80323541e-01 -1.06045043e+00 6.21462166e-01
1.50228417e+00 -3.78297567e-01 -5.71819305e-01 4.38271493e-01
5.50355852e-01 -1.07916772e+00 -1.13800859e+00 4.79345530e-01
1.04718518e+00 -8.16708446e-01 1.31168461e+00 -3.31643403e-01
2.67839909e-01 -3.20716232e-01 -6.44849092e-02 -1.17296803e+00
-2.17411935e-01 -3.05013031e-01 -3.21967512e-01 6.46330059e-01
1.99509501e-01 -2.65715420e-01 1.22138786e+00 8.17856073e-01
-5.10707088e-02 -1.00845897e+00 -1.25796616e+00 -7.26193845e-01
1.99680954e-01 -5.25631011e-01 5.16255915e-01 5.08791745e-01
5.16304821e-02 -2.86184698e-02 -5.24906993e-01 1.19841561e-01
8.83573711e-01 1.26498386e-01 1.06559777e+00 -1.15906584e+00
-3.12709123e-01 -3.93405139e-01 -7.53097475e-01 -1.16815829e+00
-1.84050277e-01 -6.24359488e-01 1.15933083e-01 -1.58167720e+00
1.27542004e-01 -2.23254114e-01 -3.54034267e-02 6.41849875e-01
2.56546177e-02 7.46903956e-01 3.54052484e-01 3.42964157e-02
-4.77300942e-01 6.09324276e-01 1.39681542e+00 1.68294739e-02
-1.11871906e-01 -1.58788428e-01 -3.71489644e-01 8.42408597e-01
7.78953969e-01 -2.38205612e-01 -3.13939989e-01 -4.98154730e-01
3.10384065e-01 1.05002830e-02 7.74682641e-01 -1.33226621e+00
4.11026888e-02 1.37301639e-01 1.14428926e+00 -7.81198025e-01
7.98329115e-01 -6.26142502e-01 1.25820667e-01 6.33283615e-01
-2.69612551e-01 2.21832052e-01 -7.59660527e-02 2.58117229e-01
2.22491935e-01 4.37754989e-01 6.35753334e-01 -2.83918798e-01
-3.71492296e-01 5.26042044e-01 1.62985951e-01 1.54149532e-01
8.55056822e-01 -3.65620822e-01 -3.82710248e-02 -4.56898540e-01
-9.04687524e-01 3.25460732e-01 8.31304491e-01 5.80579042e-01
6.37898803e-01 -1.53704762e+00 -6.23398960e-01 2.11908832e-01
1.32416815e-01 3.25295836e-01 3.26803535e-01 9.11305130e-01
-8.32181156e-01 4.79521215e-01 -3.71310294e-01 -8.25214565e-01
-1.05968678e+00 3.60675871e-01 4.91284519e-01 -1.52339637e-01
-9.37416792e-01 7.93158293e-01 9.88618061e-02 -8.07782829e-01
3.91991258e-01 -5.85836232e-01 3.07163477e-01 -1.62483931e-01
3.28670651e-01 3.51077259e-01 4.32377122e-02 -8.21096361e-01
-6.94695115e-01 1.03710079e+00 1.79412857e-01 -3.48189950e-01
1.14509261e+00 -2.33119637e-01 2.42813826e-01 2.14095116e-01
1.35987842e+00 -2.47844681e-01 -1.60214424e+00 1.00414127e-01
-4.12783861e-01 -4.22616541e-01 -1.75335988e-01 -9.74920809e-01
-1.08482397e+00 8.76171708e-01 8.02803278e-01 -4.51871634e-01
8.23017597e-01 1.52996287e-01 9.00117993e-01 1.51008308e-01
5.92731953e-01 -1.28015935e+00 3.26444060e-02 2.92646915e-01
1.20994771e+00 -1.18768454e+00 3.60301197e-01 -3.92292231e-01
-3.78511310e-01 1.03647423e+00 7.52926171e-01 -2.49799371e-01
5.34314692e-01 3.29624442e-03 1.23545900e-01 -2.66497523e-01
-1.43202692e-01 -1.13218710e-01 8.46661806e-01 6.74138546e-01
6.29844189e-01 3.81957173e-01 -2.49754667e-01 4.95334774e-01
-7.63952315e-01 -1.39758617e-01 4.29162830e-02 8.54430079e-01
-1.99287131e-01 -9.02113378e-01 -6.02979124e-01 1.96485907e-01
-3.00151080e-01 3.52699250e-01 -4.69936877e-01 9.53168452e-01
2.14030474e-01 5.71786404e-01 7.37002194e-02 -5.34676850e-01
5.31331241e-01 -1.60228778e-02 1.00605226e+00 -4.53361779e-01
-2.00778499e-01 9.63531509e-02 1.92164004e-01 -1.08221090e+00
-1.37265816e-01 -6.85680509e-01 -1.47445011e+00 -3.83821487e-01
1.47102028e-01 -2.39524886e-01 9.16037440e-01 7.09085584e-01
3.17185938e-01 2.44581729e-01 -1.22130243e-02 -1.43542039e+00
-4.65629190e-01 -9.29430008e-01 -3.59312624e-01 5.99682868e-01
1.82921574e-01 -9.56553876e-01 1.76685601e-01 -2.71569967e-01] | [6.991746425628662, -0.9270896315574646] |
eabb0e13-23ca-4591-9481-b5fcc898a2d6 | a-deep-dive-into-explainable-self-supervised | 2306.10798 | null | https://arxiv.org/abs/2306.10798v2 | https://arxiv.org/pdf/2306.10798v2.pdf | ExpPoint-MAE: Better interpretability and performance for self-supervised point cloud transformers | In this paper we delve into the properties of transformers, attained through self-supervision, in the point cloud domain. Specifically, we evaluate the effectiveness of Masked Autoencoding as a pretraining scheme, and explore Momentum Contrast as an alternative. In our study we investigate the impact of data quantity on the learned features, and uncover similarities in the transformer's behavior across domains. Through comprehensive visualiations, we observe that the transformer learns to attend to semantically meaningful regions, indicating that pretraining leads to a better understanding of the underlying geometry. Moreover, we examine the finetuning process and its effect on the learned representations. Based on that, we devise an unfreezing strategy which consistently outperforms our baseline without introducing any other modifications to the model or the training pipeline, and achieve state-of-the-art results in the classification task among transformer models. | ['Adrian Munteanu', 'Konstantinos Moustakas', 'Vlassis Fotis', 'Ioannis Romanelis'] | 2023-06-19 | null | null | null | null | ['3d-point-cloud-classification', 'explainable-artificial-intelligence'] | ['computer-vision', 'computer-vision'] | [ 4.50880527e-02 1.92874387e-01 -3.45873795e-02 -1.74490452e-01
-6.77797735e-01 -7.77009130e-01 8.57980967e-01 1.84107572e-01
-2.23532394e-01 2.99688607e-01 4.07051116e-01 -2.51149207e-01
-1.47804156e-01 -8.04004610e-01 -9.83777583e-01 -7.19044864e-01
3.58498879e-02 4.91241187e-01 4.31224883e-01 -2.22383350e-01
2.43978679e-01 6.99034095e-01 -1.34240162e+00 2.00222164e-01
8.83610487e-01 1.07463205e+00 1.10372640e-01 1.63264468e-01
1.90001160e-01 5.52472949e-01 -2.66052306e-01 -4.97055829e-01
4.12529111e-01 -1.97348464e-02 -8.02837670e-01 2.82435447e-01
6.68849587e-01 -1.76263705e-01 -5.04821062e-01 7.59111762e-01
7.83733949e-02 -6.23772503e-04 1.09163475e+00 -8.01033616e-01
-5.55200517e-01 3.32360417e-01 -4.10413295e-01 4.85464752e-01
-1.62848588e-02 2.12178126e-01 1.39147484e+00 -9.17752802e-01
6.33037090e-01 9.58224058e-01 7.38139629e-01 3.42789851e-02
-1.45459366e+00 -5.17369747e-01 3.81170869e-01 5.62034063e-02
-1.24137485e+00 -6.26255333e-01 9.07199621e-01 -5.96025467e-01
7.26992965e-01 -2.37909347e-01 6.33075893e-01 9.89912808e-01
1.21336646e-01 6.51327550e-01 1.06742418e+00 -4.55188334e-01
2.35954463e-01 2.29444847e-01 -2.38943323e-02 8.93104970e-01
2.38178581e-01 3.01675916e-01 -5.84217012e-01 8.19074810e-02
1.00354123e+00 -2.42145166e-01 -3.98867846e-01 -8.61189127e-01
-9.59533691e-01 7.99776852e-01 6.99797750e-01 1.83954030e-01
-2.32022837e-01 1.14002578e-01 1.90692395e-01 2.94481188e-01
6.42132103e-01 6.76388383e-01 -5.17281473e-01 1.27561651e-02
-8.54864478e-01 1.25822902e-01 3.42152745e-01 6.54468834e-01
8.33711982e-01 -5.61973751e-02 -1.63664848e-01 5.68619013e-01
-7.94363841e-02 1.06010437e-01 1.61514565e-01 -7.61457443e-01
5.10529578e-01 7.46250093e-01 -1.18272118e-01 -6.68636978e-01
-1.81830674e-01 -7.95996428e-01 -3.49624097e-01 2.13210315e-01
5.26003480e-01 9.16856378e-02 -9.65441704e-01 1.73844588e+00
1.68679625e-01 1.10118255e-01 -8.48241076e-02 8.00939679e-01
4.30182278e-01 1.73343867e-01 -2.06067227e-04 2.52538115e-01
1.09822774e+00 -9.70851541e-01 -1.19306915e-01 -2.77769566e-01
5.52900016e-01 -6.57040834e-01 1.28492975e+00 3.86806458e-01
-9.96885121e-01 -4.66304421e-01 -1.01790714e+00 -1.01443209e-01
-1.86367020e-01 2.69389898e-01 7.36583531e-01 3.10781717e-01
-9.57673669e-01 1.21365356e+00 -1.14619422e+00 -5.10211885e-01
7.07515597e-01 2.39620030e-01 -3.54231358e-01 3.99500877e-02
-6.20349467e-01 8.91808569e-01 7.71354958e-02 -3.52708876e-01
-1.07023299e+00 -9.19398904e-01 -7.33737171e-01 2.28499472e-01
2.98184991e-01 -6.85247600e-01 1.19380140e+00 -9.43538308e-01
-1.38430834e+00 8.69941771e-01 -1.97380632e-01 -3.58028114e-01
5.31414747e-01 -8.40133578e-02 8.09744652e-03 2.76000708e-01
-6.27660081e-02 6.85021341e-01 9.06361341e-01 -1.23868096e+00
-4.30139244e-01 -4.26355153e-01 2.37740129e-01 1.48298934e-01
-3.96886379e-01 -3.47036481e-01 -3.90761644e-01 -8.31435204e-01
1.68626070e-01 -9.65017021e-01 -4.25810851e-02 4.74016555e-02
-2.78904080e-01 -1.31511971e-01 3.95466715e-01 -3.16281140e-01
7.61598945e-01 -2.38759947e+00 2.90453494e-01 2.64511555e-01
1.36948287e-01 -1.61446333e-01 -1.74275547e-01 4.45843160e-01
-1.85904101e-01 8.84857215e-03 -1.62609324e-01 -6.71459734e-01
-3.26235145e-02 2.34880418e-01 -5.08713901e-01 5.73545039e-01
5.24468362e-01 9.06517327e-01 -4.74211246e-01 -3.21223438e-01
8.82861465e-02 4.10699159e-01 -9.66303110e-01 1.92654550e-01
-1.86280951e-01 6.74232364e-01 -5.00864506e-01 5.35459280e-01
4.45081949e-01 -6.37403488e-01 1.40434042e-01 -3.31742108e-01
-6.79720640e-02 8.90748799e-01 -5.94038725e-01 1.80383122e+00
-5.79450727e-01 6.23270929e-01 -1.06719106e-01 -1.16670382e+00
6.46775007e-01 1.55964404e-01 3.57551277e-01 -6.56009495e-01
8.41082633e-02 6.43402115e-02 6.80058375e-02 -1.57337680e-01
2.80678362e-01 -4.44164991e-01 2.89565384e-01 2.50246167e-01
2.38392010e-01 3.74777168e-02 -1.31157264e-01 1.15624204e-01
1.08204401e+00 2.38335311e-01 -1.85420141e-02 -5.98296344e-01
1.52862566e-02 1.61860913e-01 1.99066684e-01 5.69981456e-01
2.27459416e-01 7.51023114e-01 5.49391210e-01 -2.85353959e-01
-9.08474743e-01 -1.24959540e+00 -3.84048015e-01 1.03036654e+00
1.82589442e-01 -4.85937417e-01 -4.62810755e-01 -7.57284641e-01
1.36294156e-01 5.35105407e-01 -8.15316141e-01 -4.33133364e-01
-5.01764178e-01 -6.48618877e-01 2.92304248e-01 9.12902296e-01
4.86075342e-01 -5.11989534e-01 -6.93573058e-01 -1.09514967e-01
4.88346964e-02 -1.36335385e+00 -1.77687719e-01 4.03478295e-01
-1.17418969e+00 -9.96149182e-01 -3.52322489e-01 -5.54884255e-01
6.37995005e-01 1.14095695e-01 1.32068992e+00 6.71474934e-02
7.46545121e-02 3.78618240e-01 -2.27434054e-01 -6.67648166e-02
-3.91188376e-02 5.51658690e-01 -2.32682183e-01 -2.95389473e-01
-2.57167164e-02 -1.02248633e+00 -6.83096349e-01 3.45437646e-01
-4.92922843e-01 -1.05322432e-03 7.34411776e-01 6.56015337e-01
4.22540098e-01 1.14499219e-01 2.91140582e-02 -8.55359077e-01
3.59176993e-01 -2.52411753e-01 -6.50518596e-01 7.94404447e-02
-6.57663941e-01 6.03970528e-01 6.41921580e-01 -3.04459095e-01
-9.04769957e-01 -1.38976341e-02 -1.33343309e-01 -6.27149165e-01
1.32930605e-02 4.08930957e-01 -5.34722507e-02 -3.93042594e-01
4.44929332e-01 9.12266001e-02 -9.14003626e-02 -7.45718062e-01
2.71994203e-01 7.17656836e-02 3.69018555e-01 -9.72447693e-01
9.76908386e-01 6.77874088e-01 -1.22584619e-01 -5.17493904e-01
-9.93004501e-01 -2.06416890e-01 -5.69620371e-01 1.21823035e-01
7.19410479e-01 -1.08802879e+00 -4.84579831e-01 8.49228054e-02
-9.50827539e-01 -5.77742457e-01 -3.49426061e-01 2.74097621e-01
-5.84110141e-01 1.04785385e-02 -7.25981772e-01 -3.07848990e-01
1.69127434e-02 -1.38036036e+00 1.22431433e+00 -6.46369234e-02
-4.58228588e-02 -9.90538180e-01 1.30794346e-01 2.06935465e-01
3.57546985e-01 3.15217189e-02 1.17655087e+00 -6.74167037e-01
-1.04536355e+00 3.25504243e-01 -3.13819528e-01 1.17175445e-01
5.87506555e-02 -1.43203318e-01 -1.30094540e+00 -3.70311022e-01
-9.62011740e-02 -2.76148498e-01 1.25311863e+00 2.53500223e-01
1.30397701e+00 -2.13300869e-01 -4.22742665e-01 9.25574064e-01
1.35202479e+00 -4.03073549e-01 5.88255763e-01 5.47959030e-01
6.01273835e-01 5.87243736e-01 2.96719670e-01 2.15429574e-01
4.64948475e-01 7.64840543e-01 5.81469178e-01 -1.33150965e-01
-6.28302619e-02 -5.92164993e-01 1.67866275e-01 5.38994133e-01
-2.31034204e-01 3.37500987e-03 -8.30637455e-01 5.45472264e-01
-1.50317907e+00 -4.84702855e-01 4.58019227e-01 2.20571470e+00
6.58536792e-01 4.06367153e-01 2.16193870e-01 7.08784685e-02
1.81307092e-01 3.61776501e-01 -4.09068286e-01 -3.68468352e-02
4.02582064e-02 5.66186845e-01 6.32035017e-01 3.61691952e-01
-1.03599393e+00 1.07684970e+00 6.99708271e+00 6.56402528e-01
-1.57154679e+00 -4.40424904e-02 5.57836115e-01 -1.46634489e-01
-4.94050592e-01 6.74447119e-02 -5.26660264e-01 1.87620237e-01
6.33402109e-01 1.69332311e-01 5.99963844e-01 7.93819070e-01
-1.11306369e-01 1.09090686e-01 -1.43536913e+00 5.25736213e-01
-2.00343460e-01 -1.45800519e+00 9.29639786e-02 3.84947330e-01
5.90334296e-01 3.31832975e-01 2.53546208e-01 3.31089795e-01
3.62452954e-01 -9.71497178e-01 9.82495844e-01 2.31875911e-01
6.44865036e-01 -5.52854598e-01 3.62221062e-01 8.08337107e-02
-1.17271376e+00 -1.12135801e-02 -2.54210472e-01 -1.41943797e-01
-8.26165676e-02 4.71153051e-01 -8.09801280e-01 4.89858270e-01
7.83859670e-01 8.09575558e-01 -9.32503045e-01 7.56250024e-01
-3.45367819e-01 8.04863274e-01 -3.09349597e-01 4.38747913e-01
1.51709691e-01 -1.12305485e-01 4.18120593e-01 9.20203328e-01
2.05313846e-01 -6.56356886e-02 1.72878161e-01 1.01686037e+00
-1.54792830e-01 -6.22291937e-02 -4.50307935e-01 -1.78038850e-01
3.53360325e-01 9.88155186e-01 -7.54347801e-01 -1.88220143e-01
-3.46985877e-01 8.50270033e-01 8.21579933e-01 4.60651457e-01
-8.41689467e-01 1.36946157e-01 7.98775375e-01 5.94620645e-01
9.49720800e-01 -2.42380068e-01 -4.87644792e-01 -1.26015341e+00
2.32388765e-01 -8.11186671e-01 2.25520790e-01 -7.36195862e-01
-1.10487342e+00 4.75032508e-01 8.65437090e-03 -9.92103636e-01
-4.41531278e-02 -6.79737210e-01 -8.19842517e-01 5.91788530e-01
-1.85945296e+00 -1.05046129e+00 -2.03588456e-01 5.19207478e-01
2.16403171e-01 7.66881481e-02 4.66846138e-01 1.48239285e-01
-5.30380249e-01 7.07546055e-01 -3.35169286e-02 -3.39329541e-02
6.42296314e-01 -1.13472939e+00 4.42799330e-01 8.63678694e-01
3.69764894e-01 8.79929125e-01 7.72597671e-01 -3.07344764e-01
-1.19690907e+00 -9.29437876e-01 4.74313617e-01 -6.27968252e-01
8.24314475e-01 -4.88724619e-01 -1.01091409e+00 9.02889907e-01
1.02099612e-01 1.48022026e-01 3.35918248e-01 5.71504414e-01
-6.64650083e-01 -1.71927642e-02 -8.70794475e-01 2.63875902e-01
1.30915129e+00 -7.11025774e-01 -6.09561324e-01 2.10575201e-02
7.91315198e-01 -4.06498164e-01 -8.22850943e-01 5.81477463e-01
3.82199258e-01 -1.20336795e+00 1.07011795e+00 -6.79836631e-01
7.83960283e-01 -1.10862544e-02 -2.35762462e-01 -1.37874746e+00
-4.96486604e-01 -9.42597166e-02 5.92032187e-02 1.18090582e+00
4.77513283e-01 -7.41041601e-01 1.00211167e+00 3.03554177e-01
-3.26905429e-01 -1.11723280e+00 -8.15852582e-01 -7.21390605e-01
3.87613565e-01 -2.17126951e-01 6.05129957e-01 8.07195008e-01
-4.51573506e-02 4.83264387e-01 8.49184170e-02 3.95073891e-01
3.30308914e-01 5.05852938e-01 6.14525735e-01 -1.20968175e+00
-5.76520205e-01 -4.16078478e-01 -3.33600760e-01 -1.40692651e+00
3.13347429e-01 -8.12198341e-01 -2.32485652e-01 -1.17367244e+00
2.05419481e-01 -6.90040112e-01 -4.15598541e-01 4.88324136e-01
-5.08489236e-02 1.93856612e-01 2.07011759e-01 3.30079406e-01
-4.54779059e-01 8.33698213e-01 1.14499176e+00 -2.91781854e-02
1.66229978e-02 -8.76534581e-02 -9.00810838e-01 5.18218577e-01
6.62570417e-01 -3.30463827e-01 -4.84949529e-01 -7.63936758e-01
1.09825626e-01 -2.23183468e-01 5.46677589e-01 -8.59148979e-01
-1.08332500e-01 5.22429757e-02 3.82879555e-01 -1.80114314e-01
3.81812662e-01 -7.72880256e-01 -2.69486129e-01 9.09161344e-02
-3.07137877e-01 2.46037573e-01 4.06436145e-01 4.62967396e-01
-1.30127743e-01 -7.12833703e-02 7.78702378e-01 -4.41791527e-02
-4.64577287e-01 2.57854760e-01 -1.58733219e-01 3.33008498e-01
5.31605601e-01 -1.37811631e-01 -2.11386129e-01 -1.38409615e-01
-5.29470384e-01 4.22572158e-02 1.09698498e+00 1.75100327e-01
1.25806645e-01 -1.11681366e+00 -2.43050233e-01 3.03585768e-01
2.12031513e-01 1.77521467e-01 -1.96862504e-01 8.31317663e-01
-3.06791335e-01 3.45168769e-01 -1.01199895e-01 -7.50324011e-01
-7.95957983e-01 4.40663964e-01 4.23160315e-01 -4.56459522e-01
-7.38827765e-01 8.24758708e-01 5.17387509e-01 -4.09065038e-01
2.58938432e-01 -6.14098787e-01 5.71728349e-02 -1.23542741e-01
-7.52615333e-02 -4.95178401e-02 3.16518575e-01 -2.96000928e-01
-3.82669508e-01 6.43136680e-01 -1.80139437e-01 -3.35181095e-02
1.50758410e+00 5.01877926e-02 3.19530427e-01 3.10122162e-01
1.09629381e+00 2.21486866e-01 -1.63871038e+00 -1.78989977e-01
-3.89260910e-02 -3.91003102e-01 2.20434040e-01 -5.60728371e-01
-1.34241450e+00 8.71057749e-01 4.93952274e-01 -8.16578269e-02
1.05287373e+00 3.81389588e-01 4.26597804e-01 2.86512613e-01
3.02657634e-01 -6.39279962e-01 2.77649701e-01 4.24752057e-01
7.38273203e-01 -1.35441542e+00 1.08620107e-01 -4.85001773e-01
-5.73748171e-01 8.42175484e-01 4.99385387e-01 -4.98164713e-01
6.16462111e-01 2.03011498e-01 -9.70957726e-02 -4.31693912e-01
-8.78434718e-01 -2.15465218e-01 4.26594287e-01 2.98865050e-01
4.76485521e-01 -2.71388859e-01 2.61560649e-01 2.43770406e-01
-6.05690897e-01 -1.31255820e-01 1.32782608e-02 7.95290947e-01
-2.55694300e-01 -1.01263380e+00 -3.93965580e-02 3.14805955e-01
-3.11779231e-01 -1.44022137e-01 -5.16173005e-01 9.08704638e-01
-1.50981164e-02 4.47747439e-01 3.32897902e-01 -3.99173766e-01
4.33555722e-01 -1.77581306e-03 9.06637490e-01 -6.10206664e-01
-4.69051719e-01 -1.04151173e-02 5.64010404e-02 -5.16524851e-01
-2.31615186e-01 -7.11753964e-01 -9.58419859e-01 -1.73533246e-01
-4.36895400e-01 1.15831181e-01 2.56375790e-01 1.14612412e+00
4.90360677e-01 6.08791292e-01 3.76152158e-01 -8.97532880e-01
-6.10448956e-01 -8.03475797e-01 -2.67255247e-01 5.16504109e-01
4.33681816e-01 -1.05000222e+00 -4.20431435e-01 -2.28433669e-01] | [9.505839347839355, 1.7830580472946167] |
d86e3f4a-ffd5-4eb3-ae45-5ba2f9223e6d | fvor-robust-joint-shape-and-pose-optimization | 2205.07763 | null | https://arxiv.org/abs/2205.07763v1 | https://arxiv.org/pdf/2205.07763v1.pdf | FvOR: Robust Joint Shape and Pose Optimization for Few-view Object Reconstruction | Reconstructing an accurate 3D object model from a few image observations remains a challenging problem in computer vision. State-of-the-art approaches typically assume accurate camera poses as input, which could be difficult to obtain in realistic settings. In this paper, we present FvOR, a learning-based object reconstruction method that predicts accurate 3D models given a few images with noisy input poses. The core of our approach is a fast and robust multi-view reconstruction algorithm to jointly refine 3D geometry and camera pose estimation using learnable neural network modules. We provide a thorough benchmark of state-of-the-art approaches for this problem on ShapeNet. Our approach achieves best-in-class results. It is also two orders of magnitude faster than the recent optimization-based approach IDR. Our code is released at \url{https://github.com/zhenpeiyang/FvOR/} | ['QiXing Huang', 'Qi Shan', 'Zaiwei Zhang', 'Miguel Angel Bautista', 'Zhile Ren', 'Zhenpei Yang'] | 2022-05-16 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yang_FvOR_Robust_Joint_Shape_and_Pose_Optimization_for_Few-View_Object_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yang_FvOR_Robust_Joint_Shape_and_Pose_Optimization_for_Few-View_Object_CVPR_2022_paper.pdf | cvpr-2022-1 | ['object-reconstruction'] | ['computer-vision'] | [-1.94769144e-01 -5.35484590e-02 8.60079899e-02 -4.16064650e-01
-1.13246214e+00 -7.49678850e-01 4.07132506e-01 -5.15851676e-01
-6.63492605e-02 1.57591477e-01 1.32336663e-02 -3.85105647e-02
8.35425109e-02 -4.47335035e-01 -1.21245992e+00 -2.73540020e-01
4.63819712e-01 1.10139000e+00 2.27336988e-01 5.97827882e-02
1.48278981e-01 8.40646803e-01 -1.31366765e+00 3.64161544e-02
2.43279517e-01 9.70037520e-01 4.58281904e-01 8.77928376e-01
2.07869709e-01 5.11549592e-01 1.13548532e-01 -4.44447130e-01
6.00841582e-01 5.08669280e-02 -7.40761936e-01 4.96969104e-01
9.61147130e-01 -7.48135030e-01 -5.19011438e-01 1.06005418e+00
4.60011393e-01 -6.03513904e-02 4.94351536e-01 -9.65361476e-01
-6.23185277e-01 1.27464950e-01 -5.76559305e-01 -2.00687781e-01
4.23842400e-01 2.15819478e-01 9.46158051e-01 -1.41247141e+00
7.88752794e-01 1.14420974e+00 7.42869139e-01 4.60621566e-01
-1.28369868e+00 -4.20423090e-01 2.19651669e-01 1.24141842e-01
-1.56456220e+00 -6.21474862e-01 9.74998534e-01 -4.66901451e-01
9.13720787e-01 9.02428702e-02 7.42343843e-01 1.04433167e+00
-9.48228911e-02 8.80221486e-01 8.75475287e-01 -1.45334035e-01
-8.82783681e-02 -1.29005387e-01 -2.79072553e-01 8.31090868e-01
2.06757531e-01 1.90134615e-01 -4.45293427e-01 -1.39679462e-01
1.40121949e+00 3.22973639e-01 -2.88445115e-01 -8.38431835e-01
-1.29401779e+00 7.02738822e-01 5.59970498e-01 -2.78517842e-01
-4.63589638e-01 4.63398784e-01 -1.47462606e-01 -3.03901848e-03
6.37119591e-01 2.33176067e-01 -6.44053221e-01 -9.16281268e-02
-7.34842837e-01 4.00655776e-01 7.67159164e-01 1.20802808e+00
8.16047668e-01 8.02180842e-02 5.79693139e-01 7.15266109e-01
5.99954724e-01 6.02077663e-01 -1.63228273e-01 -1.38279760e+00
2.61539042e-01 3.48080844e-01 2.94651240e-01 -7.06581712e-01
-3.02583098e-01 -5.11499226e-01 -6.05557561e-01 2.86546856e-01
3.60746354e-01 1.56131700e-01 -8.43268871e-01 1.16725159e+00
6.13782823e-01 4.88182098e-01 -1.59814596e-01 1.21468318e+00
1.12352085e+00 4.92485940e-01 -6.24071658e-01 1.89509615e-01
9.55305517e-01 -1.23417187e+00 -1.42112195e-01 -5.25194228e-01
-6.80949911e-02 -1.02410495e+00 7.64950871e-01 5.34642041e-01
-1.40732968e+00 -4.12647843e-01 -7.81469762e-01 -3.13506126e-01
2.18971461e-01 3.11323792e-01 5.78185856e-01 1.71026915e-01
-1.13575101e+00 5.97710669e-01 -1.08126771e+00 -2.74205834e-01
5.80661952e-01 3.55948985e-01 -7.04708576e-01 -4.14085150e-01
-2.33662307e-01 8.90014172e-01 1.11475684e-01 1.70074761e-01
-1.34669733e+00 -7.95699120e-01 -9.34245408e-01 -3.05351287e-01
6.22697294e-01 -1.03817153e+00 1.66732919e+00 -6.59359038e-01
-1.60104012e+00 1.09421146e+00 -2.40217775e-01 -6.71721622e-02
6.63501561e-01 -5.88994920e-01 2.94064879e-01 1.45959243e-01
-1.62136972e-01 7.24132478e-01 9.05411243e-01 -1.76885438e+00
-1.88999727e-01 -5.98748386e-01 5.34897372e-02 4.05185163e-01
2.99512893e-01 -4.12238315e-02 -1.01353955e+00 -3.89312536e-01
5.48436821e-01 -1.06572580e+00 -4.97095346e-01 5.28305888e-01
-3.78717303e-01 1.01097226e-02 7.38960445e-01 -6.33060396e-01
2.81676859e-01 -1.78112662e+00 4.68618840e-01 -2.04978868e-01
2.98083574e-01 1.06087849e-01 -1.47255719e-01 2.52226532e-01
8.45237002e-02 -1.76155075e-01 -1.24993250e-02 -8.82268310e-01
-6.16098456e-02 2.28894740e-01 -2.34761506e-01 9.35820341e-01
4.84787859e-02 1.05594516e+00 -6.36729777e-01 -5.42799756e-02
5.78826725e-01 8.12280715e-01 -5.08091867e-01 5.00778377e-01
-4.45098221e-01 7.41231382e-01 -3.62118930e-01 8.06212783e-01
8.25444877e-01 -7.37919569e-01 -8.57051909e-02 -4.90600288e-01
1.65599640e-02 3.75083536e-02 -1.43702054e+00 2.16591144e+00
-3.65716338e-01 4.11985040e-01 2.47193187e-01 -7.44783938e-01
8.84151220e-01 2.13613346e-01 5.89586198e-01 1.33224751e-03
3.53520006e-01 1.86521828e-01 -4.87166554e-01 -3.14738244e-01
4.21399713e-01 -1.61923580e-02 2.47927427e-01 2.86528975e-01
1.69013798e-01 -7.35347450e-01 -3.80028993e-01 5.13368957e-02
9.18216646e-01 7.23816395e-01 5.55005968e-01 1.40030205e-01
2.62618661e-01 -4.94395271e-02 5.32814860e-01 4.56817627e-01
9.01288912e-02 1.26838040e+00 -1.59750700e-01 -7.90170372e-01
-1.33580148e+00 -1.19070315e+00 -2.62319185e-02 5.06874800e-01
3.10047239e-01 -3.86747569e-01 -4.49609160e-01 -4.91782993e-01
8.81344527e-02 3.93003881e-01 -3.34942967e-01 3.10311735e-01
-5.64492285e-01 -2.33442068e-01 7.22453818e-02 7.44979024e-01
2.82967985e-01 -7.10729718e-01 -4.45101887e-01 -2.15501934e-02
-1.59999177e-01 -1.48975837e+00 -5.77623785e-01 -1.25761643e-01
-1.04719627e+00 -1.11543512e+00 -7.22845256e-01 -6.38982177e-01
9.08697248e-01 6.40317738e-01 1.39625895e+00 1.82694897e-01
-2.20880002e-01 6.32725418e-01 -2.23239616e-01 -4.68465835e-01
-3.01778257e-01 -1.30662873e-01 1.48524866e-01 -1.51058212e-01
1.75130963e-01 -7.31720269e-01 -5.99215746e-01 3.69078577e-01
-7.20781863e-01 3.46341968e-01 5.29427290e-01 5.52048445e-01
1.12663817e+00 -4.47074473e-01 -8.79328996e-02 -6.81411326e-01
-2.01070070e-01 -2.61640668e-01 -1.10557818e+00 1.93984841e-03
-2.52368122e-01 -2.82298587e-02 4.70665127e-01 -3.51268202e-01
-8.12543690e-01 8.24093103e-01 -3.88480723e-01 -1.13758731e+00
-4.83323604e-01 1.33972853e-01 -1.66139841e-01 -3.85847390e-01
4.91347939e-01 2.71694690e-01 6.53930083e-02 -8.43994319e-01
4.34982449e-01 3.62464815e-01 5.30008674e-01 -4.63682532e-01
1.00371623e+00 8.22525203e-01 1.38455536e-02 -6.50313020e-01
-1.29603028e+00 -6.77895248e-01 -1.00774884e+00 -3.22293788e-01
7.37031996e-01 -1.30272079e+00 -5.75867534e-01 5.56825936e-01
-1.43266213e+00 -4.47415173e-01 -2.91626733e-02 5.13523757e-01
-9.86952424e-01 3.49957079e-01 -5.02565920e-01 -5.06731451e-01
-3.36080104e-01 -1.34685290e+00 1.61194956e+00 1.66794449e-01
1.53446808e-01 -8.50017965e-01 7.35250115e-02 7.28441119e-01
3.05694770e-02 2.50206202e-01 1.75760046e-01 -2.26049498e-01
-1.29116690e+00 -1.58198893e-01 -1.82239249e-01 2.41718024e-01
-3.22543196e-02 -6.29357025e-02 -9.23287153e-01 -5.38104236e-01
7.83078372e-03 -4.68928397e-01 6.63958192e-01 6.78906202e-01
1.29711223e+00 -1.58488348e-01 -2.14056447e-01 1.16509116e+00
1.82471645e+00 -2.98255265e-01 4.60836768e-01 1.15165450e-01
1.04967511e+00 1.18256897e-01 4.27049071e-01 5.41311085e-01
7.11937428e-01 9.38959837e-01 9.29610133e-01 1.21286616e-01
-2.68148124e-01 -3.53769004e-01 1.51861727e-01 7.34433949e-01
-3.31937551e-01 -1.18708871e-01 -9.09451485e-01 5.68142533e-01
-1.84490395e+00 -6.92193866e-01 -1.90935627e-01 2.02678895e+00
4.93412673e-01 -1.24969967e-01 -5.60929114e-03 -5.08770049e-01
5.19132435e-01 1.43152282e-01 -7.85845995e-01 1.51714206e-01
7.54574314e-02 -6.37136996e-02 5.75555563e-01 6.27956569e-01
-1.14926517e+00 1.07167602e+00 5.93667221e+00 4.47423697e-01
-9.10629928e-01 2.50499606e-01 4.57845390e-01 -2.42287248e-01
-2.55593896e-01 7.32801855e-02 -9.65635717e-01 -1.01045609e-01
4.84585136e-01 3.05884957e-01 7.49121308e-01 1.01680160e+00
1.14503302e-01 -2.92462185e-02 -1.26092362e+00 1.53330910e+00
5.25865197e-01 -1.71547568e+00 -1.44109502e-01 1.20530799e-01
9.46572185e-01 8.03885758e-01 -1.43631190e-01 -2.95159489e-01
5.27443349e-01 -1.00896192e+00 1.07540202e+00 6.47450805e-01
7.99157321e-01 -4.97980356e-01 4.07115459e-01 6.09143853e-01
-1.15089738e+00 2.30755627e-01 -5.62992275e-01 2.73519736e-02
5.56204557e-01 5.83140314e-01 -7.30974078e-01 6.25477314e-01
1.02692223e+00 1.01029754e+00 -5.70257843e-01 1.21610403e+00
-3.80046248e-01 3.99132192e-01 -5.50682306e-01 2.57151425e-01
2.28861743e-03 -2.71456391e-01 8.49584222e-01 7.05560327e-01
4.55103219e-01 2.17209816e-01 4.80660051e-01 1.01767993e+00
-2.89164543e-01 -3.33575308e-01 -7.91637480e-01 3.72203469e-01
3.76657933e-01 1.43412066e+00 -6.27027810e-01 -6.91628531e-02
-5.33213735e-01 1.00868940e+00 6.63253963e-01 1.06344648e-01
-8.14580977e-01 3.77988845e-01 5.34402430e-01 1.75963074e-01
6.61164522e-01 -6.45267069e-01 -2.83824265e-01 -1.35345507e+00
2.71899879e-01 -8.09434056e-01 5.23359329e-02 -1.13320506e+00
-1.39919603e+00 4.98678893e-01 1.24053191e-02 -1.36906302e+00
-1.01167336e-01 -8.35991859e-01 -2.86425233e-01 6.80224419e-01
-1.40701127e+00 -1.56717730e+00 -5.65000057e-01 4.77453768e-01
9.21125829e-01 -1.27803072e-01 7.67486334e-01 6.19228445e-02
-2.23875716e-01 7.00108707e-02 3.77451926e-02 1.15656890e-01
4.37493414e-01 -1.11586034e+00 7.08815396e-01 7.70846605e-01
5.55716693e-01 3.47691357e-01 5.91923296e-01 -5.32874346e-01
-1.98952115e+00 -1.19260323e+00 3.98498952e-01 -9.66019154e-01
3.49272430e-01 -3.71296674e-01 -7.51403511e-01 1.13783932e+00
8.79950300e-02 5.89249372e-01 2.55188733e-01 -2.22620621e-01
-4.07386690e-01 5.13450541e-02 -9.20022845e-01 3.80954593e-01
1.26374090e+00 -4.63267714e-01 -4.51553762e-01 4.66743141e-01
6.46627307e-01 -1.05729687e+00 -8.02616596e-01 3.71365577e-01
5.69595873e-01 -9.85753477e-01 1.26961064e+00 -3.87648553e-01
3.52037221e-01 -5.28417289e-01 -4.05017406e-01 -1.29693377e+00
-3.06906223e-01 -6.63061917e-01 -3.58025581e-01 7.12757111e-01
2.14695692e-01 -3.36814731e-01 9.86896276e-01 4.06285256e-01
-2.10814655e-01 -9.36696172e-01 -7.54602969e-01 -7.74191916e-01
-1.11682648e-02 -6.28327012e-01 4.30844277e-01 4.95710343e-01
-9.02104378e-01 2.94109106e-01 -5.52106261e-01 5.91894090e-01
1.01059783e+00 3.92843455e-01 1.20499432e+00 -1.20503628e+00
-5.46412289e-01 -4.02796604e-02 -3.91298980e-01 -1.43821359e+00
2.60196596e-01 -8.29995692e-01 1.69292301e-01 -1.83447576e+00
4.49165106e-01 -2.21808136e-01 3.89131665e-01 3.64592582e-01
1.09666601e-01 5.80444515e-01 3.75163227e-01 2.65965879e-01
-6.60975456e-01 5.63276052e-01 1.34033132e+00 5.18543124e-02
1.15477450e-01 2.00342074e-01 -5.37229955e-01 1.06859732e+00
6.93098664e-01 -4.75968391e-01 -1.98089808e-01 -1.03406286e+00
1.94387496e-01 1.62975177e-01 9.11531866e-01 -7.05084443e-01
2.89961785e-01 -2.80502647e-01 5.61298013e-01 -1.00371099e+00
9.08108652e-01 -9.58608747e-01 6.71701133e-01 1.15941815e-01
9.95164141e-02 2.64942646e-01 7.92072713e-02 4.97399151e-01
1.06585644e-01 -2.56952673e-01 8.68701875e-01 -6.11774266e-01
-7.54285932e-01 8.42958510e-01 1.63485318e-01 -6.62733242e-02
8.73552680e-01 -1.46838531e-01 -1.38454512e-01 -5.48884630e-01
-6.10837162e-01 1.98036600e-02 7.86996782e-01 5.52613258e-01
1.12801349e+00 -1.41928852e+00 -9.67702031e-01 1.32298037e-01
6.28577918e-02 6.84066594e-01 1.39195591e-01 7.05646992e-01
-8.38783443e-01 2.37331092e-01 3.37636508e-02 -1.05112565e+00
-1.49112606e+00 4.27036405e-01 5.67943990e-01 1.11373588e-01
-8.65886748e-01 9.64919209e-01 1.61267668e-01 -8.94542634e-01
1.75651163e-01 -1.31003007e-01 2.79423565e-01 -6.32655501e-01
4.18228477e-01 2.52408385e-01 8.90874043e-02 -9.28112090e-01
-3.63761336e-01 1.04523778e+00 1.36262849e-02 -9.44761485e-02
1.83401227e+00 -1.13680832e-01 -8.72303918e-02 2.97778934e-01
1.28557312e+00 -2.06194863e-01 -1.80037773e+00 -3.61747742e-01
-4.36169714e-01 -7.98389375e-01 3.16383392e-01 -5.27626753e-01
-1.19316649e+00 7.72422969e-01 4.02125418e-01 -4.11694497e-01
8.60133708e-01 5.72704315e-01 7.29213059e-01 4.49167490e-01
6.60821855e-01 -6.84537828e-01 2.95377344e-01 5.88612497e-01
1.29650486e+00 -1.57536352e+00 3.08912516e-01 -5.78087926e-01
-4.37717110e-01 1.11214185e+00 6.79964483e-01 -4.54375565e-01
7.81129718e-01 3.10953617e-01 8.90320241e-02 -4.09938455e-01
-5.78318834e-01 -4.79071178e-02 4.92780715e-01 5.01261353e-01
7.23784715e-02 -9.19783860e-02 5.35057366e-01 3.11272889e-01
-1.30823120e-01 -1.30544171e-01 5.54078758e-01 7.02459872e-01
-2.46115580e-01 -1.05007970e+00 -5.88950276e-01 2.11281344e-01
-3.59794378e-01 1.08077832e-01 -3.97190869e-01 6.13161802e-01
-1.86445579e-01 5.77713549e-01 -6.42988756e-02 -1.93469435e-01
3.92458826e-01 -3.41490775e-01 9.62151408e-01 -7.63178587e-01
-1.60395965e-01 3.66627276e-01 -8.19281489e-02 -8.61937523e-01
-6.47953749e-01 -9.61757243e-01 -1.06959331e+00 -2.23234698e-01
-3.73690337e-01 -5.25715470e-01 6.98523343e-01 8.19305897e-01
4.39776808e-01 2.82117631e-02 5.93297303e-01 -1.63575876e+00
-7.21537352e-01 -6.26729906e-01 -3.15861344e-01 3.19924772e-01
5.00952959e-01 -7.06841707e-01 -1.81918040e-01 1.65597647e-01] | [8.012419700622559, -2.655604362487793] |
7c306061-aa96-42f5-9282-7bb1f6fe5822 | real-time-human-detection-in-fire-scenarios | 2307.04223 | null | https://arxiv.org/abs/2307.04223v1 | https://arxiv.org/pdf/2307.04223v1.pdf | Real-time Human Detection in Fire Scenarios using Infrared and Thermal Imaging Fusion | Fire is considered one of the most serious threats to human lives which results in a high probability of fatalities. Those severe consequences stem from the heavy smoke emitted from a fire that mostly restricts the visibility of escaping victims and rescuing squad. In such hazardous circumstances, the use of a vision-based human detection system is able to improve the ability to save more lives. To this end, a thermal and infrared imaging fusion strategy based on multiple cameras for human detection in low-visibility scenarios caused by smoke is proposed in this paper. By processing with multiple cameras, vital information can be gathered to generate more useful features for human detection. Firstly, the cameras are calibrated using a Light Heating Chessboard. Afterward, the features extracted from the input images are merged prior to being passed through a lightweight deep neural network to perform the human detection task. The experiments conducted on an NVIDIA Jetson Nano computer demonstrated that the proposed method can process with reasonable speed and can achieve favorable performance with a mAP@0.5 of 95%. | ['My-Ha Le', 'Nghe-Nhan Truong', 'Truong-Dong Do'] | 2023-07-09 | null | null | null | null | ['human-detection'] | ['computer-vision'] | [ 5.51709294e-01 -5.56160629e-01 5.18192232e-01 8.60159025e-02
-1.71714097e-01 -4.16533917e-01 4.65830892e-01 -1.23057932e-01
-8.88982892e-01 4.45838332e-01 -2.88713068e-01 -5.87242767e-02
1.22525014e-01 -9.21458781e-01 -2.28391141e-01 -1.04779375e+00
4.01390076e-01 -2.52767324e-01 5.33803582e-01 5.81338480e-02
2.01193526e-01 5.24395585e-01 -1.75641882e+00 8.83476511e-02
6.53540611e-01 1.07830071e+00 5.68123162e-01 6.48298264e-01
2.49915093e-01 6.21387303e-01 -6.40284359e-01 -9.26861018e-02
5.72833061e-01 -1.16509229e-01 -4.61174063e-02 -1.57233894e-01
2.07186997e-01 -8.22273254e-01 -2.30302975e-01 1.04569900e+00
4.57681507e-01 3.07153106e-01 5.44149756e-01 -1.13296044e+00
-4.67922390e-02 -4.85365540e-02 -7.43709266e-01 3.68906140e-01
2.88380682e-01 5.79756081e-01 9.00018811e-02 -6.56817317e-01
-2.43132543e-02 1.21164358e+00 4.97754842e-01 5.55081308e-01
-5.14766157e-01 -8.83045137e-01 -5.27144015e-01 2.32236475e-01
-1.26536679e+00 -1.94960088e-01 5.97411394e-01 -3.79740775e-01
6.91045702e-01 4.12397534e-01 7.47594535e-01 9.93214846e-01
4.09388065e-01 6.34589717e-02 1.15899837e+00 -2.85500556e-01
7.46347904e-02 2.41487939e-03 -5.87834120e-02 7.58216202e-01
8.32400501e-01 4.03708607e-01 -1.22249313e-01 -1.01364031e-02
5.09268522e-01 5.22724450e-01 -3.02622974e-01 2.35070750e-01
-8.95236492e-01 6.64462268e-01 8.20850492e-01 1.57105654e-01
-4.53500777e-01 2.10910086e-02 2.22416088e-01 -4.22167689e-01
1.38940722e-01 -1.05725840e-01 2.28792086e-01 1.43760517e-01
-7.14123666e-01 -5.15082739e-02 2.60669857e-01 3.03291321e-01
4.49726969e-01 -3.05411946e-02 -3.19385044e-02 3.27031702e-01
4.00365680e-01 1.08928978e+00 6.61798101e-03 -8.55916560e-01
5.06456852e-01 7.36130118e-01 2.46734872e-01 -1.20056355e+00
-5.29632509e-01 -1.84774533e-01 -1.12904394e+00 9.28491652e-01
2.92895675e-01 -3.45735788e-01 -7.86012650e-01 1.14103806e+00
4.79630232e-01 1.22362204e-01 4.56961580e-02 1.26978528e+00
7.06049502e-01 8.98374557e-01 2.83965886e-01 -1.59984931e-01
1.58576095e+00 -5.89445412e-01 -4.42624986e-01 -4.28314924e-01
-1.47133216e-01 -6.20988846e-01 6.73848331e-01 5.50740182e-01
-4.58205462e-01 -8.29870760e-01 -1.21414578e+00 2.16675624e-01
-2.30290785e-01 2.00684890e-01 2.79588163e-01 7.93065310e-01
-4.78825510e-01 2.54617989e-01 -8.56901467e-01 -4.11663443e-01
3.78651351e-01 1.44133165e-01 -7.37282261e-02 -3.26094866e-01
-1.02237248e+00 1.03816926e+00 3.93671781e-01 5.98266661e-01
-1.00952828e+00 -3.34427685e-01 -4.50525135e-01 -1.00295857e-01
2.51935184e-01 -7.49929905e-01 8.36240232e-01 -5.36054790e-01
-9.72582936e-01 4.45442498e-01 1.82130978e-01 -2.23537460e-01
6.95535719e-01 -5.03437102e-01 -3.73148710e-01 4.51127410e-01
4.73908186e-02 2.83859402e-01 9.41649020e-01 -1.06347358e+00
-9.51992095e-01 -5.50789654e-01 -2.93961409e-02 3.93079162e-01
-4.14020330e-01 2.85675079e-01 -6.15194030e-02 -4.21680370e-03
-4.33075130e-01 -9.15954351e-01 -2.37805605e-01 1.43043250e-01
-2.80980200e-01 -9.34176072e-02 1.23137307e+00 -5.89481413e-01
8.49482358e-01 -2.04935217e+00 -2.67288059e-01 -5.52262366e-03
8.54824483e-02 8.25874388e-01 1.83172062e-01 1.96142092e-01
4.62330014e-01 -2.40354016e-01 -3.27677906e-01 -1.04418779e-02
-6.53705299e-01 -9.43662450e-02 -1.95965543e-02 5.74924290e-01
-7.87606537e-02 1.66516334e-01 -8.10331941e-01 -3.75948310e-01
6.21357203e-01 8.32568526e-01 1.34972125e-01 3.94387156e-01
2.38018751e-01 5.78498602e-01 -5.50489604e-01 3.73356968e-01
8.35776746e-01 2.75995463e-01 -3.38346750e-01 -1.81758076e-01
-6.20234549e-01 -4.12075430e-01 -1.05078197e+00 9.19566870e-01
-4.70985442e-01 4.44845408e-01 4.34609771e-01 -3.52983892e-01
9.35332060e-01 2.16985315e-01 3.24110240e-01 -6.97366178e-01
5.40161848e-01 -1.58985138e-01 -3.89190793e-01 -8.54337871e-01
2.40464315e-01 -2.13807166e-01 1.36759970e-02 3.55892062e-01
-6.60184443e-01 1.01756468e-01 -4.44340371e-02 2.88120750e-02
1.04649627e+00 -2.86779135e-01 5.85367121e-02 1.31083146e-01
5.29883385e-01 1.85399264e-01 5.79615176e-01 5.04665136e-01
-2.99689949e-01 4.42821622e-01 -4.09801960e-01 -6.07287943e-01
-9.10474241e-01 -1.23647463e+00 1.38398826e-01 7.11418927e-01
5.56313217e-01 5.06606065e-02 -8.87097776e-01 -3.70191604e-01
-3.17422628e-01 6.98171735e-01 -1.09882049e-01 -4.33617800e-01
-4.49278772e-01 -8.33112359e-01 5.17189264e-01 5.80886245e-01
1.19486463e+00 -1.19898260e+00 -1.76253891e+00 -7.53777940e-03
-2.70945787e-01 -1.03120160e+00 -8.41138810e-02 -1.62368372e-01
-4.95218843e-01 -1.37196136e+00 -5.43775678e-01 -6.35574400e-01
4.32125121e-01 9.85635400e-01 2.62802690e-01 2.35488459e-01
-9.27622318e-01 6.54391721e-02 -2.34560832e-01 -6.70792341e-01
-2.12015018e-01 -5.16575813e-01 1.19204216e-01 -3.01017575e-02
3.90644759e-01 -1.28941596e-01 -1.03607118e+00 -4.36635129e-03
-9.48918521e-01 1.25164598e-01 5.70243597e-01 1.90554231e-01
-8.08719471e-02 5.41835904e-01 1.06079303e-01 -3.24958593e-01
6.01191521e-01 -1.81519032e-01 -8.83182466e-01 9.78664979e-02
-2.15600953e-01 -4.05779153e-01 8.22072268e-01 -1.48392573e-01
-1.44548512e+00 2.85494447e-01 2.72991031e-01 -2.97975004e-01
-5.23340583e-01 -2.09014311e-01 -2.00564973e-02 6.51715025e-02
5.96907973e-01 6.04005828e-02 -1.94059648e-02 -2.19873160e-01
6.33458141e-03 9.66768384e-01 7.25215197e-01 2.57704973e-01
1.04548538e+00 8.20511520e-01 2.88665563e-01 -1.28419054e+00
-6.79668903e-01 -6.41878545e-01 -5.66372037e-01 -8.34624290e-01
1.43122351e+00 -9.31756854e-01 -1.07219481e+00 7.44735420e-01
-1.26254511e+00 2.23647356e-01 5.11290908e-01 5.57840228e-01
1.76147059e-01 6.10360742e-01 -5.11893630e-01 -1.26234841e+00
-9.76461709e-01 -8.89590979e-01 7.44576156e-01 8.45089555e-01
1.86077923e-01 -5.04333794e-01 9.39735174e-02 7.84607470e-01
4.94581670e-01 3.11706185e-01 3.70306641e-01 1.88366920e-01
-4.91325200e-01 -1.68625876e-01 -4.21523660e-01 5.34845054e-01
1.81624532e-01 1.69255152e-01 -1.01971626e+00 -2.72357285e-01
2.91670471e-01 -6.53141588e-02 1.08865929e+00 1.12042233e-01
6.44177079e-01 -1.04238593e-03 -5.14512718e-01 4.59596753e-01
1.56079507e+00 4.43692267e-01 5.68049729e-01 2.97392279e-01
8.03093314e-01 4.11193818e-01 7.26016402e-01 5.70442438e-01
-6.32561222e-02 2.62971371e-01 8.28356564e-01 -2.39373416e-01
-7.47989640e-02 1.75215080e-01 5.33527434e-01 1.98773742e-01
-4.60163087e-01 -2.44562119e-01 -8.12008679e-01 1.14446647e-01
-1.52108717e+00 -1.24423826e+00 -6.41053677e-01 2.25190949e+00
1.54779956e-01 1.29876696e-02 6.25882894e-02 1.26816913e-01
1.15283263e+00 -1.48730697e-02 -3.17801714e-01 -9.15173590e-02
3.67229640e-01 4.94628623e-02 6.67742610e-01 1.89932972e-01
-1.30741274e+00 4.89680916e-01 5.40839148e+00 2.87157506e-01
-1.04483759e+00 -1.13895908e-02 1.49990186e-01 -1.13234080e-01
3.89671296e-01 -1.52414992e-01 -7.09323585e-01 6.34719253e-01
6.22056127e-01 8.99860188e-02 4.65706557e-01 5.60485244e-01
5.62613428e-01 -5.65835476e-01 -3.12331200e-01 9.71378446e-01
1.30613297e-01 -5.27595520e-01 -1.00826152e-01 -1.57526284e-01
2.39412144e-01 -7.50775486e-02 -6.38450086e-02 -1.43037975e-01
1.46757305e-01 -8.26169908e-01 4.82844979e-01 6.70738637e-01
5.95205665e-01 -1.07876241e+00 6.86239541e-01 8.32223713e-01
-1.16695011e+00 -3.86230618e-01 -7.24110365e-01 -3.93828839e-01
3.30745727e-01 6.42005801e-01 -8.54918659e-01 3.76775980e-01
8.91581655e-01 8.85438025e-02 -5.24075747e-01 1.16921484e+00
-4.44581181e-01 3.53237480e-01 -5.32287240e-01 -2.76247323e-01
1.67257220e-01 -4.34647828e-01 4.24906522e-01 1.06147480e+00
5.58226347e-01 4.34655547e-01 2.64194012e-01 7.56283581e-01
2.50700027e-01 -4.23629433e-01 -7.74560452e-01 5.08707821e-01
1.85965806e-01 1.72230983e+00 -7.00399935e-01 -2.66620129e-01
-4.35668588e-01 9.80197668e-01 -9.28399190e-02 1.07201360e-01
-1.14938962e+00 -4.50546265e-01 1.86399192e-01 1.69053704e-01
3.72103625e-03 -3.83972168e-01 -5.16818650e-02 -8.28670800e-01
-8.04685652e-02 -1.77306578e-01 3.40728283e-01 -9.31270540e-01
-1.04880655e+00 6.80779815e-01 -9.20235589e-02 -1.06179667e+00
2.90624529e-01 -6.38479471e-01 -9.44841802e-01 7.94810295e-01
-1.14234686e+00 -9.26792085e-01 -1.07794702e+00 6.52796447e-01
4.40190792e-01 1.63911551e-01 5.62705815e-01 2.35380769e-01
-8.35079610e-01 -1.20176822e-01 -1.67575166e-01 1.62291214e-01
5.24084270e-01 -7.40305066e-01 -4.63371649e-02 1.27902019e+00
-3.17140847e-01 1.67309612e-01 5.09845793e-01 -8.82021308e-01
-1.24596560e+00 -1.28975451e+00 3.35965365e-01 -9.43556130e-02
1.15026549e-01 -1.13019325e-01 -5.67711234e-01 9.62092429e-02
4.04284626e-01 -1.37330309e-01 2.72858858e-01 -6.99249148e-01
4.49926453e-03 -3.54501128e-01 -1.29583192e+00 5.07476151e-01
6.59186363e-01 -2.36392617e-01 -6.18350506e-01 1.88898727e-01
1.96338981e-01 1.23541079e-01 -2.20567793e-01 3.05231005e-01
5.63706636e-01 -1.16579556e+00 9.43900287e-01 -2.82173660e-02
8.37999433e-02 -5.33929944e-01 1.47483826e-01 -9.61283505e-01
-3.66849542e-01 -1.32038966e-01 4.16895300e-01 1.14592946e+00
-4.52115238e-02 -3.68843794e-01 4.70382839e-01 5.57332635e-01
1.03302367e-01 5.06561548e-02 -9.44742978e-01 -5.04877090e-01
-4.88823086e-01 -2.03777090e-01 4.75035841e-03 2.91616321e-01
-1.59345433e-01 3.86430353e-01 -3.99533927e-01 6.79620385e-01
1.17655480e+00 -9.21141878e-02 5.42034268e-01 -1.28132474e+00
1.89244375e-01 1.20789170e-01 -1.52101174e-01 -2.96025544e-01
-3.90063196e-01 -5.74143410e-01 2.89976597e-01 -1.73681772e+00
5.99196613e-01 8.99940953e-02 -2.13551655e-01 3.69286269e-01
-3.52418333e-01 5.47137618e-01 4.46444482e-01 1.39271626e-02
-5.82889497e-01 3.82948965e-01 1.06424630e+00 -1.28223106e-01
1.15690380e-01 1.95399180e-01 -1.76065624e-01 9.12928581e-01
1.02928710e+00 -3.88223916e-01 -1.44051850e-01 -5.09799361e-01
-1.99440181e-01 1.13680296e-01 8.88120294e-01 -1.77006650e+00
5.63862264e-01 -2.18919009e-01 7.63712585e-01 -5.36287963e-01
4.38663751e-01 -1.20711684e+00 2.70178705e-01 1.01636541e+00
1.49101064e-01 -1.53012685e-02 3.33296582e-02 6.10052824e-01
2.51375467e-01 -2.28475079e-01 1.09319389e+00 -1.89128563e-01
-7.32446432e-01 -2.17572749e-02 -8.79579842e-01 -4.88887519e-01
1.57496262e+00 -1.44775316e-01 -6.12486243e-01 -6.23978823e-02
-1.82304978e-01 1.35799319e-01 3.01474750e-01 3.69806051e-01
9.13451135e-01 -1.00359666e+00 -5.02135873e-01 2.53327757e-01
-1.45496383e-01 -6.99555352e-02 4.40076560e-01 5.54807842e-01
-8.48967552e-01 2.01459691e-01 -4.17638719e-01 -5.01345813e-01
-1.70499671e+00 7.40800202e-01 1.58534661e-01 2.87738681e-01
-7.98538744e-01 3.98692966e-01 2.06978247e-02 2.35349461e-01
8.47898945e-02 -1.61893874e-01 -3.12806129e-01 -8.12024251e-02
7.91114151e-01 9.63944495e-01 -1.01836033e-01 -6.90277219e-01
-6.24657214e-01 6.59439981e-01 3.17224562e-01 -1.58227403e-02
1.23359621e+00 -1.14110872e-01 -2.00393889e-03 -8.09179321e-02
7.88699508e-01 -7.97221437e-02 -1.38327563e+00 2.94106603e-01
-5.73768079e-01 -4.31000054e-01 2.51348138e-01 -7.87765801e-01
-1.13139653e+00 9.83754575e-01 9.72310960e-01 8.90257210e-02
1.32916141e+00 -2.88897932e-01 9.73812282e-01 3.65698755e-01
4.56163883e-01 -9.34466124e-01 9.00872424e-02 9.00043696e-02
4.74385202e-01 -1.05755639e+00 -1.63367558e-02 -4.41325337e-01
-5.35501182e-01 1.35236728e+00 7.52439618e-01 -1.50529638e-01
2.11377785e-01 4.81131852e-01 3.60759109e-01 -6.42440915e-02
-2.31588215e-01 -3.49839687e-01 -1.20981254e-01 6.69434786e-01
-1.52346119e-01 9.74406675e-02 -1.14579216e-01 3.27733576e-01
2.58538932e-01 -8.06223005e-02 5.03709674e-01 9.66095746e-01
-1.13848484e+00 -4.64458972e-01 -1.00475740e+00 2.94012308e-01
-3.95775288e-01 2.87844568e-01 -3.11085641e-01 5.43629587e-01
5.24786472e-01 1.65748811e+00 -6.96679801e-02 -4.86066759e-01
3.43016595e-01 -2.08027437e-01 3.47110480e-01 -3.44788194e-01
-5.66329956e-01 -2.10489944e-01 -2.13830128e-01 -5.16410291e-01
-4.52700227e-01 -3.80613357e-01 -1.35827720e+00 -2.59035856e-01
-2.13809788e-01 -1.05283484e-01 6.87286973e-01 8.43120158e-01
-2.48089787e-02 4.40456569e-01 6.54059947e-01 -9.17192459e-01
-4.63608325e-01 -8.27732801e-01 -5.06672025e-01 5.26175141e-01
2.07329795e-01 -6.46238625e-01 -6.23545706e-01 -2.00840145e-01] | [9.053510665893555, -1.0377546548843384] |
dd344407-e668-4c1f-b0c7-58fc805d9f71 | neural-architectures-for-nested-ner-through-1 | 1908.06926 | null | https://arxiv.org/abs/1908.06926v1 | https://arxiv.org/pdf/1908.06926v1.pdf | Neural Architectures for Nested NER through Linearization | We propose two neural network architectures for nested named entity recognition (NER), a setting in which named entities may overlap and also be labeled with more than one label. We encode the nested labels using a linearized scheme. In our first proposed approach, the nested labels are modeled as multilabels corresponding to the Cartesian product of the nested labels in a standard LSTM-CRF architecture. In the second one, the nested NER is viewed as a sequence-to-sequence problem, in which the input sequence consists of the tokens and output sequence of the labels, using hard attention on the word whose label is being predicted. The proposed methods outperform the nested NER state of the art on four corpora: ACE-2004, ACE-2005, GENIA and Czech CNEC. We also enrich our architectures with the recently published contextual embeddings: ELMo, BERT and Flair, reaching further improvements for the four nested entity corpora. In addition, we report flat NER state-of-the-art results for CoNLL-2002 Dutch and Spanish and for CoNLL-2003 English. | ['Jan Hajič', 'Jana Straková', 'Milan Straka'] | 2019-08-19 | neural-architectures-for-nested-ner-through | https://aclanthology.org/P19-1527 | https://aclanthology.org/P19-1527.pdf | acl-2019-7 | ['hard-attention', 'nested-named-entity-recognition', 'nested-mention-recognition'] | ['methodology', 'natural-language-processing', 'natural-language-processing'] | [-1.17005862e-01 3.52254808e-01 -2.20061764e-02 -6.04158640e-01
-6.18948340e-01 -9.07762587e-01 5.92192590e-01 4.70755965e-01
-1.18876827e+00 9.46488857e-01 4.52681363e-01 -4.42369640e-01
2.93461949e-01 -6.21931791e-01 -6.02086306e-01 -4.64894831e-01
-2.52739966e-01 7.32214034e-01 -2.12611943e-01 1.33386970e-01
-1.51555255e-01 6.66516602e-01 -8.90433550e-01 2.73626834e-01
6.56088233e-01 7.12834597e-01 4.81419340e-02 7.04701722e-01
-3.74938041e-01 9.32748854e-01 -4.40525651e-01 -7.15226650e-01
6.40902519e-02 1.55308621e-03 -1.17236388e+00 -5.20024359e-01
4.27229702e-01 1.26594126e-01 -2.08060086e-01 9.36508179e-01
4.63671565e-01 4.66109097e-01 6.28716946e-01 -8.60538125e-01
-9.39903736e-01 1.03998888e+00 -2.89832018e-02 -5.94792999e-02
1.05573414e-02 -5.04610300e-01 1.41927183e+00 -1.05351794e+00
9.43596125e-01 1.22613454e+00 9.36467588e-01 6.54645264e-01
-1.19584060e+00 -4.02387768e-01 1.66872099e-01 7.04656467e-02
-1.38752317e+00 -2.66401112e-01 9.81633291e-02 -4.28218335e-01
1.40642321e+00 -3.88522446e-02 3.84573042e-02 8.76958370e-01
-1.35189161e-01 6.44523919e-01 9.04963434e-01 -7.35354662e-01
2.59111404e-01 3.25876810e-02 4.79892164e-01 6.13345861e-01
-8.33745822e-02 2.13741049e-01 2.38860250e-02 -1.57972470e-01
4.08456177e-01 -1.85854703e-01 1.06248647e-01 -3.60688604e-02
-1.28616166e+00 9.58700120e-01 5.93473673e-01 6.60079420e-01
-4.93422538e-01 1.42016470e-01 6.12618625e-01 3.07948850e-02
5.27204514e-01 6.17452264e-01 -1.12360585e+00 -1.01447548e-03
-8.88106763e-01 -1.58332542e-01 1.30186379e+00 1.14662468e+00
9.01521385e-01 6.05578907e-03 -2.85413980e-01 8.73142540e-01
2.60193020e-01 1.38136983e-01 4.40060049e-01 -5.46567261e-01
5.45677304e-01 2.55048156e-01 3.38426262e-01 -4.81956720e-01
-7.72493899e-01 -2.38492370e-01 -7.66445577e-01 -1.48964494e-01
3.09748292e-01 -6.78313434e-01 -1.19247270e+00 2.00490737e+00
2.43410990e-01 5.40437818e-01 4.92684901e-01 4.80202734e-01
1.03564274e+00 8.35021436e-01 7.16918945e-01 8.97364765e-02
1.56511629e+00 -1.38021863e+00 -8.25112224e-01 -2.07156464e-01
1.08664727e+00 -4.36276019e-01 4.91034895e-01 -8.05447027e-02
-6.84497118e-01 -5.90976357e-01 -7.00499117e-01 -3.99001896e-01
-1.09742141e+00 6.74168050e-01 1.90457955e-01 4.95130599e-01
-1.11160707e+00 7.46135294e-01 -7.71965861e-01 -4.35022295e-01
-1.53793752e-01 3.85026306e-01 -6.32297397e-01 1.06919236e-01
-1.66734171e+00 1.21984506e+00 9.25899923e-01 4.66137350e-01
-6.19262755e-01 -5.58822989e-01 -1.23405457e+00 3.21207047e-01
-4.55116807e-03 -2.58459419e-01 1.35125506e+00 -4.05519724e-01
-1.18491924e+00 1.12838721e+00 -2.30377182e-01 -6.07837379e-01
1.15609452e-01 -4.31727618e-01 -8.57065260e-01 -2.71742195e-01
2.81007081e-01 1.04957366e+00 -1.31011695e-01 -1.26984560e+00
-6.91453218e-01 -5.60265742e-02 5.63565455e-02 6.54413998e-02
-7.34672621e-02 2.77261674e-01 -3.06513421e-02 -4.43512172e-01
-3.82262409e-01 -1.00288379e+00 -6.17693961e-01 -7.49781728e-01
-7.32014835e-01 -8.03843379e-01 3.68867666e-01 -8.71709585e-01
1.31801939e+00 -2.23490024e+00 -2.68049687e-02 -1.78001136e-01
-7.66610056e-02 4.52205211e-01 -4.86604720e-01 6.87380552e-01
-4.20118272e-01 7.15595305e-01 -2.60094523e-01 -7.27032602e-01
4.84494805e-01 4.79340583e-01 -2.10942954e-01 4.29998517e-01
5.63950479e-01 9.39392567e-01 -7.79116511e-01 -5.18186808e-01
-1.94573998e-02 5.86233199e-01 -1.19192056e-01 4.99493778e-01
-7.50666559e-02 2.85632521e-01 -5.14363172e-03 2.92320132e-01
6.07878923e-01 -6.66046962e-02 3.29012811e-01 -1.24316268e-01
-6.45043969e-01 6.11057997e-01 -1.21303165e+00 1.53447258e+00
-6.16585612e-01 3.92573476e-01 -1.17475279e-01 -7.36239254e-01
9.98849511e-01 8.07460904e-01 5.61079457e-02 -3.40784192e-01
-2.12194938e-02 3.60934377e-01 -2.90948272e-01 -2.96747386e-01
8.93387198e-01 -3.43363792e-01 -5.95418930e-01 1.08242385e-01
4.95758623e-01 3.50643784e-01 5.15440702e-01 -2.35970959e-01
1.00421751e+00 2.60963261e-01 7.41553724e-01 -1.50123969e-01
4.06138778e-01 -5.51087335e-02 9.72761869e-01 7.89690316e-01
-1.90224677e-01 2.92903751e-01 5.24090588e-01 -4.95204300e-01
-1.18542016e+00 -9.50724721e-01 -3.72344136e-01 1.46653664e+00
-3.36837053e-01 -3.68226171e-01 -7.15058208e-01 -1.01479423e+00
-3.13568681e-01 1.08245516e+00 -8.15556943e-01 4.12074834e-01
-1.10990727e+00 -1.64366663e-01 1.17069137e+00 6.32284105e-01
2.99826473e-01 -1.78505933e+00 -3.05861354e-01 5.08760870e-01
-1.52317211e-01 -1.51402175e+00 -6.54278636e-01 1.10262120e+00
-5.88835478e-01 -6.56701326e-01 -5.58385909e-01 -1.44041622e+00
2.91509867e-01 -7.50857115e-01 1.34888256e+00 -3.03205699e-01
8.68984535e-02 -1.67918086e-01 -5.21328032e-01 -1.11491591e-01
-3.19817752e-01 5.33592522e-01 -8.18542764e-02 -2.58062810e-01
4.58972216e-01 -2.92335570e-01 2.66486909e-02 -2.64620353e-02
-9.32436109e-01 -4.66723323e-01 5.85287094e-01 1.12030327e+00
6.40350342e-01 -4.24317569e-01 5.58918118e-01 -1.27671254e+00
2.82977939e-01 -6.95719004e-01 -3.19069237e-01 3.79891604e-01
-1.16117321e-01 3.37352902e-01 1.06138217e+00 -4.69019830e-01
-1.23627102e+00 3.24427873e-01 -6.39845669e-01 -3.69616672e-02
-9.20928121e-01 6.76369011e-01 -3.66432399e-01 5.67412615e-01
4.73492682e-01 -1.75662294e-01 -1.16515136e+00 -8.22626054e-01
9.88330007e-01 7.38170743e-01 6.76692307e-01 -6.07660532e-01
3.19169074e-01 -1.73833836e-02 -2.09799781e-01 -7.77755857e-01
-1.05776393e+00 -5.51825047e-01 -1.20401752e+00 3.94918829e-01
1.47759402e+00 -8.35462153e-01 -5.12141585e-01 1.94768548e-01
-1.84779978e+00 -2.75362372e-01 -5.36308289e-01 5.88072777e-01
-3.04725617e-01 2.80555874e-01 -1.25498402e+00 -6.11253619e-01
-2.91933745e-01 -9.61666226e-01 1.04029751e+00 4.59185272e-01
-1.90846413e-01 -1.42886281e+00 5.65453112e-01 -3.72186899e-01
1.90361038e-01 1.65525123e-01 1.28548753e+00 -1.57713461e+00
9.14500728e-02 -1.32053658e-01 -3.33415747e-01 4.04775441e-01
-2.62963295e-01 -3.78571868e-01 -9.88978982e-01 -1.58547506e-01
-3.40481490e-01 -4.15634334e-01 7.42147386e-01 2.07165964e-02
3.88798714e-01 -2.45095089e-01 -3.74829471e-01 4.93075043e-01
1.62440956e+00 4.29438561e-01 4.72378463e-01 4.22251463e-01
8.38950753e-01 6.48102462e-01 2.62140304e-01 2.07786009e-01
4.56231505e-01 3.20109963e-01 2.43602112e-01 -3.11549604e-01
9.98061821e-02 -3.40866685e-01 2.92748600e-01 1.17337322e+00
3.23075712e-01 -6.09235823e-01 -1.05727816e+00 8.91419590e-01
-1.63469851e+00 -6.49888277e-01 -2.64226466e-01 1.85721540e+00
7.74283946e-01 -3.18555802e-01 -3.08678806e-01 -5.49326181e-01
1.15779960e+00 2.84760922e-01 -3.13596964e-01 -9.13510382e-01
-1.50369957e-01 6.17232144e-01 4.76287484e-01 4.47120994e-01
-1.64211702e+00 1.38236535e+00 5.90542698e+00 5.29873073e-01
-7.11256087e-01 4.48094219e-01 4.25105065e-01 6.78906918e-01
-1.98295973e-02 2.04571158e-01 -1.27102125e+00 1.13726333e-01
1.65339434e+00 4.06285465e-01 7.34987780e-02 7.90489197e-01
-2.01164916e-01 2.67127484e-01 -1.46994674e+00 4.44237202e-01
-1.48373634e-01 -9.69040036e-01 -3.34077418e-01 8.93889144e-02
8.38481128e-01 4.63265032e-01 -4.43501711e-01 8.11062992e-01
9.02405679e-01 -1.17854559e+00 6.51574850e-01 5.12996137e-01
8.91642392e-01 -6.55920923e-01 1.20307899e+00 2.33690828e-01
-1.51896465e+00 1.10254996e-01 -4.57679749e-01 2.64153600e-01
5.91758966e-01 4.52917457e-01 -7.45035112e-01 7.41952837e-01
4.38584954e-01 5.35600245e-01 -2.96356589e-01 1.01112270e+00
-6.47077978e-01 7.62271941e-01 -2.66344398e-01 -1.99030325e-01
6.56727970e-01 -1.26815513e-01 2.68143922e-01 2.08595133e+00
2.29992732e-01 -7.47762099e-02 3.62413675e-01 7.02940226e-01
-5.39403021e-01 2.11513340e-01 -4.16121155e-01 -2.50042081e-01
6.48476899e-01 1.39263237e+00 -5.45822382e-01 -4.16437387e-01
-5.19417286e-01 1.08630323e+00 1.14609182e+00 3.92494470e-01
-7.02349603e-01 -8.82023096e-01 4.22978044e-01 -6.85909808e-01
7.42958784e-01 -3.54092419e-01 -1.08111873e-01 -1.16678607e+00
-2.49441147e-01 -3.32238793e-01 5.49859107e-01 -4.79313731e-01
-1.66973698e+00 1.19391930e+00 -3.60955775e-01 -6.73335969e-01
-2.63112813e-01 -8.45399916e-01 -3.94777507e-01 9.84471619e-01
-1.57212818e+00 -1.18655670e+00 4.30097312e-01 1.61116213e-01
1.85117751e-01 1.27266780e-01 1.47691572e+00 5.69011092e-01
-5.19322336e-01 4.18958098e-01 4.15154338e-01 8.32950532e-01
7.64509737e-01 -1.58098686e+00 9.34764326e-01 7.28993297e-01
5.58397174e-01 4.71379042e-01 2.52107084e-01 -5.58533251e-01
-4.53781337e-01 -1.30707943e+00 1.98802054e+00 -2.09361464e-01
9.13575828e-01 -7.03159690e-01 -6.98552549e-01 1.12817967e+00
7.12950826e-01 2.00908184e-01 7.82463849e-01 3.42867613e-01
-5.50198138e-01 4.73890036e-01 -1.04928207e+00 3.56267422e-01
8.47811103e-01 -7.08218515e-01 -1.06028581e+00 1.39774278e-01
1.20102859e+00 -2.31167287e-01 -1.21034980e+00 1.60865948e-01
1.82090014e-01 -4.84755725e-01 5.33285320e-01 -1.08960795e+00
2.59119272e-01 -2.31584013e-01 -4.14217889e-01 -1.57219565e+00
-5.33196032e-01 -1.39257982e-01 3.13329756e-01 1.73025990e+00
8.94655943e-01 -4.28981543e-01 4.08107519e-01 3.71194839e-01
-4.24740285e-01 -5.17757893e-01 -1.12575018e+00 -7.22386122e-01
5.21765292e-01 -2.79654473e-01 6.44665480e-01 1.24850690e+00
-1.00958049e-01 5.12699008e-01 -3.23530555e-01 3.41762811e-01
1.88275367e-01 -3.46138686e-01 -9.56817716e-02 -1.32126129e+00
2.00172309e-02 4.90414649e-02 -2.92953342e-01 -1.04955840e+00
8.83213401e-01 -1.15950155e+00 6.38474762e-01 -1.73945057e+00
-4.09823842e-02 -6.24336004e-01 -5.68711817e-01 9.64292288e-01
4.14473051e-03 -2.61527058e-02 3.67379576e-01 -1.76663190e-01
-7.49837935e-01 4.43378776e-01 5.05631208e-01 9.26213115e-02
-9.19620544e-02 -4.73833144e-01 -2.37754986e-01 6.09286249e-01
6.23730540e-01 -9.61256802e-01 6.25298083e-01 -6.40358984e-01
2.18064100e-01 2.34195232e-01 -3.47284637e-02 -8.42758417e-01
4.07414258e-01 1.41470909e-01 1.77271858e-01 -6.92116439e-01
1.47430331e-01 -9.13888097e-01 2.48767622e-02 1.42601416e-01
-7.20895410e-01 3.90272707e-01 1.19507112e-01 3.24593663e-01
-4.16054606e-01 -7.21126854e-01 6.74961030e-01 -1.36840537e-01
-8.39509606e-01 1.35518655e-01 -5.44458508e-01 4.92732793e-01
7.20395267e-01 3.65708351e-01 -3.58485341e-01 2.29041591e-01
-1.31893265e+00 2.42480963e-01 -7.41033703e-02 3.39198738e-01
1.65372208e-01 -1.43392742e+00 -7.91010320e-01 -1.00083627e-01
-3.43391038e-02 -7.34825432e-02 1.40735000e-01 3.05238217e-01
-3.57999951e-01 6.42833710e-01 -5.47652170e-02 -1.02198087e-01
-8.33234429e-01 6.28389359e-01 3.28096032e-01 -1.03553557e+00
-2.13708311e-01 8.48635375e-01 1.07073516e-01 -1.46657848e+00
2.32652798e-01 -3.45504999e-01 -5.18702924e-01 3.08778435e-01
1.79198101e-01 3.96210074e-01 1.14030875e-01 -1.06096220e+00
-4.79381323e-01 3.24797839e-01 -1.20047428e-01 -1.51746243e-01
1.51896417e+00 2.09760815e-02 -3.75259012e-01 7.50318587e-01
1.57307339e+00 5.92489317e-02 -7.01249838e-01 -2.79166251e-01
6.84767962e-01 5.25021493e-01 -3.20633084e-01 -9.13518667e-01
-6.90794349e-01 9.42527473e-01 5.67456663e-01 1.09401464e-01
6.46636307e-01 1.02330573e-01 7.77922392e-01 6.27775550e-01
1.85025215e-01 -9.37993884e-01 -6.33549988e-01 1.33979547e+00
2.61486799e-01 -8.59714031e-01 -6.63199663e-01 -7.55200014e-02
-6.98976398e-01 1.16330826e+00 5.43817937e-01 -2.09438547e-01
7.01566935e-01 3.94513458e-01 2.03794986e-01 2.12515984e-02
-5.08559406e-01 -4.32690799e-01 1.50003225e-01 3.97067100e-01
7.84049153e-01 3.85679603e-01 -4.95505661e-01 8.50618005e-01
-3.60940434e-02 -2.21347943e-01 4.34339106e-01 7.37496674e-01
-2.26806760e-01 -1.33086503e+00 -8.51126015e-02 1.22772805e-01
-7.29139328e-01 -7.24338472e-01 -2.50905156e-01 9.47295427e-01
6.28683925e-01 1.09004903e+00 1.70793071e-01 -1.96648911e-01
6.16897881e-01 7.14222372e-01 -1.68661326e-01 -9.80007708e-01
-1.11023712e+00 -1.73253983e-01 5.46868026e-01 -1.83411673e-01
-4.37904328e-01 -5.58135211e-01 -1.66247022e+00 2.35750139e-01
-4.80657846e-01 7.33322859e-01 6.65283859e-01 1.11093199e+00
3.05695772e-01 4.28969622e-01 3.70341688e-01 -7.75429666e-01
-5.37221491e-01 -1.24867249e+00 -8.53124678e-01 3.92372757e-01
2.11663783e-01 -3.85665178e-01 -3.72373372e-01 3.78611982e-02] | [9.809535026550293, 9.723176956176758] |
f3d60001-4530-4191-8f3c-c319fb9c8cd5 | a-unified-multi-view-multi-person-tracking | 2302.03820 | null | https://arxiv.org/abs/2302.03820v1 | https://arxiv.org/pdf/2302.03820v1.pdf | A Unified Multi-view Multi-person Tracking Framework | Although there is a significant development in 3D Multi-view Multi-person Tracking (3D MM-Tracking), current 3D MM-Tracking frameworks are designed separately for footprint and pose tracking. Specifically, frameworks designed for footprint tracking cannot be utilized in 3D pose tracking, because they directly obtain 3D positions on the ground plane with a homography projection, which is inapplicable to 3D poses above the ground. In contrast, frameworks designed for pose tracking generally isolate multi-view and multi-frame associations and may not be robust to footprint tracking, since footprint tracking utilizes fewer key points than pose tracking, which weakens multi-view association cues in a single frame. This study presents a Unified Multi-view Multi-person Tracking framework to bridge the gap between footprint tracking and pose tracking. Without additional modifications, the framework can adopt monocular 2D bounding boxes and 2D poses as the input to produce robust 3D trajectories for multiple persons. Importantly, multi-frame and multi-view information are jointly employed to improve the performance of association and triangulation. The effectiveness of our framework is verified by accomplishing state-of-the-art performance on the Campus and Shelf datasets for 3D pose tracking, and by comparable results on the WILDTRACK and MMPTRACK datasets for 3D footprint tracking. | ['Shan Jiang', 'Shoichi Masui', 'Hiroaki Fujimoto', 'Sosuke Yamao', 'Shigeyuki Odashima', 'Fan Yang'] | 2023-02-08 | null | null | null | null | ['multiple-people-tracking', 'pose-tracking', '3d-multi-person-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-3.59385580e-01 -4.93077278e-01 -2.59363085e-01 3.75197679e-02
-5.76672912e-01 -7.93798685e-01 3.32083523e-01 -2.42579356e-01
-2.42958590e-01 4.30735797e-01 1.48298085e-01 1.69128194e-01
3.11680771e-02 -7.43285120e-01 -5.17525136e-01 -3.95100296e-01
2.42900640e-01 5.45355916e-01 4.11609650e-01 -1.74504489e-01
-2.89995581e-01 7.24834919e-01 -1.44832253e+00 -2.10581675e-01
3.93607527e-01 4.09477413e-01 -2.67096698e-01 6.63259625e-01
1.96869344e-01 -3.06286633e-01 -5.20592690e-01 -4.60467458e-01
4.85323578e-01 3.49088460e-02 2.68178768e-02 2.66647786e-01
1.46340430e+00 -6.47975028e-01 -3.98654401e-01 8.13841164e-01
8.29616129e-01 -9.93859097e-02 2.93225378e-01 -1.30873561e+00
-4.04433459e-01 -2.67046511e-01 -8.80845547e-01 2.62658950e-02
1.06715024e+00 -9.60113928e-02 4.78666395e-01 -1.17716920e+00
7.67162025e-01 1.47866654e+00 1.36538959e+00 5.98267674e-01
-1.04009640e+00 -8.03619325e-01 5.11486948e-01 -4.57762241e-01
-1.65610003e+00 -2.36048669e-01 4.63481963e-01 -5.33779502e-01
6.28721774e-01 2.92918891e-01 1.15537989e+00 1.27481782e+00
1.54227749e-01 6.67535722e-01 9.40583110e-01 -1.11410618e-01
-5.31176507e-01 -3.63044515e-02 1.70659810e-01 8.87237430e-01
7.69784153e-01 4.76432353e-01 -6.92951679e-01 -2.38320023e-01
1.15595841e+00 5.07757366e-01 -1.89779639e-01 -8.91796947e-01
-1.41365623e+00 4.03259248e-01 3.17311078e-01 -1.33849800e-01
-9.98611748e-02 4.86908667e-02 1.49600148e-01 -1.19297363e-01
4.88019824e-01 -2.38356054e-01 -1.87141359e-01 -1.01033881e-01
-1.04884207e+00 6.21589780e-01 3.18365008e-01 1.64058113e+00
4.43546176e-01 1.88159719e-02 -2.77573168e-01 3.37499082e-01
7.02596307e-01 1.41475308e+00 -3.61897588e-01 -7.17816353e-01
6.35681510e-01 8.51651013e-01 4.67314512e-01 -1.13315320e+00
-7.10975528e-01 -5.79690456e-01 -4.14448500e-01 1.80029962e-03
7.09989965e-01 -2.89982378e-01 -7.05227375e-01 1.55791771e+00
8.33495259e-01 1.22831367e-01 -4.84807312e-01 1.02809775e+00
9.48324859e-01 6.26600906e-02 -2.30267853e-01 -1.39817759e-01
1.45114791e+00 -7.94978857e-01 -7.67682254e-01 -1.74582884e-01
4.73011762e-01 -7.82744050e-01 6.34086847e-01 8.56220350e-02
-1.10750341e+00 -9.13805664e-01 -9.00436044e-01 1.90949023e-01
-3.23166221e-01 2.22980544e-01 2.20885485e-01 1.06375182e+00
-6.47350132e-01 -1.86114199e-02 -7.77958810e-01 -6.56165123e-01
1.47041366e-01 3.53977084e-01 -4.08431262e-01 -1.56225339e-01
-9.70343828e-01 9.86854196e-01 3.64646465e-02 2.87375122e-01
-5.84169745e-01 -9.29437637e-01 -8.00292969e-01 -4.69985723e-01
4.98085558e-01 -1.11318815e+00 7.57144809e-01 2.05353141e-01
-1.00634539e+00 9.47135031e-01 -3.69290859e-01 1.27134904e-01
9.13807750e-01 -8.33695292e-01 -6.75267339e-01 5.00278026e-02
2.41092876e-01 3.55879158e-01 7.32003510e-01 -1.48520184e+00
-5.84268928e-01 -8.42162490e-01 2.40787282e-03 2.99673349e-01
-1.74880505e-01 6.84046559e-03 -9.21104550e-01 -6.96796715e-01
3.71411711e-01 -1.34323883e+00 -3.69497985e-02 3.36882621e-01
-4.13270950e-01 2.24308193e-01 1.20466816e+00 -5.24250507e-01
1.26104558e+00 -1.90602493e+00 2.35045865e-01 1.32485092e-01
1.91464961e-01 2.24795923e-01 2.92534709e-01 2.77162760e-01
3.65998596e-01 -1.72355503e-01 5.75267971e-01 -7.12829351e-01
1.36343002e-01 8.48078169e-03 7.76315928e-02 8.78960967e-01
-4.02355254e-01 8.33726764e-01 -1.03486848e+00 -6.23948455e-01
4.89707291e-01 7.38825142e-01 -3.93148094e-01 -3.22188474e-02
2.36906633e-01 5.52212596e-01 -4.67563689e-01 1.07411492e+00
9.42816854e-01 -2.71130741e-01 3.18676792e-02 -4.16641533e-01
-3.18173170e-01 -3.75246316e-01 -1.72532082e+00 1.88621044e+00
1.88234635e-02 1.05429910e-01 1.07149303e-01 1.78337410e-01
8.26641619e-01 4.76031423e-01 8.61929476e-01 -1.86212882e-01
-4.98704205e-04 3.48382629e-02 -5.84827721e-01 -2.29209382e-02
6.33414209e-01 7.22092986e-02 -2.44931772e-01 3.23650956e-01
-1.85463563e-01 4.54291940e-01 6.38777688e-02 5.02349623e-02
4.49044287e-01 7.30725348e-01 2.49225542e-01 -2.86403876e-02
5.02384722e-01 -1.28035331e-02 7.87903786e-01 6.82891488e-01
-4.17815983e-01 7.58411825e-01 -4.88803148e-01 -5.12966871e-01
-7.95353830e-01 -1.57934892e+00 -2.15225175e-01 8.64038229e-01
5.61503291e-01 -7.68365383e-01 -4.16522622e-01 -7.92053640e-01
5.56995392e-01 -7.97062144e-02 -4.26389456e-01 1.95796609e-01
-8.19845080e-01 -5.48799694e-01 6.56710505e-01 6.97070539e-01
5.47742486e-01 -1.04235739e-01 -6.25835299e-01 6.93530515e-02
-2.20784485e-01 -1.40713549e+00 -1.05089939e+00 -6.38193130e-01
-9.60453272e-01 -1.32538652e+00 -9.19437408e-01 -2.68968642e-01
7.79438972e-01 9.47174668e-01 8.45428288e-01 1.11652710e-01
7.60236233e-02 9.97533262e-01 -2.65510410e-01 -3.96566242e-01
1.01237386e-01 -1.71535105e-01 7.93864548e-01 -2.09370509e-01
4.67745364e-01 -3.08255315e-01 -7.25052953e-01 9.34959412e-01
-4.72484678e-02 3.84471826e-02 2.02943474e-01 4.23220068e-01
7.17413723e-01 -4.10422802e-01 2.16641501e-01 -4.73695159e-01
2.63319761e-02 -3.48277874e-02 -6.26750231e-01 3.85191768e-01
-3.86437982e-01 -8.21744382e-01 8.62338692e-02 -4.37828600e-01
-8.72413933e-01 2.13429525e-01 1.49560809e-01 -9.38592017e-01
-2.26891115e-01 -1.19695716e-01 -2.57427454e-01 -4.16846037e-01
4.34565425e-01 4.17601094e-02 1.03343902e-02 -5.48758209e-01
3.89255524e-01 2.68250257e-01 4.26597744e-01 -6.17110908e-01
1.24940562e+00 7.89636850e-01 2.38632351e-01 -6.97114646e-01
-1.03474891e+00 -8.33752513e-01 -1.37740803e+00 -7.70960152e-01
9.73810375e-01 -1.31175280e+00 -8.44188988e-01 3.76413554e-01
-9.95131791e-01 4.95270699e-01 1.29644915e-01 5.93294024e-01
-3.76702160e-01 4.90050465e-01 -3.42718631e-01 -9.75736082e-01
-2.26580635e-01 -9.77702379e-01 1.58076131e+00 2.13052571e-01
-3.63412470e-01 -9.38263834e-01 1.25746503e-01 5.34863293e-01
-1.55763969e-01 5.58556855e-01 9.90982577e-02 -3.83727662e-02
-6.56489193e-01 -5.88477314e-01 6.02296069e-02 -3.61410022e-01
1.62867755e-01 -1.21742822e-01 -8.54775488e-01 -8.42781126e-01
-5.28594077e-01 1.81700602e-01 2.37538815e-01 6.07157409e-01
3.09550047e-01 3.03648800e-01 -8.54146600e-01 8.13115597e-01
1.13137674e+00 -9.82339233e-02 2.35001549e-01 4.30906922e-01
1.15050054e+00 2.17167690e-01 9.92691338e-01 4.58869129e-01
7.02538192e-01 1.27194691e+00 2.43794560e-01 -6.80201948e-02
-3.35887969e-01 -6.34376824e-01 4.12497044e-01 6.37284398e-01
-5.24234056e-01 1.21556006e-01 -8.61659884e-01 2.03443915e-01
-1.91161299e+00 -1.17488849e+00 -6.15215063e-01 2.42100358e+00
1.87841073e-01 4.50078323e-02 7.97546268e-01 -1.75221443e-01
9.02352035e-01 1.05745912e-01 -4.41012621e-01 6.02749586e-01
-2.06474215e-02 -4.70687360e-01 6.36695623e-01 3.06502938e-01
-1.28836787e+00 7.23337054e-01 6.66147804e+00 1.95384488e-01
-5.52584112e-01 2.38844112e-01 -5.04750490e-01 -4.10902023e-01
4.34150621e-02 -2.00260893e-01 -1.82731009e+00 3.62254769e-01
3.50806803e-01 2.82425225e-01 -2.00358957e-01 6.32374406e-01
8.36407095e-02 6.38952255e-02 -9.92535949e-01 1.35346758e+00
1.59824848e-01 -1.16338277e+00 -8.80153701e-02 4.23034549e-01
8.10970902e-01 -3.68810892e-01 9.00780335e-02 2.53147274e-01
-8.75432342e-02 -4.91740227e-01 9.26755250e-01 3.95963103e-01
6.44555867e-01 -6.75527811e-01 4.23610985e-01 4.92651880e-01
-1.95045805e+00 2.24657118e-01 -1.15923658e-01 1.18342400e-01
7.09277868e-01 2.96512812e-01 -4.15631890e-01 1.13589585e+00
8.20358634e-01 8.72740865e-01 -5.69715440e-01 1.09938753e+00
1.08184427e-01 -1.34145260e-01 -4.43725646e-01 3.85290980e-01
-1.41426608e-01 -1.74576774e-01 9.31936026e-01 1.10528827e+00
4.70949560e-01 -1.20989010e-01 9.06169951e-01 6.15129471e-01
3.75494123e-01 -2.85189003e-01 -8.23968947e-01 5.16076863e-01
8.16594660e-01 1.09837437e+00 -6.16181254e-01 -1.57639161e-01
-7.41279364e-01 6.48976564e-01 -7.10530281e-02 3.34814906e-01
-1.22699714e+00 3.80563080e-01 8.12794149e-01 4.77321684e-01
3.53555083e-01 -7.37066746e-01 -2.02531770e-01 -1.33914983e+00
3.63735296e-02 -5.96132576e-01 6.00844920e-01 -5.35879254e-01
-1.19479978e+00 2.01871827e-01 3.55949253e-01 -1.88178635e+00
-2.89830882e-02 -4.24984843e-01 -1.02793097e-01 8.79815698e-01
-1.09252238e+00 -1.84329331e+00 -5.16370058e-01 8.12023461e-01
2.63977110e-01 7.28539610e-03 7.19937325e-01 7.20408976e-01
-6.11637890e-01 8.25550914e-01 -3.03423822e-01 1.85612693e-01
9.24974799e-01 -1.02118158e+00 4.60170716e-01 1.07406116e+00
1.65347144e-01 1.05567074e+00 5.85396945e-01 -1.24375761e+00
-1.92504609e+00 -1.24194491e+00 4.87596959e-01 -1.22597694e+00
8.20747092e-02 -5.98792434e-01 -3.85902464e-01 1.09493291e+00
-3.97739887e-01 3.19156766e-01 8.17657113e-01 4.71197635e-01
-2.87219167e-01 1.46970809e-01 -1.07458782e+00 6.38514698e-01
1.70538628e+00 -3.86054873e-01 -4.91506785e-01 4.43304852e-02
3.71225178e-01 -1.13275063e+00 -1.36520624e+00 4.79941279e-01
1.00769877e+00 -8.27233076e-01 1.75305831e+00 -2.85107762e-01
-6.15336776e-01 -8.67878914e-01 -2.88119674e-01 -8.57234538e-01
-4.56135631e-01 -5.80820024e-01 -7.46331334e-01 1.17874050e+00
-1.05603449e-01 -3.76156747e-01 1.07892859e+00 4.12008643e-01
-1.17276475e-01 -4.91660744e-01 -8.28576446e-01 -1.13348114e+00
-4.46493924e-01 -2.88558513e-01 6.45482600e-01 7.61763453e-01
-4.77007896e-01 8.02300200e-02 -7.61625648e-01 7.41865158e-01
1.22607398e+00 4.17762280e-01 1.43729043e+00 -1.44687295e+00
6.66755661e-02 -7.54160807e-02 -2.51708150e-01 -1.46975029e+00
-1.48786068e-01 -6.45643830e-01 -2.58727938e-01 -1.42918193e+00
2.37288937e-01 -3.52919251e-01 1.97621524e-01 1.18446611e-01
-2.40586042e-01 4.92306471e-01 6.06612206e-01 2.82374084e-01
-7.02321053e-01 4.83796328e-01 1.52990973e+00 8.45310837e-02
-1.61654323e-01 1.47338122e-01 -4.38379258e-01 8.53584826e-01
1.48657843e-01 -3.53981942e-01 -2.41647974e-01 -4.86498445e-01
1.41083607e-02 1.36006474e-01 7.30560005e-01 -1.13037503e+00
3.94159526e-01 -5.35713322e-02 9.06837046e-01 -1.51882720e+00
7.93098688e-01 -1.00718713e+00 6.50608182e-01 3.93306136e-01
5.27130306e-01 4.82490212e-01 2.61617303e-01 7.46135652e-01
2.98947722e-01 2.44096205e-01 4.52558160e-01 -3.34165752e-01
-6.25742316e-01 8.14607322e-01 2.11830452e-01 -7.62829557e-02
1.17659056e+00 -1.07415605e+00 -2.87990212e-01 3.25277937e-03
-9.26654458e-01 4.22102362e-01 8.58086526e-01 6.26389682e-01
7.31403053e-01 -1.88275003e+00 -4.80254173e-01 2.49700353e-01
5.98669872e-02 5.24137318e-02 3.79572481e-01 1.14855409e+00
-1.37237728e-01 6.04719639e-01 -1.97180405e-01 -1.22583795e+00
-1.69700980e+00 4.49783444e-01 3.33368510e-01 -1.19483456e-01
-9.48949993e-01 4.44979966e-01 5.79382926e-02 -7.09545016e-01
2.22946718e-01 1.14731146e-02 1.54036814e-02 1.07483320e-01
6.28229201e-01 7.96272695e-01 -1.78367928e-01 -1.15771759e+00
-6.59036756e-01 1.42175019e+00 7.30402619e-02 9.52757075e-02
9.19237137e-01 -5.71121991e-01 4.07403827e-01 3.64234984e-01
7.07883537e-01 3.53356689e-01 -1.49697220e+00 -1.10623598e-01
-2.89248466e-01 -9.34809864e-01 -2.04776973e-01 -3.46988976e-01
-9.61627781e-01 5.32552958e-01 7.10564017e-01 -1.59762770e-01
5.37196398e-01 -2.92354375e-01 9.58246648e-01 7.79967010e-02
8.83353353e-01 -8.95261407e-01 1.16990395e-01 5.11055231e-01
5.80885470e-01 -1.06850743e+00 3.92488807e-01 -7.85955667e-01
-1.88984215e-01 1.01973951e+00 9.33494210e-01 7.09046945e-02
6.07761025e-01 3.02644968e-01 1.47167419e-03 -5.20647705e-01
5.09778038e-03 -2.85154581e-01 6.12864017e-01 9.18054521e-01
3.12236100e-01 -7.16807246e-02 1.99474365e-01 3.58667612e-01
6.60632923e-02 -1.43763646e-01 4.77409847e-02 1.13582420e+00
-2.26974472e-01 -1.03450263e+00 -1.10329425e+00 2.51675453e-02
-6.62813857e-02 5.72938204e-01 -1.84109479e-01 1.25313950e+00
2.41007522e-01 8.28115046e-01 -1.62453979e-01 -5.17509222e-01
8.74646127e-01 1.58455074e-02 1.04887795e+00 -6.82955563e-01
-6.38574064e-01 4.01729643e-01 1.39908135e-01 -6.60032213e-01
-7.47555137e-01 -1.00532866e+00 -9.89074588e-01 -7.31024861e-01
-4.33315188e-01 -3.42116028e-01 1.24312349e-01 7.85878420e-01
3.48664045e-01 4.27016526e-01 2.09759697e-01 -1.26207018e+00
-3.14171910e-01 -4.46280301e-01 -4.09604639e-01 3.80977213e-01
2.98209876e-01 -1.36597872e+00 2.52066642e-01 -1.30124226e-01] | [7.02747106552124, -1.0172698497772217] |
39498949-deb0-45a6-a8b9-743657cef03e | transformer-patcher-one-mistake-worth-one | 2301.09785 | null | https://arxiv.org/abs/2301.09785v1 | https://arxiv.org/pdf/2301.09785v1.pdf | Transformer-Patcher: One Mistake worth One Neuron | Large Transformer-based Pretrained Language Models (PLMs) dominate almost all Natural Language Processing (NLP) tasks. Nevertheless, they still make mistakes from time to time. For a model deployed in an industrial environment, fixing these mistakes quickly and robustly is vital to improve user experiences. Previous works formalize such problems as Model Editing (ME) and mostly focus on fixing one mistake. However, the one-mistake-fixing scenario is not an accurate abstraction of the real-world challenge. In the deployment of AI services, there are ever-emerging mistakes, and the same mistake may recur if not corrected in time. Thus a preferable solution is to rectify the mistakes as soon as they appear nonstop. Therefore, we extend the existing ME into Sequential Model Editing (SME) to help develop more practical editing methods. Our study shows that most current ME methods could yield unsatisfying results in this scenario. We then introduce Transformer-Patcher, a novel model editor that can shift the behavior of transformer-based models by simply adding and training a few neurons in the last Feed-Forward Network layer. Experimental results on both classification and generation tasks show that Transformer-Patcher can successively correct up to thousands of errors (Reliability) and generalize to their equivalent inputs (Generality) while retaining the model's accuracy on irrelevant inputs (Locality). Our method outperforms previous fine-tuning and HyperNetwork-based methods and achieves state-of-the-art performance for Sequential Model Editing (SME). The code is available at https://github.com/ZeroYuHuang/Transformer-Patcher. | ['Zhang Xiong', 'Wenge Rong', 'Jie zhou', 'Xiaofeng Zhang', 'Yikang Shen', 'Zeyu Huang'] | 2023-01-24 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [ 3.59444022e-01 1.78901628e-01 6.08713254e-02 -5.10852039e-01
-4.25576895e-01 -4.65510726e-01 2.91350573e-01 -1.63983151e-01
-3.94310713e-01 6.03259623e-01 -4.37717199e-01 -5.27599275e-01
2.36147959e-02 -7.64403343e-01 -1.00427055e+00 -3.09792340e-01
3.82048130e-01 5.22989511e-01 1.66131571e-01 -4.67934638e-01
8.82663429e-02 3.50372523e-01 -1.34246695e+00 5.67441523e-01
1.14157915e+00 7.74843097e-01 3.64766151e-01 3.41019243e-01
-3.84831667e-01 7.94434726e-01 -6.85199559e-01 -6.31104469e-01
2.51577914e-01 -1.84664428e-01 -7.73225784e-01 -5.30707002e-01
3.16143662e-01 -2.99333155e-01 3.90242338e-02 1.13113451e+00
2.99083173e-01 1.06089629e-01 1.47304550e-01 -1.51833165e+00
-8.67320120e-01 1.04955149e+00 -1.74277529e-01 -1.37929656e-02
-3.35414372e-02 1.87412396e-01 6.47467494e-01 -1.11770880e+00
2.80671328e-01 1.16967630e+00 9.54441607e-01 8.22433531e-01
-1.16047025e+00 -9.92177486e-01 4.75934058e-01 2.04650417e-01
-1.46151340e+00 -5.01946807e-01 6.94845378e-01 -1.63559705e-01
1.53149974e+00 3.93532246e-01 3.28402042e-01 1.24436498e+00
3.87708068e-01 8.03394556e-01 5.96879005e-01 -5.23044229e-01
7.69324154e-02 4.21678662e-01 1.69254541e-01 6.47363424e-01
3.24229568e-01 -8.15822706e-02 -3.77970397e-01 2.29566414e-02
7.43852496e-01 1.42493084e-01 -2.20189407e-01 -5.02810776e-02
-1.21266174e+00 4.30916190e-01 3.73978645e-01 5.93286932e-01
-4.26794559e-01 1.34921059e-01 1.86114505e-01 6.94790900e-01
3.81589323e-01 8.97295296e-01 -9.77917075e-01 -1.28925726e-01
-9.44630623e-01 3.11475694e-01 8.22737515e-01 1.17697275e+00
6.52372539e-01 4.57351327e-01 -1.09439366e-01 1.06987035e+00
-1.97285399e-01 2.26669908e-01 6.69629157e-01 -7.64942229e-01
4.34622109e-01 7.63293505e-01 5.93734197e-02 -1.07755637e+00
-3.46000612e-01 -8.30800712e-01 -1.16357124e+00 3.00719049e-02
2.29694515e-01 -1.15605801e-01 -9.08706605e-01 1.86679530e+00
-1.19109228e-01 1.55937776e-01 -1.22229800e-01 5.34513235e-01
5.47113597e-01 7.10656822e-01 8.14205557e-02 -5.56903481e-02
8.54112446e-01 -1.17860878e+00 -6.99687958e-01 -6.26696289e-01
7.62538373e-01 -5.83780289e-01 1.50219190e+00 7.27646172e-01
-1.11155427e+00 -5.71109414e-01 -8.84802163e-01 1.87871121e-02
-6.16197169e-01 3.07510048e-01 4.88758534e-01 4.48961109e-01
-1.01911533e+00 1.02191186e+00 -7.90087461e-01 -3.01997632e-01
1.26452774e-01 5.19138932e-01 -2.74238944e-01 -1.56918839e-02
-1.36333597e+00 1.11140025e+00 5.00691533e-01 4.35629249e-01
-5.56586802e-01 -1.18242574e+00 -6.78434730e-01 1.95229158e-01
5.63328683e-01 -8.64410520e-01 1.68404269e+00 -1.29766071e+00
-1.53545415e+00 3.77010465e-01 -2.78951436e-01 -4.83257830e-01
5.57922840e-01 -3.15415829e-01 -5.66389382e-01 -5.79662681e-01
-2.00993523e-01 7.31036127e-01 6.84220850e-01 -1.19437802e+00
-5.47406256e-01 1.07286505e-01 2.54703194e-01 -6.76079839e-02
-5.59422851e-01 7.08157243e-03 -3.36767972e-01 -8.11040163e-01
-6.57060444e-02 -6.78716183e-01 -2.60859519e-01 5.64014688e-02
-3.64526242e-01 -3.35530668e-01 5.69731832e-01 -6.39358640e-01
1.61687493e+00 -1.93518877e+00 3.04832421e-02 7.87832737e-02
2.16711029e-01 7.42897928e-01 -4.28695083e-01 4.54910338e-01
-2.88421124e-01 4.49580789e-01 -3.85705709e-01 -4.54705507e-01
9.15004686e-02 4.11557615e-01 -4.37382281e-01 -2.47255221e-01
3.20836693e-01 1.13197458e+00 -8.77371073e-01 -6.89151743e-03
1.11619920e-01 4.66340601e-01 -6.87956095e-01 7.07512945e-02
-3.17081660e-01 1.52035266e-01 -2.54414976e-02 5.74686587e-01
5.29422700e-01 -2.79421210e-01 -6.11999892e-02 -2.21871644e-01
3.98656912e-02 3.65404218e-01 -1.19914842e+00 1.49888361e+00
-7.70561695e-01 4.79310393e-01 -2.01524258e-01 -9.38621759e-01
9.07577932e-01 2.85864294e-01 1.13210984e-01 -7.95398951e-01
8.53377208e-02 2.56728977e-01 6.36138692e-02 -3.00833046e-01
4.87595856e-01 5.54469191e-02 1.50513828e-01 4.27757233e-01
1.27042169e-02 2.64323819e-02 2.71050543e-01 1.05112486e-01
1.07135451e+00 4.44649979e-02 2.05718622e-01 -1.32141769e-01
2.85238713e-01 -1.07481517e-01 9.07001555e-01 1.02794468e+00
2.75661852e-02 4.74868715e-01 2.79968411e-01 -6.31721199e-01
-9.95674670e-01 -9.43223834e-01 2.81520218e-01 1.13854289e+00
-2.57325619e-01 -6.52617693e-01 -9.06167567e-01 -5.26609242e-01
-1.55584782e-01 1.32661223e+00 -4.26704139e-01 -6.05214655e-01
-5.44010401e-01 -6.49835110e-01 7.79519320e-01 6.64003849e-01
5.26759505e-01 -1.25270128e+00 -4.78605330e-01 4.62771744e-01
-2.21249700e-01 -9.13334727e-01 -3.78252566e-01 2.79795676e-01
-8.16935539e-01 -6.94560051e-01 -3.86924803e-01 -7.36871779e-01
8.83073866e-01 1.22264691e-01 1.26341116e+00 4.21739668e-01
2.80655362e-02 -3.36209722e-02 -2.80335963e-01 -7.01537073e-01
-5.94505906e-01 1.99960589e-01 3.42803538e-01 -2.94552922e-01
4.98361379e-01 -7.51439691e-01 -4.50968780e-02 2.84150839e-01
-1.04221213e+00 4.08685386e-01 6.35731339e-01 9.01500940e-01
4.74622786e-01 3.03149223e-01 6.44792378e-01 -1.13969672e+00
9.88557100e-01 -2.26013362e-01 -4.78650689e-01 5.52286327e-01
-9.98026490e-01 4.06544050e-03 1.10007000e+00 -8.86601329e-01
-9.27496612e-01 -7.61887282e-02 -1.72210798e-01 -5.88196099e-01
-1.39485687e-01 5.87052107e-01 -1.20452091e-01 -2.23896820e-02
7.17571676e-01 3.06638211e-01 -3.72408688e-01 -7.08470762e-01
1.32252961e-01 5.14816523e-01 4.31557834e-01 -5.63657165e-01
8.83191884e-01 -1.80635244e-01 -5.73148191e-01 -4.66585755e-01
-7.60655761e-01 2.22728457e-02 -4.57302928e-01 2.62556914e-02
2.06513945e-02 -7.35374153e-01 -4.74491566e-01 6.26532912e-01
-1.50603139e+00 -6.20060802e-01 -2.45505199e-01 3.70208882e-02
-1.61099970e-01 6.77845478e-02 -5.16589701e-01 -5.85703850e-01
-5.41488528e-01 -1.04991162e+00 6.72672510e-01 1.17172077e-01
-4.83740866e-01 -8.78180623e-01 -2.50587523e-01 5.96347905e-04
8.69732618e-01 -2.10701704e-01 1.13853073e+00 -7.29466558e-01
-2.14230552e-01 -2.40867496e-01 3.74322087e-02 6.92230284e-01
1.51926711e-01 2.75960714e-02 -9.93560970e-01 -1.53651118e-01
-9.39629301e-02 2.59734560e-02 6.16620004e-01 1.02890909e-01
1.41205990e+00 -6.80423975e-01 -2.99089760e-01 4.82369244e-01
1.07194817e+00 2.96274722e-01 5.53711593e-01 2.29401097e-01
6.53328180e-01 3.23789507e-01 2.94043869e-01 2.01988041e-01
2.37388998e-01 5.47849596e-01 1.74677908e-01 -2.43607581e-01
1.17344655e-01 -4.95411128e-01 5.11561632e-01 9.85041380e-01
9.76646692e-02 -3.75609875e-01 -1.08035398e+00 5.49651623e-01
-1.85716701e+00 -7.97348440e-01 6.75629601e-02 2.21674371e+00
1.19486690e+00 4.27715272e-01 -3.61458451e-01 2.95694262e-01
5.63950360e-01 -3.97800386e-01 -6.13257945e-01 -7.36567974e-01
1.05553217e-01 2.70052791e-01 3.03545326e-01 5.33185065e-01
-7.21448839e-01 1.23978937e+00 6.22071886e+00 8.55673194e-01
-1.34693861e+00 1.81428805e-01 5.45202792e-01 -2.02470541e-01
-3.85608256e-01 -1.52684063e-01 -8.26418698e-01 3.19911957e-01
1.05445743e+00 -4.25101966e-01 8.39138508e-01 1.04707038e+00
3.29715520e-01 2.98718333e-01 -1.38168871e+00 1.00768507e+00
-1.93834919e-02 -1.23800480e+00 4.17818159e-01 -4.47037578e-01
5.84156573e-01 -4.35326248e-03 -7.93384481e-03 7.15094268e-01
2.79482752e-01 -1.02528679e+00 7.57650912e-01 6.14189804e-01
5.95078528e-01 -6.81710958e-01 7.03047931e-01 8.13596010e-01
-8.75269532e-01 -1.47165954e-01 -5.16152024e-01 -2.64630318e-01
7.66891167e-02 8.80542815e-01 -8.62455130e-01 3.83772701e-01
8.74365509e-01 6.17778659e-01 -6.87127948e-01 8.96908462e-01
-3.73054147e-01 6.03030503e-01 -4.28843021e-01 4.78287786e-02
-3.35154235e-02 1.62178919e-01 4.56284493e-01 1.15390861e+00
3.70811909e-01 -7.80598447e-02 -1.38403088e-01 1.28252816e+00
-2.12925360e-01 -4.45356183e-02 -6.26597106e-01 -1.35380700e-01
7.08184719e-01 1.08366239e+00 -2.73853421e-01 -4.23801035e-01
-1.79693118e-01 1.07306480e+00 5.18123806e-01 5.05405366e-01
-9.12438750e-01 -5.85271835e-01 6.08394682e-01 1.11440584e-01
3.50931101e-02 -1.45890284e-02 -4.38319296e-01 -1.20583987e+00
4.14579451e-01 -1.20070064e+00 -5.73538914e-02 -9.40274119e-01
-1.27530706e+00 9.49153662e-01 -9.29962993e-02 -9.78566945e-01
-3.27766180e-01 -6.23616695e-01 -5.86411178e-01 7.09621549e-01
-1.46583855e+00 -1.03823626e+00 -1.17656380e-01 3.58184606e-01
7.78346121e-01 -4.83044721e-02 1.09533834e+00 5.11441886e-01
-8.12807381e-01 1.04088509e+00 -1.00616500e-01 7.36132357e-03
7.24779010e-01 -9.58375692e-01 6.74193144e-01 9.91151869e-01
9.68758687e-02 1.11973023e+00 7.14276075e-01 -5.30823886e-01
-1.11314225e+00 -1.39032423e+00 1.49339592e+00 -5.17167032e-01
5.00447214e-01 -5.12918890e-01 -1.19541156e+00 1.09730601e+00
8.12492594e-02 -2.45946780e-01 2.85323113e-01 -4.53641725e-04
-4.65112478e-01 -2.91663915e-01 -1.01455379e+00 7.89965868e-01
1.13172007e+00 -2.94393837e-01 -6.14598691e-01 3.23031694e-01
9.37584162e-01 -5.06708682e-01 -6.51884675e-01 5.28708994e-01
4.67196673e-01 -7.49588132e-01 7.56450593e-01 -7.47898877e-01
3.98156434e-01 -2.16353953e-01 1.80021822e-01 -1.57093585e+00
-4.18013752e-01 -7.15509355e-01 -4.83839922e-02 1.23981857e+00
7.79978275e-01 -8.08648288e-01 3.69876027e-01 9.47634876e-01
-3.24934453e-01 -8.23038876e-01 -7.01427460e-01 -9.53618288e-01
2.86680877e-01 -7.25356758e-01 9.45340633e-01 1.13890362e+00
-1.13369346e-01 1.25889003e-01 -4.24703568e-01 4.97533903e-02
1.74091130e-01 -2.75903851e-01 6.17053628e-01 -1.21476102e+00
-3.50186110e-01 -6.54696703e-01 9.89270210e-02 -9.67716455e-01
1.11023277e-01 -8.23319376e-01 1.64530396e-01 -1.39486384e+00
-6.70277774e-02 -4.71817374e-01 -5.19313276e-01 1.21741092e+00
-4.81533408e-02 2.80864798e-02 2.74814814e-01 5.92343397e-02
-3.93373519e-01 3.82395744e-01 8.62432718e-01 -3.21122468e-01
-3.45150560e-01 1.56469252e-02 -8.46889794e-01 7.53664255e-01
1.11766982e+00 -7.13723600e-01 -5.30587971e-01 -9.50586736e-01
5.98612249e-01 -6.11050844e-01 3.78485799e-01 -1.06794596e+00
4.84237731e-01 -2.78939009e-01 1.52271122e-01 -1.97577149e-01
9.18959156e-02 -9.46184456e-01 3.92119020e-01 3.47354144e-01
-4.78711247e-01 4.32524323e-01 6.17741644e-01 1.04013979e-01
-1.10027485e-01 -2.12155193e-01 5.49498379e-01 -2.48610750e-01
-6.62028015e-01 1.26203910e-01 -3.91941577e-01 -1.42600477e-01
7.53457189e-01 -1.62169829e-01 -3.36426795e-01 -2.18681678e-01
-5.31715691e-01 2.68821627e-01 2.22443745e-01 8.26135993e-01
7.00517535e-01 -1.10784781e+00 -6.24175012e-01 5.22641897e-01
9.41613615e-02 1.46649964e-02 9.38831046e-02 6.99949682e-01
-1.72072455e-01 5.19849598e-01 -9.51021910e-02 -2.72510409e-01
-1.04248810e+00 4.10147339e-01 6.61942244e-01 -4.07363236e-01
-4.76820916e-01 9.80959654e-01 1.50003284e-02 -9.66661274e-01
5.32567084e-01 -7.07169116e-01 4.78119627e-02 -4.20604944e-01
7.14982569e-01 2.43564248e-02 5.22272885e-01 -4.48090695e-02
-2.87018180e-01 2.17671633e-01 -4.32841331e-01 3.26438636e-01
1.44177747e+00 1.10007592e-01 -2.10690990e-01 6.41773880e-01
6.51256382e-01 -4.44062293e-01 -7.62907624e-01 -4.12203431e-01
-1.66525617e-01 -1.63844913e-01 1.44152353e-02 -1.32763982e+00
-1.09543478e+00 9.75121975e-01 3.37513417e-01 -3.42294760e-02
1.20085478e+00 -5.79820216e-01 8.31144273e-01 9.72498357e-01
5.41171789e-01 -1.08173871e+00 -2.54504442e-01 7.47691154e-01
1.14192188e+00 -8.66047144e-01 -4.06647801e-01 -3.24318707e-01
-6.17247999e-01 9.62140858e-01 9.78325367e-01 1.13861568e-01
5.12484789e-01 4.51342821e-01 8.85186270e-02 2.34361365e-01
-1.24230468e+00 4.79112029e-01 4.72559705e-02 5.39599121e-01
4.36576396e-01 -1.64990406e-02 -4.31318879e-02 1.14875412e+00
-3.57869744e-01 4.52119231e-01 4.17043060e-01 9.14850414e-01
-2.09052369e-01 -1.29636598e+00 -1.62125573e-01 4.98232573e-01
-1.93858549e-01 -5.76228440e-01 -4.95109200e-01 6.95401013e-01
2.07370445e-01 9.63566840e-01 -1.10748827e-01 -6.42354012e-01
6.94349766e-01 3.47132176e-01 2.51170158e-01 -6.32705927e-01
-9.81336296e-01 -4.06989753e-01 -2.35046912e-02 -7.30430961e-01
7.31265843e-02 -1.82660788e-01 -1.25003803e+00 -5.67974865e-01
-2.98740119e-01 -4.48659956e-02 5.39338648e-01 9.91712034e-01
6.37784302e-01 8.31506789e-01 1.90798700e-01 -6.05831921e-01
-9.52082396e-01 -1.04807031e+00 -1.76254109e-01 2.33101428e-01
1.46715596e-01 -4.94352013e-01 -2.56778657e-01 2.02277750e-01] | [9.698269844055176, 7.477914810180664] |
07500b27-8dd3-4626-9869-c6b78693bfe4 | font-flow-guided-one-shot-talking-head | 2303.17789 | null | https://arxiv.org/abs/2303.17789v1 | https://arxiv.org/pdf/2303.17789v1.pdf | FONT: Flow-guided One-shot Talking Head Generation with Natural Head Motions | One-shot talking head generation has received growing attention in recent years, with various creative and practical applications. An ideal natural and vivid generated talking head video should contain natural head pose changes. However, it is challenging to map head pose sequences from driving audio since there exists a natural gap between audio-visual modalities. In this work, we propose a Flow-guided One-shot model that achieves NaTural head motions(FONT) over generated talking heads. Specifically, the head pose prediction module is designed to generate head pose sequences from the source face and driving audio. We add the random sampling operation and the structural similarity constraint to model the diversity in the one-to-many mapping between audio-visual modality, thus predicting natural head poses. Then we develop a keypoint predictor that produces unsupervised keypoints from the source face, driving audio and pose sequences to describe the facial structure information. Finally, a flow-guided occlusion-aware generator is employed to produce photo-realistic talking head videos from the estimated keypoints and source face. Extensive experimental results prove that FONT generates talking heads with natural head poses and synchronized mouth shapes, outperforming other compared methods. | ['Jizhong Han', 'Jiao Dai', 'Cai Yu', 'Yesheng Chai', 'Xiaomeng Fu', 'Xi Wang', 'Jin Liu'] | 2023-03-31 | null | null | null | null | ['talking-head-generation', 'pose-prediction'] | ['computer-vision', 'computer-vision'] | [ 1.71201080e-02 2.03759789e-01 1.32931709e-01 -5.49191952e-01
-1.02243853e+00 -1.50363669e-01 6.00037277e-01 -7.35047460e-01
4.03732568e-01 4.27234203e-01 8.56424809e-01 6.21706367e-01
2.27182031e-01 -1.16043538e-01 -6.72783911e-01 -8.01841319e-01
-3.54599543e-02 2.36247241e-01 -4.36457954e-02 -1.92303762e-01
2.14900494e-01 3.42783093e-01 -2.24792576e+00 1.75087787e-02
7.58633435e-01 1.03016579e+00 4.52239752e-01 9.33858395e-01
1.98215228e-02 8.04679990e-01 -5.54308295e-01 -2.76744217e-01
2.99059283e-02 -6.34676039e-01 -2.81079918e-01 5.70615172e-01
7.52026498e-01 -5.61060786e-01 -2.75668681e-01 7.84458756e-01
1.14024305e+00 2.92990088e-01 5.03060341e-01 -1.58006918e+00
-1.66721165e-01 2.79878736e-01 -7.53857791e-01 -2.28819788e-01
1.23471344e+00 4.28593963e-01 6.32568479e-01 -1.22782755e+00
8.20955634e-01 1.79557002e+00 4.23858315e-01 7.94863462e-01
-8.82693887e-01 -9.37501192e-01 9.94576812e-02 5.00422835e-01
-1.69194102e+00 -1.14180470e+00 1.09400439e+00 -3.40456665e-01
3.19937825e-01 2.46092960e-01 7.32892334e-01 1.10178721e+00
6.01244345e-02 8.77084970e-01 4.87579226e-01 -3.60236198e-01
7.85054117e-02 -9.94961336e-03 -5.25602162e-01 4.93440896e-01
-5.39973736e-01 3.75030302e-02 -1.10101020e+00 -1.41320661e-01
4.23088044e-01 -1.33497357e-01 -7.83075929e-01 -4.06361759e-01
-1.05676484e+00 5.60590446e-01 1.52391791e-01 6.23278841e-02
-4.44701165e-01 -1.11282445e-01 2.80456066e-01 -1.48580194e-01
2.26540402e-01 -9.07947794e-02 7.92855695e-02 -1.93543240e-01
-1.12305927e+00 3.77798289e-01 7.79031634e-01 1.19976330e+00
6.52814984e-01 2.54113495e-01 -3.25594902e-01 9.22924459e-01
3.81112397e-01 7.95575559e-01 5.57312787e-01 -1.09516585e+00
2.64611006e-01 1.86122581e-01 -2.22921949e-02 -1.19745433e+00
-3.10768515e-01 7.08899200e-02 -6.72923744e-01 -2.98436880e-01
-9.36320145e-03 -2.06614479e-01 -5.54674923e-01 1.84268391e+00
8.10382247e-01 3.74183714e-01 -1.96716145e-01 1.15062976e+00
9.66803670e-01 8.35093558e-01 -8.18007439e-02 -8.06565344e-01
1.33090377e+00 -7.62713552e-01 -1.27375698e+00 -3.30019325e-01
1.23832524e-01 -9.02366102e-01 9.37773347e-01 3.20957124e-01
-1.25187624e+00 -8.05542946e-01 -6.05124056e-01 -5.45128074e-04
4.05603856e-01 -1.59779936e-02 5.97455725e-02 3.96221191e-01
-9.74289000e-01 -5.77152483e-02 -3.88526499e-01 -2.83487439e-01
4.37143631e-02 2.03372762e-01 -4.12704647e-01 -9.91335809e-02
-1.08499289e+00 4.34265733e-01 1.68942079e-01 3.40788588e-02
-9.63714659e-01 -6.41795218e-01 -1.16869533e+00 -7.78006837e-02
4.90211964e-01 -6.54940546e-01 1.47110724e+00 -1.01332068e+00
-1.79483318e+00 6.01230085e-01 -8.05850685e-01 -9.38261524e-02
6.05264366e-01 -3.13629866e-01 -4.50444400e-01 3.40754986e-01
1.70066684e-01 1.08250403e+00 1.45930827e+00 -1.48551989e+00
-5.76821864e-01 -5.31499565e-01 -6.59111500e-01 5.66585779e-01
-3.87910724e-01 1.37387961e-01 -7.84574628e-01 -5.61217606e-01
3.14149037e-02 -8.84486854e-01 2.70269871e-01 1.40255377e-01
-6.77340031e-01 -1.93167284e-01 1.01331711e+00 -8.00488114e-01
1.13575280e+00 -2.36924624e+00 4.13734049e-01 -1.49139926e-01
9.87346023e-02 -1.77263632e-01 5.12634888e-02 2.11446792e-01
-1.16589174e-01 -6.34742260e-01 5.10832258e-02 -9.06627357e-01
-6.34793565e-02 -1.77808981e-02 -4.09684360e-01 5.95776558e-01
-1.63268838e-02 5.74402988e-01 -7.98682630e-01 -9.17364538e-01
2.64982641e-01 8.72858047e-01 -8.10427248e-01 6.57302797e-01
-6.48016334e-02 7.40393937e-01 -4.83478457e-02 7.03963876e-01
8.47548187e-01 2.50764787e-01 1.51868174e-02 -3.88038456e-01
-1.03741139e-01 -1.79906040e-01 -1.37242198e+00 1.80589652e+00
-3.33704770e-01 6.49130523e-01 3.18684161e-01 -3.58616918e-01
1.16044402e+00 6.47402763e-01 4.06565160e-01 -5.17097116e-01
3.93404543e-01 3.45047973e-02 -3.40785831e-01 -9.21133339e-01
4.98703033e-01 -2.35020995e-01 -5.46283722e-02 1.43453360e-01
2.28335842e-01 -5.29496670e-01 -3.73685993e-02 2.31957948e-03
3.80230039e-01 -2.74244952e-03 9.15460438e-02 1.16288573e-01
9.00840521e-01 -6.80593789e-01 5.78253806e-01 -8.86130556e-02
-1.90014839e-01 1.05759645e+00 3.04381132e-01 -8.09860528e-02
-7.57888317e-01 -8.18787396e-01 1.35787249e-01 1.31581032e+00
1.13932721e-01 -4.15474683e-01 -1.19786465e+00 4.85247076e-02
-3.87622774e-01 6.15992188e-01 -5.60154915e-01 -1.79981321e-01
-6.90423369e-01 4.73564193e-02 2.92911887e-01 2.67476112e-01
3.51716250e-01 -1.13480294e+00 -3.74075204e-01 1.11871593e-01
-6.16471708e-01 -1.15121758e+00 -1.19698608e+00 -5.37065804e-01
-2.80675888e-01 -8.23679388e-01 -1.25029707e+00 -1.04158354e+00
5.08221805e-01 2.48927265e-01 7.51336098e-01 -4.54517931e-01
-4.49371725e-01 4.36893076e-01 -1.56558931e-01 -3.50651324e-01
-3.55122924e-01 -3.25217634e-01 5.25734782e-01 6.54775858e-01
9.12207216e-02 -5.21211743e-01 -8.07653129e-01 4.84931588e-01
-5.86452425e-01 1.22486286e-01 3.08805883e-01 6.05778575e-01
4.19963807e-01 -3.64297450e-01 5.08885622e-01 -2.11757347e-02
4.42645371e-01 -3.86036634e-01 -1.77536860e-01 -2.46618986e-02
2.44057402e-02 -2.88590521e-01 7.85454273e-01 -7.49542296e-01
-1.28937817e+00 4.28716451e-01 -1.87438071e-01 -9.46309626e-01
-2.32848331e-01 -2.11928695e-01 -8.70996892e-01 2.45759174e-01
4.85351920e-01 6.07992530e-01 3.15461248e-01 -3.08568090e-01
4.81243789e-01 8.71436417e-01 1.14792168e+00 -2.70874232e-01
7.08019435e-01 5.40401280e-01 -2.01745555e-01 -1.33892345e+00
-4.45501298e-01 -4.22835708e-01 -5.29427528e-01 -7.16538310e-01
9.53783572e-01 -1.01685286e+00 -9.11656260e-01 6.05458796e-01
-1.29441500e+00 1.15462348e-01 3.30946334e-02 3.81213516e-01
-9.39516604e-01 3.14407140e-01 -2.45891452e-01 -9.93269622e-01
-4.76479977e-01 -1.35477555e+00 1.56290245e+00 3.66613954e-01
-4.22511965e-01 -4.63525742e-01 1.24070436e-01 3.82714361e-01
4.68506590e-02 1.43154293e-01 2.16159821e-01 -1.07410423e-01
-4.66178149e-01 -1.64522409e-01 2.50140935e-01 2.36023241e-03
2.22393945e-01 1.28038511e-01 -1.25049222e+00 -3.37562919e-01
5.22225499e-02 -2.39477783e-01 1.61148995e-01 7.96178401e-01
8.21580291e-01 -7.21508741e-01 -2.03031003e-01 7.70381868e-01
7.11753130e-01 2.42455274e-01 5.88522136e-01 -3.15919667e-01
7.85121083e-01 1.30947101e+00 6.36673450e-01 8.31995845e-01
4.73323017e-01 9.45404172e-01 2.32593223e-01 2.26837888e-01
-1.50595590e-01 -7.50167787e-01 5.72451353e-01 8.87540042e-01
3.63220751e-01 -1.46130860e-01 -6.54118598e-01 5.17074704e-01
-1.41431689e+00 -9.86572444e-01 3.36320400e-01 2.11950731e+00
7.07722187e-01 -1.46478087e-01 4.55025196e-01 2.68044323e-01
1.10842252e+00 1.62863165e-01 -7.09700227e-01 -6.22481592e-02
6.49715587e-02 -2.32972547e-01 -3.38504285e-01 5.47226667e-01
-6.20079696e-01 8.08290005e-01 5.35351753e+00 8.45776856e-01
-1.34295654e+00 -1.52172402e-01 4.82033283e-01 -3.13020229e-01
-2.37583816e-01 -3.41852248e-01 -1.02150071e+00 5.75121045e-01
6.85210764e-01 -5.03576100e-01 7.22651929e-02 8.80426705e-01
7.43290246e-01 1.66560620e-01 -1.12309837e+00 1.47015262e+00
6.73582137e-01 -9.09386337e-01 1.03089429e-01 -7.49527961e-02
4.75339621e-01 -6.91952527e-01 8.80178288e-02 -2.07040161e-02
-3.23362499e-01 -8.86107683e-01 9.36336040e-01 7.07202911e-01
1.16051018e+00 -9.80776429e-01 7.43733346e-02 4.83050525e-01
-1.59047401e+00 -2.33051002e-01 -5.31713590e-02 3.04107487e-01
6.85320437e-01 6.97890967e-02 -9.62428212e-01 1.60644874e-01
7.45002925e-01 6.41389966e-01 -2.20990404e-01 1.01536918e+00
-3.98848876e-02 1.75702542e-01 -9.07519385e-02 1.37275845e-01
-3.33588928e-01 2.06797704e-01 9.31946337e-01 1.05590999e+00
4.75912243e-01 1.22459777e-01 2.31986433e-01 5.77091098e-01
4.89589721e-02 3.12823355e-01 -7.57166624e-01 4.45171148e-01
8.10859025e-01 1.16938102e+00 -2.62903929e-01 -1.76789910e-01
-1.84481531e-01 8.95739257e-01 -3.89865845e-01 1.63817018e-01
-8.37685466e-01 -3.75931144e-01 8.08934867e-01 4.21127260e-01
9.10492614e-02 1.23552315e-01 3.09532344e-01 -8.55017185e-01
4.11994457e-02 -9.79751229e-01 -1.63465180e-02 -1.13083637e+00
-7.87016094e-01 7.36681759e-01 1.01069808e-02 -1.57035899e+00
-9.33385015e-01 3.14990841e-02 -8.39712977e-01 7.13364780e-01
-1.04461014e+00 -1.02069318e+00 -6.54473960e-01 9.57267940e-01
1.15632665e+00 2.09703092e-02 4.98622864e-01 2.35574469e-01
-6.28275633e-01 9.57246304e-01 -3.54042947e-01 -6.70661330e-02
9.78704214e-01 -6.91657901e-01 2.80478418e-01 3.91082734e-01
-3.85230273e-01 5.03868878e-01 1.16341913e+00 -4.57453132e-01
-1.61408770e+00 -9.62324142e-01 7.91484356e-01 5.47556058e-02
1.89247712e-01 -4.74418670e-01 -9.51512575e-01 3.11409205e-01
1.85421035e-01 -5.61015159e-02 4.98898625e-01 -6.21178687e-01
-2.16126591e-02 -3.40020716e-01 -1.01088560e+00 6.19513392e-01
9.57463384e-01 -4.97398674e-01 -5.85805655e-01 1.21958725e-01
7.14871645e-01 -4.13988918e-01 -5.38215160e-01 1.32946759e-01
7.02721179e-01 -1.05280530e+00 9.26619887e-01 1.12238832e-01
2.92104691e-01 -1.63951263e-01 -7.22128106e-03 -1.18742228e+00
4.09172336e-03 -1.27919269e+00 -1.86405331e-01 1.58261061e+00
-1.60524845e-01 -1.95028409e-01 7.39849150e-01 4.57865387e-01
-1.73688620e-01 -3.84370297e-01 -9.08618510e-01 -5.27376294e-01
-3.88434500e-01 -2.32628495e-01 7.45751798e-01 5.71763456e-01
1.33157507e-01 4.94637489e-01 -8.15020144e-01 -2.15034164e-03
8.05315793e-01 -7.58212153e-03 1.30001867e+00 -1.19383204e+00
-6.97733536e-02 -1.11823507e-01 -4.54808354e-01 -1.30636084e+00
3.71217132e-01 -2.49993473e-01 5.03525436e-01 -1.10752642e+00
1.78726226e-01 2.74328858e-01 6.19878590e-01 -8.51934329e-02
-1.95355698e-01 6.84507862e-02 2.66732186e-01 3.33478987e-01
-4.13102746e-01 1.07014132e+00 1.51422477e+00 6.80879951e-02
-4.68040645e-01 4.12372053e-02 -5.27892351e-01 9.16625917e-01
3.38427663e-01 -3.20622399e-02 -6.18166268e-01 -7.68360421e-02
-2.20419437e-01 8.42262149e-01 2.47951001e-01 -1.22194326e+00
5.08235157e-01 1.79760382e-02 2.47809306e-01 -8.34556043e-01
8.80470693e-01 -5.96374035e-01 2.35876143e-01 2.31977195e-01
-2.39197522e-01 -5.03280610e-02 1.10848881e-02 5.26663184e-01
-2.87904710e-01 2.36390740e-01 9.09990728e-01 2.23650523e-02
-5.74151337e-01 6.12083673e-01 -1.84196904e-01 1.72188163e-01
1.08195293e+00 -5.06676137e-01 2.30457455e-01 -1.06705427e+00
-7.57873893e-01 2.87537336e-01 4.06748861e-01 7.54891515e-01
8.78619134e-01 -1.42590177e+00 -8.38426054e-01 5.84497094e-01
1.78408504e-01 2.10673735e-01 6.76628292e-01 5.93088388e-01
-1.57164812e-01 2.37811685e-01 -1.53448358e-01 -1.06274855e+00
-1.49747932e+00 4.71307337e-01 2.24939182e-01 6.58468843e-01
-6.24346137e-01 9.79208767e-01 6.58573389e-01 -9.08535942e-02
6.56914234e-01 1.35053828e-01 -4.44637060e-01 4.14498389e-01
9.98033702e-01 4.98427451e-01 -2.23265246e-01 -1.40847003e+00
-2.34104410e-01 8.89210165e-01 2.70106401e-02 -2.44004235e-01
1.07156122e+00 -6.84328020e-01 2.60293394e-01 4.74873483e-01
1.57864022e+00 8.60056356e-02 -1.61706579e+00 -1.46634355e-01
-5.33221364e-01 -6.63248301e-01 -2.33578607e-01 3.28716226e-02
-1.05463088e+00 1.03691041e+00 6.65414691e-01 -2.12139785e-01
1.29331386e+00 1.43391922e-01 1.07001936e+00 -1.81501936e-02
3.87619615e-01 -9.30638671e-01 4.51634586e-01 2.14067951e-01
1.18447506e+00 -9.92010355e-01 -4.28141683e-01 -4.69992757e-01
-9.32228804e-01 9.72336173e-01 8.07781219e-01 2.86155283e-01
4.71533567e-01 1.91132784e-01 6.78598061e-02 -2.18713582e-02
-7.44738698e-01 2.89564747e-02 3.05890232e-01 8.55479479e-01
2.84143806e-01 -1.52311087e-01 2.16696426e-01 4.72298980e-01
-8.65305305e-01 1.30034760e-02 3.10654819e-01 6.82625651e-01
-6.12041235e-01 -4.93188977e-01 -1.00288737e+00 -5.79667054e-02
-1.43961608e-01 1.56360909e-01 -2.97604978e-01 2.98238188e-01
-6.74293339e-02 1.08701062e+00 1.28540799e-01 -5.15771985e-01
4.71816689e-01 2.33830929e-01 3.24720114e-01 -4.37341005e-01
-4.02987935e-02 5.69233179e-01 -4.48258579e-01 -5.65293491e-01
-1.41503856e-01 -7.57950544e-01 -1.13756871e+00 -4.65856582e-01
-3.67442369e-01 1.10287309e-01 4.14208263e-01 7.21671760e-01
4.09511119e-01 3.20806801e-01 1.15158522e+00 -1.74221432e+00
-3.11125547e-01 -1.07179058e+00 -6.35653913e-01 4.73167032e-01
6.46279633e-01 -7.79141903e-01 -6.40406370e-01 3.82846653e-01] | [13.225131034851074, -0.42579248547554016] |
7fcd09e0-fd92-4be5-82a0-616d044ae05a | dformer-diffusion-guided-transformer-for | 2306.03437 | null | https://arxiv.org/abs/2306.03437v2 | https://arxiv.org/pdf/2306.03437v2.pdf | DFormer: Diffusion-guided Transformer for Universal Image Segmentation | This paper introduces an approach, named DFormer, for universal image segmentation. The proposed DFormer views universal image segmentation task as a denoising process using a diffusion model. DFormer first adds various levels of Gaussian noise to ground-truth masks, and then learns a model to predict denoising masks from corrupted masks. Specifically, we take deep pixel-level features along with the noisy masks as inputs to generate mask features and attention masks, employing diffusion-based decoder to perform mask prediction gradually. At inference, our DFormer directly predicts the masks and corresponding categories from a set of randomly-generated masks. Extensive experiments reveal the merits of our proposed contributions on different image segmentation tasks: panoptic segmentation, instance segmentation, and semantic segmentation. Our DFormer outperforms the recent diffusion-based panoptic segmentation method Pix2Seq-D with a gain of 3.6% on MS COCO val2017 set. Further, DFormer achieves promising semantic segmentation performance outperforming the recent diffusion-based method by 2.2% on ADE20K val set. Our source code and models will be publicly on https://github.com/cp3wan/DFormer | ['Yanwei Pang', 'Fahad Shahbaz Khan', 'Jin Xie', 'Rao Muhammad Anwer', 'Jiale Cao', 'Hefeng Wang'] | 2023-06-06 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 4.95569289e-01 1.33570224e-01 -1.12185948e-01 -5.73233783e-01
-9.72185552e-01 -7.23282933e-01 4.80302483e-01 -4.50255185e-01
-4.78281111e-01 3.69240850e-01 8.95989761e-02 -2.51304835e-01
3.91111046e-01 -7.89544880e-01 -8.56372833e-01 -7.08365381e-01
2.93345183e-01 5.17761767e-01 3.44610482e-01 1.57602042e-01
1.07031219e-01 1.38485491e-01 -1.12980330e+00 2.69897938e-01
1.17897868e+00 1.23379409e+00 6.08575702e-01 1.01377058e+00
-1.89012751e-01 4.80212539e-01 -5.81999183e-01 -4.40173239e-01
6.71822011e-01 -4.52255934e-01 -7.87488341e-01 3.42421740e-01
5.72214007e-01 -4.72564280e-01 -3.66928339e-01 1.41820526e+00
3.10685933e-01 -2.47934051e-02 6.31837845e-01 -7.18549669e-01
-8.18857431e-01 8.40602517e-01 -8.89458001e-01 3.47149849e-01
-5.46866596e-01 6.97805822e-01 9.95775163e-01 -6.22346580e-01
4.97857839e-01 1.27494431e+00 5.96645653e-01 6.38756394e-01
-1.42922926e+00 -7.54456639e-01 2.68113762e-01 -2.94925660e-01
-1.30136979e+00 -3.36407244e-01 2.55517721e-01 -4.34090137e-01
7.11883903e-01 3.02164942e-01 4.46957052e-01 8.70712161e-01
6.05087318e-02 1.17667484e+00 1.24625778e+00 1.40944421e-01
1.60096854e-01 -2.42993578e-01 8.25772882e-02 6.76923752e-01
-6.33646622e-02 -2.62097884e-02 9.22740158e-03 2.50797659e-01
1.01123822e+00 -1.88239336e-01 -2.23908976e-01 5.31764805e-01
-1.19390488e+00 7.61632979e-01 6.49351358e-01 6.01933114e-02
-6.55427754e-01 3.85940433e-01 9.34951976e-02 1.06225707e-01
1.02042603e+00 3.02516788e-01 -6.14735544e-01 1.77217394e-01
-1.38550580e+00 2.00267091e-01 6.45974100e-01 7.11405873e-01
9.94624376e-01 7.27214739e-02 -5.83182156e-01 9.09151256e-01
2.85606980e-01 7.59194970e-01 2.99758166e-01 -1.31597769e+00
3.72987419e-01 1.98727950e-01 5.76017089e-02 -7.57477462e-01
-9.40833017e-02 -5.78281999e-01 -9.29421604e-01 -9.18783024e-02
2.81908691e-01 -6.24263585e-01 -1.73297334e+00 1.58322334e+00
3.59641194e-01 5.51582754e-01 -1.64554462e-01 1.09574759e+00
8.27643752e-01 1.13530815e+00 -2.95380764e-02 2.70113617e-01
1.11453164e+00 -1.41134846e+00 -5.05924582e-01 -4.27591711e-01
1.76747575e-01 -8.01493406e-01 8.43960285e-01 4.44150925e-01
-1.03617179e+00 -6.36532128e-01 -7.09662437e-01 -3.16285603e-02
-2.11582154e-01 3.17069441e-01 3.96717638e-01 4.66532320e-01
-1.27585852e+00 5.41491807e-01 -1.04936624e+00 -5.65986224e-02
9.05266881e-01 3.24680209e-01 2.34829500e-01 -5.49942441e-02
-8.79769266e-01 3.49159390e-01 2.94941127e-01 3.80232900e-01
-1.27248108e+00 -7.44541109e-01 -7.68573046e-01 -1.20658269e-02
5.02474904e-01 -6.20767415e-01 1.38379514e+00 -1.11227739e+00
-1.48611474e+00 1.03742433e+00 -2.64821619e-01 -8.54654849e-01
6.68415308e-01 -3.81476313e-01 -1.67025715e-01 2.45513394e-01
3.43022108e-01 1.52011633e+00 9.74660993e-01 -1.25087106e+00
-7.38818467e-01 -9.59520042e-02 -2.16932312e-01 7.33828638e-03
1.80372477e-01 -3.79602283e-01 -1.01378918e+00 -8.86076748e-01
1.52416185e-01 -7.47604907e-01 -7.40311205e-01 6.29196502e-03
-9.83020663e-01 -1.67471152e-02 5.91381431e-01 -8.94697726e-01
9.94587779e-01 -2.06634974e+00 2.30523378e-01 1.65338784e-01
4.17451143e-01 4.20919687e-01 -5.28983533e-01 -3.06953222e-01
3.50853026e-01 4.62850183e-01 -1.06352735e+00 -6.20809317e-01
8.98537338e-02 3.87065500e-01 -3.43744934e-01 3.32767993e-01
3.57087821e-01 1.11055148e+00 -6.83328152e-01 -3.46449524e-01
4.34940428e-01 4.81835634e-01 -5.60658276e-01 3.22942108e-01
-8.77211690e-01 6.13362372e-01 -3.63559663e-01 7.39107311e-01
1.11812901e+00 -2.12913483e-01 -3.50673348e-01 -4.49240841e-02
-2.00325936e-01 3.63802090e-02 -9.22845066e-01 1.78181934e+00
-2.30379328e-01 6.28307223e-01 5.57970345e-01 -9.31819260e-01
8.26481402e-01 -1.07118011e-01 2.82747120e-01 -7.07644641e-01
2.21148729e-01 2.04735488e-01 -2.90950149e-01 -2.98446923e-01
5.14212489e-01 1.59095556e-01 -1.79625768e-02 2.75705904e-01
1.53245986e-01 -5.18822610e-01 2.06990451e-01 1.83455989e-01
8.76831949e-01 2.54560500e-01 -3.58967841e-01 -3.91932070e-01
3.73523533e-01 5.60527146e-02 8.38153839e-01 9.72429931e-01
-3.27806503e-01 9.79368925e-01 6.66872621e-01 -1.64392292e-01
-8.26541841e-01 -1.12059581e+00 -1.67785883e-01 7.15157688e-01
2.27180824e-01 -1.10783122e-01 -1.23266399e+00 -7.66253650e-01
7.46467337e-02 6.31777167e-01 -8.27946305e-01 3.87476593e-01
-4.09211010e-01 -1.01139760e+00 6.62886858e-01 2.16528282e-01
1.01185930e+00 -1.06571233e+00 -1.77575171e-01 8.37343261e-02
-1.67290315e-01 -1.14633572e+00 -8.85506034e-01 1.48982346e-01
-7.81756103e-01 -8.06262672e-01 -9.71839786e-01 -7.96072304e-01
5.15232921e-01 1.53020889e-01 1.04266262e+00 2.27405876e-02
-3.51769656e-01 -7.66251516e-03 -7.36464038e-02 -1.24941140e-01
-2.96740860e-01 3.15957934e-01 -6.19923830e-01 2.54450679e-01
1.85782671e-01 -3.83565634e-01 -1.21452451e+00 2.79981583e-01
-1.14555681e+00 2.48301476e-01 5.94972193e-01 5.60617924e-01
1.20327842e+00 -4.26809378e-02 3.38204175e-01 -1.22859907e+00
3.00409079e-01 -8.07299078e-01 -9.72388744e-01 -2.24780990e-03
-4.16052848e-01 -1.16876163e-01 4.31692511e-01 -1.34690344e-01
-1.16176903e+00 1.53264925e-01 -6.02153122e-01 -5.67805409e-01
-3.01584810e-01 2.03569055e-01 -1.11459412e-01 3.03810388e-01
3.63589048e-01 2.20934197e-01 -1.39101490e-01 -7.77455032e-01
6.74476564e-01 7.58109450e-01 8.39244306e-01 -3.74986768e-01
5.70694923e-01 5.62491357e-01 -4.87964392e-01 -7.53780663e-01
-1.10099781e+00 -3.44465435e-01 -4.79348868e-01 -3.42453681e-02
1.29964519e+00 -1.10089409e+00 -1.64373860e-01 9.53740299e-01
-1.02057838e+00 -1.09192240e+00 -2.73391724e-01 7.50561804e-03
-3.61720413e-01 9.67258736e-02 -9.16602910e-01 -4.73209828e-01
-5.13320208e-01 -1.46239948e+00 1.23300636e+00 5.16964912e-01
1.70658052e-01 -9.16084647e-01 -1.31634355e-01 6.30429089e-01
5.35900533e-01 2.26541340e-01 4.50819224e-01 -3.28619987e-01
-9.76505995e-01 1.01808995e-01 -5.52927256e-01 8.64272654e-01
2.40521118e-01 6.95627034e-02 -1.08458900e+00 1.17196590e-01
-1.72063768e-01 -2.43477166e-01 1.64940155e+00 9.88445163e-01
1.75851154e+00 -1.99087679e-01 -2.91753054e-01 1.07157540e+00
1.49472535e+00 -1.56432092e-02 6.30808532e-01 -1.22098602e-01
7.92968750e-01 4.64835703e-01 2.55177200e-01 2.50753760e-01
4.76983219e-01 1.10612750e-01 6.15262568e-01 -4.59884375e-01
-4.86147791e-01 -1.72111258e-01 1.24315143e-01 6.03340626e-01
4.25677299e-01 -6.67658031e-01 -8.51603210e-01 9.04315650e-01
-1.68204117e+00 -4.83262897e-01 -2.67318904e-01 1.67622232e+00
9.01071846e-01 2.18255535e-01 -4.94399555e-02 -5.69945395e-01
9.24284697e-01 5.20635009e-01 -1.00203323e+00 -2.45022774e-01
-3.51990104e-01 5.30301332e-01 8.83038223e-01 8.96179020e-01
-1.48049355e+00 1.65773618e+00 5.42650938e+00 1.10483885e+00
-1.22155976e+00 2.83684313e-01 1.42205215e+00 -9.28397849e-02
-3.60561818e-01 -1.26538783e-01 -6.85420632e-01 7.00060129e-01
8.38386238e-01 3.21606636e-01 5.69476604e-01 2.99284309e-01
4.06564891e-01 -2.61785597e-01 -4.48892057e-01 7.20749140e-01
-2.20828876e-01 -1.49752581e+00 4.55462895e-02 -7.82419965e-02
1.19758987e+00 8.57124090e-01 4.22525287e-01 1.51604071e-01
7.91723073e-01 -1.15866518e+00 6.44725680e-01 5.70974588e-01
7.47157693e-01 -4.95238304e-01 6.07062995e-01 1.56364411e-01
-8.95697176e-01 2.60086477e-01 -4.07962829e-01 2.46119991e-01
2.58087575e-01 9.94340599e-01 -3.75506878e-01 2.38952696e-01
7.04653263e-01 9.68595028e-01 -4.75842208e-01 1.01851630e+00
-3.52825314e-01 1.31630397e+00 -5.23338735e-01 5.49139977e-01
7.20161617e-01 -4.64073986e-01 3.82793665e-01 1.46181059e+00
1.32455915e-01 1.78230822e-01 1.95400417e-01 1.27548754e+00
-5.57376206e-01 -1.56143188e-01 -1.08083837e-01 -1.09804899e-01
3.09522659e-01 1.39129579e+00 -1.07530940e+00 -5.77507496e-01
-2.49627322e-01 1.42102540e+00 -7.73498788e-02 5.11367202e-01
-9.36322212e-01 -2.34983593e-01 8.71936560e-01 -2.25444555e-01
8.21758866e-01 1.64769739e-02 -6.24134243e-01 -1.15750015e+00
-3.09060305e-01 -7.22970366e-01 2.25546658e-01 -7.15620279e-01
-1.22385299e+00 7.12578177e-01 -2.80449092e-01 -6.44008458e-01
3.79359871e-01 -3.52994084e-01 -7.15653658e-01 9.64022458e-01
-1.58538091e+00 -1.03614664e+00 -1.17911272e-01 1.52643189e-01
1.04080153e+00 2.00194687e-01 2.67727464e-01 2.47165754e-01
-9.53193307e-01 2.01283768e-01 2.93110818e-01 2.21518993e-01
4.00011599e-01 -1.46115267e+00 1.07630670e+00 1.05228174e+00
1.52917400e-01 1.86181557e-03 5.57895720e-01 -7.21415579e-01
-8.15677702e-01 -1.78289497e+00 4.43841934e-01 -2.89237857e-01
5.63061237e-01 -3.94324243e-01 -9.57500041e-01 6.89817071e-01
5.68794549e-01 1.54566795e-01 9.57960933e-02 -6.14602864e-01
-2.36149952e-02 -1.41282693e-01 -1.23682058e+00 5.50589383e-01
1.06205666e+00 -1.50496230e-01 -1.69709787e-01 4.33411658e-01
1.17104459e+00 -5.62595069e-01 -6.34821177e-01 2.83906668e-01
2.19036192e-01 -7.95305550e-01 7.59850025e-01 -1.03665786e-02
7.52852261e-01 -3.07336301e-01 -1.81209460e-01 -1.39150238e+00
-1.31612927e-01 -7.89794683e-01 1.85410261e-01 1.26852810e+00
7.21158028e-01 -5.90457082e-01 8.22387338e-01 1.37219369e-01
-2.07248583e-01 -8.05394113e-01 -7.86637485e-01 -3.69486183e-01
4.59774256e-01 -6.55949831e-01 7.07399726e-01 6.03092074e-01
-1.19456565e+00 3.92071754e-02 -2.86276937e-01 3.73785943e-01
6.84911430e-01 2.07128197e-01 5.61875403e-01 -8.01890910e-01
-3.33703309e-01 -6.04736209e-01 2.50200123e-01 -1.83235395e+00
1.19790271e-01 -1.07277715e+00 4.02974427e-01 -1.67955136e+00
-6.66425452e-02 -3.71938229e-01 -3.12014222e-01 4.28338438e-01
-3.40688050e-01 6.74611509e-01 1.54556453e-01 6.71246871e-02
-5.87483883e-01 5.80140412e-01 1.59725475e+00 -4.81303453e-01
-2.53236115e-01 9.45141166e-02 -7.60525525e-01 6.74284875e-01
1.06577587e+00 -5.71806312e-01 -2.29575351e-01 -1.03995669e+00
-2.58300304e-01 -1.57699227e-01 4.06066924e-01 -7.48246849e-01
3.50697599e-02 -1.17153637e-01 2.55620748e-01 -7.55586565e-01
2.04907581e-01 -3.91265422e-01 -1.03536874e-01 4.78180408e-01
-2.59400576e-01 -1.84553206e-01 3.52829427e-01 5.70470989e-01
-1.35895178e-01 -3.97806661e-03 8.13144803e-01 -3.03572088e-01
-8.90170097e-01 5.25485456e-01 -5.27277946e-01 3.98604929e-01
7.34882236e-01 -3.01593728e-02 -3.05334121e-01 -1.09519176e-01
-9.00508225e-01 6.98123634e-01 3.03569764e-01 3.16554338e-01
4.80096638e-01 -6.36588156e-01 -9.55290139e-01 2.67604172e-01
-5.33395946e-01 6.44396186e-01 3.27892601e-01 8.42625916e-01
-7.99954176e-01 2.44311929e-01 2.18862310e-01 -8.86676669e-01
-7.46621788e-01 1.07168173e-02 5.96081376e-01 -6.49265125e-02
-6.67360961e-01 1.39739811e+00 6.02137685e-01 -5.99287212e-01
1.40915990e-01 -7.61603951e-01 2.31996879e-01 -4.30694856e-02
2.36511290e-01 1.21679567e-01 -2.76927173e-01 -4.87976164e-01
-9.46792588e-02 4.83184963e-01 -7.70997107e-02 -2.14575931e-01
1.38716710e+00 -3.07819068e-01 -1.95021316e-01 1.70292884e-01
1.03031886e+00 -2.97993988e-01 -1.99240148e+00 -4.81463015e-01
-1.07566483e-01 -3.55163693e-01 5.02789259e-01 -1.22261918e+00
-1.75657392e+00 8.70489955e-01 6.50322795e-01 -1.38776943e-01
1.28612173e+00 8.55681598e-02 1.30762911e+00 -1.61866501e-01
-2.14693800e-01 -1.01042449e+00 -2.23160386e-01 4.51025993e-01
6.80200636e-01 -1.33572257e+00 -4.41368610e-01 -4.91095394e-01
-7.32854605e-01 5.27661204e-01 6.83576941e-01 -3.18439007e-01
7.98161566e-01 4.16515529e-01 2.87233740e-01 -1.59650162e-01
-6.56214714e-01 -5.41513622e-01 2.25481421e-01 3.55027556e-01
6.58893287e-02 3.88055742e-01 1.60586985e-03 6.97645903e-01
-1.89153299e-01 -3.03629618e-02 2.05869853e-01 3.24620485e-01
-5.91791809e-01 -7.85203099e-01 -3.28065902e-01 7.49427974e-01
-6.22145951e-01 -4.67560887e-01 -3.28874797e-01 4.22600389e-01
5.57445467e-01 9.50911403e-01 4.92534995e-01 -2.73625225e-01
5.60920350e-02 -1.84043199e-01 1.30997777e-01 -6.17465615e-01
-6.09362423e-01 3.99161607e-01 -2.29259953e-01 -6.42154396e-01
-3.68271738e-01 -6.09918416e-01 -1.19377613e+00 -2.45969355e-01
1.88062061e-02 -1.44225694e-02 5.48573196e-01 8.72086942e-01
3.69686306e-01 6.11907482e-01 4.93966103e-01 -8.43368948e-01
-2.80382574e-01 -1.03417981e+00 -3.68817359e-01 1.69206947e-01
5.10319889e-01 -1.11125819e-01 -3.48199278e-01 2.23868892e-01] | [9.560737609863281, 0.28495633602142334] |
27fba2a0-698b-4a77-812d-f7a91452e4e1 | wman-weakly-supervised-moment-alignment | null | null | https://openreview.net/forum?id=BJx4rerFwB | https://openreview.net/pdf?id=BJx4rerFwB | wMAN: WEAKLY-SUPERVISED MOMENT ALIGNMENT NETWORK FOR TEXT-BASED VIDEO SEGMENT RETRIEVAL | Given a video and a sentence, the goal of weakly-supervised video moment retrieval is to locate the video segment which is described by the sentence without having access to temporal annotations during training. Instead, a model must learn how to identify the correct segment (i.e. moment) when only being provided with video-sentence pairs. Thus, an inherent challenge is automatically inferring the latent correspondence between visual and language representations. To facilitate this alignment, we propose our Weakly-supervised Moment Alignment Network (wMAN) which exploits a multi-level co-attention mechanism to learn richer multimodal representations. The aforementioned mechanism is comprised of a Frame-By-Word interaction module as well as a novel Word-Conditioned Visual Graph (WCVG). Our approach also incorporates a novel application of positional encodings, commonly used in Transformers, to learn visual-semantic representations that contain contextual information of their relative positions in the temporal sequence through iterative message-passing. Comprehensive experiments on the DiDeMo and Charades-STA datasets demonstrate the effectiveness of our learned representations: our combined wMAN model not only outperforms the state-of-the-art weakly-supervised method by a significant margin but also does better than strongly-supervised state-of-the-art methods on some metrics. | ['Bryan A. Plummer', 'Kate Saenko', 'Huijuan Xu', 'Reuben Tan'] | 2019-09-25 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [ 1.83601588e-01 -1.37332186e-01 -5.64734459e-01 -5.00543356e-01
-9.05218959e-01 -4.85461086e-01 8.48118722e-01 2.28005335e-01
-5.30862927e-01 3.17754924e-01 4.82066810e-01 -6.49892911e-02
2.65902787e-01 -4.62203264e-01 -9.68273044e-01 -4.62968379e-01
-1.86373711e-01 2.22424030e-01 1.87080309e-01 -6.20931499e-02
2.25279450e-01 5.29490896e-02 -1.37897992e+00 7.68751919e-01
4.25150573e-01 8.87554288e-01 5.19736409e-01 6.46901906e-01
-2.59453267e-01 1.21992743e+00 -1.83768973e-01 -2.99099416e-01
-1.07216410e-01 -6.87134504e-01 -9.88068938e-01 4.27340120e-01
4.99936849e-01 -4.10862833e-01 -6.55750394e-01 7.82577455e-01
1.33580267e-01 3.37480903e-01 6.99869871e-01 -1.11594021e+00
-5.89505672e-01 5.65860271e-01 -7.18883574e-01 5.23092210e-01
6.19360209e-01 1.68972686e-01 1.20038009e+00 -9.23131287e-01
9.04473841e-01 1.11112165e+00 1.96454659e-01 3.65139157e-01
-1.16962171e+00 -3.02810848e-01 5.39738655e-01 7.36339211e-01
-1.40030348e+00 -5.37514210e-01 1.03849006e+00 -4.91074353e-01
1.01861703e+00 -4.01008502e-02 6.93723977e-01 1.34232688e+00
-5.82873002e-02 1.03949499e+00 5.78070819e-01 -3.27557772e-01
4.20873798e-02 -1.43570349e-01 -8.18848833e-02 1.10682356e+00
-4.90008324e-01 -2.54497498e-01 -9.45343018e-01 1.70833334e-01
6.86909616e-01 5.24383299e-02 -3.06847215e-01 -5.39964497e-01
-1.62790632e+00 6.73387706e-01 3.17280293e-01 4.46803033e-01
-3.22589368e-01 3.73924077e-01 5.76376379e-01 2.23072574e-01
6.34502053e-01 1.15587495e-01 -2.08633631e-01 -3.88405889e-01
-1.07285857e+00 -5.97439148e-02 5.29803872e-01 8.47388744e-01
7.51835525e-01 -9.75756720e-02 -3.69419873e-01 5.09847283e-01
4.41939175e-01 2.12255046e-01 3.57880235e-01 -9.05710101e-01
7.15533018e-01 5.10591149e-01 -3.61915007e-02 -1.20025408e+00
-1.18915319e-01 -2.78699309e-01 -5.67022204e-01 -3.45083833e-01
3.64596158e-01 3.08374226e-01 -7.95445383e-01 1.89143384e+00
2.61537641e-01 7.37918079e-01 -3.13567594e-02 1.01859617e+00
7.72923946e-01 9.44875240e-01 2.13841170e-01 -3.80781144e-01
1.32627118e+00 -1.09566402e+00 -7.60985553e-01 -4.29924339e-01
6.44423485e-01 -6.33646250e-01 1.01706553e+00 -1.03875175e-01
-1.14472139e+00 -6.88618898e-01 -1.01169872e+00 -2.92743295e-01
-1.63741395e-01 1.20440155e-01 4.53906208e-01 -8.79055485e-02
-9.56417024e-01 4.82890427e-01 -9.45926189e-01 -5.49390495e-01
2.87186384e-01 -9.93238911e-02 -5.45428336e-01 -1.53696358e-01
-1.04476094e+00 6.89426482e-01 2.85444677e-01 1.00169569e-01
-1.09637880e+00 -5.04771650e-01 -1.31489611e+00 4.18321975e-02
4.62594807e-01 -8.02727044e-01 1.15181375e+00 -1.33177459e+00
-1.40572822e+00 1.07477510e+00 -6.46700680e-01 -5.06995976e-01
3.83990079e-01 -2.61877060e-01 -3.03867996e-01 8.90215874e-01
1.40600845e-01 8.73304188e-01 1.04201233e+00 -1.22466874e+00
-4.88626480e-01 -1.00739539e-01 3.69260103e-01 3.63028556e-01
-3.06121826e-01 5.14194444e-02 -9.11273956e-01 -6.35123312e-01
7.37741068e-02 -7.18385458e-01 4.45061922e-02 2.15769023e-01
-3.08850437e-01 -2.96757132e-01 8.24306667e-01 -8.29364955e-01
1.28397167e+00 -2.20566082e+00 5.44013083e-01 4.31314204e-03
1.68815792e-01 -2.61314940e-02 -4.50363904e-01 6.60149515e-01
-1.24865368e-01 -1.26037002e-01 -1.01581134e-01 -7.72026181e-01
-9.73907411e-02 1.84304282e-01 -4.37066525e-01 7.83336103e-01
3.59346926e-01 9.78798211e-01 -1.27661276e+00 -7.79405236e-01
3.73638421e-01 5.38013935e-01 -4.42840189e-01 5.75295448e-01
-4.63635832e-01 6.37623906e-01 -2.90129095e-01 5.82341313e-01
1.01975344e-01 -7.04098821e-01 2.65130788e-01 -5.70267797e-01
-6.32449985e-02 3.00806314e-01 -6.49667978e-01 2.46081519e+00
-4.35491085e-01 8.85799646e-01 -3.99374187e-01 -1.28683472e+00
4.86555874e-01 4.11991656e-01 5.47701180e-01 -7.29759276e-01
1.55032203e-02 -2.17641845e-01 -4.00714844e-01 -9.39999759e-01
3.86376470e-01 2.97913373e-01 -7.70531297e-02 5.53669453e-01
2.79091686e-01 4.53660905e-01 3.76450717e-01 7.31594026e-01
9.73976254e-01 5.31862199e-01 2.58904040e-01 3.38148102e-02
5.42213023e-01 -2.03556627e-01 4.31005508e-01 6.32334352e-01
-2.24861935e-01 6.00602865e-01 6.67375028e-01 -3.42597812e-01
-9.68395710e-01 -9.44064498e-01 4.05187786e-01 1.32915413e+00
2.61306614e-01 -7.73231089e-01 -5.94338477e-01 -7.62199461e-01
-3.39182556e-01 5.21854579e-01 -7.45212376e-01 -1.26824901e-01
-5.12836993e-01 4.59190784e-03 2.35706076e-01 5.55848241e-01
3.99629027e-01 -8.30453992e-01 -5.32098711e-01 1.58675313e-02
-6.75283074e-01 -1.54489338e+00 -7.24117100e-01 -2.76634961e-01
-6.76312208e-01 -1.08407402e+00 -6.82904005e-01 -8.36354971e-01
8.32450390e-01 5.09169519e-01 1.13323021e+00 1.70283481e-01
-1.84826870e-02 7.74745107e-01 -5.48320293e-01 3.33296835e-01
-1.08427882e-01 -6.14795946e-02 -1.98789403e-01 3.23688179e-01
2.90065318e-01 -5.35351336e-01 -7.85289288e-01 1.54203102e-01
-8.84459257e-01 5.23702025e-01 5.02325118e-01 7.93762684e-01
6.94699705e-01 -5.04757762e-01 2.57031113e-01 -6.44166052e-01
5.53214699e-02 -7.37492561e-01 -3.05896312e-01 3.90851825e-01
-1.63565651e-01 1.64464355e-01 4.21088398e-01 -5.45578718e-01
-8.72816443e-01 2.71147281e-01 1.03348859e-01 -8.06413114e-01
-7.51717910e-02 7.66113758e-01 -1.13227800e-01 3.03668529e-01
1.93617940e-01 5.00905335e-01 -6.31181970e-02 -2.83776253e-01
6.73335433e-01 3.53637338e-01 7.18696535e-01 -4.73208785e-01
7.89173245e-01 7.40061104e-01 -9.75681469e-02 -8.89634073e-01
-1.10781634e+00 -7.52333820e-01 -8.73866856e-01 -4.85055029e-01
1.18804693e+00 -1.21338832e+00 -6.79584980e-01 1.33519173e-01
-1.39747548e+00 -3.67023677e-01 -1.08181760e-02 4.69141036e-01
-7.47496128e-01 7.49729335e-01 -5.94660223e-01 -5.62385023e-01
-1.05458722e-02 -9.50089991e-01 1.17964447e+00 -1.29934803e-01
-2.10091963e-01 -1.02089405e+00 4.93644066e-02 6.01263225e-01
5.64507134e-02 2.15528354e-01 8.48546207e-01 -5.32741666e-01
-7.73599863e-01 -5.05642556e-02 -2.71495163e-01 1.57187041e-02
-1.99134052e-02 -1.75124053e-02 -8.33422899e-01 -3.93162608e-01
-3.20949376e-01 -4.24356222e-01 1.03891766e+00 2.11450055e-01
1.32421136e+00 -4.00371045e-01 -3.24408501e-01 6.76525652e-01
1.18032467e+00 -1.63652226e-01 5.21083891e-01 3.14568788e-01
9.34494078e-01 5.73660016e-01 5.87795198e-01 4.03150797e-01
7.77159214e-01 8.01100910e-01 5.96576035e-01 -1.15673486e-02
-1.06108278e-01 -5.17428398e-01 5.61623454e-01 1.05164182e+00
-8.13287050e-02 -2.52804458e-01 -7.71703899e-01 6.48127913e-01
-2.31111169e+00 -1.33730364e+00 2.49319479e-01 1.92221999e+00
7.58470714e-01 3.95138860e-02 8.43246505e-02 -4.96669151e-02
5.80130100e-01 7.67604470e-01 -4.44177091e-01 3.34798843e-02
3.59803587e-02 -3.44313800e-01 5.80531880e-02 4.34085071e-01
-1.26342666e+00 9.51736212e-01 5.60541296e+00 6.28630519e-01
-1.01946080e+00 2.15372071e-01 5.98383129e-01 -1.71885401e-01
-2.25866452e-01 7.39095062e-02 -4.05505687e-01 4.51466620e-01
9.64453042e-01 3.12996730e-02 5.38353264e-01 6.09362125e-01
4.57330048e-01 -1.08403854e-01 -1.46394515e+00 1.13884366e+00
6.07922077e-01 -1.61443758e+00 2.02698544e-01 -2.52713412e-01
5.76193094e-01 -1.41116664e-01 -7.85022415e-03 2.26324618e-01
-3.26617390e-01 -7.89394617e-01 9.77061987e-01 7.59168923e-01
8.41818392e-01 -6.35482252e-01 4.37615365e-01 2.65775055e-01
-1.64999914e+00 9.85831320e-02 -1.30500272e-01 -3.34374234e-02
3.95740479e-01 3.16510320e-01 -5.87656379e-01 5.37025213e-01
4.90106225e-01 1.33091354e+00 -5.56833506e-01 7.78046012e-01
-4.74258363e-01 5.20950317e-01 1.54850781e-01 1.78581566e-01
4.48814809e-01 -1.26294019e-02 4.99908835e-01 1.43352497e+00
1.84149370e-01 5.27411252e-02 3.86189163e-01 6.66368902e-01
-2.00050741e-01 8.28071237e-02 -7.01009154e-01 -2.83422530e-01
3.36698741e-01 1.14047110e+00 -6.68204248e-01 -4.34948474e-01
-7.61976719e-01 1.41164768e+00 5.58455944e-01 6.18652821e-01
-9.65319872e-01 -8.80793855e-02 5.89855969e-01 7.78394863e-02
4.93950814e-01 -4.74726170e-01 3.38698208e-01 -1.37778735e+00
-6.46221498e-03 -6.04480326e-01 4.93377119e-01 -1.04969299e+00
-1.22882855e+00 5.92856884e-01 4.03218009e-02 -1.50718999e+00
-5.47600567e-01 -2.27473035e-01 -5.67358077e-01 4.98696506e-01
-1.68919480e+00 -1.43679428e+00 -3.19826335e-01 8.35867047e-01
8.92280221e-01 -3.71133536e-02 5.99160910e-01 3.36314976e-01
-4.63088632e-01 3.97910953e-01 -1.70954421e-01 4.17599410e-01
7.60852575e-01 -9.64345574e-01 1.06705293e-01 1.00704336e+00
5.74553132e-01 5.36132455e-01 6.28698289e-01 -5.21820426e-01
-1.65415716e+00 -1.20938396e+00 1.00611246e+00 -4.05500144e-01
8.47844660e-01 -4.47817892e-01 -9.21871185e-01 8.38051677e-01
4.73396629e-01 1.62730768e-01 6.54982865e-01 -1.01645432e-01
-6.31005466e-01 -5.58187924e-02 -4.48182225e-01 6.31472826e-01
1.26774549e+00 -1.21614623e+00 -7.54631519e-01 4.66605574e-01
8.65111828e-01 -4.18450832e-01 -5.10144770e-01 9.35879424e-02
3.77120912e-01 -8.52929294e-01 1.03499484e+00 -7.61014044e-01
7.92084992e-01 -3.33432049e-01 -2.34168768e-01 -1.01250780e+00
-1.84371740e-01 -6.04519367e-01 -6.18586123e-01 1.25417197e+00
2.26315901e-01 4.08750289e-04 5.21818757e-01 1.57227322e-01
-1.19274400e-01 -7.26296067e-01 -9.52157259e-01 -5.30342698e-01
-6.00047827e-01 -6.32012784e-01 1.15988903e-01 1.08145320e+00
2.89143920e-01 5.53789735e-01 -6.49430394e-01 2.54170388e-01
5.32097399e-01 2.50148207e-01 6.70563102e-01 -7.36111224e-01
-3.36158723e-01 -2.84085661e-01 -5.90583861e-01 -1.51573491e+00
6.53269827e-01 -9.73998427e-01 1.75890192e-01 -1.53598058e+00
5.62300324e-01 3.20579618e-01 -4.76849765e-01 3.81639034e-01
-9.21252370e-02 1.77434832e-01 3.09673876e-01 3.21617156e-01
-1.33290911e+00 6.04176998e-01 1.06174386e+00 -3.22352290e-01
8.44443515e-02 -4.05510098e-01 -3.07352960e-01 6.62035227e-01
3.44932288e-01 -3.10084164e-01 -5.63851953e-01 -7.37195671e-01
3.90216887e-01 3.72443169e-01 5.63103855e-01 -7.16700673e-01
4.43558633e-01 -1.08691067e-01 3.29925984e-01 -7.12341666e-01
5.42088807e-01 -7.21015811e-01 -8.38011801e-02 1.96368799e-01
-6.63698316e-01 2.83068627e-01 -6.97029009e-02 9.92745519e-01
-3.96773100e-01 5.81936277e-02 3.51983666e-01 -8.15081671e-02
-1.14555192e+00 4.83606130e-01 -4.30505276e-01 -4.00087312e-02
1.02646184e+00 -8.56521800e-02 -4.20648962e-01 -6.80388570e-01
-6.36341512e-01 2.46836692e-01 4.93333787e-01 5.50547898e-01
9.10203159e-01 -1.33538067e+00 -4.63828415e-01 1.43924644e-02
5.24175227e-01 -3.10730517e-01 3.52564156e-01 9.89097238e-01
-2.57761687e-01 3.51920873e-01 5.15904091e-03 -8.75859857e-01
-1.23701155e+00 6.17814541e-01 6.40725046e-02 -1.41132697e-01
-7.60908902e-01 7.06793487e-01 2.41760284e-01 1.61525771e-01
5.35165191e-01 -2.25921631e-01 -2.90224373e-01 2.28685394e-01
6.50790751e-01 1.08360732e-02 -3.43605310e-01 -1.07238483e+00
-4.23528492e-01 6.04325950e-01 2.91158701e-03 -1.79422855e-01
1.24906445e+00 -4.29405034e-01 -1.05269812e-01 8.45958650e-01
1.50477862e+00 -3.08340222e-01 -1.56219471e+00 -5.02987742e-01
-2.22913045e-02 -4.98402655e-01 1.22161061e-02 -3.80198359e-01
-9.98587728e-01 1.02420461e+00 3.96994054e-01 -1.82335123e-01
9.87071753e-01 3.14028412e-01 7.41268754e-01 4.75139260e-01
1.19080164e-01 -8.74572277e-01 6.65207624e-01 4.94731992e-01
7.33124197e-01 -1.45177352e+00 -9.29575562e-02 -1.73528582e-01
-8.57351065e-01 1.08859551e+00 5.04265904e-01 3.51551501e-03
4.85255927e-01 -2.96583027e-01 1.66767295e-02 -1.81449845e-01
-1.06625533e+00 -3.65505874e-01 6.76257014e-01 3.58726442e-01
4.41094577e-01 -3.21635127e-01 -1.46439016e-01 3.45887363e-01
5.00211418e-01 -8.87017995e-02 2.44355753e-01 1.01849222e+00
-3.12684774e-01 -7.78014898e-01 8.28947425e-02 6.54504970e-02
-3.15648913e-01 -9.88480821e-02 -1.44279897e-01 6.01295710e-01
-2.39302918e-01 9.03723419e-01 3.75440747e-01 -4.23450112e-01
-7.19898287e-03 4.28653099e-02 4.57510054e-01 -6.18241131e-01
-2.41645783e-01 2.36735493e-01 4.55484614e-02 -1.06061399e+00
-9.23521578e-01 -5.16274869e-01 -1.27717996e+00 -1.23804100e-01
1.78746000e-01 1.05735332e-01 5.34956217e-01 1.27723300e+00
3.80195707e-01 5.82556605e-01 6.63179040e-01 -1.18801975e+00
-1.10004313e-01 -7.74869144e-01 -2.32503533e-01 7.72029221e-01
4.92951125e-01 -6.10006928e-01 -2.58952022e-01 5.24487376e-01] | [10.107789039611816, 0.7890902757644653] |
35a26d98-14ce-44b4-b868-d26f0fa4d6a9 | label-penet-sequential-label-propagation-and | 1910.02624 | null | https://arxiv.org/abs/1910.02624v3 | https://arxiv.org/pdf/1910.02624v3.pdf | Label-PEnet: Sequential Label Propagation and Enhancement Networks for Weakly Supervised Instance Segmentation | Weakly-supervised instance segmentation aims to detect and segment object instances precisely, given imagelevel labels only. Unlike previous methods which are composed of multiple offline stages, we propose Sequential Label Propagation and Enhancement Networks (referred as Label-PEnet) that progressively transform image-level labels to pixel-wise labels in a coarse-to-fine manner. We design four cascaded modules including multi-label classification, object detection, instance refinement and instance segmentation, which are implemented sequentially by sharing the same backbone. The cascaded pipeline is trained alternatively with a curriculum learning strategy that generalizes labels from high-level images to low-level pixels gradually with increasing accuracy. In addition, we design a proposal calibration module to explore the ability of classification networks to find key pixels that identify object parts, which serves as a post validation strategy running in the inverse order. We evaluate the efficiency of our Label-PEnet in mining instance masks on standard benchmarks: PASCAL VOC 2007 and 2012. Experimental results show that Label-PEnet outperforms the state-of-the-art algorithms by a clear margin, and obtains comparable performance even with the fully-supervised approaches. | ['Weilin Huang', 'Sheng Guo', 'Matthew R. Scott', 'Weifeng Ge'] | 2019-10-07 | label-penet-sequential-label-propagation-and-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Ge_Label-PEnet_Sequential_Label_Propagation_and_Enhancement_Networks_for_Weakly_Supervised_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Ge_Label-PEnet_Sequential_Label_Propagation_and_Enhancement_Networks_for_Weakly_Supervised_ICCV_2019_paper.pdf | iccv-2019-10 | ['weakly-supervised-instance-segmentation', 'image-level-supervised-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 7.51461267e-01 4.07421380e-01 -4.50840175e-01 -6.74472034e-01
-9.00008738e-01 -6.05223298e-01 4.29241300e-01 2.08938271e-01
-6.22665882e-01 5.03185749e-01 -6.14173710e-01 -9.96276736e-02
2.21809745e-01 -6.78752244e-01 -9.69685376e-01 -6.47575498e-01
1.41138017e-01 6.62189186e-01 6.92866623e-01 4.72383529e-01
2.26056740e-01 3.92025858e-01 -1.60062206e+00 5.84865510e-01
7.12297618e-01 1.19075477e+00 5.72629683e-02 6.20829582e-01
-3.26494724e-01 9.97823954e-01 -3.46518219e-01 -3.24806839e-01
3.22391033e-01 -2.87756473e-01 -1.13978410e+00 6.48713231e-01
7.87972093e-01 -1.92002013e-01 2.76064247e-01 1.09932613e+00
1.00482330e-01 -1.73743039e-01 5.72829187e-01 -1.06438506e+00
-3.77349377e-01 6.11077130e-01 -1.00964105e+00 -1.24091581e-01
-2.12842599e-01 1.25730023e-01 9.06070173e-01 -1.03658271e+00
5.14560044e-01 1.15610373e+00 7.19111145e-01 6.45769596e-01
-1.41416979e+00 -7.60884285e-01 5.73662758e-01 -7.59586245e-02
-1.25131011e+00 -4.45515588e-02 7.08180547e-01 -4.24025834e-01
4.64827776e-01 1.50932372e-01 4.11022753e-01 6.31415188e-01
-4.97394919e-01 1.13174593e+00 1.41114926e+00 -3.82542163e-01
1.01235561e-01 4.28699076e-01 4.41008806e-01 9.83981550e-01
-2.67797932e-02 3.65752652e-02 -2.92405218e-01 2.51647532e-01
7.21581221e-01 7.48344138e-02 -4.01190575e-03 -3.96784097e-01
-1.17045736e+00 6.94674015e-01 6.78239703e-01 3.92738916e-02
-3.48066002e-01 2.14543730e-01 3.16000372e-01 4.52531986e-02
6.07790828e-01 3.68220925e-01 -7.85094619e-01 5.27192175e-01
-1.32860553e+00 1.98783949e-02 6.03320658e-01 1.11836362e+00
1.17258561e+00 -4.24799919e-01 -4.86303300e-01 9.04815555e-01
2.87044227e-01 2.71360832e-03 1.52632281e-01 -1.11924934e+00
2.49529809e-01 9.94643033e-01 -2.69709341e-02 -2.22934186e-01
-4.84470457e-01 -9.39576924e-01 -6.29389405e-01 3.50605786e-01
4.62571055e-01 -1.99482962e-01 -1.38390148e+00 1.43168545e+00
7.55592287e-01 6.09633803e-01 -2.02359274e-01 8.90469670e-01
9.15554404e-01 5.41291118e-01 4.20801878e-01 -6.54347613e-02
1.41984868e+00 -1.66357672e+00 -2.82630533e-01 -3.97992134e-01
6.25882506e-01 -4.78267103e-01 9.76787984e-01 2.79200941e-01
-1.21122539e+00 -9.02540922e-01 -8.72266948e-01 -5.74370883e-02
-2.69868165e-01 4.24634248e-01 5.55731654e-01 5.46507716e-01
-9.36247468e-01 6.49934053e-01 -7.03754306e-01 -5.66774458e-02
1.06080711e+00 3.47621411e-01 -8.70083272e-02 -3.69639359e-02
-6.88454807e-01 5.01209319e-01 8.00208569e-01 1.79815870e-02
-1.29761827e+00 -9.98119712e-01 -9.31627512e-01 -9.02651623e-03
6.39423132e-01 -4.11797106e-01 1.28628623e+00 -1.32146800e+00
-1.45967317e+00 1.40992463e+00 -1.11440822e-01 -4.89071995e-01
6.34309769e-01 -2.16770601e-02 1.44761056e-01 2.34865278e-01
1.89892218e-01 1.45506883e+00 8.78733099e-01 -1.70380747e+00
-1.31225741e+00 -1.79054677e-01 1.92302272e-01 1.56727374e-01
-9.75410119e-02 3.47179510e-02 -8.72130036e-01 -3.56681883e-01
2.76957899e-01 -7.32485056e-01 -4.75357383e-01 5.46712801e-02
-5.81843317e-01 -4.65302080e-01 8.72542799e-01 -3.36178333e-01
9.97960091e-01 -1.96816218e+00 -1.07002437e-01 1.85935035e-01
2.04900160e-01 1.87469423e-01 -1.54764265e-01 -3.29911679e-01
-8.04208666e-02 1.59305837e-02 -6.38105452e-01 -7.49417305e-01
-1.42861575e-01 1.15533181e-01 2.37336755e-02 5.11639178e-01
7.27677584e-01 1.04652858e+00 -9.90438461e-01 -8.58993649e-01
1.85841873e-01 1.53006434e-01 -4.33631897e-01 2.00009659e-01
-6.56330347e-01 6.81338012e-01 -3.47854853e-01 1.01204252e+00
7.02632248e-01 -7.01629877e-01 1.00752339e-02 -1.88379109e-01
-1.68363094e-01 1.08484761e-03 -1.17897749e+00 1.71721971e+00
-3.53735089e-01 3.41971368e-01 1.57936230e-01 -1.15478146e+00
7.09034920e-01 8.32995698e-02 3.42580259e-01 -5.62955201e-01
1.41192034e-01 8.07552636e-02 -3.38438570e-01 -2.77065754e-01
2.44853809e-01 1.70911178e-02 -2.77409945e-02 3.22966903e-01
1.70435444e-01 6.61875457e-02 4.40512061e-01 9.72054228e-02
8.11314762e-01 5.39584339e-01 -1.41687691e-01 -2.02499524e-01
7.33197689e-01 2.13637248e-01 7.72768259e-01 8.59658480e-01
-2.32577831e-01 6.96762145e-01 4.55728114e-01 -2.67250091e-01
-7.97599912e-01 -7.91994631e-01 -4.55190003e-01 1.44472289e+00
3.93891573e-01 -1.07802436e-01 -1.07206416e+00 -1.29482603e+00
-2.47526895e-02 5.04714251e-01 -8.43738794e-01 2.26017669e-01
-5.38634121e-01 -7.70194530e-01 3.42899024e-01 6.79176390e-01
8.27437878e-01 -1.36025763e+00 -6.21869266e-01 1.37765646e-01
1.57481089e-01 -1.24794674e+00 -4.15204197e-01 6.93157732e-01
-7.73491263e-01 -1.10137236e+00 -8.27264428e-01 -1.24786651e+00
1.10983622e+00 -1.11835338e-01 1.27873933e+00 3.42731714e-01
-3.24769825e-01 8.80892761e-03 -1.87095672e-01 -2.54404098e-01
-2.64082819e-01 2.37809867e-01 -6.47041142e-01 2.46534795e-01
9.29332599e-02 -1.26594722e-01 -7.59205401e-01 3.99851322e-01
-8.49110544e-01 3.20966482e-01 9.08168614e-01 7.74024487e-01
1.15921986e+00 1.05918013e-01 4.69999284e-01 -1.58165312e+00
-1.95030570e-01 -2.22424313e-01 -7.88126647e-01 4.21556652e-01
-7.59291291e-01 -1.01561405e-01 4.62589741e-01 -4.68904674e-01
-1.05613649e+00 6.63569748e-01 -1.03633136e-01 -3.55641901e-01
-6.56230986e-01 8.30498263e-02 -1.96465269e-01 -1.86697617e-01
4.01250035e-01 7.76869208e-02 -3.69728625e-01 -4.70399350e-01
5.80663383e-01 5.28666139e-01 6.85000241e-01 -5.32635212e-01
8.00436139e-01 5.72145283e-01 -1.15554571e-01 -2.81349123e-01
-1.52648914e+00 -8.03054214e-01 -8.45863998e-01 -3.47408831e-01
1.18490613e+00 -9.94055450e-01 -5.49480557e-01 6.02068961e-01
-9.55040097e-01 -7.67023087e-01 -4.91716623e-01 -1.99298020e-02
-4.36508387e-01 -2.01854520e-02 -9.38973904e-01 -3.91365975e-01
-1.87874377e-01 -1.18984830e+00 1.43599010e+00 5.20722270e-01
1.70598850e-01 -7.57065296e-01 -2.98910201e-01 5.68819106e-01
-6.72029853e-02 2.69009560e-01 6.44583046e-01 -5.79602301e-01
-7.47821271e-01 -1.09592956e-02 -7.11292326e-01 4.85871881e-01
-2.53172636e-01 -1.15946382e-01 -1.28505158e+00 -2.30631158e-01
-4.07593995e-01 -6.36604130e-01 1.23419237e+00 4.38572556e-01
1.64161372e+00 -3.45303230e-02 -4.91318077e-01 8.02958667e-01
1.46784520e+00 -6.73437789e-02 3.97486359e-01 3.49180788e-01
9.43204880e-01 7.87100434e-01 7.60217965e-01 -3.51032093e-02
3.08168530e-01 3.76669496e-01 6.19537115e-01 -7.70017028e-01
-3.60689431e-01 -1.11255065e-01 -5.48708886e-02 1.90350696e-01
3.03300798e-01 -1.87345035e-02 -6.76547766e-01 6.05458736e-01
-1.77165473e+00 -5.32689929e-01 -2.94936538e-01 1.91457272e+00
1.13977659e+00 4.60862756e-01 2.10279271e-01 6.32306114e-02
8.86007309e-01 -4.24550846e-02 -7.62722254e-01 -7.11719841e-02
1.30606115e-01 3.31224263e-01 8.44070256e-01 3.72442901e-01
-1.50421953e+00 1.25903857e+00 5.65954590e+00 8.41430664e-01
-8.87026668e-01 3.10097873e-01 1.18294322e+00 1.80883884e-01
-1.91181013e-03 4.85451408e-02 -1.03302383e+00 2.51419187e-01
5.28562486e-01 7.47546196e-01 -1.32640311e-03 8.81084561e-01
-1.32780537e-01 -5.20244315e-02 -1.27238882e+00 5.76090872e-01
-1.23050265e-01 -1.33519852e+00 -1.29941419e-01 -2.10346922e-01
1.26297629e+00 -1.00237243e-01 2.75747571e-03 4.36667055e-01
3.51668119e-01 -7.73373067e-01 9.29636598e-01 3.17146182e-01
8.05295706e-01 -7.16306627e-01 5.69301248e-01 4.98787135e-01
-1.26355004e+00 -1.88124672e-01 -2.02697739e-01 2.83054769e-01
1.05801135e-01 6.37687802e-01 -6.81853771e-01 2.63419300e-01
7.50990748e-01 8.15910161e-01 -7.27099359e-01 1.07090569e+00
-5.88893354e-01 8.83064091e-01 -8.94233733e-02 3.27696085e-01
5.58475375e-01 3.81568400e-03 -4.46195118e-02 1.34704328e+00
-3.36186737e-01 7.90445060e-02 5.97281098e-01 1.06454217e+00
-3.62710029e-01 -1.59270540e-01 2.00373113e-01 3.20192128e-01
2.97978997e-01 1.73945856e+00 -1.36003602e+00 -5.75111270e-01
-4.26885575e-01 1.15981245e+00 4.46667075e-01 3.50458324e-01
-9.56325829e-01 -1.63249359e-01 6.06096461e-02 2.09840298e-01
5.79640269e-01 2.43262723e-01 -4.28456903e-01 -6.18348181e-01
-1.13915391e-01 -6.06459737e-01 3.79188001e-01 -4.84393716e-01
-1.17045128e+00 3.68791103e-01 -2.05521494e-01 -9.32274699e-01
1.99838251e-01 -5.71902990e-01 -5.50584733e-01 7.30657041e-01
-2.06268287e+00 -1.43562412e+00 -4.96061087e-01 3.63030046e-01
8.51550937e-01 3.30650628e-01 5.79151690e-01 4.36629057e-01
-8.29065263e-01 6.79400623e-01 -2.32843280e-01 3.50315720e-01
5.49520671e-01 -1.55099225e+00 1.82130858e-01 7.54806638e-01
2.22156331e-01 1.33475199e-01 1.16677895e-01 -5.34813344e-01
-5.57113051e-01 -1.69006121e+00 4.50473160e-01 -3.48538220e-01
3.51577342e-01 -4.82465237e-01 -8.68570328e-01 6.81805909e-01
9.88859963e-03 4.88264143e-01 3.48199010e-01 -1.24567943e-02
-2.34993950e-01 -8.86722058e-02 -1.12016916e+00 1.86849147e-01
9.95427489e-01 -3.32532495e-01 -2.19048440e-01 5.34925878e-01
7.77595162e-01 -5.33004582e-01 -6.64274156e-01 7.48515368e-01
2.83760607e-01 -8.03940952e-01 8.59501898e-01 -3.83200794e-01
5.02136230e-01 -5.87410033e-01 2.38575146e-01 -7.98173189e-01
-3.94371510e-01 -1.79389715e-01 7.99803659e-02 1.46768022e+00
6.96002185e-01 -2.54630983e-01 1.17741668e+00 2.67161340e-01
-2.19220370e-01 -1.08036888e+00 -4.58385527e-01 -4.37680185e-01
-9.52623188e-02 -3.51498634e-01 3.87004346e-01 9.13986564e-01
-6.98342919e-01 3.77204031e-01 3.96894999e-02 4.36874926e-01
9.14054930e-01 4.09367770e-01 4.46919203e-01 -1.16564345e+00
-4.14351821e-01 -5.27776420e-01 -1.23933665e-01 -1.32033634e+00
4.01249141e-01 -9.98736680e-01 4.48279262e-01 -1.55895126e+00
5.45583189e-01 -9.00470495e-01 -6.50493085e-01 7.25526869e-01
-5.41154265e-01 7.74824381e-01 -6.86634630e-02 1.75953582e-01
-1.13970983e+00 1.43909946e-01 1.36336076e+00 -5.04836559e-01
-1.95349589e-01 1.89842373e-01 -5.44304907e-01 9.76663888e-01
5.61585903e-01 -6.70445025e-01 -2.80652791e-01 -2.63275445e-01
-2.41673872e-01 -2.61942983e-01 4.92640465e-01 -1.00274467e+00
1.81059167e-01 2.28971597e-02 5.42187393e-01 -8.56013000e-01
-3.32056656e-02 -9.01000977e-01 -1.99870825e-01 3.72577488e-01
-8.17443967e-01 -5.02112925e-01 4.82270867e-02 5.29436409e-01
-1.03989869e-01 -4.51418400e-01 1.13029349e+00 -2.38834679e-01
-9.39940035e-01 4.80479300e-01 9.61833894e-02 2.45384257e-02
1.35108829e+00 -2.39205167e-01 -1.49611309e-01 3.43814611e-01
-8.32820833e-01 6.39521837e-01 2.91988790e-01 1.84270471e-01
3.63289922e-01 -9.13496494e-01 -7.08849251e-01 1.60289347e-01
2.72395998e-01 6.31336391e-01 7.16447756e-02 7.00031400e-01
-4.77722853e-01 7.68162608e-02 2.23421708e-01 -1.06324780e+00
-1.17047799e+00 5.49951255e-01 5.40170729e-01 -4.66872096e-01
-6.29598558e-01 1.35746968e+00 4.81545419e-01 -6.92215741e-01
6.73020482e-01 -2.91574746e-01 -2.71357507e-01 -2.67651100e-02
3.98907214e-01 1.12117007e-01 2.83930823e-02 -4.59731191e-01
-2.49489293e-01 7.14131773e-01 -3.22671860e-01 2.50572115e-01
1.08216012e+00 -4.28991243e-02 -1.36698231e-01 2.68809646e-01
1.18233192e+00 -4.38943624e-01 -1.86871898e+00 -3.03024650e-01
2.32322201e-01 -2.65885144e-02 1.69118762e-01 -1.09012711e+00
-1.43855083e+00 6.96824431e-01 7.98880816e-01 -1.94569211e-02
1.24589264e+00 3.05443168e-01 6.30258441e-01 1.77216139e-02
1.66478500e-01 -1.06237268e+00 1.94043815e-01 3.09574753e-01
2.03660384e-01 -1.49296319e+00 -1.27920240e-01 -7.72109866e-01
-4.00516421e-01 7.73902416e-01 9.25148249e-01 -6.34537414e-02
5.55974960e-01 3.88813436e-01 2.02492326e-01 -1.62380755e-01
-5.10132432e-01 -5.38593948e-01 4.59708929e-01 3.51730675e-01
2.74212390e-01 1.84999108e-02 -1.02336831e-01 4.46689963e-01
4.35608804e-01 -8.96074176e-02 -2.09963899e-02 8.92665327e-01
-6.39725327e-01 -1.19388402e+00 -3.32501531e-01 5.12486637e-01
-5.06044984e-01 -1.04391880e-01 -2.54265785e-01 7.58155286e-01
7.40681231e-01 8.59412491e-01 1.70073077e-01 -9.67062116e-02
2.33808517e-01 6.77226782e-02 4.07820851e-01 -9.45111096e-01
-9.49675977e-01 1.64430454e-01 -7.65611380e-02 -5.94208002e-01
-7.21206844e-01 -6.25370502e-01 -1.64871550e+00 3.97429377e-01
-5.71096659e-01 3.81176434e-02 6.79061115e-01 1.04139173e+00
5.46042388e-03 9.31342900e-01 6.10111773e-01 -8.85788620e-01
-2.77481794e-01 -9.62249756e-01 -4.31462169e-01 4.28214550e-01
2.94037253e-01 -3.67378026e-01 -1.67707294e-01 4.27017868e-01] | [9.48993968963623, 0.5899838209152222] |
4e6bb683-72b9-42c1-833a-6ebceb3d6ad2 | towards-a-gold-standard-corpus-for-variable | null | null | https://aclanthology.org/L18-1084 | https://aclanthology.org/L18-1084.pdf | Towards a Gold Standard Corpus for Variable Detection and Linking in Social Science Publications | null | ['Peter Mutschke', 'Andrea Zielinski'] | 2018-05-01 | towards-a-gold-standard-corpus-for-variable-1 | https://aclanthology.org/L18-1084 | https://aclanthology.org/L18-1084.pdf | lrec-2018-5 | ['variable-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.373117446899414, 3.747725009918213] |
f2156713-cb06-452b-ade9-c210b840ce99 | ray-space-motion-compensation-for-lenslet | 2207.00522 | null | https://arxiv.org/abs/2207.00522v1 | https://arxiv.org/pdf/2207.00522v1.pdf | Ray-Space Motion Compensation for Lenslet Plenoptic Video Coding | Plenoptic images and videos bearing rich information demand a tremendous amount of data storage and high transmission cost. While there has been much study on plenoptic image coding, investigations into plenoptic video coding have been very limited. We investigate the motion compensation for plenoptic video coding from a slightly different perspective by looking at the problem in the ray-space domain instead of in the conventional pixel domain. Here, we develop a novel motion compensation scheme for lenslet video under two sub-cases of ray-space motion, that is, integer ray-space motion and fractional ray-space motion. The proposed new scheme of light field motion-compensated prediction is designed such that it can be easily integrated into well-known video coding techniques such as HEVC. Experimental results compared to relevant existing methods have shown remarkable compression efficiency with an average gain of 19.63% and a peak gain of 29.1%. | ['Byeungwoo Jeon', 'Jonghoon Yim', 'Vinh Van Duong', 'Thuc Nguyen Huu'] | 2022-07-01 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 5.29641092e-01 -1.32192761e-01 2.54999194e-02 -9.55853835e-02
-1.63912550e-01 -2.18620881e-01 3.12837899e-01 -4.08292383e-01
-2.02527866e-01 7.63441205e-01 2.20690057e-01 -1.80342972e-01
-1.82715833e-01 -7.15095460e-01 -3.89219463e-01 -1.05455542e+00
1.62580714e-01 2.08238419e-02 6.22355342e-01 5.84865473e-02
4.91391569e-01 4.15850043e-01 -1.29291320e+00 2.01428220e-01
5.97516537e-01 1.24566007e+00 3.65841597e-01 8.47199559e-01
1.34657875e-01 1.04032755e+00 -2.54962176e-01 -4.11545873e-01
5.98341107e-01 -5.08277535e-01 -3.41873944e-01 4.58270937e-01
1.49394512e-01 -6.71271384e-01 -6.73871040e-01 1.01130402e+00
4.31056142e-01 5.57507835e-02 3.81228238e-01 -1.00757933e+00
-5.74820578e-01 4.21694368e-02 -8.83040667e-01 3.36130947e-01
3.33599716e-01 -2.35941634e-02 3.63550574e-01 -6.39186800e-01
7.57519305e-01 8.22716057e-01 2.75177240e-01 4.59616572e-01
-8.01810086e-01 -4.27153438e-01 -7.61851370e-01 6.81100309e-01
-9.75101054e-01 -4.01685178e-01 1.19570351e+00 -4.11784321e-01
5.75582147e-01 4.02478278e-01 7.69076943e-01 2.01589018e-01
4.95849192e-01 5.32179892e-01 1.04136837e+00 -5.88013411e-01
7.86268339e-02 7.58581907e-02 -2.56091654e-01 6.82834744e-01
2.89470702e-01 3.14309187e-02 -3.77581865e-01 2.41275784e-02
1.17830515e+00 6.97477460e-02 -7.94148564e-01 -6.40330613e-01
-1.24661219e+00 6.57877088e-01 -1.39714792e-01 5.60431957e-01
-2.69649595e-01 2.31009543e-01 1.70371860e-01 1.42368466e-01
4.24798101e-01 1.56904340e-01 -5.64761227e-03 -9.98908356e-02
-1.20121157e+00 3.48503329e-02 7.01468647e-01 1.13956678e+00
3.92896712e-01 2.23412469e-01 1.00538120e-01 6.53145552e-01
3.00994009e-01 4.63695049e-01 2.51282483e-01 -1.35393095e+00
4.57380921e-01 1.02464899e-01 2.72562116e-01 -1.00274801e+00
-2.82107890e-02 -2.62764275e-01 -8.45525682e-01 4.04766440e-01
3.18849295e-01 4.57401164e-02 -4.71996546e-01 1.18845832e+00
1.56883553e-01 3.35997164e-01 2.35541552e-01 8.72285426e-01
6.95353925e-01 1.11924493e+00 -5.56451559e-01 -9.41804767e-01
1.05220473e+00 -1.04426336e+00 -1.03601027e+00 3.55795532e-01
1.86462194e-01 -1.26912594e+00 3.56530279e-01 6.44408345e-01
-1.91329539e+00 -2.78285801e-01 -1.09984326e+00 -1.54390574e-01
4.23850238e-01 5.99846914e-02 5.16974568e-01 7.62030661e-01
-1.12700737e+00 2.18896076e-01 -5.10047853e-01 1.03036752e-02
1.54100731e-02 2.07015768e-01 -1.26155332e-01 -3.41347635e-01
-7.13915050e-01 4.70487535e-01 2.68596619e-01 -1.75528407e-01
-3.59418869e-01 -7.42756784e-01 -3.01212221e-01 1.40818432e-01
2.79156238e-01 -6.14589036e-01 9.48639274e-01 -4.29918647e-01
-1.46760809e+00 6.53398514e-01 -1.56194568e-01 -3.08079600e-01
7.03359485e-01 2.67510772e-01 -5.36972642e-01 8.09030414e-01
-2.40423784e-01 3.67743343e-01 1.03157258e+00 -1.24057996e+00
-6.07722998e-01 -1.58029407e-01 9.68982279e-02 1.99334428e-01
-2.10969806e-01 6.88387379e-02 -4.95601505e-01 -7.88979828e-01
4.15130168e-01 -8.39139700e-01 6.04998097e-02 4.76832539e-01
1.07163899e-01 2.24020049e-01 1.13554835e+00 -6.30591452e-01
1.30967176e+00 -1.98883414e+00 -5.71981743e-02 -3.44877303e-01
3.56682241e-01 4.78662640e-01 1.80960253e-01 4.40092057e-01
-2.13947281e-01 -3.63469630e-01 -2.69661933e-01 6.15695752e-02
-5.28192222e-01 -9.65362787e-02 -2.54473269e-01 6.59520924e-01
-4.56490934e-01 5.25379539e-01 -7.39541233e-01 -4.97482389e-01
6.49765372e-01 6.97360635e-01 -9.65578735e-01 -3.59733999e-02
6.89687133e-02 6.00054383e-01 -5.02674162e-01 7.82835007e-01
1.28580022e+00 -2.08164528e-01 -2.16452137e-01 -5.10988235e-01
-4.98291343e-01 -4.03755546e-01 -9.62986946e-01 1.26185083e+00
-3.24905634e-01 1.07894611e+00 3.33902061e-01 -9.49827969e-01
6.92715287e-01 6.17180467e-01 1.09130347e+00 -5.81932604e-01
1.91789374e-01 3.40575337e-01 -1.99964941e-01 -8.60872686e-01
7.80387580e-01 -3.87832820e-01 5.56349993e-01 2.54255831e-01
-3.90330344e-01 -2.61240333e-01 3.14276248e-01 -1.29181221e-01
1.06120312e+00 -1.69980451e-01 3.08127880e-01 -1.03979968e-01
9.90081131e-01 -1.83781371e-01 3.53909254e-01 2.66261011e-01
-1.18592516e-01 6.58731759e-01 1.24627560e-01 -2.42207378e-01
-1.42273414e+00 -8.01950514e-01 -3.06695342e-01 3.27043653e-01
5.61737418e-01 -4.23347026e-01 -5.13410270e-01 1.12174988e-01
-4.66705322e-01 5.94601989e-01 1.85824722e-01 2.29082257e-01
-1.01784587e+00 -7.55668819e-01 -3.37036066e-02 -7.23173842e-02
1.19114459e+00 -7.41050363e-01 -9.29783165e-01 1.61152035e-01
-1.57070920e-01 -1.41391647e+00 -2.64820457e-01 -5.60066938e-01
-9.35466886e-01 -1.11740160e+00 -1.36747098e+00 -1.04564404e+00
5.27852714e-01 8.89285386e-01 8.40193033e-01 -6.62506893e-02
-3.23165059e-01 7.88229287e-01 -4.35972601e-01 -1.13034742e-02
-4.40058321e-01 -7.62845755e-01 -3.34129006e-01 2.27480829e-01
1.02118917e-01 -7.31625199e-01 -1.19931448e+00 5.14414370e-01
-1.18435287e+00 2.92378724e-01 5.02125323e-01 6.06677055e-01
2.88945794e-01 4.76052910e-01 9.43158865e-02 -5.68827510e-01
2.31533468e-01 -2.13768914e-01 -8.77152145e-01 1.32856354e-01
-5.70283890e-01 -2.78911501e-01 5.04780114e-01 -9.32408124e-02
-1.27967954e+00 -1.44222856e-01 -4.99904118e-02 -4.24707115e-01
2.31295198e-01 -2.54646298e-02 1.14605777e-01 -7.93026626e-01
1.17111884e-01 6.26922309e-01 -1.63652137e-01 -2.83591300e-01
1.25312414e-02 6.53473794e-01 5.94604969e-01 -1.60980020e-02
8.63576949e-01 9.39020991e-01 4.79556203e-01 -1.21206248e+00
5.39569892e-02 -6.10050738e-01 -1.44375443e-01 -5.12252331e-01
1.13514006e+00 -7.44446874e-01 -1.01728845e+00 3.98856401e-01
-1.34326208e+00 1.92894727e-01 -1.24200262e-01 9.70943630e-01
-9.85249639e-01 9.95609105e-01 -7.43931234e-01 -5.77052653e-01
-9.32186767e-02 -1.32992136e+00 7.76411831e-01 8.76168981e-02
5.03092349e-01 -1.08083677e+00 -1.42960981e-01 7.68353462e-01
5.46429455e-01 1.70937285e-01 9.04257596e-01 3.85721028e-01
-1.45952606e+00 -1.95583612e-01 -3.56785536e-01 3.00161839e-01
-1.15983136e-01 -3.58840525e-01 -7.57495284e-01 -3.47256154e-01
1.01585138e+00 1.22006081e-01 7.74871230e-01 7.31310785e-01
1.24430907e+00 -3.10073316e-01 -1.89074486e-01 1.01856148e+00
1.84689891e+00 6.94970608e-01 9.70879138e-01 2.40836605e-01
4.05684888e-01 3.53948981e-01 4.70439762e-01 6.68326735e-01
-9.34483260e-02 8.30410182e-01 6.07983410e-01 1.27971038e-01
-4.03861135e-01 1.91448927e-01 2.01854832e-03 1.07740259e+00
-4.33198512e-01 -7.53722310e-01 -3.43906462e-01 3.01152378e-01
-1.49104929e+00 -1.21575582e+00 -3.83040398e-01 2.16958213e+00
5.13890326e-01 -2.17083335e-01 -4.82823133e-01 4.59790170e-01
7.79995561e-01 5.40394068e-01 -2.60859638e-01 -2.16223046e-01
-1.30359828e-01 -1.78311571e-01 9.79719341e-01 7.21858263e-01
-7.47403145e-01 3.27485740e-01 6.17254686e+00 1.17620265e+00
-1.17659485e+00 1.87652305e-01 3.24277073e-01 -1.58009827e-01
-4.80031610e-01 1.63511559e-01 -5.12157798e-01 6.04702532e-01
3.08973253e-01 -3.00303012e-01 2.95570374e-01 4.28836644e-01
3.35777313e-01 -5.41347742e-01 -7.34725118e-01 1.65940249e+00
2.95673519e-01 -1.34375286e+00 1.49542794e-01 2.47444257e-01
8.84215951e-01 -5.89485288e-01 2.01157957e-01 -5.01215398e-01
-6.75120831e-01 -5.30485570e-01 3.56908113e-01 4.16987598e-01
8.18227410e-01 -4.48848426e-01 3.40904564e-01 2.95196295e-01
-9.86203015e-01 -1.73674114e-02 -6.55890822e-01 1.70007721e-01
7.93020904e-01 8.49342406e-01 -6.47368133e-01 5.19481897e-01
2.33058378e-01 8.85077953e-01 -5.53199425e-02 1.48769701e+00
4.06501025e-01 4.36724782e-01 4.78297248e-02 2.41925672e-01
9.69610587e-02 -5.50955772e-01 1.10793114e+00 6.54990017e-01
8.77785206e-01 6.37778699e-01 -4.71010268e-01 5.58093250e-01
2.66659170e-01 -3.00039593e-02 -4.79129225e-01 2.16632202e-01
-1.18480004e-01 8.90035748e-01 -7.49031544e-01 -2.83753395e-01
-7.99752712e-01 1.02753878e+00 -7.35580444e-01 4.04579610e-01
-7.79187322e-01 -2.89150745e-01 1.14542924e-01 4.96760815e-01
5.08906007e-01 -4.87155408e-01 -5.96154146e-02 -1.35982680e+00
6.79121017e-02 -1.47274911e-01 -1.37881458e-01 -1.29292309e+00
-6.31946325e-01 3.61142784e-01 2.15053737e-01 -1.71273816e+00
-2.45036766e-01 -5.56313038e-01 -2.86050320e-01 5.46180844e-01
-1.75342619e+00 -8.94736230e-01 -3.16037863e-01 7.58940160e-01
7.91040719e-01 -2.87962079e-01 3.33419472e-01 4.97249901e-01
1.07036181e-01 1.50524899e-01 7.27148950e-01 -3.87116194e-01
4.54237491e-01 -5.64454019e-01 -3.64336938e-01 7.63708651e-01
-4.46201921e-01 1.39898583e-01 1.02249539e+00 -5.92403173e-01
-1.64655960e+00 -5.78996956e-01 9.17626143e-01 8.14459100e-02
3.69933426e-01 1.59499031e-02 -6.04561210e-01 3.79048705e-01
4.68183845e-01 -3.13940160e-02 4.82392222e-01 -1.22227788e+00
1.44846246e-01 -1.39006674e-01 -1.46324956e+00 3.89020860e-01
1.00502026e+00 -1.43692121e-01 -2.70828485e-01 5.51222742e-01
6.15994573e-01 -3.78110856e-01 -8.53269517e-01 5.33484697e-01
4.77954477e-01 -1.62177896e+00 1.31522286e+00 2.63948262e-01
7.92107821e-01 -3.62466931e-01 -4.24271733e-01 -7.61595130e-01
-3.69282216e-01 -9.31689620e-01 4.78204675e-02 8.88660491e-01
-3.20355088e-01 -4.71192688e-01 1.09555793e+00 4.45780665e-01
-2.13929415e-01 -6.86713994e-01 -9.22211051e-01 -7.30172575e-01
-5.21277249e-01 -2.69035488e-01 5.63524850e-02 8.01241398e-01
-7.37784281e-02 -1.39340073e-01 -7.42089748e-01 -8.45346693e-03
9.77391899e-01 2.64821857e-01 3.73844296e-01 -9.51257169e-01
-5.46161056e-01 -2.99450427e-01 -8.12518060e-01 -1.46229804e+00
-2.47655466e-01 -7.06183434e-01 -4.70187217e-01 -1.46026874e+00
2.44678274e-01 -2.83861309e-01 2.62281030e-01 -4.85712975e-01
5.26471496e-01 7.01052785e-01 2.28180185e-01 4.45563406e-01
-2.95726210e-01 5.55849969e-01 1.79949558e+00 4.22056355e-02
-7.74058029e-02 2.45278910e-01 -6.52530119e-02 7.17982948e-01
5.32806158e-01 6.28823116e-02 -8.60599995e-01 -5.37214816e-01
3.26368749e-01 8.64651740e-01 3.62728626e-01 -1.26254022e+00
4.42737967e-01 -2.03088954e-01 1.84126303e-01 -8.39272261e-01
8.49837124e-01 -1.19511878e+00 4.45400268e-01 4.32478368e-01
-1.16223581e-01 -1.65412545e-01 -1.95387512e-01 7.63987184e-01
-4.23716992e-01 -4.97124583e-01 9.79018092e-01 -3.25797260e-01
-7.79047847e-01 1.69791594e-01 -3.81960273e-01 -1.97294429e-01
1.38890827e+00 -7.16292500e-01 -4.74428236e-01 -5.95151305e-01
-6.18983746e-01 -4.43526983e-01 6.83443308e-01 3.09084170e-03
9.53696668e-01 -1.27501237e+00 -5.23412168e-01 4.84721035e-01
-1.75920799e-01 -4.30434108e-01 4.64603513e-01 9.78778005e-01
-1.27342296e+00 8.95238042e-01 -4.79304135e-01 -6.27343059e-01
-1.42556214e+00 5.68207264e-01 -1.20304517e-01 3.24604176e-02
-8.04212272e-01 5.56977212e-01 5.01555085e-01 5.84524810e-01
-7.07456246e-02 -1.14879414e-01 -2.01306939e-01 -3.88077199e-01
6.00578666e-01 7.66928196e-01 -1.82724282e-01 -8.00533295e-01
4.00386788e-02 1.25330865e+00 3.09956163e-01 -6.47979528e-02
1.33502340e+00 -4.55174536e-01 -4.11889523e-01 1.79835986e-02
1.42863500e+00 3.86984408e-01 -1.34555173e+00 4.41278517e-02
-5.15924871e-01 -1.19273388e+00 2.10178107e-01 -2.53240407e-01
-1.35454071e+00 7.95800805e-01 4.32429820e-01 4.72930700e-01
1.54592466e+00 -1.11037888e-01 1.12407947e+00 -1.09600328e-01
6.31478369e-01 -7.51952469e-01 -9.81190279e-02 6.30114302e-02
8.17415953e-01 -9.37482655e-01 3.29430073e-01 -9.57162619e-01
-1.48755267e-01 1.37845278e+00 -3.27848736e-03 -2.45947242e-01
8.28191042e-01 1.13262229e-01 -5.28542101e-01 -3.91098857e-02
-6.72521889e-01 2.78899550e-01 1.81360692e-01 5.72840929e-01
5.69004357e-01 -3.25165838e-01 -9.72417414e-01 -2.01890975e-01
1.76072512e-02 5.31497717e-01 1.08786881e+00 8.50770831e-01
-6.76500916e-01 -9.52171266e-01 -5.65128386e-01 1.55448809e-01
-6.63102150e-01 -1.32874204e-02 4.78280008e-01 5.86885035e-01
3.28911357e-02 1.01445007e+00 -1.17183607e-02 -2.22034939e-02
-1.19524047e-01 -4.05900955e-01 7.74011672e-01 -3.33446302e-02
1.54444024e-01 4.32812989e-01 -1.64582610e-01 -4.90196884e-01
-9.55770731e-01 -4.30835009e-01 -9.63296771e-01 -5.09852052e-01
-1.13532441e-02 7.00938031e-02 8.21505010e-01 4.75181460e-01
-1.47847924e-02 2.50518084e-01 7.00724602e-01 -7.97824025e-01
-1.19028524e-01 -4.02515560e-01 -7.74873316e-01 1.67121723e-01
6.47005200e-01 -4.54116821e-01 -5.70001304e-01 2.26706386e-01] | [11.105462074279785, -2.156357526779175] |
62552830-b9e9-4dea-9a8b-323be67e7040 | open-source-german-distant-speech-recognition | null | null | https://link.springer.com/chapter/10.1007/978-3-319-24033-6_54 | https://download.hrz.tu-darmstadt.de/pub/FB20/Dekanat/Publikationen/LangTech/Radeck-ArnethEtAl_TSD2015_SpeechCorpus.pdf | Open Source German Distant Speech Recognition: Corpus and Acoustic Model | We present a new freely available corpus for German distant speech recognition and report speaker-independent word error rate (WER) results for two open source speech recognizers trained on this corpus. The corpus has been recorded in a controlled environment with three different microphones at a distance of one meter. It comprises 180 different speakers with a total of 36 hours of audio recordings. We show recognition results with the open source toolkit Kaldi (20.5% WER) and PocketSphinx (39.6% WER) and make a complete open source solution for German distant speech recognition possible. | ['and Chris Biemann', 'Max Mühlhäuser', 'Stefan Radomski', 'Evandro Gouvea', 'Arvid Lange', 'Benjamin Milde', 'Stephan Radeck-Arneth'] | 2015-12-11 | null | null | null | international-conference-on-text-speech-and | ['distant-speech-recognition'] | ['speech'] | [-1.21274009e-01 8.10818449e-02 3.80642205e-01 -6.15763307e-01
-1.47756982e+00 -6.67441547e-01 5.98430753e-01 -2.60281771e-01
-5.81284404e-01 3.37079167e-01 5.42995334e-01 -6.14725947e-01
2.84155697e-01 -1.39487118e-01 -1.33409619e-01 -7.16698527e-01
-1.97615623e-01 5.31249404e-01 1.80971697e-01 -2.25817055e-01
-1.18632138e-01 3.91394764e-01 -1.40455008e+00 2.22902566e-01
3.30272257e-01 7.06819415e-01 3.24884057e-01 1.57938731e+00
1.62046939e-01 5.41764557e-01 -1.26134634e+00 -4.73764777e-01
2.46917918e-01 -7.05225691e-02 -1.16458070e+00 -6.26813853e-03
3.28721255e-01 -2.74958402e-01 -3.52528244e-01 6.10866368e-01
1.21523702e+00 4.07685012e-01 9.91529822e-02 -5.39958537e-01
-6.60930634e-01 6.98884726e-01 3.58119547e-01 8.16338003e-01
6.20405436e-01 -9.09205601e-02 7.99617589e-01 -9.26264346e-01
1.83525532e-01 1.14188540e+00 2.91655481e-01 8.67920935e-01
-9.64086175e-01 -4.65665042e-01 -2.16322735e-01 -6.87094927e-02
-1.99571323e+00 -1.45443392e+00 1.47759885e-01 -1.03343450e-01
1.80194974e+00 7.78869212e-01 1.58079803e-01 1.51640499e+00
-4.54705387e-01 5.52450538e-01 8.01046848e-01 -4.65428501e-01
2.49937847e-01 3.26496631e-01 7.56927654e-02 1.44605801e-01
-5.38610578e-01 5.19564152e-02 -8.69026721e-01 -3.17708880e-01
4.06303257e-01 -5.80139160e-01 -4.78455096e-01 5.22774100e-01
-1.45330679e+00 5.49323976e-01 -8.80115926e-02 6.83899224e-01
-1.37771443e-01 -2.90740639e-01 3.60330105e-01 7.28237569e-01
6.47095382e-01 -9.57986191e-02 -5.66609681e-01 -8.36436510e-01
-7.96999514e-01 -3.91284287e-01 1.22488320e+00 1.02818477e+00
1.43644899e-01 5.37922621e-01 1.36630908e-01 1.46722174e+00
4.43931311e-01 9.38777268e-01 1.04849851e+00 -4.92225766e-01
6.39885187e-01 -3.82785916e-01 1.30087072e-02 -4.24765557e-01
1.68237135e-01 -2.53947377e-01 -3.67358953e-01 -1.39022604e-01
3.98568809e-01 -3.67771566e-01 -8.29261661e-01 1.19887602e+00
1.41958788e-01 3.10960829e-01 4.58194345e-01 7.23957717e-01
1.02484512e+00 1.15958917e+00 -1.70819297e-01 2.60576960e-02
1.08810735e+00 -1.24446082e+00 -7.30181098e-01 -4.18655902e-01
6.44029975e-01 -1.13186955e+00 1.14783621e+00 5.84942937e-01
-1.03279638e+00 -5.09432495e-01 -1.00807881e+00 -1.17161348e-02
-3.51194471e-01 -4.70524877e-02 -6.84833946e-03 1.26685810e+00
-1.41103733e+00 5.23313731e-02 -8.12855661e-01 -3.13017309e-01
-2.45044351e-01 2.77858615e-01 -5.69756091e-01 1.55839279e-01
-1.12599754e+00 8.43571365e-01 -7.77361095e-02 1.24384783e-01
-8.87521386e-01 -1.94007680e-01 -6.04117095e-01 5.69119900e-02
-1.32941946e-01 -3.70796360e-02 1.66532052e+00 -7.26782560e-01
-2.22104502e+00 1.17692983e+00 -4.26754594e-01 -3.83827716e-01
2.05420613e-01 -2.77134001e-01 -1.23580003e+00 -1.62624136e-01
-3.01547408e-01 5.54935150e-02 5.46715021e-01 -5.34512103e-01
-4.87736970e-01 -3.32619458e-01 -5.89974880e-01 1.24354303e-01
-3.72925848e-01 8.75410199e-01 -2.11954936e-01 -7.97688603e-01
-9.27323923e-02 -7.25554228e-01 1.58541471e-01 -6.91845894e-01
-2.69714952e-01 -3.38642567e-01 2.75334150e-01 -1.16296852e+00
1.38887107e+00 -2.54109859e+00 -8.70835967e-03 -6.62488192e-02
-2.69576967e-01 7.27806032e-01 -2.18561620e-01 5.83262563e-01
-8.36891755e-02 2.09732093e-02 2.43433379e-02 -6.82887316e-01
2.85811514e-01 2.04036206e-01 -3.34958553e-01 3.89772117e-01
-4.06888813e-01 4.24682677e-01 -4.95806783e-01 1.23599209e-01
5.03496468e-01 8.11032295e-01 -3.81875336e-02 5.28373063e-01
5.65587759e-01 2.04352841e-01 9.35971662e-02 6.07844293e-01
5.28611541e-01 5.64149499e-01 -2.70557284e-01 7.26608932e-01
-1.55575708e-01 9.99636054e-01 -1.30385649e+00 1.69254577e+00
-7.97943592e-01 8.23759317e-01 5.81204116e-01 -5.47981739e-01
1.21294320e+00 8.59790742e-01 -4.36796099e-01 -2.75596499e-01
1.53813317e-01 6.18676305e-01 -1.20187029e-01 -3.07161212e-01
2.91289389e-01 -1.80862099e-01 -6.50041213e-04 1.82195336e-01
3.19231033e-01 -7.97179267e-02 -3.48338455e-01 -1.19578861e-01
1.27959633e+00 -7.77454257e-01 2.57222354e-01 -9.81982872e-02
6.87167645e-01 -6.05404198e-01 8.88985321e-02 5.94804645e-01
-6.76826239e-01 7.25668430e-01 -2.89670080e-01 -1.78750604e-01
-6.36942089e-01 -1.43906677e+00 -3.58336687e-01 1.38572991e+00
-6.13514483e-01 -6.02592230e-01 -7.61574388e-01 -1.62852794e-01
-4.55187321e-01 7.14466929e-01 8.10036063e-03 1.86064497e-01
-4.26537395e-01 8.37280825e-02 1.23943496e+00 4.59867507e-01
4.44144011e-01 -1.10200453e+00 3.73289406e-01 1.79245353e-01
-1.54578522e-01 -1.28847861e+00 -7.49631703e-01 2.96168983e-01
-3.67051810e-01 -3.66554260e-01 -1.06890619e+00 -1.18135488e+00
-5.45912273e-02 2.41441220e-01 1.03833640e+00 -2.68313736e-01
9.47130471e-02 3.84149134e-01 -4.41116095e-01 -7.78625831e-02
-6.45367682e-01 2.71699816e-01 2.72962540e-01 2.34110765e-02
7.09118485e-01 -3.58125091e-01 -1.11926183e-01 5.15036523e-01
-3.05141717e-01 -7.14158416e-01 7.60259107e-02 5.80949187e-01
1.30288452e-01 -4.69041079e-01 6.09863520e-01 -2.34193534e-01
8.35075498e-01 -4.31291223e-01 -5.91498315e-01 8.06425363e-02
-1.59992903e-01 -3.32872033e-01 4.11280990e-01 -2.90984511e-01
-1.04520714e+00 6.83831377e-03 -1.13841712e+00 1.82513092e-02
-7.35503137e-01 1.31455943e-01 -4.23147231e-01 2.22493410e-01
8.10095310e-01 3.84779483e-01 -2.90688604e-01 -9.45399761e-01
1.16663657e-01 2.03153014e+00 7.93116570e-01 6.11939700e-04
3.09216559e-01 -3.63172382e-01 -1.14184082e+00 -1.45432854e+00
-2.30468005e-01 -9.68396306e-01 -2.51800656e-01 2.50710964e-01
7.75630355e-01 -1.29839170e+00 -7.17882872e-01 9.27103877e-01
-1.04875040e+00 -5.49492300e-01 -9.44904909e-02 6.09770954e-01
-1.97827518e-01 2.28169516e-01 -7.62873709e-01 -9.63480830e-01
-5.45301676e-01 -9.14616346e-01 1.22539735e+00 -1.94559395e-01
-4.88410830e-01 -1.19658363e+00 2.85932779e-01 7.62304842e-01
6.48859262e-01 -5.66719830e-01 -4.03415896e-02 -1.13395965e+00
1.38067141e-01 -2.76969522e-01 4.05611575e-01 9.73904610e-01
3.92349005e-01 -1.36789769e-01 -1.51857233e+00 -2.55558312e-01
1.55085484e-02 -2.47511208e-01 4.14107293e-01 -1.73588786e-02
7.00880766e-01 -3.68990183e-01 5.26971892e-02 3.53964269e-01
7.16114461e-01 3.07401031e-01 5.64375043e-01 1.28825873e-01
4.38414097e-01 4.21736687e-01 -1.05964482e-01 3.84298950e-01
2.22366899e-01 8.83558571e-01 -1.43254936e-01 3.63460302e-01
-2.94934601e-01 -1.15873687e-01 9.22965229e-01 1.46459627e+00
4.12997097e-01 -7.08653688e-01 -1.33599770e+00 9.50218737e-01
-1.13275445e+00 -8.72144282e-01 -1.76991150e-01 2.52390099e+00
9.86711264e-01 -3.02122027e-01 2.78773546e-01 3.72223675e-01
8.44175816e-01 3.38243395e-01 1.13014102e-01 -1.04880977e+00
-1.46277189e-01 5.50210834e-01 2.25908875e-01 1.33136272e+00
-7.29795039e-01 1.00033331e+00 7.51661968e+00 8.20949435e-01
-1.14694393e+00 2.86762178e-01 3.07490647e-01 -5.08385241e-01
4.62261476e-02 -4.01507527e-01 -8.81250620e-01 5.53890765e-01
2.12525105e+00 -3.28657746e-01 6.81050420e-01 9.84631062e-01
1.15351737e-01 1.88363969e-01 -9.31492686e-01 1.40801227e+00
2.00005993e-01 -9.81606960e-01 -3.29852343e-01 1.00308485e-01
3.37366015e-01 9.48823810e-01 -1.04744434e-01 2.74544984e-01
2.88095504e-01 -1.12458599e+00 8.34920168e-01 -5.99077046e-02
9.15076792e-01 -7.84997404e-01 6.57838523e-01 5.56861937e-01
-8.86382103e-01 2.39396423e-01 -3.15737128e-01 -6.40941560e-02
2.74595410e-01 5.54789841e-01 -1.12508535e+00 2.44015306e-02
8.16796601e-01 3.11124176e-01 -3.10237110e-01 7.32556999e-01
-1.92676052e-01 1.15754819e+00 -6.24303520e-01 -2.58257657e-01
8.04682747e-02 1.82047471e-01 6.70935750e-01 1.55025947e+00
3.85558665e-01 -1.32962033e-01 -2.66258150e-01 -2.82634906e-02
-8.25440064e-02 2.34925926e-01 -4.68559921e-01 6.91610575e-02
7.43153095e-01 8.27642679e-01 -9.31906998e-02 -2.98686385e-01
-2.85158545e-01 1.44901669e+00 1.86957046e-01 3.43814790e-01
-4.84293729e-01 -9.07918870e-01 1.10695171e+00 -9.65554491e-02
4.51412648e-01 -5.55829048e-01 1.01432502e-01 -1.11927521e+00
2.72032529e-01 -9.12794352e-01 9.87523273e-02 -8.16892982e-01
-1.17683601e+00 1.22311115e+00 -5.97292900e-01 -7.07982600e-01
-5.37197709e-01 -6.74720705e-01 -4.34314400e-01 1.48308647e+00
-1.18672693e+00 -4.89050984e-01 5.17412461e-02 8.59352887e-01
9.33749080e-01 -4.46461111e-01 1.55285668e+00 6.42611206e-01
-5.93413293e-01 8.00377727e-01 4.04305995e-01 2.45480761e-01
8.41323555e-01 -1.36500239e+00 1.18019736e+00 6.45706892e-01
4.77207392e-01 7.68098950e-01 4.31922257e-01 -4.11637463e-02
-1.18056631e+00 -7.54560530e-01 1.88010263e+00 -6.38817310e-01
7.44527459e-01 -8.77356589e-01 -9.10720170e-01 8.29804242e-01
4.30156171e-01 3.26218307e-01 1.12327671e+00 3.82811189e-01
-4.33349967e-01 -2.31792673e-01 -1.05288434e+00 3.22757691e-01
1.06919718e+00 -9.88392949e-01 -7.05090046e-01 4.79536742e-01
6.54139757e-01 -5.02811134e-01 -8.52276146e-01 -2.99718052e-01
4.20478702e-01 -8.97907019e-01 6.35353744e-01 -2.11352229e-01
-5.21980822e-01 -1.99835990e-02 -5.41593015e-01 -1.63647902e+00
-1.75452437e-02 -1.16117287e+00 1.11399293e-01 1.60305655e+00
7.13448584e-01 -1.22518253e+00 2.84470052e-01 5.22966981e-01
-4.36896086e-01 -7.19279721e-02 -1.66291344e+00 -1.19792688e+00
2.11760193e-01 -9.35026705e-01 5.54781318e-01 6.81931078e-01
3.65857750e-01 5.65402567e-01 -2.03785792e-01 2.80406982e-01
-1.33926541e-01 -7.03749537e-01 6.75019205e-01 -7.68561542e-01
-4.47415292e-01 -1.81486860e-01 -6.55514598e-01 -1.54576349e+00
1.82841375e-01 -7.53980696e-01 1.58682093e-01 -1.13086665e+00
-5.06821215e-01 -1.84458539e-01 -6.62205461e-03 5.03123701e-01
1.55223519e-01 1.16189510e-01 -1.98807597e-01 -5.52332476e-02
-3.97833496e-01 3.93206865e-01 9.31139141e-02 -1.92517444e-01
5.01452107e-03 2.24376813e-01 -4.20555770e-01 3.46367866e-01
8.51838946e-01 -3.85398924e-01 -2.77345806e-01 -9.05530274e-01
-3.17545623e-01 6.08624518e-02 8.27100724e-02 -1.03000844e+00
1.55104131e-01 3.69924307e-01 -1.39942974e-01 -3.24686617e-02
3.96312624e-01 -5.90672016e-01 8.76691192e-03 1.96165502e-01
-2.81097084e-01 -4.23875917e-03 3.12226057e-01 2.18418762e-01
-5.81041634e-01 1.57434084e-02 5.65941393e-01 1.29376665e-01
-6.39292836e-01 -2.64500260e-01 -8.04195702e-01 9.30398107e-02
7.79582977e-01 -2.00307980e-01 -1.72482893e-01 -7.41456866e-01
-1.12024570e+00 -2.63918400e-01 3.28041129e-02 9.16288614e-01
4.59176004e-01 -1.33605134e+00 -1.06527126e+00 7.65350163e-01
2.25183144e-01 -6.06337726e-01 4.18679267e-02 3.48637938e-01
-4.52189296e-01 7.17170656e-01 5.13169646e-01 -4.21193242e-01
-1.75170076e+00 2.14769766e-02 5.88088572e-01 4.41542327e-01
-4.46560562e-01 1.43598986e+00 -5.07286668e-01 -7.22112596e-01
5.73047578e-01 -3.43293846e-01 2.99979001e-01 -4.43623483e-01
1.05756021e+00 5.36173165e-01 8.22076678e-01 -1.04582155e+00
-8.42793763e-01 3.35170925e-02 1.33747116e-01 -6.95904851e-01
1.00262702e+00 -4.06920582e-01 3.20522994e-01 7.29865193e-01
1.51445925e+00 6.84947371e-01 -6.50395155e-01 -2.02262342e-01
-1.59166843e-01 -4.72616673e-01 3.56498063e-01 -9.67083931e-01
-6.63532436e-01 7.12911129e-01 7.42979467e-01 5.35799384e-01
7.35048354e-01 1.54877722e-01 1.08855712e+00 5.16277850e-01
3.66760463e-01 -1.12076986e+00 -5.60329616e-01 9.10577059e-01
9.50939894e-01 -1.23865008e+00 -9.27510381e-01 -3.08207303e-01
-5.11112928e-01 8.11624765e-01 -1.20118668e-03 9.55038965e-02
7.22702146e-01 4.08057123e-01 8.19023073e-01 2.61315793e-01
-7.48205423e-01 -1.26312569e-01 1.78613529e-01 8.18299711e-01
9.79661226e-01 3.05902600e-01 2.03622758e-01 5.82927227e-01
-8.06505442e-01 -6.71589732e-01 2.02865899e-01 7.02901423e-01
-4.02802557e-01 -1.09360409e+00 -4.86757487e-01 -3.15139383e-01
-6.79373622e-01 -3.82650614e-01 -7.61735320e-01 1.80859163e-01
-4.33712035e-01 1.86813927e+00 8.60272571e-02 -3.63910943e-01
4.37927753e-01 4.32678223e-01 9.36677381e-02 -7.64583528e-01
-7.79908717e-01 8.62902328e-02 7.42104828e-01 -5.27151942e-01
4.49193083e-02 -7.29174256e-01 -1.16833222e+00 -6.17382824e-01
-3.72716516e-01 4.41465735e-01 1.04257166e+00 6.65323794e-01
6.21051490e-01 2.45950997e-01 8.40342045e-01 -4.14189994e-01
-7.50608802e-01 -1.37448609e+00 -8.17506194e-01 -1.84360698e-01
6.45705998e-01 1.27063349e-01 -9.80783522e-01 7.98875187e-03] | [14.52806282043457, 6.548745632171631] |
925ff465-90c7-442c-a683-908d210421f9 | cross-domain-few-shot-learning-via-meta | 2202.05713 | null | https://arxiv.org/abs/2202.05713v3 | https://arxiv.org/pdf/2202.05713v3.pdf | Cross Domain Few-Shot Learning via Meta Adversarial Training | Few-shot relation classification (RC) is one of the critical problems in machine learning. Current research merely focuses on the set-ups that both training and testing are from the same domain. However, in practice, this assumption is not always guaranteed. In this study, we present a novel model that takes into consideration the afore-mentioned cross-domain situation. Not like previous models, we only use the source domain data to train the prototypical networks and test the model on target domain data. A meta-based adversarial training framework (MBATF) is proposed to fine-tune the trained networks for adapting to data from the target domain. Empirical studies confirm the effectiveness of the proposed model. | ['Yongyi Mao', 'Chune Li', 'Richong Zhang', 'Jirui Qi'] | 2022-02-11 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning', 'few-shot-relation-classification', 'relation-classification', 'few-shot-relation-classification'] | ['computer-vision', 'computer-vision', 'methodology', 'natural-language-processing', 'natural-language-processing'] | [ 3.70690405e-01 1.93509549e-01 -2.22833529e-01 -3.47377062e-01
-2.83767372e-01 -3.26166034e-01 7.15700686e-01 -1.10839084e-01
-3.71552020e-01 1.04297280e+00 -2.59873033e-01 -1.89457491e-01
-1.03570364e-01 -1.12751114e+00 -5.21685183e-01 -5.61727643e-01
3.39323968e-01 4.40761000e-01 5.67236483e-01 -4.82782960e-01
3.04857027e-02 3.94032896e-01 -1.32209527e+00 1.90894067e-01
1.11378610e+00 8.58796477e-01 3.49482335e-02 4.32984114e-01
-4.39882800e-02 7.69570410e-01 -9.37658548e-01 -7.38648057e-01
2.50616401e-01 -7.56645560e-01 -7.13077605e-01 -6.62686825e-02
-1.46162689e-01 -2.17003822e-01 -3.74278307e-01 1.28005815e+00
5.48732042e-01 5.71511090e-01 8.79315495e-01 -1.47603297e+00
-8.74671102e-01 4.91847932e-01 -1.96523681e-01 4.69920516e-01
1.96980923e-01 -1.82353705e-01 6.01328969e-01 -4.71790433e-01
7.10593760e-01 8.76891792e-01 4.91798937e-01 8.19945633e-01
-9.20051694e-01 -7.33006239e-01 1.16050810e-01 4.56813753e-01
-1.37549007e+00 -3.72050494e-01 1.05361557e+00 -1.36300549e-01
7.84218252e-01 -4.64416631e-02 4.29096103e-01 1.45671654e+00
3.20980906e-01 4.16243970e-01 1.09729052e+00 -7.21256793e-01
5.27942657e-01 4.86079574e-01 9.99725387e-02 -2.51139030e-02
1.19687766e-01 2.98069835e-01 -6.61095828e-02 -1.43323734e-01
5.68946838e-01 -1.52177468e-01 -2.75885850e-01 -4.06868339e-01
-7.10402787e-01 8.82461071e-01 1.93212524e-01 7.21630752e-01
-2.79417634e-01 -5.86558759e-01 5.87951005e-01 6.27146661e-01
6.37895048e-01 3.55569929e-01 -3.32755774e-01 -3.56521830e-02
-6.16315484e-01 7.36360401e-02 8.81360590e-01 1.04941094e+00
4.02808547e-01 2.10065663e-01 -2.08143070e-02 8.65206003e-01
-6.73086941e-02 1.04492873e-01 9.62183237e-01 -4.55047131e-01
5.08434713e-01 6.03213131e-01 8.64521116e-02 -8.40096235e-01
-1.12325348e-01 -3.81297171e-01 -7.22341776e-01 2.77771503e-01
3.85094792e-01 -2.00661331e-01 -7.61748910e-01 1.71903610e+00
6.18444443e-01 8.15966368e-01 6.73048496e-01 8.05595815e-01
9.28161323e-01 5.26280999e-01 2.68819064e-01 -3.49613398e-01
9.30659592e-01 -9.41265523e-01 -8.47345114e-01 -1.28809199e-01
4.78266865e-01 -5.50898194e-01 1.04709148e+00 1.61602885e-01
-5.88825941e-01 -7.24594712e-01 -1.43913460e+00 3.16615164e-01
-7.51368344e-01 -3.41515541e-01 3.06429982e-01 7.48822689e-01
-3.68196726e-01 7.28304625e-01 -3.84665102e-01 -5.28650999e-01
1.61032617e-01 1.32838324e-01 -4.93710309e-01 -1.92134559e-01
-1.87239707e+00 1.14202011e+00 8.13195467e-01 -9.64017063e-02
-7.52374947e-01 -3.71346861e-01 -5.82518637e-01 -1.27876475e-01
4.18054312e-01 -2.55754739e-01 1.24637771e+00 -1.36210704e+00
-1.66005123e+00 7.91958570e-01 3.29090089e-01 -6.94215655e-01
6.64204299e-01 -1.10240027e-01 -8.60731184e-01 -7.73341060e-02
-1.56049535e-01 -5.34356199e-02 7.71367908e-01 -1.11152434e+00
-5.17008483e-01 -3.53992321e-02 3.25068414e-01 2.29833677e-01
-4.06252980e-01 2.29384266e-02 -7.46803805e-02 -7.59740829e-01
-2.87394047e-01 -5.47255397e-01 1.03457823e-01 -4.19186234e-01
-2.49990523e-01 -2.45144337e-01 1.03286660e+00 -4.04766917e-01
1.13192368e+00 -2.24862552e+00 -1.40483379e-02 7.05892742e-02
-2.50190824e-01 8.80098104e-01 -7.74651617e-02 4.28214878e-01
-3.40907782e-01 -2.55594701e-01 -3.07013124e-01 6.68848082e-02
-1.66146457e-01 3.80706728e-01 -3.47714096e-01 3.56510043e-01
2.29202479e-01 7.96688199e-01 -9.70987320e-01 -6.00638509e-01
1.60596818e-01 2.61878788e-01 -1.94607422e-01 5.52691758e-01
-1.83838412e-01 4.95676577e-01 -6.12420797e-01 3.82732600e-01
7.35359490e-01 8.73356164e-02 3.20466459e-01 -6.74311668e-02
3.78258228e-01 -4.35165167e-02 -1.26730084e+00 1.51457298e+00
-3.09205264e-01 2.52677083e-01 -4.55311775e-01 -1.53270912e+00
1.17834854e+00 5.95677912e-01 2.38215864e-01 -6.54811621e-01
4.55334067e-01 7.91134834e-02 3.68820161e-01 -7.08972216e-01
1.44927233e-01 -5.97368777e-01 -1.66659072e-01 3.45491529e-01
2.53086001e-01 2.20104028e-02 2.54681353e-02 -2.83678442e-01
1.12166691e+00 4.46207672e-01 8.60338092e-01 1.17620282e-01
7.26434469e-01 2.13966817e-01 8.37779045e-01 4.78071094e-01
-6.20931327e-01 5.20044386e-01 3.41908991e-01 -3.46019059e-01
-1.18910408e+00 -8.68081629e-01 -1.76309749e-01 1.05396807e+00
1.88656151e-01 1.00034505e-01 -8.17279160e-01 -1.05886483e+00
-1.80019036e-01 1.12225938e+00 -8.23592305e-01 -6.79197550e-01
-3.12437296e-01 -5.49457729e-01 7.47943223e-01 3.09860080e-01
8.42974722e-01 -1.25667298e+00 -6.40809596e-01 3.93079489e-01
-5.05811572e-02 -1.18287003e+00 -5.33508807e-02 1.46797970e-01
-6.56576753e-01 -1.22643745e+00 -6.93134010e-01 -8.88889074e-01
4.97973859e-01 9.81616303e-02 9.46613073e-01 -1.11288629e-01
1.14367805e-01 2.67854333e-03 -8.61513257e-01 -4.84796524e-01
-8.36285949e-01 1.94957316e-01 1.12153687e-01 1.73552483e-01
8.88821423e-01 -8.23034525e-01 -1.15593538e-01 3.32630783e-01
-1.01939964e+00 -2.46366203e-01 4.99396920e-01 1.03912675e+00
3.40007156e-01 5.52157819e-01 1.07782114e+00 -1.41070580e+00
8.55811417e-01 -6.93945646e-01 -2.87138283e-01 6.34886801e-01
-5.27531385e-01 -2.80358166e-01 1.01034737e+00 -1.05443084e+00
-1.46516633e+00 -2.25485027e-01 -5.07218838e-02 -7.04497576e-01
-5.22292256e-01 4.24580097e-01 -5.45104742e-01 -5.55936620e-02
8.52326512e-01 2.35759467e-01 -1.72020838e-01 -3.03310275e-01
1.20653309e-01 8.82314920e-01 4.96666074e-01 -5.18517375e-01
9.10194278e-01 9.84845087e-02 -2.56165385e-01 -4.32186484e-01
-9.07651186e-01 9.80983023e-03 -9.28194046e-01 -1.94880068e-01
7.07976818e-01 -5.54131210e-01 -9.97740850e-02 4.18407738e-01
-1.10380328e+00 -3.84339690e-01 -4.43240762e-01 3.98983836e-01
-5.57507277e-01 1.38813108e-01 -2.91106045e-01 -9.21809316e-01
-2.43175298e-01 -7.49996066e-01 2.74049282e-01 3.69072407e-01
-9.68716517e-02 -1.15019941e+00 2.64622480e-01 7.64640570e-02
2.95221865e-01 5.09062588e-01 9.18784559e-01 -1.32167208e+00
1.31482422e-01 -4.05764371e-01 4.20547947e-02 6.08035743e-01
3.49415720e-01 -1.83893099e-01 -1.06725430e+00 1.21316351e-02
4.08337057e-01 -4.68522400e-01 1.81126148e-01 -5.38820438e-02
7.52661645e-01 -2.02263817e-02 -2.11110070e-01 3.26918185e-01
1.48943627e+00 6.71202779e-01 7.85438001e-01 5.21485746e-01
3.40130657e-01 6.30335808e-01 1.25110757e+00 2.22833604e-01
1.05308376e-01 6.89042509e-01 2.39147961e-01 3.12766194e-01
7.30913654e-02 -2.87919343e-01 1.01495057e-01 5.23447394e-01
-1.63235739e-02 -3.76786858e-01 -8.54483485e-01 5.00845790e-01
-1.86453080e+00 -1.05806518e+00 3.09823781e-01 2.17602062e+00
8.75593066e-01 4.49077338e-01 -1.05117507e-01 4.04054046e-01
1.02128983e+00 -1.07666880e-01 -6.86969340e-01 -5.52826166e-01
5.84041551e-02 4.16085958e-01 6.38589785e-02 3.23121518e-01
-1.31007469e+00 1.05678880e+00 5.67100525e+00 9.89977300e-01
-1.21001005e+00 4.38759923e-01 3.45816135e-01 1.03402980e-01
6.69260323e-02 -4.19783928e-02 -4.56274390e-01 4.68342990e-01
9.79043722e-01 -3.97762865e-01 3.18853706e-01 1.09619462e+00
-8.42654184e-02 2.90405713e-02 -1.00477076e+00 4.89350736e-01
1.22094803e-01 -7.46555328e-01 -1.05600744e-01 -2.64368176e-01
6.58949196e-01 -4.90434110e-01 -1.89140677e-01 7.37370729e-01
3.24203163e-01 -8.34406257e-01 3.12344253e-01 4.29414272e-01
7.91613817e-01 -9.93112028e-01 1.04995286e+00 5.74647069e-01
-8.73858333e-01 -7.73874670e-02 -6.26369476e-01 -5.70833944e-02
4.45591770e-02 4.17014062e-01 -8.06448400e-01 9.32138443e-01
4.46921349e-01 5.08011937e-01 -3.76132131e-01 8.16485584e-01
-2.06739843e-01 5.10014653e-01 4.30086590e-02 3.93690132e-02
-8.36588964e-02 -7.83357918e-02 4.17742997e-01 8.86754334e-01
2.41565526e-01 3.36547583e-01 -5.51974885e-02 5.02177238e-01
6.18852191e-02 1.86945558e-01 -1.09609354e+00 -1.20208301e-01
6.74094379e-01 9.82489347e-01 -4.95740473e-01 -3.93629313e-01
-5.96631527e-01 1.01836419e+00 5.32951713e-01 3.54123473e-01
-1.01563203e+00 -6.06673658e-01 3.75720650e-01 -8.40989649e-02
2.84701467e-01 1.17571197e-01 -8.43058079e-02 -1.17747927e+00
-1.19385839e-01 -9.79400337e-01 4.71080840e-01 -4.27327514e-01
-1.70175087e+00 7.39846528e-01 1.88188910e-01 -1.63245249e+00
-2.93676257e-01 -3.72090608e-01 -8.42875242e-01 8.67929757e-01
-1.58611929e+00 -1.10538971e+00 -2.15483814e-01 7.75615036e-01
4.66005206e-01 -5.17162859e-01 1.06814003e+00 5.00499964e-01
-6.86824083e-01 9.17541921e-01 1.70693174e-01 3.46377790e-01
8.93339813e-01 -8.29970479e-01 2.30635062e-01 1.02981997e+00
-8.39181617e-02 5.93970299e-01 9.83656228e-01 -8.46204340e-01
-8.45835567e-01 -1.11207795e+00 7.27564096e-01 -1.36872634e-01
7.01148868e-01 -2.24180058e-01 -1.31441414e+00 6.99652374e-01
2.29837954e-01 2.81544209e-01 9.17785227e-01 -1.48832813e-01
-3.61801833e-01 -2.17183724e-01 -1.60528314e+00 3.91000658e-01
9.59870040e-01 -4.17358607e-01 -1.20482111e+00 1.76408693e-01
6.84667706e-01 -4.82224196e-01 -1.05987000e+00 5.06596327e-01
3.71411413e-01 -8.52245867e-01 7.97117651e-01 -9.49750483e-01
4.91496414e-01 -8.51140395e-02 -2.49645978e-01 -1.42790163e+00
-8.70048702e-02 -1.55350551e-01 -2.31012926e-01 1.55732310e+00
1.59398615e-01 -8.48435163e-01 7.27216601e-01 5.98838210e-01
7.72608817e-02 -4.50449258e-01 -1.20610046e+00 -1.07812452e+00
1.76688835e-01 -2.71396369e-01 6.87173188e-01 1.29966259e+00
1.18347276e-02 4.63810921e-01 -3.08259696e-01 1.73846185e-01
3.92464310e-01 1.45631328e-01 7.04775035e-01 -1.26867151e+00
-4.86787915e-01 -4.51869071e-02 -4.99057025e-01 -2.75060743e-01
4.44693238e-01 -6.53451383e-01 8.57344083e-03 -1.19086206e+00
-6.21734411e-02 -5.23402750e-01 -5.56104422e-01 2.58346111e-01
-2.76857436e-01 8.13457668e-02 2.37505627e-03 9.18409005e-02
-2.11980984e-01 7.25077033e-01 1.21382236e+00 7.80544281e-02
-1.66747034e-01 3.85242790e-01 -5.65361977e-01 4.96109366e-01
1.07132661e+00 -6.00617170e-01 -9.51885879e-01 1.10452279e-01
-2.74781197e-01 -3.23062614e-02 1.42341092e-01 -1.19262075e+00
-1.01909377e-02 -5.35774529e-01 1.99503526e-01 -7.69676343e-02
3.99580508e-01 -1.19908786e+00 1.90025270e-01 3.34898561e-01
-3.01944405e-01 -2.78240412e-01 8.22645947e-02 6.99806511e-01
-3.08453798e-01 -7.41722345e-01 9.90998387e-01 -4.90990691e-02
-8.85079265e-01 1.02580398e-01 -3.20719965e-02 2.03701198e-01
1.68898606e+00 -2.34052658e-01 -3.47187638e-01 -2.71776319e-01
-6.34015620e-01 -1.96246549e-01 3.95884901e-01 5.30591846e-01
5.35640776e-01 -1.53064203e+00 -5.47146797e-01 1.25374898e-01
2.73434222e-01 -8.56443495e-02 2.83020169e-01 3.51752758e-01
-1.98087081e-01 -3.67471855e-03 -4.50192988e-01 4.48554419e-02
-1.10408831e+00 1.08269632e+00 4.07882959e-01 -3.86940151e-01
-5.95631897e-01 5.11634111e-01 -6.38105348e-02 -5.93890429e-01
3.57438028e-02 3.35865170e-01 -6.35570943e-01 3.90920043e-02
4.43857282e-01 1.33056238e-01 -8.08455497e-02 -4.60945964e-01
-1.68482319e-01 9.98389199e-02 -1.65891066e-01 -1.30141243e-01
1.24260008e+00 1.35386527e-01 2.37948999e-01 6.63870811e-01
8.77844095e-01 -3.37607116e-01 -1.05197859e+00 -5.67265749e-01
-4.42585722e-02 -4.02053833e-01 -3.60162586e-01 -6.69207335e-01
-7.55677521e-01 8.14379394e-01 5.79601288e-01 3.10147077e-01
1.16194916e+00 -3.62973809e-01 5.84271848e-01 2.93588966e-01
4.35527861e-01 -1.35121381e+00 3.91494632e-02 4.34791386e-01
6.44447446e-01 -1.16761708e+00 -4.96953018e-02 -3.24333817e-01
-8.43843520e-01 9.96343851e-01 1.09440100e+00 -5.06279707e-01
7.03742385e-01 4.13813032e-02 2.92535797e-02 2.35021651e-01
-6.68579638e-01 -1.41895056e-01 3.20983790e-02 1.07487595e+00
3.05719912e-01 -1.57262862e-01 -5.36146700e-01 8.77316475e-01
-7.60574313e-03 2.71591306e-01 4.53173310e-01 1.18521214e+00
-3.05294156e-01 -1.37800622e+00 -3.38581234e-01 3.32365513e-01
-4.56543505e-01 1.34585112e-01 -3.13165545e-01 1.01530898e+00
3.30077767e-01 1.05133164e+00 -9.06614959e-02 -6.87715590e-01
4.26941633e-01 4.63897705e-01 2.74106950e-01 -7.05668747e-01
-5.10539114e-01 -4.63142663e-01 2.00549752e-01 -1.13553867e-01
-5.77286720e-01 -4.21072602e-01 -9.20138597e-01 -2.69551814e-01
-4.08699840e-01 2.59194016e-01 2.41244480e-01 1.20288277e+00
5.60580157e-02 8.04107606e-01 6.51696324e-01 -4.70882744e-01
-6.91518664e-01 -1.21070731e+00 -6.72091126e-01 5.74484825e-01
5.27925463e-03 -1.01631224e+00 -3.23753119e-01 -7.98583254e-02] | [10.216848373413086, 3.1290273666381836] |
4343b255-ddb8-48ae-9091-7344f0043d8b | contextual-reasoning-for-scene-generation | 2305.02255 | null | https://arxiv.org/abs/2305.02255v1 | https://arxiv.org/pdf/2305.02255v1.pdf | Contextual Reasoning for Scene Generation (Technical Report) | We present a continuation to our previous work, in which we developed the MR-CKR framework to reason with knowledge overriding across contexts organized in multi-relational hierarchies. Reasoning is realized via ASP with algebraic measures, allowing for flexible definitions of preferences. In this paper, we show how to apply our theoretical work to real autonomous-vehicle scene data. Goal of this work is to apply MR-CKR to the problem of generating challenging scenes for autonomous vehicle learning. In practice, most of the scene data for AV learning models common situations, thus it might be difficult to capture cases where a particular situation occurs (e.g. partial occlusions of a crossing pedestrian). The MR-CKR model allows for data organization exploiting the multi-dimensionality of such data (e.g., temporal and spatial). Reasoning over multiple contexts enables the verification and configuration of scenes, using the combination of different scene ontologies. We describe a framework for semantically guided data generation, based on a combination of MR-CKR and Algebraic Measures. The framework is implemented in a proof-of-concept prototype exemplifying some cases of scene generation. | ['Daria Stepanova', 'Rafael Kiesel', 'Thomas Eiter', 'Loris Bozzato'] | 2023-05-03 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 9.13592929e-04 5.28353751e-01 8.76650885e-02 -7.06240594e-01
-2.45391473e-01 -6.01658165e-01 9.77545142e-01 4.58668262e-01
-2.10552320e-01 6.62046194e-01 8.15608278e-02 -3.08030277e-01
-6.45791709e-01 -1.27061164e+00 -8.16772103e-01 -2.42575690e-01
-1.27482563e-01 5.92714548e-01 8.72451782e-01 -4.50381070e-01
6.67251274e-02 9.38422978e-01 -2.57850933e+00 6.19435251e-01
4.81322497e-01 7.76018858e-01 3.86955589e-01 5.19523501e-01
-3.63139123e-01 1.09319460e+00 -4.32989925e-01 -4.98407841e-01
2.71673024e-01 2.40623131e-01 -9.87061799e-01 4.55801100e-01
6.23788655e-01 2.25852653e-02 2.60851663e-02 8.31258833e-01
-5.75265512e-02 3.40249419e-01 4.82523561e-01 -1.80191994e+00
-1.39881581e-01 5.22849798e-01 4.34086710e-01 -2.41865069e-01
6.27461016e-01 1.57486349e-02 9.80016708e-01 -6.11590505e-01
9.80338812e-01 1.58495045e+00 2.33456701e-01 4.74821329e-01
-1.26837182e+00 -1.95287517e-03 4.23771441e-01 8.55292380e-01
-1.40909398e+00 -2.12168366e-01 6.53455913e-01 -6.54602468e-01
1.00188518e+00 5.02252042e-01 9.71693039e-01 7.43522584e-01
-1.91151351e-01 9.11512494e-01 1.02788520e+00 -5.44113219e-01
6.25560403e-01 6.10683143e-01 2.93865234e-01 2.86898345e-01
3.53487760e-01 1.52268171e-01 -3.95045310e-01 -3.72240059e-02
5.07654965e-01 -1.88010171e-01 2.51317263e-01 -1.27879119e+00
-9.65569973e-01 7.94656575e-01 1.64580837e-01 3.20337057e-01
-6.49516210e-02 -5.93126714e-02 4.90340978e-01 3.47887069e-01
-1.09097257e-01 4.97100651e-01 -4.87438381e-01 1.02348864e-01
-5.63969493e-01 7.54735172e-01 9.37600493e-01 1.57389736e+00
1.05337167e+00 -2.74782658e-01 -4.47835140e-02 5.76162279e-01
6.24817848e-01 2.51730949e-01 -7.50128478e-02 -1.46700180e+00
1.94972947e-01 7.98540235e-01 2.38152698e-01 -8.56415987e-01
-4.40127045e-01 1.67993873e-01 -2.02217519e-01 5.79970717e-01
1.42623901e-01 1.84996560e-01 -5.55222213e-01 1.67917395e+00
6.31445050e-01 7.36984909e-02 6.75655186e-01 6.99458420e-01
7.43678927e-01 2.84309417e-01 2.68479049e-01 -1.96964100e-01
1.29150045e+00 -5.20210862e-01 -6.28705263e-01 5.30985110e-02
8.45683515e-01 -2.42295131e-01 9.06323552e-01 3.37802351e-01
-9.00100708e-01 -5.29474497e-01 -1.02838898e+00 -3.25686447e-02
-1.06552327e+00 -2.67144352e-01 8.15674543e-01 6.91890776e-01
-1.21618843e+00 2.40542755e-01 -3.66233259e-01 -7.48895228e-01
1.50324494e-01 1.68279469e-01 -3.53750676e-01 -1.77862182e-01
-1.34695745e+00 1.15675986e+00 9.85617757e-01 -2.48098522e-01
-9.87596631e-01 -6.81288421e-01 -1.17157328e+00 -2.51705408e-01
8.09085667e-01 -7.12440908e-01 1.21639144e+00 -5.74998617e-01
-1.34478509e+00 8.74267459e-01 1.71357378e-01 -6.59929693e-01
7.01645851e-01 -9.04054046e-02 -7.08372355e-01 -5.48169501e-02
1.74565479e-01 6.93590283e-01 2.75744051e-01 -1.64414227e+00
-9.44133461e-01 -2.64538795e-01 1.00996387e+00 3.15189600e-01
2.67870903e-01 -2.72227954e-02 -1.23863339e-01 -1.49413859e-02
-1.77134454e-01 -6.96510434e-01 -5.39719820e-01 -2.63905853e-01
-1.09407209e-01 -1.93381980e-01 1.02372420e+00 1.76970199e-01
8.64936054e-01 -2.13827443e+00 1.45634130e-01 5.03287971e-01
-1.08040646e-01 5.68985343e-02 8.24645907e-02 6.01091385e-01
3.13569084e-02 1.81702495e-01 -1.79902494e-01 5.71184121e-02
4.34359550e-01 6.50837898e-01 -3.44615012e-01 -1.72188506e-02
1.88056007e-01 3.39469880e-01 -1.00631618e+00 -7.97938168e-01
7.39595354e-01 3.40213090e-01 -6.24158859e-01 1.12987988e-01
-8.14858139e-01 1.13909423e-01 -5.76377034e-01 4.04984593e-01
6.54787838e-01 3.44261944e-01 5.24379969e-01 -1.78917944e-01
-3.39547694e-01 -8.79245102e-02 -1.88347685e+00 1.65350842e+00
-5.52766383e-01 3.78843307e-01 -1.99980795e-01 -7.39411175e-01
8.55974019e-01 3.23839486e-01 4.85845655e-01 -4.20482188e-01
-2.10331082e-01 -6.32271767e-02 -3.69006425e-01 -8.67842376e-01
8.11953366e-01 9.53101143e-02 -2.76262969e-01 1.91143584e-02
-2.55992673e-02 -7.23687947e-01 6.95862830e-01 3.96173745e-01
8.57604802e-01 4.41449642e-01 5.03929555e-01 -4.51661885e-01
8.37546408e-01 4.49683636e-01 4.37688142e-01 7.16431379e-01
4.96354662e-02 1.43289492e-01 6.21036291e-01 -5.97075045e-01
-1.01363671e+00 -1.05135584e+00 -3.75648141e-01 8.82974148e-01
3.71999115e-01 -8.52866888e-01 -7.42385983e-01 -5.61924577e-01
-6.35714084e-02 1.28868616e+00 -3.57064068e-01 3.50925364e-02
-3.11757326e-01 -2.39380881e-01 3.65713388e-01 2.66444951e-01
2.93621868e-01 -1.14765918e+00 -1.38102245e+00 2.53848672e-01
9.58273783e-02 -1.40123236e+00 5.77056229e-01 -1.05076537e-01
-6.45845950e-01 -1.27086723e+00 3.68401140e-01 -2.23752201e-01
2.71239728e-01 1.43486813e-01 1.36572969e+00 2.69761402e-02
-3.83764476e-01 1.07510757e+00 -6.04117155e-01 -5.93895972e-01
-6.01517200e-01 -4.20501590e-01 -9.81236026e-02 5.33028692e-02
4.97848004e-01 -3.99802208e-01 -8.94559473e-02 4.48513567e-01
-1.26561546e+00 1.50000706e-01 1.79619730e-01 3.72874171e-01
7.82793880e-01 5.71232557e-01 7.51092434e-02 -9.29078996e-01
3.74275684e-01 -4.70011085e-01 -9.70774174e-01 5.95784009e-01
-2.47391880e-01 1.15069807e-01 2.30635494e-01 -2.59242266e-01
-1.20814085e+00 2.34615341e-01 2.88631469e-01 -3.66597563e-01
-8.24642420e-01 4.84548151e-01 -9.26580727e-01 3.50046873e-01
5.45141280e-01 -2.77981579e-01 -4.47910637e-01 4.33182828e-02
8.65149856e-01 3.63363594e-01 1.45317212e-01 -1.13760626e+00
4.73711312e-01 4.36961472e-01 4.19117898e-01 -8.86183679e-01
-5.18611312e-01 -4.45312738e-01 -8.64159226e-01 -5.16541660e-01
9.79693472e-01 -7.67421365e-01 -7.12237179e-01 5.03842607e-02
-1.08225894e+00 -4.82873619e-01 -8.70247662e-01 4.27996397e-01
-1.21543407e+00 1.12161167e-01 1.69749960e-01 -9.73352671e-01
6.38526678e-01 -1.40224504e+00 1.03344929e+00 -9.85632688e-02
-4.59205098e-02 -9.09044564e-01 -5.22636846e-02 2.77501464e-01
1.79194793e-01 4.21676576e-01 9.97398436e-01 -6.35042787e-01
-9.39504981e-01 1.39700651e-01 2.11001728e-02 1.15892375e-02
-1.62135035e-01 3.36377412e-01 -9.68682826e-01 2.25188509e-01
-2.31552988e-01 -2.59087622e-01 2.99511403e-01 3.36018391e-03
1.06514812e+00 -1.45081431e-01 -2.65831202e-01 2.11059973e-02
1.66027451e+00 2.94379622e-01 8.62494171e-01 6.17626011e-01
6.18860662e-01 1.19803488e+00 1.22575653e+00 4.78975475e-01
9.30472255e-01 1.16854095e+00 6.85791671e-01 2.83194870e-01
-2.34373715e-02 -2.02205442e-02 4.94501600e-03 -9.07352716e-02
-1.54343903e-01 9.65291634e-02 -1.06110334e+00 7.00227559e-01
-2.27234101e+00 -1.34001625e+00 -3.74824345e-01 2.24988794e+00
5.34083664e-01 -8.93160552e-02 1.16954997e-01 1.16755150e-01
4.49818701e-01 -1.52684404e-02 -2.75916547e-01 -5.94398439e-01
-2.00467125e-01 -2.52722114e-01 3.68606776e-01 8.95077407e-01
-1.00680220e+00 9.61887479e-01 6.05145502e+00 4.76634830e-01
-3.62578720e-01 -8.93861242e-03 -1.60225928e-02 3.69815230e-02
-5.97185493e-01 4.78150010e-01 -9.11355078e-01 -1.20455362e-01
9.83125806e-01 -1.15776330e-01 4.88833338e-01 1.13900912e+00
-2.49239849e-03 -1.31366909e-01 -1.50171304e+00 7.38826454e-01
8.01036730e-02 -1.29461396e+00 2.42352381e-01 5.45274466e-02
4.33454216e-01 -2.21460119e-01 -4.36979115e-01 2.91225076e-01
8.50193858e-01 -7.54968345e-01 1.12694359e+00 7.50843048e-01
4.51138079e-01 -6.87919915e-01 4.55939710e-01 3.33399147e-01
-1.21833682e+00 -3.04813176e-01 -3.09881747e-01 2.70056635e-01
1.58342734e-01 5.20655632e-01 -8.20723236e-01 1.00988817e+00
7.48144269e-01 6.15196109e-01 -6.67846620e-01 9.84813571e-01
-1.10825365e-02 -2.18182355e-01 -3.22094977e-01 1.63179994e-01
-8.43155012e-02 -1.39333248e-01 7.36129344e-01 1.42956090e+00
3.26490641e-01 -3.74187790e-02 3.03102940e-01 7.43539989e-01
5.18763185e-01 1.62087187e-01 -1.13220561e+00 5.82659304e-01
4.29573238e-01 1.16380072e+00 -4.17383522e-01 -6.30408943e-01
-3.69732380e-01 1.22290745e-01 2.01754510e-01 2.91134834e-01
-7.86400676e-01 -1.40446156e-01 8.10326397e-01 2.45020971e-01
2.11905316e-01 -1.72838554e-01 3.57912816e-02 -1.12683284e+00
1.78898145e-02 -9.30308700e-01 5.37104309e-01 -1.09837878e+00
-1.01562417e+00 5.49404383e-01 1.01647174e+00 -1.37083125e+00
-4.42426056e-01 -6.98469162e-01 -1.03543419e-02 4.16040093e-01
-1.50020552e+00 -1.36104774e+00 -4.29319292e-01 8.84257913e-01
4.88539696e-01 -1.82940111e-01 9.05611157e-01 2.15971544e-02
1.07543627e-02 -1.84992135e-01 -6.76245570e-01 -5.17002642e-01
2.14767620e-01 -1.53749216e+00 -1.41691089e-01 7.51757205e-01
1.10601850e-01 4.60675776e-01 9.04839396e-01 -5.50804496e-01
-1.24369240e+00 -1.23258901e+00 7.47281015e-01 -5.78174233e-01
6.94826066e-01 -3.54140043e-01 -8.10743332e-01 8.68060768e-01
1.08831555e-01 1.94230322e-02 6.09198511e-01 2.77550798e-02
-6.12790942e-01 -3.50301892e-01 -1.50965071e+00 8.98817956e-01
1.29145467e+00 -4.48408365e-01 -7.65918851e-01 4.41352010e-01
7.34567940e-01 -3.68735820e-01 -1.09197628e+00 6.98782980e-01
2.34933123e-01 -1.25741422e+00 1.05805707e+00 -6.00358844e-01
-4.61135656e-02 -9.64998603e-01 -9.23517406e-01 -1.13729978e+00
-1.09164283e-01 -2.55985439e-01 -1.24365844e-01 1.16280901e+00
3.27909946e-01 -4.95073795e-01 2.71127135e-01 9.81165349e-01
-3.36296946e-01 -1.95561752e-01 -9.65562165e-01 -6.94465339e-01
-2.50632375e-01 -1.11661386e+00 1.22425675e+00 7.32223570e-01
1.12464927e-01 -1.52457142e-02 3.40496637e-02 5.51577330e-01
6.27929032e-01 7.97681063e-02 1.02794993e+00 -1.51611471e+00
-2.62351900e-01 -3.28198254e-01 -1.13443625e+00 -4.27617848e-01
2.81086147e-01 -6.36013091e-01 -6.95022717e-02 -1.75990558e+00
-1.30642936e-01 -7.25899577e-01 1.23916622e-02 3.88782382e-01
4.99889791e-01 -4.28202927e-01 5.42826772e-01 -1.13005318e-01
-8.81333172e-01 3.13502550e-01 9.74411905e-01 -1.92720398e-01
-2.20374689e-01 -1.53626472e-01 -3.59583288e-01 7.38269031e-01
8.97692680e-01 -1.59976363e-01 -9.77661788e-01 -2.65105426e-01
6.16232395e-01 -6.31235540e-03 6.32262647e-01 -1.26834095e+00
2.34302774e-01 -8.07592094e-01 -1.90877095e-01 -5.29979467e-01
6.17937505e-01 -1.57189298e+00 7.48327553e-01 -8.16137642e-02
-3.47909778e-01 -1.17900088e-01 2.42835939e-01 4.11300719e-01
-3.37271452e-01 -4.35471982e-01 4.64046687e-01 -3.03550333e-01
-1.32156491e+00 -1.25657812e-01 -4.38274562e-01 -2.32828319e-01
1.67423749e+00 -4.03892428e-01 -3.73665899e-01 5.34396023e-02
-1.02916718e+00 4.29902583e-01 6.26351893e-01 5.78624487e-01
6.50693297e-01 -1.22921193e+00 -4.69727576e-01 1.03814036e-01
6.17740393e-01 1.98526010e-01 2.76520312e-01 2.67307490e-01
-3.64669532e-01 3.58146966e-01 -3.52735609e-01 -7.54694939e-01
-1.36694312e+00 9.45233881e-01 5.63490570e-01 -2.29462180e-02
-5.63596368e-01 6.64430037e-02 7.39321634e-02 -6.48485005e-01
1.84569910e-01 -4.13928956e-01 -6.92741692e-01 1.60333589e-01
4.70285118e-01 2.83007294e-01 9.79377180e-02 -9.37999368e-01
-3.65748316e-01 6.64737642e-01 1.72153220e-01 -2.79397577e-01
1.09205270e+00 -2.84528047e-01 -2.03555509e-01 7.50573277e-01
4.87465143e-01 -8.66526291e-02 -1.16971803e+00 -6.09429292e-02
3.26690108e-01 -4.77874428e-01 -2.95951605e-01 -6.94877684e-01
-5.25877118e-01 5.40872216e-01 5.68286598e-01 5.14246941e-01
9.22733903e-01 3.81861269e-01 -2.89027542e-01 7.52164900e-01
1.02636826e+00 -1.30480242e+00 -2.16218203e-01 5.54787099e-01
1.12040091e+00 -9.72541928e-01 6.33177441e-03 -6.01747155e-01
-9.68012571e-01 1.20925307e+00 7.60326684e-01 3.20239991e-01
7.47953176e-01 2.89874941e-01 4.77760322e-02 -5.84325850e-01
-8.77454996e-01 -6.84911847e-01 1.29435705e-02 1.03464866e+00
-1.18321896e-01 4.04760957e-01 -4.35223877e-02 6.22715876e-02
-3.23666364e-01 8.93961117e-02 9.05561030e-01 1.11528766e+00
-5.12839973e-01 -1.31855202e+00 -3.55120450e-01 1.02987476e-01
-2.42774598e-02 3.56038183e-01 -2.17116505e-01 9.89810467e-01
6.55355930e-01 1.21265984e+00 8.71166214e-02 -1.85712263e-01
8.91147077e-01 3.59067507e-02 5.06896973e-01 -7.87619233e-01
-2.48122633e-01 -4.01472569e-01 4.59180385e-01 -7.19876230e-01
-9.87865984e-01 -1.02678430e+00 -1.26955092e+00 1.15840934e-01
6.68963864e-02 1.25734344e-01 7.49926567e-01 8.35203767e-01
2.32670769e-01 4.63636965e-01 2.02934533e-01 -5.61598957e-01
-5.80077022e-02 -4.70626831e-01 -5.22368014e-01 6.56448185e-01
2.43650496e-01 -8.59433651e-01 -1.30579050e-03 2.85679370e-01] | [8.768560409545898, 6.7373881340026855] |
3bfa6cdb-5171-4bb3-950b-7f0540395a72 | arsc-net-adventitious-respiratory-sound | null | null | https://ieeexplore.ieee.org/document/9669787/ | https://ieeexplore.ieee.org/document/9669787/ | ARSC-Net: Adventitious Respiratory Sound Classification Network Using Parallel Paths with Channel-Spatial Attention | Automatic identification of adventitious respiratory sound has still been a challenging problem in recent years. To address this challenge, we propose an adventitious respiratory sound classification network (ARSC-Net), which combines residual block with channel-spatial attention for accurate classification. Specifically, we extract two types of features from adventitious respiratory sound, including Mel-Frequency Cepstral Coefficients (MFCCs) and Mel-spectrogram. The two types of features are entered into the parallel encoders paths with residual attention for extracting feature representation, and then fused into a channel-spatial attention module to adaptively focus on the important features between channel and spatial part for the classification task. Moreover, the channel-spatial attention can enhance the feature representation, in which the channel attention explores the inter-channel relationship of the spectrums, and then the inter-spatial correlation mapping is generated by the spatial attention introduced serially. We evaluate our proposed method on ICBHI 2017 database. Experimental results show that our proposed method achieves encouraging predictive performance with an accuracy of 80.0% for identifying abnormal sounds from normal sounds, and with an accuracy of 92.4% for distinguishing crackles from wheezes. In addition, our method also achieves a score of 56.76% for the four-class sound classification of adventitious sounds and outperforms several state-of-the-art methods. | ['Jianxin Wang', 'Fan Wu', 'Hulin Kuang', 'Jin Liu', 'Jianhong Cheng', 'Lei Xu'] | 2022-01-14 | null | null | null | ieee-international-conference-on-3 | ['sound-classification'] | ['audio'] | [ 1.33786062e-02 -5.62647164e-01 1.82971731e-01 1.34455889e-01
-9.74741161e-01 -2.25763902e-01 -3.54480296e-02 1.06472924e-01
-3.10723782e-01 2.48053864e-01 3.25897813e-01 -2.45745227e-01
-2.95703381e-01 -4.27322596e-01 -2.17795536e-01 -7.58915722e-01
-1.09584682e-01 -3.05477351e-01 4.77150291e-01 5.83108477e-02
2.29949221e-01 2.34410018e-01 -1.58698034e+00 3.83848131e-01
6.37880862e-01 1.40081596e+00 3.18124741e-01 1.18653917e+00
-2.62894183e-01 6.72715843e-01 -7.27119327e-01 1.49991557e-01
-2.43578181e-01 -6.11077666e-01 -7.08228171e-01 -5.52572727e-01
-1.11588255e-01 -1.39562860e-01 -2.56239295e-01 9.74490404e-01
9.46886897e-01 3.04941207e-01 4.60983157e-01 -9.58051920e-01
-4.75658208e-01 3.61807525e-01 -4.68668014e-01 8.91317606e-01
2.79115975e-01 9.14697200e-02 1.13057995e+00 -9.51093018e-01
-2.88291574e-01 1.03387618e+00 7.28002548e-01 3.57286245e-01
-5.42114913e-01 -1.12285829e+00 -1.32660672e-01 5.34877837e-01
-1.61599398e+00 -3.26336563e-01 8.25694799e-01 -4.65442270e-01
9.42335904e-01 6.20584548e-01 5.25152922e-01 5.38904965e-01
3.50277461e-02 4.86059546e-01 7.97844470e-01 -2.19384626e-01
-1.94223419e-01 3.17743495e-02 4.28661108e-02 4.06177342e-01
-1.97477102e-01 8.81155059e-02 -5.39758980e-01 -1.00565314e-01
6.07249618e-01 1.68729842e-01 -6.02485538e-01 7.35972345e-01
-1.13268960e+00 6.13397777e-01 5.44683456e-01 6.83630109e-01
-4.75502998e-01 7.04786628e-02 4.08933669e-01 5.68650067e-02
2.44669378e-01 3.78084093e-01 -4.40658629e-01 -3.62919003e-01
-7.59317160e-01 -9.46656987e-02 4.92451221e-01 4.47935492e-01
2.07648754e-01 1.60665601e-01 -3.66770893e-01 1.07650042e+00
5.40169537e-01 5.01498818e-01 8.72506857e-01 -5.39128482e-01
6.04750454e-01 1.85117975e-01 -2.08065137e-01 -9.48531985e-01
-6.31142318e-01 -7.77504742e-01 -9.68356788e-01 -4.70982760e-01
-2.33398438e-01 -1.80360258e-01 -5.29954016e-01 1.43385696e+00
3.91181082e-01 6.85608685e-01 -8.75844620e-03 9.58651662e-01
1.16358280e+00 9.51459467e-01 1.58527464e-01 -5.80723584e-01
1.52500248e+00 -8.95311058e-01 -9.28737164e-01 2.78922677e-01
1.77260280e-01 -1.04076314e+00 9.65108871e-01 2.30138138e-01
-1.09842324e+00 -1.11014652e+00 -9.28025961e-01 1.53515220e-01
-5.11409119e-02 2.43969992e-01 -1.02665655e-01 4.72384363e-01
-6.94870353e-01 5.23444533e-01 -6.43786430e-01 1.36351004e-01
2.72574216e-01 4.37761635e-01 -5.70252258e-03 4.01311070e-01
-1.51936054e+00 1.75933376e-01 2.48174667e-01 1.37418196e-01
-4.98773873e-01 -8.07273567e-01 -4.72490281e-01 3.57354552e-01
-8.68309103e-03 -4.39680666e-01 1.28435791e+00 -3.64094287e-01
-1.44918835e+00 2.22492307e-01 -8.38636607e-02 -2.36848190e-01
-9.51115340e-02 -7.74077028e-02 -9.95985806e-01 4.35227215e-01
-8.62300023e-02 2.14534134e-01 8.59576344e-01 -6.69436336e-01
-1.05119240e+00 -1.37110129e-01 -3.61341238e-01 3.42719644e-01
-6.44233942e-01 2.21082404e-01 -3.70088100e-01 -7.98816681e-01
2.39188477e-01 -5.70618570e-01 1.32080719e-01 -4.30721462e-01
-3.24097157e-01 -4.34865862e-01 7.74541855e-01 -8.23047280e-01
1.86731625e+00 -2.66386318e+00 -2.38010108e-01 5.31262606e-02
2.97299713e-01 5.89090049e-01 -1.18581451e-01 1.87716663e-01
-2.19561562e-01 2.95460314e-01 -5.92220500e-02 -1.67359225e-02
-2.49588147e-01 -3.36367399e-01 -3.16264570e-01 2.04691872e-01
1.70838624e-01 5.25602460e-01 -8.74987841e-01 -5.73681414e-01
3.35872680e-01 7.91630208e-01 -4.98932600e-01 6.87982082e-01
4.06372011e-01 5.64968526e-01 -5.96525013e-01 3.49014014e-01
6.12056792e-01 -3.71816829e-02 -3.55655611e-01 -3.07512641e-01
-2.29500845e-01 6.76740348e-01 -1.17960083e+00 1.20000756e+00
-7.65797794e-01 3.86683077e-01 -1.89182069e-02 -6.03364944e-01
9.66499686e-01 7.85780191e-01 5.85814297e-01 -5.97358763e-01
1.70770422e-01 3.66341114e-01 4.64052171e-01 -1.02958727e+00
9.07855257e-02 -2.43413046e-01 2.43615717e-01 2.72976607e-01
-1.37836924e-02 -3.34542021e-02 -3.70414019e-01 -2.45090097e-01
8.96740258e-01 -5.26798069e-01 2.13629022e-01 2.01398903e-03
1.09350121e+00 -6.83108807e-01 3.72518331e-01 3.57743740e-01
-3.30399603e-01 7.17404604e-01 -3.25356498e-02 2.04382297e-02
-4.87668514e-01 -8.33735704e-01 -4.11213487e-01 1.09702754e+00
1.55392112e-02 -4.00211930e-01 -7.04243302e-01 -5.76163471e-01
-1.95403308e-01 4.23215717e-01 -4.65044260e-01 -3.08362782e-01
-5.78898191e-01 -6.53504848e-01 7.71034777e-01 8.35096836e-01
6.35540724e-01 -1.27912617e+00 -3.63056004e-01 3.57490420e-01
-2.97443330e-01 -8.35042119e-01 -7.85922468e-01 2.33789831e-01
-5.01915097e-01 -9.97924805e-01 -6.24859810e-01 -8.08252573e-01
6.56418502e-02 3.93999577e-01 6.72314286e-01 3.87888134e-01
-4.86900926e-01 2.83801228e-01 -5.83804190e-01 -4.62061107e-01
-9.14492682e-02 8.18606690e-02 -4.36722748e-02 3.44875991e-01
1.48197502e-01 -5.97363770e-01 -1.03626919e+00 2.34191120e-01
-7.17561722e-01 -3.75547171e-01 4.41568881e-01 8.04849446e-01
6.84825778e-01 3.40552390e-01 9.36250567e-01 -3.80849183e-01
8.37509394e-01 -8.05749118e-01 -4.70370613e-02 -2.40963474e-01
-2.56305039e-01 -4.53463554e-01 1.11485291e+00 -3.68920416e-01
-7.69025803e-01 -1.47845030e-01 -8.92497778e-01 -5.17170012e-01
-1.61618620e-01 2.15643525e-01 8.78366083e-02 2.04503745e-01
3.87965858e-01 5.72461128e-01 -3.72668624e-01 -7.07286179e-01
-4.32028547e-02 1.43742013e+00 6.06898248e-01 -1.11781672e-01
7.40751624e-01 6.70838505e-02 -2.12206095e-01 -9.30028260e-01
-6.44310772e-01 -9.87781465e-01 -3.76115173e-01 3.60321850e-02
1.25155818e+00 -8.67290020e-01 -7.54589856e-01 4.79438573e-01
-1.06458688e+00 1.08482525e-01 -2.46714577e-01 1.03033757e+00
-2.27885991e-01 3.24068248e-01 -5.63475251e-01 -1.10716295e+00
-8.08215261e-01 -1.16843200e+00 9.69223261e-01 3.78343523e-01
-1.46662951e-01 -7.44165063e-01 4.58145030e-02 2.66329169e-01
5.65062821e-01 -2.58902550e-01 8.56236100e-01 -8.57056260e-01
-1.67923138e-01 -6.57293573e-02 -2.50640273e-01 4.55424219e-01
3.18801612e-01 -9.27872211e-02 -1.33735144e+00 -1.88106552e-01
1.73264906e-01 -5.59267923e-02 8.24714243e-01 1.89432129e-01
1.91219580e+00 -2.79640138e-01 -3.99649404e-02 7.01134861e-01
9.00820017e-01 8.12257528e-01 3.90354425e-01 -2.41948828e-01
7.66979158e-01 2.86310226e-01 5.82873702e-01 3.99103791e-01
2.91097224e-01 5.95985413e-01 4.67516363e-01 -7.43944347e-02
-3.78334820e-01 -2.70589024e-01 3.18563692e-02 1.62556124e+00
-1.75240077e-02 -2.43954152e-01 -6.49301589e-01 6.27459109e-01
-1.19713414e+00 -7.72747934e-01 -3.21557373e-01 2.18143725e+00
8.28572392e-01 -1.94414541e-01 1.30001098e-01 6.28486276e-01
8.41839015e-01 2.89816469e-01 -3.55043173e-01 -3.71967047e-01
5.02666414e-01 5.86343527e-01 -1.10220611e-02 4.64158833e-01
-1.22298837e+00 2.82073349e-01 5.58386230e+00 1.16243708e+00
-1.33420861e+00 2.38741010e-01 5.11884689e-01 -1.78115778e-02
-1.27768919e-01 -6.56895816e-01 -6.23245180e-01 8.74397159e-01
1.14547229e+00 1.29199877e-01 4.98461246e-01 5.69113851e-01
9.48777273e-02 5.18725991e-01 -6.05328858e-01 1.17902637e+00
-7.43109360e-02 -8.72023880e-01 -2.69039363e-01 -3.06284726e-01
3.00980479e-01 -1.70605741e-02 3.08510005e-01 2.29577914e-01
-4.56642747e-01 -1.12322295e+00 2.50666052e-01 4.29417223e-01
1.21500254e+00 -9.91934299e-01 7.64904618e-01 2.47568220e-01
-1.88971436e+00 -3.12579900e-01 -3.16318125e-01 3.04470062e-01
-2.31967866e-01 4.20124918e-01 -8.16997349e-01 5.84758878e-01
8.62406969e-01 4.51601118e-01 -2.77757317e-01 1.22338891e+00
-1.27182817e-02 1.08853149e+00 -2.59019673e-01 -2.20566452e-01
1.30789295e-01 3.36195201e-01 5.07427692e-01 1.46121216e+00
5.40699899e-01 4.28659260e-01 -1.10948846e-01 5.01615107e-01
-3.28596383e-02 5.91315567e-01 -2.17537254e-01 2.07898673e-02
7.33802199e-01 1.21814513e+00 -3.73006731e-01 -2.60238171e-01
-3.21349770e-01 7.58112252e-01 -1.95754230e-01 1.60996005e-01
-9.86396134e-01 -1.05519962e+00 4.18243587e-01 -7.55570531e-02
3.61028522e-01 3.92834634e-01 4.92097959e-02 -5.65392315e-01
-1.37814805e-01 -8.64461303e-01 4.67229277e-01 -7.83338189e-01
-1.07360983e+00 1.01837754e+00 -4.76291209e-01 -1.42320192e+00
3.63794602e-02 -2.29095265e-01 -1.05381441e+00 1.15457428e+00
-1.67965209e+00 -6.96868300e-01 -4.96045113e-01 7.69519031e-01
7.08051503e-01 -1.75244287e-01 8.65368962e-01 6.92069829e-01
-7.36401081e-01 9.07199979e-01 -5.41592352e-02 6.23886473e-02
5.20100236e-01 -1.23897386e+00 2.47395277e-01 5.37302732e-01
2.06477448e-01 6.13982141e-01 9.77213383e-02 -2.75761694e-01
-1.02319026e+00 -1.48108900e+00 8.54017794e-01 2.04424746e-02
4.62917894e-01 9.24435258e-02 -1.15839326e+00 4.65770289e-02
3.45736519e-02 3.63003284e-01 8.82492959e-01 -2.89674699e-01
-3.04162621e-01 -3.10425222e-01 -8.45408559e-01 9.15221348e-02
6.66349530e-01 -8.76256168e-01 -4.37391967e-01 2.44849294e-01
1.13397586e+00 -4.44982350e-01 -1.06220317e+00 4.10984963e-01
5.70340216e-01 -7.74533451e-01 1.07907248e+00 -3.07130575e-01
2.19662935e-01 -5.46793818e-01 -2.22659200e-01 -1.29311979e+00
-5.99671245e-01 -5.73092163e-01 3.20456340e-03 1.31151950e+00
2.15846881e-01 -6.74565077e-01 4.32031043e-02 -2.72940487e-01
-4.15369660e-01 -9.82631326e-01 -8.44861150e-01 -3.92441660e-01
-1.03652477e-01 -6.82565391e-01 7.51827955e-01 6.87475681e-01
-1.41956076e-01 4.37591761e-01 -2.68458754e-01 2.80455649e-01
-3.20753194e-02 -1.30881155e-02 1.76223889e-01 -1.04617941e+00
-4.68921006e-01 -4.65829372e-01 -2.59637147e-01 -1.05843747e+00
-1.35708615e-01 -8.63503039e-01 2.14404136e-01 -1.23230481e+00
6.14482015e-02 -4.77131546e-01 -1.03639150e+00 2.20155910e-01
-5.97099185e-01 4.39345658e-01 1.39445111e-01 1.41315207e-01
-4.73248482e-01 4.20827568e-01 1.38053489e+00 3.11448779e-02
-4.46533978e-01 4.47274297e-01 -6.00101769e-01 8.35157514e-01
9.17450666e-01 -4.14258689e-01 -5.22643864e-01 3.04823760e-02
-2.40689397e-01 3.65565032e-01 3.11432213e-01 -1.30060518e+00
1.89186573e-01 7.79620111e-02 3.18386674e-01 -8.05695772e-01
3.40028495e-01 -7.52971828e-01 -7.56338909e-02 6.48017406e-01
-4.72005546e-01 -1.21380851e-01 2.71937519e-01 6.52279198e-01
-5.05754590e-01 -1.23658374e-01 9.64750409e-01 1.64333507e-01
-2.86397368e-01 4.57724214e-01 -2.86919594e-01 2.22321257e-01
7.35597491e-01 2.77974129e-01 -2.55038261e-01 -1.95757523e-01
-5.95260739e-01 -2.86392309e-02 -5.20092845e-01 4.38666463e-01
8.21274638e-01 -1.40455306e+00 -6.99186921e-01 6.69830620e-01
-4.96160286e-03 -8.61827657e-02 7.73981094e-01 1.03008807e+00
-4.82991636e-01 4.34094608e-01 1.14106029e-01 -6.32530749e-01
-1.42758000e+00 4.38701689e-01 5.41483760e-01 -1.05745234e-01
-4.30647790e-01 1.23202312e+00 4.45243239e-01 -7.42300823e-02
3.26879442e-01 -4.72330123e-01 -7.07361817e-01 1.03315845e-01
7.16490149e-01 5.81652820e-01 8.75248015e-02 -7.33382523e-01
-6.09447241e-01 9.58913147e-01 1.76912412e-01 1.22544162e-01
9.89828050e-01 -1.97046116e-01 -9.84312594e-02 3.96416754e-01
1.65362942e+00 3.41012150e-01 -7.00627208e-01 -1.04990393e-01
-6.12797737e-01 -3.58419508e-01 1.72018632e-01 -8.17943454e-01
-1.24569035e+00 1.54169273e+00 9.27322149e-01 5.84079802e-01
1.48794448e+00 -1.82299435e-01 1.34920180e+00 -3.94232683e-02
-3.46610844e-01 -7.57412732e-01 3.62437248e-01 3.84949028e-01
8.61573398e-01 -8.89104724e-01 -4.39743251e-01 -2.33714014e-01
-5.16216755e-01 1.12900805e+00 5.31124294e-01 -1.06021734e-02
9.35914874e-01 1.36989489e-01 4.52710800e-02 -3.24945860e-02
-6.55853570e-01 -3.48000348e-01 5.89726329e-01 5.14049828e-01
5.78092158e-01 1.20045245e-01 -2.11303234e-01 1.22949111e+00
-3.26140374e-01 -4.63272035e-01 -1.08764410e-01 2.50749677e-01
-7.02582955e-01 -6.96272314e-01 -5.51987350e-01 5.34500599e-01
-1.09541893e+00 -3.81477028e-01 -6.71285391e-02 3.51684809e-01
5.57223201e-01 1.31033218e+00 2.84825623e-01 -9.70839322e-01
3.66304517e-01 1.61218986e-01 -1.13669038e-01 -5.54468155e-01
-9.38469708e-01 6.12869680e-01 -2.05494642e-01 -2.73367018e-01
-9.66997817e-02 -3.93898726e-01 -1.57970595e+00 2.84974694e-01
-4.63491380e-01 6.07446492e-01 4.61451113e-01 7.01328754e-01
2.60368228e-01 1.23250794e+00 1.30151987e+00 -3.89664054e-01
-4.81100231e-01 -1.08923888e+00 -5.63587010e-01 3.49558055e-01
8.81775379e-01 -3.50923002e-01 -6.14944696e-01 -1.81974426e-01] | [15.041692733764648, 4.889824390411377] |
c829eca7-1fa5-46c5-a8a0-e369c39d0806 | continuous-time-q-learning-for-mckean-vlasov | 2306.16208 | null | https://arxiv.org/abs/2306.16208v2 | https://arxiv.org/pdf/2306.16208v2.pdf | Continuous Time q-learning for McKean-Vlasov Control Problems | This paper studies the q-learning, recently coined as the continuous time counterpart of Q-learning by Jia and Zhou (2023), for continuous time Mckean-Vlasov control problems in the setting of entropy-regularized reinforcement learning. In contrast to the single agent's control problem in Jia and Zhou (2023), the mean-field interaction of agents renders the definition of the q-function more subtle, for which we reveal that two distinct q-functions naturally arise: (i) the integrated q-function (denoted by $q$) as the first-order approximation of the integrated Q-function introduced in Gu, Guo, Wei and Xu (2023), which can be learnt by a weak martingale condition involving test policies; and (ii) the essential q-function (denoted by $q_e$) that is employed in the policy improvement iterations. We show that two q-functions are related via an integral representation under all test policies. Based on the weak martingale condition and our proposed searching method of test policies, some model-free learning algorithms are devised. In two examples, one in LQ control framework and one beyond LQ control framework, we can obtain the exact parameterization of the optimal value function and q-functions and illustrate our algorithms with simulation experiments. | ['Xiang Yu', 'Xiaoli Wei'] | 2023-06-28 | null | null | null | null | ['q-learning'] | ['methodology'] | [-2.83282399e-01 3.59740824e-01 -3.94216746e-01 2.65842885e-01
-9.04244661e-01 -4.42169130e-01 2.10485771e-01 2.96739340e-01
-8.72668207e-01 1.63866389e+00 -1.90096125e-01 -3.17731827e-01
-8.18406522e-01 -6.73593760e-01 -8.95615101e-01 -1.08594549e+00
-5.51723897e-01 1.77297890e-01 -9.27459672e-02 -2.40866318e-01
4.36361253e-01 -3.18569057e-02 -1.24557507e+00 -7.18159199e-01
1.29793167e+00 1.15591383e+00 2.43512884e-01 7.49162793e-01
1.27557591e-01 6.27892971e-01 -5.70871115e-01 3.09735704e-02
5.90852976e-01 -7.80944288e-01 -5.34881234e-01 1.44826531e-01
-1.64098650e-01 -2.65254140e-01 -2.19306126e-02 1.49615300e+00
5.92838943e-01 5.70094585e-01 5.91819942e-01 -1.41504002e+00
-5.36645949e-01 2.72047371e-01 -6.00309789e-01 9.08749774e-02
1.07234873e-01 5.15122235e-01 8.69638205e-01 -2.69418329e-01
5.03079057e-01 1.25290871e+00 5.00694036e-01 7.10728347e-01
-1.09061301e+00 -3.65851521e-01 1.69009387e-01 1.56791121e-01
-1.01977360e+00 2.66906977e-01 3.77778828e-01 -4.22333598e-01
5.43136895e-01 -5.24421446e-02 9.76367176e-01 5.21672666e-01
7.61365891e-01 8.34756374e-01 1.53241694e+00 -5.28033137e-01
7.70710707e-01 8.12446028e-02 -2.38307953e-01 7.48127401e-01
3.19272369e-01 6.96585000e-01 1.95456311e-01 -2.37357140e-01
8.69297743e-01 -1.89374965e-02 -2.01170906e-01 -6.65168107e-01
-9.22531784e-01 1.13674009e+00 -1.07540451e-01 3.58744785e-02
-8.06495786e-01 2.24020258e-01 3.74766797e-01 7.78346181e-01
3.34611326e-01 5.31889081e-01 -5.82234979e-01 -1.84248492e-01
-5.01853168e-01 7.07136273e-01 8.25164616e-01 8.82951438e-01
5.69219232e-01 4.11942989e-01 -5.18957615e-01 3.97849470e-01
3.34674925e-01 6.92327976e-01 4.08681363e-01 -1.35816383e+00
4.98534590e-01 6.58826530e-02 8.56502771e-01 -2.18289033e-01
-1.93376645e-01 -6.00415826e-01 -5.60051322e-01 3.71792197e-01
6.32901609e-01 -7.99329996e-01 -3.45089972e-01 2.09195924e+00
4.23245192e-01 1.05598547e-01 1.59245893e-01 6.30514681e-01
-2.74722606e-01 5.54613113e-01 -1.30856633e-01 -1.16099298e+00
8.18604529e-01 -8.02358091e-01 -1.00001383e+00 3.07366252e-01
4.66971129e-01 -1.50808036e-01 9.84361529e-01 3.57093543e-01
-1.45761704e+00 -4.05020893e-01 -9.96187210e-01 8.86464834e-01
-7.93502033e-02 -5.68285108e-01 1.70438290e-01 6.32055402e-01
-1.19485688e+00 8.36766958e-01 -3.90626401e-01 -4.04762030e-01
2.64411867e-01 3.50730985e-01 1.56395078e-01 4.11771059e-01
-1.44659865e+00 8.90408933e-01 3.42182010e-01 -1.12112664e-01
-1.28439987e+00 -6.49724841e-01 -5.28759241e-01 -1.24336310e-01
1.00176132e+00 -4.83900845e-01 1.55888903e+00 -9.07428622e-01
-1.98050964e+00 1.41110510e-01 4.08942431e-01 -4.76025820e-01
9.70162332e-01 -7.85405338e-02 -2.65288986e-02 3.50238264e-01
3.84429663e-01 1.67967901e-01 9.25351620e-01 -1.12337804e+00
-8.67311716e-01 -2.81443775e-01 1.82349265e-01 3.40091914e-01
3.16763707e-02 -3.24336946e-01 5.56451142e-01 -5.90029418e-01
-6.64544165e-01 -7.78214276e-01 -5.08037090e-01 -2.09020346e-01
-5.91215380e-02 -4.61205930e-01 4.41024870e-01 -5.29693782e-01
1.06034827e+00 -1.83153570e+00 5.73947951e-02 1.98830321e-01
-2.25296915e-01 1.96044639e-01 -1.83850259e-01 7.08408773e-01
-3.24306376e-02 1.75208107e-01 -4.05211776e-01 2.48894125e-01
3.03889960e-01 2.01236382e-01 -1.75798699e-01 7.73572385e-01
1.28170297e-01 9.16766286e-01 -1.19671655e+00 -4.80456978e-01
7.46969581e-02 -1.15822494e-01 -5.86658239e-01 3.09128582e-01
-4.57562506e-01 6.59687042e-01 -8.92314017e-01 1.82964757e-01
5.35968900e-01 1.30617812e-01 -4.54722047e-02 4.83620077e-01
-4.02843922e-01 -4.45703447e-01 -1.54765856e+00 1.05527771e+00
-2.49200940e-01 -1.02570504e-01 5.99489987e-01 -1.62820864e+00
7.95138478e-01 4.66472149e-01 8.59082878e-01 -7.84457684e-01
1.03622735e-01 1.37383580e-01 -2.60973215e-01 -6.97493374e-01
1.04549170e-01 -7.50710309e-01 -2.01107636e-01 5.49268842e-01
-3.49132973e-03 -3.41587305e-01 4.51397121e-01 -1.58181280e-01
8.48563135e-01 2.61284322e-01 5.66167295e-01 -9.13152754e-01
8.47443402e-01 -2.70578951e-01 7.45696127e-01 1.08041942e+00
-6.72499955e-01 -2.11111471e-01 8.32907498e-01 -1.43390251e-02
-1.04392374e+00 -9.69504058e-01 -1.48129463e-01 7.67674267e-01
1.07811503e-01 1.42066061e-01 -6.63181782e-01 -7.16937661e-01
3.91068757e-01 7.00462461e-01 -7.27353811e-01 -2.02765688e-01
-4.82917517e-01 -3.90855372e-01 -1.74494553e-02 1.08119011e-01
7.74217069e-01 -1.11892545e+00 -8.76756549e-01 5.63293159e-01
2.43613854e-01 -4.02313977e-01 -8.14635456e-01 3.09779555e-01
-8.74580860e-01 -1.29936600e+00 -8.74661028e-01 -3.80672783e-01
2.37694800e-01 -3.41189772e-01 7.87742376e-01 -3.01998198e-01
1.32665589e-01 1.06058753e+00 -1.07304826e-01 -4.93690938e-01
-3.66159558e-01 -3.76532793e-01 3.48319292e-01 3.68297324e-02
-2.20699206e-01 -2.90450811e-01 -7.09955871e-01 2.43288711e-01
-9.39865112e-01 -5.76949537e-01 3.18454802e-01 1.09870064e+00
7.23552108e-01 2.60536313e-01 1.34084094e+00 -3.93521637e-01
1.01669061e+00 -4.18150485e-01 -1.27751386e+00 3.31680834e-01
-9.67599988e-01 4.82796490e-01 8.86840284e-01 -5.58165073e-01
-9.48060751e-01 -3.92264932e-01 1.16581313e-01 -1.97392181e-01
2.95629442e-01 3.04676712e-01 1.12642173e-03 -1.18559502e-01
5.84212504e-02 3.58367383e-01 4.33945239e-01 -2.74755061e-01
2.95457393e-01 4.15498018e-01 5.77254221e-02 -1.01078594e+00
6.78323627e-01 6.24082312e-02 2.77045041e-01 -5.36809981e-01
-7.13900864e-01 -9.77512300e-02 -2.46152058e-01 -2.62312502e-01
9.95600522e-01 -4.50171590e-01 -1.47090983e+00 1.11780345e-01
-6.93934500e-01 -4.60794032e-01 -1.00144553e+00 7.18408465e-01
-1.49904931e+00 2.97309726e-01 -5.77070713e-01 -1.55128121e+00
-6.17587715e-02 -9.43186820e-01 4.19227362e-01 5.14618516e-01
6.07728958e-01 -1.24290490e+00 5.77344537e-01 -2.62639314e-01
2.92854369e-01 4.80544031e-01 8.86379540e-01 -1.96683794e-01
-2.15807825e-01 6.79686815e-02 2.40085438e-01 5.29948473e-01
-2.65177995e-01 -4.20139700e-01 -2.63638169e-01 -8.34989727e-01
5.25110900e-01 -5.46326578e-01 4.45607960e-01 8.94773841e-01
8.51816416e-01 -6.51682377e-01 3.82455736e-02 2.75191486e-01
1.78735065e+00 1.04811370e+00 2.12414235e-01 5.48257411e-01
1.39378225e-02 4.60621715e-01 9.43821847e-01 7.73653090e-01
1.73898876e-01 3.73012275e-01 5.92190683e-01 4.27871466e-01
8.35396051e-01 -3.31774473e-01 5.39466560e-01 6.03727102e-01
-9.91321281e-02 2.00465843e-01 -3.30862999e-01 3.26661319e-01
-2.18540788e+00 -1.15717435e+00 4.03605074e-01 2.67607307e+00
8.93945098e-01 -1.07933044e-01 6.01307034e-01 -2.08608285e-01
9.28876638e-01 -7.70224333e-02 -9.53958869e-01 -6.27716839e-01
-2.93061212e-02 2.04835534e-01 7.58527279e-01 8.23125362e-01
-8.09189439e-01 3.12338144e-01 6.75616121e+00 9.40948009e-01
-6.42186880e-01 1.88827351e-01 7.24724054e-01 1.66005284e-01
-1.59429058e-01 -1.04208104e-01 -5.69964230e-01 6.37221873e-01
8.59811008e-01 -6.32899821e-01 5.02459347e-01 8.07208300e-01
6.61651909e-01 -3.24690878e-01 -7.44572222e-01 6.12619400e-01
-4.87064362e-01 -7.95995176e-01 -6.54681504e-01 4.41616237e-01
1.09337080e+00 -4.74588156e-01 -6.20932784e-03 8.01007569e-01
4.57654655e-01 -8.91330123e-01 6.87444687e-01 6.04052246e-01
5.47447681e-01 -1.11179280e+00 6.73822641e-01 4.90553468e-01
-1.20695901e+00 -6.08793139e-01 -3.47445935e-01 -2.53792882e-01
7.06402287e-02 1.20443739e-01 -1.40600771e-01 6.82389855e-01
1.57392859e-01 4.50196713e-01 -1.21428119e-02 1.16622317e+00
-2.21266504e-02 5.24385571e-01 -7.94138238e-02 -5.93085229e-01
6.17110431e-01 -7.76148975e-01 6.45775974e-01 5.73541522e-01
4.83731717e-01 1.54458603e-03 5.03573954e-01 8.91344249e-01
3.40392411e-01 1.48689270e-01 -4.94166195e-01 -6.49914369e-02
1.85051441e-01 1.00639462e+00 -5.11375964e-01 -1.57343701e-01
-2.42478594e-01 3.94384950e-01 2.25087795e-02 7.07685173e-01
-7.23834872e-01 -5.68191886e-01 5.42119324e-01 -1.95198655e-01
5.11858642e-01 -1.37317210e-01 3.44286472e-01 -9.48870063e-01
2.52865944e-02 -7.11423397e-01 4.47494268e-01 2.71923523e-02
-1.37256908e+00 -8.51641782e-03 3.26851219e-01 -1.19434464e+00
-6.15409136e-01 -4.68116760e-01 -5.74941158e-01 7.47631073e-01
-1.46052659e+00 -3.21638227e-01 6.58260763e-01 7.98384726e-01
2.48899013e-01 -2.61599481e-01 2.82528996e-01 -2.48251073e-02
-5.18188238e-01 2.40597844e-01 9.12838101e-01 -3.65076691e-01
1.36431351e-01 -1.70075631e+00 -4.07883644e-01 3.83787334e-01
-4.97053027e-01 3.19447517e-02 9.22058642e-01 -6.62369430e-01
-1.64484072e+00 -8.16447258e-01 1.54794052e-01 -3.23752500e-02
1.00792563e+00 1.07307643e-01 -6.31354153e-01 3.84709477e-01
4.03371811e-01 -2.81568873e-03 -5.89524843e-02 -4.72115785e-01
5.84229410e-01 -3.81269693e-01 -1.36095965e+00 3.12679529e-01
6.86818600e-01 -1.18921056e-01 -4.27689463e-01 3.28511208e-01
8.37559521e-01 1.45526782e-01 -9.39274490e-01 2.05146030e-01
3.43627036e-01 -7.76812434e-01 6.03962183e-01 -9.30845499e-01
9.58353840e-03 -4.19466197e-02 1.21131577e-02 -1.65499616e+00
-1.18920736e-01 -1.32898068e+00 -1.16691463e-01 8.14463556e-01
-1.31186947e-01 -1.02148092e+00 5.06827176e-01 1.40000075e-01
8.96360949e-02 -1.11555731e+00 -1.34949625e+00 -1.34711254e+00
7.31762826e-01 -5.69908582e-02 3.57224107e-01 6.01704001e-01
1.85730740e-01 -1.05150178e-01 -3.24261069e-01 -2.90269017e-01
1.18525970e+00 -8.74094889e-02 1.93644226e-01 -7.34167635e-01
-6.28229678e-01 -5.95016122e-01 7.46357441e-02 -9.01117384e-01
2.40473807e-01 -4.25102949e-01 2.50934839e-01 -1.37410152e+00
4.90667187e-02 -3.18714559e-01 -6.02714837e-01 -1.89869538e-01
-1.27815813e-01 -6.10439420e-01 5.34825265e-01 -2.16682017e-01
-7.25211442e-01 9.83281553e-01 1.79193914e+00 3.36387791e-02
-4.37219411e-01 4.18371201e-01 -4.73808914e-01 4.49456930e-01
7.12302268e-01 -3.57841313e-01 -7.18286991e-01 3.13753843e-01
2.53131568e-01 9.59976673e-01 2.48152539e-01 -6.12023652e-01
4.89124507e-02 -8.25367510e-01 -1.25867978e-01 -3.95057797e-01
-8.88587832e-02 -5.78898728e-01 -1.80540055e-01 1.10180283e+00
-4.90694284e-01 2.15869308e-01 -1.89348161e-01 9.13405061e-01
-5.65505810e-02 -7.16918707e-01 8.69571269e-01 -3.82761896e-01
-2.38316372e-01 5.10587454e-01 -5.65954030e-01 6.31707370e-01
1.19405770e+00 3.23080011e-02 -1.07259974e-01 -7.69279599e-01
-7.51278758e-01 6.23548567e-01 -5.62040098e-02 -1.79786518e-01
4.51068401e-01 -1.47869575e+00 -5.75372934e-01 4.76689115e-02
-4.92613524e-01 -3.25709909e-01 7.96785951e-02 9.58756268e-01
7.56092817e-02 3.89232457e-01 -2.30882213e-01 -1.79462552e-01
-2.86718488e-01 7.24312186e-01 7.40157187e-01 -6.27455354e-01
-2.71062165e-01 2.73950636e-01 4.67649214e-02 -1.39795259e-01
2.63980895e-01 -2.47035667e-01 -1.32338807e-01 1.31705388e-01
4.35950279e-01 6.02329433e-01 -5.36635339e-01 -7.77135864e-02
-3.67179862e-03 7.23274469e-01 3.91837329e-01 -6.49497628e-01
1.36986303e+00 -4.52216148e-01 1.77016467e-01 5.64454079e-01
1.04634571e+00 -4.52349275e-01 -1.94977593e+00 -1.49908900e-01
2.61685997e-01 -2.06328735e-01 -1.77093714e-01 -5.63378572e-01
-8.30290020e-01 6.50306582e-01 8.39915454e-01 6.72429621e-01
9.65120196e-01 -3.71848464e-01 4.40758258e-01 3.58079016e-01
7.10543036e-01 -1.63008809e+00 2.23899111e-01 5.85201025e-01
7.80487537e-01 -1.04319990e+00 -2.34159186e-01 4.80949402e-01
-6.75899029e-01 1.02868032e+00 4.73577440e-01 -5.54623902e-01
7.72043228e-01 1.19325779e-01 -4.64517146e-01 3.48526359e-01
-9.34238553e-01 -4.53648597e-01 -1.31053746e-01 7.07507610e-01
-2.08511669e-02 1.40203357e-01 -9.08468425e-01 3.91518652e-01
3.06270659e-01 2.05668285e-01 5.41750729e-01 1.02228677e+00
-6.77711546e-01 -9.20848072e-01 -4.15426582e-01 3.31455767e-01
-4.52872008e-01 3.20096493e-01 4.52170193e-01 1.04840255e+00
1.61149781e-02 8.61921787e-01 -9.32453796e-02 2.69260287e-01
3.30242783e-01 1.21867180e-01 6.80878699e-01 -2.63690770e-01
-4.08428311e-01 3.31737965e-01 -3.10041159e-01 -5.70687711e-01
-5.90714335e-01 -5.51517963e-01 -1.03399587e+00 -6.14816844e-02
-2.39133969e-01 7.83964276e-01 3.37963074e-01 1.25410569e+00
-1.26290485e-01 2.30892524e-01 1.08229387e+00 -4.52393025e-01
-1.67602479e+00 -8.74221683e-01 -1.13812387e+00 1.19777620e-01
6.40989661e-01 -7.91407824e-01 -5.52520156e-01 -4.81849223e-01] | [4.195976734161377, 2.5038814544677734] |
eaf6d1b3-4a89-4fae-a7f2-6322ca35bb46 | using-gaussian-processes-for-rumour-stance | 1609.01962 | null | http://arxiv.org/abs/1609.01962v1 | http://arxiv.org/pdf/1609.01962v1.pdf | Using Gaussian Processes for Rumour Stance Classification in Social Media | Social media tend to be rife with rumours while new reports are released
piecemeal during breaking news. Interestingly, one can mine multiple reactions
expressed by social media users in those situations, exploring their stance
towards rumours, ultimately enabling the flagging of highly disputed rumours as
being potentially false. In this work, we set out to develop an automated,
supervised classifier that uses multi-task learning to classify the stance
expressed in each individual tweet in a rumourous conversation as either
supporting, denying or questioning the rumour. Using a classifier based on
Gaussian Processes, and exploring its effectiveness on two datasets with very
different characteristics and varying distributions of stances, we show that
our approach consistently outperforms competitive baseline classifiers. Our
classifier is especially effective in estimating the distribution of different
types of stance associated with a given rumour, which we set forth as a desired
characteristic for a rumour-tracking system that will warn both ordinary users
of Twitter and professional news practitioners when a rumour is being rebutted. | ['Michal Lukasik', 'Kalina Bontcheva', 'Trevor Cohn', 'Rob Procter', 'Maria Liakata', 'Arkaitz Zubiaga'] | 2016-09-07 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-9.45324749e-02 3.75568718e-01 -3.40638012e-01 -3.00316662e-01
-7.68000364e-01 -5.96707106e-01 1.30299556e+00 6.63882792e-01
-1.43401980e-01 7.42102265e-01 5.67302942e-01 -4.44271624e-01
2.49721631e-01 -6.95137262e-01 -3.50047261e-01 -5.77246726e-01
9.32828337e-02 7.70305932e-01 3.52249593e-01 -5.79416990e-01
8.46631885e-01 2.31913298e-01 -1.18409336e+00 7.93410957e-01
3.73443305e-01 8.23992491e-01 -2.57274717e-01 6.20100558e-01
-2.54334390e-01 1.71781266e+00 -1.13728213e+00 -4.02957022e-01
-4.56196040e-01 -3.78878862e-01 -1.21459758e+00 2.52721012e-01
2.32172087e-01 8.64325315e-02 2.14522466e-01 1.00629592e+00
1.36605650e-01 -1.12645425e-01 7.17818499e-01 -1.28042245e+00
-3.26504409e-01 8.84315610e-01 -7.92053223e-01 7.59858191e-01
8.26552451e-01 -4.02431190e-01 9.29374039e-01 -7.33345032e-01
6.85695112e-01 1.32944548e+00 9.50361013e-01 1.79332674e-01
-1.21982121e+00 -6.86625719e-01 -1.44722238e-01 7.38533959e-02
-8.54084730e-01 -5.74860811e-01 8.58845949e-01 -8.73830378e-01
2.52046317e-01 4.40563798e-01 2.36343458e-01 1.68025529e+00
5.70849299e-01 6.76178038e-01 1.69809484e+00 6.26682267e-02
1.88903019e-01 7.74218023e-01 4.31011051e-01 2.83253014e-01
-1.66617900e-01 -4.78572845e-01 -7.95832694e-01 -1.05437255e+00
1.38211912e-02 -1.53900728e-01 -3.67045999e-01 2.90369600e-01
-1.07642627e+00 1.28849030e+00 2.25731209e-01 3.34593713e-01
-8.00443232e-01 -5.69043219e-01 7.78580725e-01 8.33827972e-01
1.37366629e+00 5.33387661e-01 -4.26903367e-02 -2.17235640e-01
-1.23190069e+00 6.20809674e-01 1.53868163e+00 2.61517733e-01
5.47561288e-01 -5.46093225e-01 -1.46851689e-02 1.08325255e+00
-9.04888473e-03 1.32718772e-01 5.81206203e-01 -5.26772618e-01
1.31811246e-01 3.34663063e-01 5.85786581e-01 -1.47774005e+00
-6.39323235e-01 -3.79161239e-01 -4.82578099e-01 8.54120627e-02
2.93594837e-01 -5.65251589e-01 1.50761958e-02 1.19568849e+00
4.10766810e-01 -1.36134192e-01 -4.76122834e-02 7.27356315e-01
6.37833178e-01 6.57939374e-01 -3.47027928e-01 -4.68493134e-01
1.56121469e+00 -6.62083745e-01 -8.92589688e-01 -1.67516455e-01
2.93823510e-01 -1.10035539e+00 6.50809348e-01 6.29193008e-01
-9.94692564e-01 -6.75657690e-02 -1.02208436e+00 3.07985753e-01
-1.48338735e-01 -6.52669251e-01 9.66233313e-02 4.47242051e-01
-5.70531845e-01 7.92465508e-01 -2.85167247e-01 -2.40451202e-01
2.84860969e-01 -3.93088847e-01 -9.57412720e-02 2.63518631e-01
-1.37432361e+00 1.17695773e+00 1.23926744e-01 -2.74910122e-01
-7.74789810e-01 -3.66711497e-01 -3.68209988e-01 -4.70177859e-01
5.24834394e-01 -2.23584205e-01 1.52528977e+00 -9.23476458e-01
-1.29626036e+00 1.15251231e+00 -3.84391882e-02 -6.34979486e-01
1.04068506e+00 -1.29776493e-01 -7.13853121e-01 -5.62106632e-02
4.92283702e-01 -2.57311344e-01 1.37355280e+00 -1.09522772e+00
-6.28069460e-01 -3.82871807e-01 -8.88838917e-02 -5.91968857e-02
6.87982887e-02 7.69879580e-01 7.66452789e-01 -3.86047482e-01
1.62882745e-01 -8.66391242e-01 2.17206880e-01 -7.28736520e-01
-7.60944486e-01 -4.59517628e-01 1.15943980e+00 -4.79737312e-01
8.18727195e-01 -1.80821395e+00 -1.98903888e-01 -5.22921719e-02
5.01895368e-01 -1.18030263e-02 6.65068984e-01 7.56366730e-01
2.21067429e-01 3.21624577e-01 2.23821700e-01 -2.70027131e-01
-2.95584619e-01 7.05289021e-02 -7.39118695e-01 9.53748822e-01
6.36744052e-02 3.75651777e-01 -1.21390271e+00 -2.76547223e-01
-5.55985749e-01 1.43606991e-01 4.70039528e-03 2.43551403e-01
-1.15916252e-01 4.12689209e-01 -5.21917284e-01 3.25753659e-01
5.59019148e-01 -5.08370280e-01 9.12805125e-02 1.99843511e-01
-3.40707421e-01 9.09675181e-01 -6.88195288e-01 4.23769802e-01
-3.95574152e-01 9.84297216e-01 4.14915085e-01 -7.03008235e-01
1.31144023e+00 4.61622447e-01 -9.81540512e-03 3.64608914e-02
4.36610252e-01 2.73109645e-01 -3.41172874e-01 -4.10078198e-01
7.24256098e-01 -7.36645639e-01 -3.04215491e-01 1.20534468e+00
-2.83865899e-01 -2.08132580e-01 -2.80245125e-01 1.47945747e-01
1.01953816e+00 -5.49443245e-01 5.23770571e-01 -5.10672569e-01
4.21038926e-01 5.96155487e-02 8.94992277e-02 9.46263671e-01
-3.89245868e-01 3.66850853e-01 1.07739663e+00 -5.32676280e-01
-8.46678972e-01 -5.29854834e-01 -3.32093090e-01 1.45266199e+00
-7.26142824e-02 2.57228296e-02 -3.47392589e-01 -6.14477873e-01
7.80657828e-02 8.90416205e-01 -8.24200034e-01 2.85677373e-01
-1.97268084e-01 -1.07015443e+00 6.95613980e-01 -1.51644900e-01
3.09325367e-01 -1.03340113e+00 -4.46460426e-01 4.68348145e-01
-6.20613933e-01 -1.01356399e+00 -6.23883829e-02 2.32679233e-01
-3.81790757e-01 -1.17092061e+00 -5.77104151e-01 -3.79880130e-01
1.39350258e-02 4.77546066e-01 9.37813878e-01 5.12877628e-02
5.90181291e-01 -1.59565911e-01 -4.48566139e-01 -6.75920844e-01
-1.27270567e+00 -1.68525949e-02 1.72714710e-01 3.50268632e-01
1.76043510e-01 -6.51037335e-01 -1.79729834e-02 5.68573713e-01
-7.40337491e-01 -1.50724962e-01 -2.73112934e-02 6.67375445e-01
-3.14860016e-01 -5.27233109e-02 8.51199448e-01 -1.50040555e+00
1.45709062e+00 -1.42790580e+00 1.29563034e-01 -4.06250149e-01
-5.34186661e-01 -3.18194240e-01 3.17844421e-01 -5.99536538e-01
-9.10084784e-01 -1.04667115e+00 -5.29757366e-02 1.09128237e-01
-2.82879055e-01 6.59434378e-01 7.08649099e-01 7.77784660e-02
1.25360513e+00 1.09873697e-01 3.94373000e-01 -5.34848690e-01
9.75782797e-02 1.24611223e+00 2.97659189e-01 -2.90572971e-01
4.96539623e-01 8.13251197e-01 -5.63075423e-01 -9.84830916e-01
-1.63117909e+00 -1.04940486e+00 -2.25707859e-01 -2.90897995e-01
3.37253124e-01 -8.29409838e-01 -6.73456252e-01 6.34467661e-01
-1.38602662e+00 1.44668713e-01 4.47255254e-01 -3.96956578e-02
-5.65390050e-01 3.01358521e-01 -1.16816831e+00 -9.54943419e-01
-3.70889872e-01 -6.18974090e-01 4.12901133e-01 -2.48409107e-01
-1.02212620e+00 -1.08818305e+00 5.04671454e-01 7.22122252e-01
6.32334292e-01 3.97639215e-01 4.75925267e-01 -1.55641305e+00
3.86133403e-01 -4.81527627e-01 -1.77745342e-01 1.27326280e-01
2.65122235e-01 -1.21137604e-01 -1.09395576e+00 -1.71338692e-01
7.97187924e-01 -7.35280633e-01 5.90212405e-01 -1.51518568e-01
4.25912738e-01 -9.24330771e-01 -2.09216684e-01 -2.56834865e-01
7.08468854e-01 -4.44135815e-01 2.50968993e-01 8.29931557e-01
4.83841076e-02 9.32282507e-01 5.26042402e-01 6.25635564e-01
2.19950601e-01 4.49073195e-01 3.18181336e-01 2.90063918e-01
4.58971530e-01 -8.88501853e-02 6.80334687e-01 5.49582422e-01
1.19594559e-01 -1.81821674e-01 -9.59381998e-01 5.50348580e-01
-1.75851059e+00 -1.30394006e+00 -5.49314499e-01 1.71226525e+00
1.13495421e+00 7.00010240e-01 5.71081042e-01 1.78526506e-01
1.02154469e+00 7.18111634e-01 -4.15975088e-03 -8.25142324e-01
-8.62081572e-02 -5.05033195e-01 2.69814044e-01 6.26734495e-01
-1.13802314e+00 3.97552490e-01 6.28225613e+00 2.70438433e-01
-1.29564202e+00 5.50865114e-01 6.01784825e-01 -3.31608690e-02
-3.16940136e-02 -2.58065760e-01 -7.57411242e-01 5.47298312e-01
1.25321138e+00 -3.44767064e-01 -5.64998798e-02 8.92111719e-01
5.19330561e-01 -2.61895865e-01 -7.46969521e-01 3.54292572e-01
5.02761185e-01 -1.49410844e+00 -2.99241960e-01 4.85447282e-03
7.69146264e-01 4.93539870e-01 -1.18637949e-01 2.08162412e-01
3.90662044e-01 -6.82079196e-01 1.00296772e+00 2.42270619e-01
-6.18410781e-02 -5.68691254e-01 1.07727718e+00 1.03113687e+00
8.10500309e-02 5.81291951e-02 -3.93984988e-02 -4.89290088e-01
4.80420053e-01 9.31700110e-01 -1.60333490e+00 -5.09801842e-02
3.42219561e-01 6.22498810e-01 -2.00326845e-01 6.46957994e-01
-4.34245378e-01 8.48295510e-01 -4.93798628e-02 -2.86193550e-01
3.80066276e-01 2.75989801e-01 1.01647210e+00 1.53197122e+00
-1.86217710e-01 -8.57059583e-02 3.14584762e-01 8.37190628e-01
-1.30820736e-01 8.73829797e-02 -5.08661985e-01 -5.55653349e-02
2.98005402e-01 1.29350603e+00 -6.80288434e-01 -4.04767841e-01
-2.81161759e-02 9.13983285e-01 3.40606242e-01 -4.77120914e-02
-6.56113088e-01 1.82636365e-01 5.13071477e-01 6.75011337e-01
-6.14939667e-02 1.15353264e-01 -1.39120072e-01 -1.00597870e+00
-4.03949082e-01 -9.39109087e-01 1.46488115e-01 -6.52825296e-01
-1.71482909e+00 7.32078254e-01 -2.16764867e-01 -7.80008554e-01
-3.52125287e-01 -1.67274147e-01 -1.01319492e+00 8.52851152e-01
-1.52047491e+00 -8.60668480e-01 -1.35512561e-01 1.98618397e-01
5.19770026e-01 9.05169547e-02 4.82904553e-01 -1.48355752e-01
-1.80191219e-01 -3.09153907e-02 -2.10543692e-01 1.86512306e-01
1.00168908e+00 -1.15152752e+00 3.79272163e-01 1.19903341e-01
-7.06031024e-02 5.42415798e-01 1.64271724e+00 -6.55450523e-01
-6.88028991e-01 -8.14149261e-01 1.14424598e+00 -7.28071570e-01
1.66387677e+00 -1.61261484e-01 -1.44057584e+00 7.30026841e-01
1.67653114e-01 -4.47702736e-01 9.25060272e-01 6.78044260e-01
-6.00618899e-01 6.94565415e-01 -1.16379213e+00 1.58217043e-01
1.88227072e-01 -8.76156867e-01 -1.33224988e+00 8.95047843e-01
2.13541031e-01 -2.59640336e-01 -6.43654168e-01 -1.99557140e-01
3.73032898e-01 -1.18488622e+00 3.64849836e-01 -9.90721941e-01
7.87905335e-01 1.10826999e-01 1.47553295e-01 -1.55506837e+00
-2.95960724e-01 -8.60923707e-01 -5.66499978e-02 1.06226242e+00
4.47920769e-01 -7.56603956e-01 5.33716142e-01 -6.30967617e-02
2.21189901e-01 -4.54048008e-01 -6.55772626e-01 -4.97813642e-01
1.38252392e-01 -3.77403259e-01 9.10998285e-02 1.55233884e+00
6.47729218e-01 6.73917890e-01 -7.61454344e-01 1.43745631e-01
4.95513886e-01 2.51608133e-01 8.38282704e-01 -1.51573014e+00
4.51574195e-03 -5.78681350e-01 -2.41627470e-01 -8.01601410e-01
1.24608353e-01 -6.16796196e-01 1.92210972e-01 -8.25215280e-01
3.54121357e-01 -2.79671222e-01 -2.02878546e-02 8.99912193e-02
-6.00803457e-02 1.64744854e-01 -9.71002653e-02 8.78010392e-01
-4.88010228e-01 1.04860082e-01 8.30704749e-01 -3.02009489e-02
-1.43954858e-01 6.53782606e-01 -9.02026415e-01 1.16832089e+00
7.02873349e-01 -7.70691812e-01 1.95592552e-01 5.79753101e-01
7.01518953e-01 3.20713848e-01 5.16948938e-01 -5.41671097e-01
3.02777857e-01 -1.77244917e-01 -3.64142776e-01 -5.86758018e-01
2.17973098e-01 -3.95819470e-02 -1.71799585e-01 2.27171808e-01
-6.31191552e-01 -1.71155706e-01 -8.03746283e-02 8.21618974e-01
-3.17402482e-01 -3.20317566e-01 1.02665555e+00 -2.39374056e-01
-2.38004103e-01 -2.53282517e-01 -9.32317495e-01 2.04891860e-01
8.27049136e-01 1.60149828e-01 -5.99960029e-01 -9.09329474e-01
-8.76524568e-01 5.26508167e-02 4.41134959e-01 5.39702177e-01
3.27370167e-01 -6.40907288e-01 -1.47650695e+00 -1.62329227e-01
1.09920233e-01 -6.54034972e-01 -7.99306110e-02 1.23655570e+00
-1.81545228e-01 7.22193792e-02 -1.28528833e-01 -4.22383815e-01
-1.35211301e+00 4.59352046e-01 2.05530137e-01 -1.43418834e-01
-6.16810679e-01 7.94216275e-01 -4.16867137e-01 -2.69985259e-01
-2.33495057e-01 9.81234983e-02 -4.81062740e-01 9.44478095e-01
1.13360143e+00 4.99847144e-01 2.54436404e-01 -8.42705965e-01
-6.22626096e-02 -5.57807684e-01 -5.83586693e-01 -1.41491115e-01
1.11734402e+00 -3.69493663e-01 -4.60268646e-01 1.28631914e+00
1.09878564e+00 4.38871711e-01 -6.98577464e-01 -4.37160969e-01
4.26952153e-01 -3.36310685e-01 2.06746683e-01 -6.51079774e-01
-1.11437887e-01 3.70734543e-01 -5.31031072e-01 1.19549918e+00
1.06216028e-01 4.54765767e-01 8.43898773e-01 1.34524196e-01
4.41002190e-01 -7.42778778e-01 1.48342401e-01 6.66081965e-01
1.20819962e+00 -1.42934883e+00 2.21788511e-01 -3.31963390e-01
-8.85033607e-01 1.14930296e+00 -2.01409191e-01 -2.87032723e-01
6.83042645e-01 9.03558284e-02 2.54815221e-01 -7.74009407e-01
-9.28457797e-01 5.23181975e-01 1.99489780e-02 7.75723085e-02
4.82293725e-01 2.81198293e-01 -4.27610904e-01 3.35613728e-01
-4.66983736e-01 -1.56094894e-01 1.26365221e+00 7.98912883e-01
-9.78335142e-01 -3.92601460e-01 -7.87782669e-01 4.91222084e-01
-1.06878841e+00 1.36690617e-01 -8.22890639e-01 5.29522955e-01
-2.62534231e-01 1.21596956e+00 -6.24066591e-02 -3.22443217e-01
-1.37796700e-01 2.83164561e-01 -1.62620515e-01 -9.44020808e-01
-8.47948790e-01 -2.18464285e-01 8.94619703e-01 -1.46975175e-01
-4.51969504e-01 -1.02695835e+00 -6.49409056e-01 -7.07277596e-01
-5.29702485e-01 5.96971512e-01 6.57091498e-01 1.40400875e+00
-1.80006504e-01 2.12125145e-02 8.87263656e-01 -6.50041282e-01
-1.02759874e+00 -1.37511659e+00 -7.02293813e-01 5.76426268e-01
9.12012041e-01 -7.57267892e-01 -7.30557084e-01 -2.64898032e-01] | [8.231082916259766, 10.106389045715332] |
a1b159d8-1e3a-4838-af11-a3701e50b328 | boosting-video-text-retrieval-with-explicit | 2208.04215 | null | https://arxiv.org/abs/2208.04215v2 | https://arxiv.org/pdf/2208.04215v2.pdf | Boosting Video-Text Retrieval with Explicit High-Level Semantics | Video-text retrieval (VTR) is an attractive yet challenging task for multi-modal understanding, which aims to search for relevant video (text) given a query (video). Existing methods typically employ completely heterogeneous visual-textual information to align video and text, whilst lacking the awareness of homogeneous high-level semantic information residing in both modalities. To fill this gap, in this work, we propose a novel visual-linguistic aligning model named HiSE for VTR, which improves the cross-modal representation by incorporating explicit high-level semantics. First, we explore the hierarchical property of explicit high-level semantics, and further decompose it into two levels, i.e. discrete semantics and holistic semantics. Specifically, for visual branch, we exploit an off-the-shelf semantic entity predictor to generate discrete high-level semantics. In parallel, a trained video captioning model is employed to output holistic high-level semantics. As for the textual modality, we parse the text into three parts including occurrence, action and entity. In particular, the occurrence corresponds to the holistic high-level semantics, meanwhile both action and entity represent the discrete ones. Then, different graph reasoning techniques are utilized to promote the interaction between holistic and discrete high-level semantics. Extensive experiments demonstrate that, with the aid of explicit high-level semantics, our method achieves the superior performance over state-of-the-art methods on three benchmark datasets, including MSR-VTT, MSVD and DiDeMo. | ['Errui Ding', 'Jungong Han', 'Zhong Ji', 'Fu Li', 'Dongliang He', 'Di Xu', 'Haoran Wang'] | 2022-08-08 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 1.70055926e-01 -2.19361767e-01 -5.23426890e-01 -1.37018457e-01
-7.63024986e-01 -4.40524578e-01 7.90287077e-01 3.04870158e-01
-1.91247895e-01 9.98021215e-02 6.22021377e-01 -3.85028832e-02
6.15580231e-02 -6.05006516e-01 -6.03008151e-01 -3.72726053e-01
2.77458489e-01 2.24031001e-01 4.40015286e-01 -1.72164172e-01
9.86785144e-02 -1.31362230e-01 -1.68688965e+00 6.60760403e-01
6.52746320e-01 1.07943559e+00 3.17824543e-01 1.85966536e-01
-4.39113349e-01 1.06001532e+00 -1.79995850e-01 -3.63626868e-01
-1.32980198e-01 -5.20534456e-01 -7.67604172e-01 4.01798189e-01
3.79771262e-01 -2.60152221e-01 -6.74710989e-01 1.15361214e+00
1.40550196e-01 3.72510433e-01 6.09985828e-01 -1.45184422e+00
-7.58938313e-01 4.69595283e-01 -8.52534890e-01 -1.08011574e-01
8.45493555e-01 1.96001381e-01 1.25952721e+00 -1.01712668e+00
9.25126135e-01 1.52653074e+00 1.04148969e-01 3.81807894e-01
-8.69577050e-01 -2.78734207e-01 5.94533026e-01 4.02516663e-01
-1.53169298e+00 -2.17557922e-01 1.01090229e+00 -5.03102362e-01
6.31227911e-01 1.54018283e-01 7.32251287e-01 1.16038609e+00
-2.81883806e-01 1.22241127e+00 8.80562782e-01 -2.60438561e-01
8.34445655e-02 2.71103960e-02 1.87941417e-01 7.22882926e-01
-1.32232169e-02 -3.77041101e-01 -7.32920170e-01 -8.50796849e-02
5.95743537e-01 2.76059031e-01 -4.89861488e-01 -6.32643878e-01
-1.44716775e+00 6.63189948e-01 3.26296687e-01 3.93329322e-01
-4.45798099e-01 1.16010107e-01 7.04121351e-01 -7.89065808e-02
1.91915348e-01 -1.91764653e-01 -3.98630388e-02 9.64249074e-02
-9.43672419e-01 1.83067620e-02 4.73130226e-01 1.26805067e+00
8.54182541e-01 -1.63842559e-01 -5.85653186e-01 8.30054998e-01
7.09053218e-01 4.48239118e-01 5.60353994e-01 -6.55830085e-01
8.40103507e-01 8.74058723e-01 -9.68702883e-02 -1.33261204e+00
5.43901511e-02 -7.45809451e-02 -8.22055161e-01 -4.82393533e-01
-1.57145157e-01 4.13344413e-01 -1.15892160e+00 1.62420547e+00
4.48653013e-01 2.98700243e-01 3.08720052e-01 1.19270492e+00
1.31264031e+00 8.86961579e-01 3.83381218e-01 -3.30433428e-01
1.77939141e+00 -1.24611378e+00 -8.73861372e-01 -3.27030092e-01
5.27371347e-01 -5.73501885e-01 1.34978831e+00 -1.60275564e-01
-9.58897173e-01 -3.66909772e-01 -9.52973425e-01 -3.99745017e-01
-4.98195738e-01 1.74400911e-01 4.71343324e-02 -4.61097574e-03
-8.78500402e-01 6.46334365e-02 -4.37246978e-01 -5.14660180e-01
1.25769228e-01 -2.07529798e-01 -4.61171329e-01 -4.38723952e-01
-1.45878720e+00 4.37687904e-01 7.57021666e-01 -8.36368427e-02
-9.75623548e-01 -4.70200002e-01 -1.12344074e+00 7.76697397e-02
1.04245961e+00 -9.26689684e-01 8.79529536e-01 -1.00791824e+00
-1.09615827e+00 1.00311160e+00 -3.89792174e-01 -6.76572323e-02
3.98491621e-01 -1.22890331e-01 -5.03953993e-01 7.85118043e-01
2.57517755e-01 6.48270488e-01 9.23682034e-01 -1.50463164e+00
-6.42490506e-01 -5.11459053e-01 3.27483743e-01 6.36481464e-01
-6.42017901e-01 1.84140831e-01 -1.31013489e+00 -8.83550882e-01
9.74129289e-02 -5.28516769e-01 9.39839631e-02 -8.34775716e-02
-3.91152740e-01 -3.38104188e-01 9.49538469e-01 -8.62035573e-01
1.53189456e+00 -2.29337382e+00 5.29782355e-01 2.51920342e-01
5.23180604e-01 -5.66364899e-02 -1.05139926e-01 6.06232524e-01
1.85988680e-01 1.25537217e-01 -5.68050481e-02 -3.63984168e-01
1.71995535e-01 7.13156983e-02 -3.46677005e-01 1.60046250e-01
-1.73844974e-02 1.01831031e+00 -1.13207173e+00 -1.02994943e+00
2.19689682e-01 5.56376100e-01 -3.38566154e-01 1.81553930e-01
-5.21424770e-01 8.29799101e-02 -8.99628520e-01 8.87289703e-01
3.01428139e-01 -6.55948699e-01 3.81708294e-01 -6.84683800e-01
5.70542738e-02 -1.50535867e-01 -1.03943133e+00 1.92747605e+00
-2.00838640e-01 3.49449068e-01 -1.54750273e-01 -8.78382206e-01
6.99132025e-01 3.48974228e-01 6.14112675e-01 -7.27002323e-01
8.34447443e-02 6.33040518e-02 -6.39139354e-01 -6.30165219e-01
7.70460188e-01 -9.77299828e-03 -3.20081949e-01 2.12111160e-01
4.63754237e-02 3.60641181e-01 1.98101625e-01 7.31580138e-01
8.26729357e-01 3.83003026e-01 3.39910686e-01 8.81344005e-02
8.06898296e-01 2.31292248e-01 3.52430075e-01 4.19837147e-01
-2.05744326e-01 6.61506236e-01 5.48875213e-01 -7.77183399e-02
-7.36570239e-01 -8.72871041e-01 4.62925494e-01 1.18524742e+00
1.03154814e+00 -9.40506995e-01 -6.87932670e-01 -5.86466312e-01
-1.85880944e-01 5.96011758e-01 -6.07690811e-01 -1.61811069e-01
-4.36577201e-01 -3.05390328e-01 3.30076903e-01 4.79314536e-01
5.88545561e-01 -9.36736226e-01 -3.92011464e-01 -1.11968376e-01
-7.87405133e-01 -1.63811731e+00 -8.14390779e-01 -5.50051987e-01
-5.59741199e-01 -9.48201537e-01 -8.10746968e-01 -8.00107479e-01
5.27813375e-01 6.94088757e-01 1.09174478e+00 2.51126796e-01
-8.13897401e-02 8.62857163e-01 -7.86254287e-01 3.19594681e-01
-3.42738293e-02 -2.41619840e-01 -3.32330614e-01 4.20979559e-01
4.61424828e-01 -1.94678202e-01 -7.66864955e-01 2.30837584e-01
-1.23580408e+00 4.97758567e-01 6.00070536e-01 6.20683968e-01
1.01886058e+00 -1.32085592e-01 2.10818321e-01 -5.29429734e-01
3.77382457e-01 -6.95863128e-01 -1.24608919e-01 7.73881555e-01
-2.26368129e-01 -3.82563435e-02 6.89360499e-01 -3.91603142e-01
-1.01918566e+00 -3.42435464e-02 3.63383681e-01 -1.08697474e+00
-1.21317327e-01 8.43385279e-01 -4.63404417e-01 3.62552851e-01
-1.75001135e-03 6.78868890e-01 -3.60025704e-01 -3.65007102e-01
6.86536193e-01 6.35124803e-01 5.79262853e-01 -7.73727775e-01
7.42948711e-01 6.15180254e-01 -4.47563492e-02 -7.93079317e-01
-8.94697189e-01 -8.54866803e-01 -4.10592794e-01 -5.49001217e-01
1.26924026e+00 -1.18637538e+00 -4.35387075e-01 2.19822466e-01
-9.37748373e-01 3.27152088e-02 7.74352765e-03 3.48667711e-01
-6.60342395e-01 7.21106231e-01 -3.98112327e-01 -4.85637575e-01
-3.08218509e-01 -1.15006018e+00 1.48175108e+00 1.50203511e-01
2.00940192e-01 -1.04499829e+00 -3.03949028e-01 6.72492564e-01
-1.66389972e-01 2.83306271e-01 9.25401509e-01 -6.78724349e-01
-7.50115454e-01 7.92395230e-03 -5.63072622e-01 -1.73750892e-01
-5.43109365e-02 3.05858329e-02 -5.29205143e-01 -2.93437570e-01
-2.21528411e-01 -4.90656555e-01 8.56331825e-01 -1.45900790e-02
1.19797599e+00 -3.07553709e-01 -5.19734025e-01 3.52724344e-01
1.50356567e+00 -5.13326423e-03 5.93790233e-01 3.84375662e-01
1.15251255e+00 7.40494609e-01 9.80841339e-01 4.04609174e-01
9.68553483e-01 8.16611707e-01 3.28232288e-01 8.07218328e-02
-2.53640801e-01 -7.58487403e-01 4.54023540e-01 9.43834305e-01
2.76485402e-02 -3.84024918e-01 -9.37822580e-01 5.86969316e-01
-2.34719729e+00 -1.08022773e+00 -1.07164398e-01 1.99356747e+00
7.07182229e-01 -1.89250827e-01 1.46079555e-01 -2.56549805e-01
1.00534201e+00 5.99703848e-01 -4.36187238e-01 2.98859715e-01
-1.72374353e-01 -5.55796862e-01 1.16161741e-01 1.81182489e-01
-9.67655420e-01 1.06579709e+00 4.45473671e+00 1.23795915e+00
-9.19794858e-01 6.63980916e-02 3.22903097e-01 3.84587310e-02
-5.91136277e-01 1.54711053e-01 -6.19882524e-01 4.78176445e-01
4.13543582e-01 -3.20579022e-01 3.22888911e-01 6.20811522e-01
1.24907762e-01 7.74989128e-02 -1.07492125e+00 1.18967772e+00
4.95873421e-01 -1.30497551e+00 6.61578536e-01 -1.85862735e-01
5.05057156e-01 -3.31730872e-01 -1.03774838e-01 2.77337670e-01
-2.33247932e-02 -6.73314095e-01 1.00959945e+00 5.49518824e-01
7.97975421e-01 -5.70547640e-01 3.75375569e-01 2.27121815e-01
-1.98365390e+00 3.86615954e-02 -5.65139391e-02 5.88964045e-01
3.36860090e-01 1.01815559e-01 -1.18023813e-01 1.21534979e+00
7.23018110e-01 1.18489492e+00 -6.24806702e-01 5.70571721e-01
-6.86527044e-02 1.48546532e-01 -2.68004499e-02 1.25740662e-01
4.33595866e-01 -2.65062600e-01 6.13797903e-01 1.20323157e+00
1.93227887e-01 2.96000481e-01 5.58324158e-01 6.50168240e-01
-2.27673411e-01 3.81497830e-01 -4.70890373e-01 -2.60169446e-01
5.64807296e-01 1.23224711e+00 -7.51705229e-01 -6.43273175e-01
-7.47500181e-01 1.07966363e+00 2.22856030e-01 6.46414638e-01
-1.07112134e+00 -2.25619003e-01 3.04876328e-01 2.21259966e-02
3.31915736e-01 -1.95492730e-02 9.54547748e-02 -1.69785011e+00
3.71046573e-01 -8.89226139e-01 8.38762939e-01 -1.24010742e+00
-1.15183198e+00 4.88046765e-01 3.16947132e-01 -1.44483101e+00
-1.25448152e-01 -3.73408198e-01 -2.55572200e-01 5.51635265e-01
-1.61501980e+00 -1.62196612e+00 -5.71196079e-01 1.01246548e+00
7.84390569e-01 6.13867827e-02 2.79110134e-01 3.31556559e-01
-5.48195064e-01 3.83940548e-01 -1.23694643e-01 2.14148819e-01
6.24760926e-01 -8.81377578e-01 -8.63787904e-02 8.56590748e-01
2.28574559e-01 5.36264002e-01 5.29977560e-01 -8.34878862e-01
-1.75654042e+00 -1.29492676e+00 7.68042922e-01 -1.17191009e-01
9.51312363e-01 -2.31292039e-01 -9.12991226e-01 7.15970337e-01
2.54034638e-01 -4.09517884e-02 5.49237967e-01 -3.27459633e-01
-6.34149134e-01 1.23041175e-01 -7.67477334e-01 9.66623485e-01
1.23088396e+00 -9.17942166e-01 -9.76179004e-01 1.83446676e-01
1.14277494e+00 -3.13730836e-01 -8.83682132e-01 4.69695628e-01
4.16593730e-01 -7.26918221e-01 1.11511457e+00 -5.11663258e-01
7.57354558e-01 -6.30540848e-01 -4.86399561e-01 -8.48913014e-01
-1.54640023e-02 -3.62438291e-01 -4.45885360e-01 1.48527241e+00
-1.59610845e-02 -1.74780250e-01 5.18260598e-01 3.55107337e-01
-1.88049749e-01 -7.24982262e-01 -5.85525036e-01 -6.94707632e-01
-3.77849162e-01 -3.36040735e-01 3.69425297e-01 1.09980273e+00
7.22077042e-02 5.19548535e-01 -4.82363760e-01 2.20008746e-01
7.55723059e-01 5.83662808e-01 5.25900960e-01 -9.01357293e-01
-7.81978816e-02 -4.35464203e-01 -4.84550029e-01 -1.29245174e+00
2.98121870e-01 -8.82951975e-01 -3.38594988e-02 -1.80049753e+00
6.13969803e-01 5.87392710e-02 -3.67303163e-01 5.41312516e-01
-3.03343266e-01 1.61591515e-01 4.75999147e-01 4.93899524e-01
-1.23505318e+00 8.62986267e-01 1.30485010e+00 -2.34880164e-01
3.30725685e-02 -7.45375335e-01 -6.15438223e-01 7.41769016e-01
3.75518233e-01 -2.26761490e-01 -6.50556922e-01 -4.71102893e-01
1.60660982e-01 3.87743831e-01 6.61161482e-01 -6.35829747e-01
4.18493271e-01 -4.77872133e-01 5.64946048e-02 -7.28267670e-01
4.31168169e-01 -9.36036885e-01 1.32765666e-01 2.79530901e-02
-4.76312339e-01 -4.88518029e-02 -9.27392021e-02 9.61931944e-01
-7.03293443e-01 -3.70543636e-02 3.79033953e-01 -9.92693007e-02
-1.28264832e+00 5.87021172e-01 -5.33910096e-02 3.44536543e-01
1.18865681e+00 -2.71468550e-01 -4.96505499e-01 -3.00144374e-01
-7.10902512e-01 7.16839731e-01 6.87537730e-01 7.79417872e-01
9.30673659e-01 -1.64943767e+00 -4.57011789e-01 -1.85394838e-01
7.56601095e-01 -1.54471263e-01 6.01103842e-01 8.26687515e-01
-1.72809422e-01 3.27769697e-01 1.13416221e-02 -7.33053803e-01
-1.33358002e+00 9.83152866e-01 -2.97093485e-02 -2.20305637e-01
-7.16502011e-01 2.47582898e-01 6.14977539e-01 4.70668226e-02
2.53440857e-01 -2.44905241e-02 -4.45422083e-01 3.24373633e-01
4.93454486e-01 1.40534535e-01 -4.38101143e-01 -1.21366251e+00
-4.20328468e-01 9.02970791e-01 1.12912461e-01 -6.74986690e-02
9.30375338e-01 -6.74046278e-01 -2.05479667e-01 5.45892239e-01
1.26963091e+00 -2.95116812e-01 -9.53161180e-01 -4.53631938e-01
5.88471256e-02 -5.79724014e-01 2.43548770e-02 -4.46984261e-01
-1.03006804e+00 9.17703569e-01 9.32131186e-02 2.04671651e-01
1.33451152e+00 3.79217565e-01 8.88864279e-01 1.56565085e-01
2.42438003e-01 -1.01055014e+00 4.75332290e-01 2.88938761e-01
7.49928117e-01 -1.10629427e+00 -5.34228459e-02 -6.62039161e-01
-9.80609834e-01 9.85174000e-01 8.07614684e-01 2.78869033e-01
3.15298676e-01 -4.57966894e-01 -1.87808737e-01 -4.58993167e-01
-8.12253356e-01 -5.15407562e-01 7.18035579e-01 3.24947506e-01
1.90520078e-01 -1.10086098e-01 -2.38724083e-01 6.21409655e-01
4.69877928e-01 9.31232050e-03 2.35514969e-01 9.43366349e-01
-3.97663683e-01 -8.32551479e-01 -1.67649910e-01 1.42692626e-01
-3.52355778e-01 -2.59030223e-01 -3.66351336e-01 7.91389167e-01
-2.21751109e-01 8.47600579e-01 -3.72896902e-02 -4.30937499e-01
2.27067664e-01 -3.21171731e-02 1.39290228e-01 -4.52757180e-01
-2.79037565e-01 3.48953396e-01 -3.17761265e-02 -7.23286450e-01
-8.57312143e-01 -4.86692309e-01 -1.46739447e+00 -1.03903800e-01
-6.24832418e-03 1.25158131e-01 2.27779850e-01 1.06289434e+00
4.02793765e-01 4.91364479e-01 5.31496942e-01 -6.06168985e-01
-2.24437565e-01 -4.80802268e-01 -4.30939883e-01 7.41708636e-01
1.78028330e-01 -6.46123111e-01 -3.38954300e-01 4.60760504e-01] | [10.304486274719238, 1.041745662689209] |
4b195700-2470-4a34-8ce5-5e7ba5e1a34c | self-supervised-learning-of-object | 2304.04325 | null | https://arxiv.org/abs/2304.04325v1 | https://arxiv.org/pdf/2304.04325v1.pdf | Self-Supervised Learning of Object Segmentation from Unlabeled RGB-D Videos | This work proposes a self-supervised learning system for segmenting rigid objects in RGB images. The proposed pipeline is trained on unlabeled RGB-D videos of static objects, which can be captured with a camera carried by a mobile robot. A key feature of the self-supervised training process is a graph-matching algorithm that operates on the over-segmentation output of the point cloud that is reconstructed from each video. The graph matching, along with point cloud registration, is able to find reoccurring object patterns across videos and combine them into 3D object pseudo labels, even under occlusions or different viewing angles. Projected 2D object masks from 3D pseudo labels are used to train a pixel-wise feature extractor through contrastive learning. During online inference, a clustering method uses the learned features to cluster foreground pixels into object segments. Experiments highlight the method's effectiveness on both real and synthetic video datasets, which include cluttered scenes of tabletop objects. The proposed method outperforms existing unsupervised methods for object segmentation by a large margin. | ['Kostas Bekris', 'Abdeslam Boularias', 'Yunfu Deng', 'Shiyang Lu'] | 2023-04-09 | null | null | null | null | ['point-cloud-registration', 'graph-matching'] | ['computer-vision', 'graphs'] | [ 7.42425501e-01 8.48974660e-02 -3.41475070e-01 -4.97895688e-01
-5.53223491e-01 -7.19741881e-01 3.93054157e-01 -1.61267594e-02
-3.08044106e-01 3.03520001e-02 -5.13217807e-01 1.38342157e-01
-3.29965819e-03 -6.58576965e-01 -1.05791879e+00 -6.87092364e-01
-7.64266402e-02 8.14003646e-01 7.53678977e-01 3.20272267e-01
4.17549103e-01 8.53395343e-01 -1.93977332e+00 2.61769325e-01
5.10931611e-01 1.09602094e+00 7.09512413e-01 6.03694916e-01
-3.78666759e-01 5.43366969e-01 -2.06298441e-01 -1.46012623e-02
7.14883089e-01 -8.14814419e-02 -8.52462530e-01 1.11705601e+00
7.08693445e-01 -2.50426948e-01 -1.52073681e-01 1.21927142e+00
-1.87048659e-01 2.50431418e-01 6.06396914e-01 -1.39637780e+00
1.80613995e-01 3.41442406e-01 -5.12999535e-01 -9.60898325e-02
6.62661016e-01 2.72083014e-01 8.37900758e-01 -8.37761760e-01
9.16412115e-01 1.16161466e+00 3.84065121e-01 3.30788463e-01
-1.11154222e+00 -3.14494967e-01 9.75056961e-02 1.88142389e-01
-1.31028831e+00 -1.52947724e-01 1.13351095e+00 -6.49573684e-01
7.07795799e-01 3.06366563e-01 9.18465734e-01 6.83928370e-01
-3.83757576e-02 9.58640039e-01 1.13151288e+00 -3.93285692e-01
5.73632717e-01 1.22722745e-01 1.24125719e-01 8.90867293e-01
8.19131210e-02 -1.88067332e-01 -4.95012879e-01 4.62015904e-02
9.88156319e-01 3.95329207e-01 -6.68015927e-02 -9.64832604e-01
-1.34330630e+00 2.04978883e-01 4.05854046e-01 1.35468841e-01
-3.80638927e-01 -3.48022506e-02 -1.46128044e-01 1.69302300e-02
2.91892558e-01 9.58477240e-03 -5.10788798e-01 1.42817557e-01
-1.19446504e+00 -3.92538458e-02 6.97167754e-01 1.36567008e+00
1.18786287e+00 -2.58333325e-01 2.48588413e-01 1.93944320e-01
6.54360831e-01 6.81963384e-01 3.10150892e-01 -9.88212883e-01
3.09627861e-01 1.25833070e+00 9.70042404e-03 -1.19763923e+00
-5.06872296e-01 6.46774918e-02 -3.19734454e-01 4.23652619e-01
4.26395863e-01 3.65594864e-01 -1.24092150e+00 8.81060839e-01
8.74128401e-01 3.85221034e-01 1.01382285e-02 1.09371054e+00
8.02425981e-01 6.14401698e-01 -2.52738953e-01 -3.69119853e-01
1.08000624e+00 -9.13566649e-01 -5.46205878e-01 -3.83643597e-01
2.61734217e-01 -7.77171671e-01 8.33987653e-01 4.65605468e-01
-1.07713783e+00 -8.34179342e-01 -7.91464686e-01 6.42099306e-02
-2.79200613e-01 2.88693104e-02 5.27154505e-01 4.74749297e-01
-6.99680448e-01 7.22340047e-01 -1.20736802e+00 -4.75213021e-01
5.33638597e-01 7.04043090e-01 -5.22976518e-01 -1.23477623e-01
-3.49292457e-01 4.46883708e-01 7.72633791e-01 1.92678615e-01
-1.04477930e+00 -2.35668913e-01 -8.59053671e-01 -2.97987014e-01
5.50581515e-01 -3.92421931e-01 7.41114378e-01 -1.46627343e+00
-1.51121759e+00 1.28081107e+00 -7.46267885e-02 -2.13619843e-01
4.10334736e-01 -2.37446010e-01 5.35421558e-02 5.72382033e-01
-6.58145249e-02 5.81270933e-01 1.23962533e+00 -1.70240581e+00
-8.72710168e-01 -7.11183488e-01 -1.87192753e-01 6.75113976e-01
1.63862929e-01 -1.55208737e-01 -1.07050467e+00 -1.17651060e-01
8.65168631e-01 -1.15261853e+00 -2.67717779e-01 -3.57666053e-02
-5.87522805e-01 -1.74751341e-01 1.31794858e+00 -3.79879743e-01
4.81581956e-01 -2.03275800e+00 4.34552968e-01 6.25281096e-01
-7.39842132e-02 -8.52831453e-02 2.67888576e-01 -5.13936207e-02
6.82230815e-02 -5.41121602e-01 -5.06577969e-01 -4.74700511e-01
-1.92794695e-01 3.93447369e-01 9.51634645e-02 8.16768765e-01
9.73841399e-02 6.98753595e-01 -8.80706728e-01 -8.97350252e-01
6.83886170e-01 7.24537671e-02 -2.92843521e-01 5.86022973e-01
-7.63331056e-01 7.31577516e-01 -4.50889736e-01 9.64727938e-01
8.27224851e-01 -2.29267463e-01 1.61785930e-01 -3.36094558e-01
-1.00624248e-01 -1.09609477e-01 -1.65726197e+00 2.29071879e+00
3.08100849e-01 3.65717441e-01 -6.49774522e-02 -1.05151618e+00
9.08921301e-01 -2.01381609e-01 8.31575274e-01 -1.86539307e-01
3.54018152e-01 1.11635081e-01 -3.82405072e-01 -8.81991625e-01
3.64730030e-01 4.58485901e-01 1.26133889e-01 4.65102524e-01
3.10862839e-01 -5.12940347e-01 1.08747445e-01 4.15644273e-02
9.09125030e-01 6.65860415e-01 -1.13991551e-01 -1.18532307e-01
5.47925234e-01 3.98952752e-01 3.90667707e-01 4.40988600e-01
3.03788204e-02 7.55488694e-01 4.59044054e-02 -4.14229274e-01
-8.09270024e-01 -1.01186299e+00 -1.37138218e-01 7.19502509e-01
8.07275653e-01 1.66742597e-02 -8.68576348e-01 -8.26716125e-01
-6.75199227e-03 3.23925674e-01 -2.49796361e-01 1.49614066e-01
-5.05026281e-01 -4.98823971e-01 -2.22267717e-01 1.95484743e-01
3.31017554e-01 -1.04382586e+00 -1.08934760e+00 -1.10068627e-01
2.66896449e-02 -1.38604379e+00 -5.59566766e-02 4.65960175e-01
-1.22854114e+00 -1.36117315e+00 -3.51591259e-01 -1.18522346e+00
1.15412354e+00 5.36736012e-01 8.03736925e-01 2.62872670e-02
-3.14741790e-01 9.12762582e-01 -3.76055121e-01 4.53341268e-02
-3.31297517e-01 -1.21555090e-01 1.14816271e-01 4.48261440e-01
3.93013537e-01 -3.39250267e-01 -6.00108862e-01 5.35528243e-01
-8.25361133e-01 1.64080247e-01 5.15118062e-01 2.35730499e-01
1.23176014e+00 2.10491985e-01 -3.72704953e-01 -8.25401425e-01
-3.89775574e-01 -2.03931704e-01 -9.24133718e-01 2.45856062e-01
-2.44417325e-01 -1.31829053e-01 1.14740938e-01 -3.81400049e-01
-1.01874018e+00 1.08961535e+00 3.87304366e-01 -8.04640710e-01
-5.63552320e-01 -4.15948406e-02 -4.68147635e-01 -1.02257498e-01
3.76650393e-01 1.86892405e-01 8.29388946e-02 -2.68390089e-01
6.12797916e-01 7.14213908e-01 7.03021169e-01 -2.44942129e-01
1.01477349e+00 7.43626952e-01 -1.30108688e-02 -8.75743985e-01
-5.91024458e-01 -8.59518051e-01 -1.40509820e+00 -7.79546678e-01
1.30313814e+00 -9.16066885e-01 -4.78308618e-01 5.42921185e-01
-9.82322276e-01 -5.29682875e-01 -2.92212218e-01 5.92571259e-01
-8.67544472e-01 3.78072381e-01 -3.24652165e-01 -8.21304858e-01
5.48076369e-02 -1.19702840e+00 1.44726753e+00 4.18804824e-01
1.43819079e-01 -8.94951761e-01 -2.62079418e-01 6.61459982e-01
-4.43648547e-01 4.20144051e-01 5.14022410e-01 -6.77752435e-01
-1.13948166e+00 -1.31143734e-01 8.56097639e-02 4.27930087e-01
3.24192464e-01 2.71295130e-01 -8.87896001e-01 -8.08069762e-03
1.70912653e-01 -1.39515281e-01 4.42329854e-01 4.04506147e-01
1.07707870e+00 -1.02163345e-01 -6.31620288e-01 7.42871046e-01
1.49899030e+00 2.14668721e-01 4.81092006e-01 2.64626622e-01
1.24624562e+00 7.39939690e-01 1.06814659e+00 1.61682189e-01
1.81004971e-01 5.91747403e-01 8.41094375e-01 -1.70977965e-01
-4.78437692e-02 -9.86227244e-02 2.77423382e-01 6.97141230e-01
-6.25427663e-02 1.72765166e-01 -9.51981008e-01 1.48987904e-01
-1.84814250e+00 -6.40990019e-01 -4.05879319e-01 2.34786129e+00
5.74622393e-01 3.48263294e-01 1.35664880e-01 2.92787343e-01
9.12612617e-01 -1.39810100e-01 -7.72072554e-01 2.30822176e-01
-9.42979753e-02 5.62893078e-02 6.48786843e-01 3.62258852e-01
-1.40819347e+00 1.18988478e+00 5.09315968e+00 3.85787010e-01
-1.05031526e+00 -1.70747057e-01 4.46044028e-01 1.48790389e-01
8.50876272e-02 8.26581046e-02 -6.22411609e-01 2.86438644e-01
3.14530849e-01 4.12080020e-01 3.42621297e-01 1.09940135e+00
9.37558934e-02 -4.03491378e-01 -1.34693170e+00 1.17992020e+00
4.55439538e-01 -1.06767416e+00 -1.18387170e-01 -2.47786298e-01
9.74271655e-01 6.65806085e-02 -1.99241191e-01 -3.52959484e-01
-1.20956674e-01 -6.03951871e-01 7.10547805e-01 6.29213333e-01
3.03801388e-01 -4.81917530e-01 4.07654107e-01 3.90175968e-01
-9.46505964e-01 -1.79594532e-02 -3.53924870e-01 1.66307360e-01
9.78849083e-02 3.88720870e-01 -9.79025602e-01 4.40066606e-01
8.11099708e-01 9.97809768e-01 -6.49021029e-01 1.00546956e+00
-2.05815226e-01 2.88667381e-01 -5.48911631e-01 2.07675427e-01
8.51237774e-02 -6.60552084e-01 5.87277174e-01 1.06380033e+00
4.56925407e-02 1.82352141e-01 4.11626011e-01 6.92657053e-01
2.12045029e-01 2.09610872e-02 -5.66187024e-01 1.58523962e-01
2.82478452e-01 1.57777572e+00 -1.59248614e+00 -3.76398265e-01
-2.20977634e-01 1.35244107e+00 -6.59246892e-02 2.69086987e-01
-7.01400042e-01 6.68230131e-02 2.28493549e-02 1.74629718e-01
5.31454146e-01 -5.15009940e-01 -4.18570042e-02 -9.47719634e-01
1.76160522e-02 -6.99984908e-01 2.23697513e-01 -1.02691019e+00
-8.92017245e-01 4.05893981e-01 1.79227382e-01 -1.52129006e+00
-1.23622507e-01 -7.11708009e-01 -3.23641628e-01 1.55662835e-01
-1.23146176e+00 -1.18317354e+00 -6.65211022e-01 9.99284327e-01
7.71798491e-01 -1.25124946e-01 4.89322066e-01 1.16020732e-01
-4.69327718e-01 -2.07186401e-01 4.50706035e-02 4.02178727e-02
3.30923021e-01 -1.31013954e+00 -1.84462452e-03 8.71738553e-01
6.96673930e-01 3.58837396e-01 5.00357330e-01 -8.22381377e-01
-1.88906050e+00 -1.17368472e+00 2.29640245e-01 -5.48408747e-01
1.67733684e-01 -6.00764394e-01 -8.11335623e-01 7.59168327e-01
-6.71636239e-02 4.26542252e-01 3.65685135e-01 -5.44388473e-01
9.19874534e-02 -2.05944672e-01 -1.13673139e+00 2.37663195e-01
1.23265469e+00 -1.94674432e-01 -7.69369841e-01 8.56843948e-01
6.11267209e-01 -9.64571595e-01 -7.44367540e-01 3.56656224e-01
4.01355952e-01 -8.74453187e-01 9.22439218e-01 -4.28204060e-01
2.23946124e-01 -7.87000060e-01 -2.16043338e-01 -7.29072869e-01
1.93908691e-01 -6.29147470e-01 -1.87950701e-01 1.11182511e+00
3.85931544e-02 -3.57820690e-02 1.15578139e+00 6.15391374e-01
-5.50335273e-02 -2.89253831e-01 -8.48693311e-01 -5.99619985e-01
-7.96253920e-01 -5.62327027e-01 1.70635611e-01 8.23000252e-01
-3.00173819e-01 1.18129686e-01 2.21990556e-01 6.08925283e-01
1.05366063e+00 4.22658980e-01 1.16995025e+00 -1.30113161e+00
-1.30773753e-01 5.18711619e-02 -9.05295432e-01 -1.11793375e+00
2.77480811e-01 -1.01806247e+00 2.43996799e-01 -1.47066998e+00
6.48520663e-02 -7.08784521e-01 3.57850008e-02 3.35127950e-01
6.36050329e-02 3.72288823e-01 1.73064008e-01 6.03194177e-01
-9.97632682e-01 1.27053812e-01 1.02809763e+00 -2.36281157e-01
-5.42710900e-01 2.28999987e-01 2.20440164e-01 1.11256146e+00
6.21792436e-01 -6.51059806e-01 -3.27321202e-01 -3.71798515e-01
-1.22783117e-01 -1.11020133e-01 3.88305962e-01 -1.14952588e+00
5.03785312e-01 -2.06162944e-01 5.20040154e-01 -1.07279468e+00
3.00821036e-01 -1.53200960e+00 3.34800869e-01 4.90128338e-01
-5.69897965e-02 -1.28060598e-02 -2.17206310e-02 7.31123269e-01
5.46408333e-02 -4.14547145e-01 5.25591314e-01 -3.44342262e-01
-9.96178865e-01 2.73155153e-01 -2.10861459e-01 -3.09853613e-01
1.49799740e+00 -7.68417597e-01 3.41812253e-01 1.26587182e-01
-7.63586640e-01 9.99006554e-02 6.92539752e-01 4.05468166e-01
8.95138443e-01 -1.02811468e+00 -1.27859131e-01 5.57854891e-01
7.08657131e-02 8.20321202e-01 1.22220464e-01 6.29572988e-01
-9.14305151e-01 -9.77378339e-02 -2.35785261e-01 -1.60401165e+00
-1.44399202e+00 5.78136206e-01 2.07565308e-01 3.92331749e-01
-7.83748209e-01 7.26718009e-01 -2.33502492e-01 -3.94449055e-01
5.83807290e-01 -6.60017312e-01 -1.31889164e-01 3.83300036e-02
1.62464395e-01 3.10582101e-01 2.80754343e-02 -9.06704247e-01
-4.22589898e-01 1.06637013e+00 2.34796226e-01 -7.39605650e-02
1.38595653e+00 -2.53003865e-01 -2.13630468e-01 8.46895516e-01
1.21727657e+00 -2.33203560e-01 -1.59453177e+00 -4.13745433e-01
1.61808729e-01 -6.43954396e-01 -8.72866735e-02 -2.67885596e-01
-1.20024216e+00 4.64062244e-01 9.63713288e-01 6.91521987e-02
1.00445759e+00 2.81218976e-01 4.19711471e-01 4.92846310e-01
5.83690286e-01 -1.26859403e+00 3.46269011e-01 7.31924921e-02
4.09479499e-01 -1.36761880e+00 3.54375899e-01 -6.08347654e-01
-6.32872939e-01 1.25160384e+00 4.65351403e-01 -3.49907666e-01
4.08392936e-01 4.03094962e-02 1.01600647e-01 -3.91654015e-01
-4.10574451e-02 -4.74888444e-01 3.48519683e-01 6.33616388e-01
-3.12915802e-01 -2.21683055e-01 1.29814073e-01 -6.04159497e-02
1.35674153e-03 -2.75178134e-01 3.03639293e-01 1.05149531e+00
-5.83534777e-01 -8.45861316e-01 -6.10951126e-01 2.25504249e-01
-1.11811586e-01 4.88070637e-01 -3.43636692e-01 8.31200540e-01
3.78251135e-01 8.35599005e-01 4.99374747e-01 -3.46368790e-01
3.01909864e-01 -4.17640246e-03 7.26887345e-01 -7.69383311e-01
-4.42725241e-01 3.46367896e-01 -4.91099238e-01 -8.54628921e-01
-1.21379328e+00 -9.59736168e-01 -1.62286246e+00 6.38972223e-01
-4.72523540e-01 -1.81189612e-01 1.06392312e+00 9.51024950e-01
7.08170459e-02 2.59524047e-01 9.98586416e-01 -1.36579394e+00
1.01721972e-01 -4.85203862e-01 -5.52957654e-01 7.20269263e-01
1.21345215e-01 -6.75681174e-01 -4.16254103e-01 8.22411001e-01] | [7.732198715209961, -2.677171230316162] |
b9025547-bfad-4db5-aa20-c7d3f7b429ff | self-supervised-generalisation-with-meta | 1901.08933 | null | https://arxiv.org/abs/1901.08933v3 | https://arxiv.org/pdf/1901.08933v3.pdf | Self-Supervised Generalisation with Meta Auxiliary Learning | Learning with auxiliary tasks can improve the ability of a primary task to generalise. However, this comes at the cost of manually labelling auxiliary data. We propose a new method which automatically learns appropriate labels for an auxiliary task, such that any supervised learning task can be improved without requiring access to any further data. The approach is to train two neural networks: a label-generation network to predict the auxiliary labels, and a multi-task network to train the primary task alongside the auxiliary task. The loss for the label-generation network incorporates the loss of the multi-task network, and so this interaction between the two networks can be seen as a form of meta learning with a double gradient. We show that our proposed method, Meta AuXiliary Learning (MAXL), outperforms single-task learning on 7 image datasets, without requiring any additional data. We also show that MAXL outperforms several other baselines for generating auxiliary labels, and is even competitive when compared with human-defined auxiliary labels. The self-supervised nature of our method leads to a promising new direction towards automated generalisation. Source code can be found at https://github.com/lorenmt/maxl. | ['Shikun Liu', 'Edward Johns', 'Andrew J. Davison'] | 2019-01-25 | self-supervised-generalisation-with-meta-1 | http://papers.nips.cc/paper/8445-self-supervised-generalisation-with-meta-auxiliary-learning | http://papers.nips.cc/paper/8445-self-supervised-generalisation-with-meta-auxiliary-learning.pdf | neurips-2019-12 | ['auxiliary-learning'] | ['methodology'] | [ 7.10358441e-01 6.43976748e-01 -8.49305093e-02 -6.11674309e-01
-1.35419559e+00 -7.51109004e-01 8.53240013e-01 1.50291324e-01
-7.67973185e-01 8.10288429e-01 1.46395832e-01 -2.83565909e-01
1.57405198e-01 -2.91341364e-01 -7.73595333e-01 -9.59862649e-01
3.13158333e-01 7.37490177e-01 1.47542253e-01 -1.83371052e-01
1.38606951e-01 1.93241835e-01 -1.22546017e+00 3.88090283e-01
6.96993232e-01 9.72799480e-01 3.51150423e-01 7.41089165e-01
1.83779478e-01 6.18012488e-01 -4.34389412e-01 -4.92927581e-01
2.98892021e-01 -4.40295577e-01 -1.29120708e+00 1.63191110e-01
3.87045234e-01 1.20481357e-01 1.69213727e-01 7.99056590e-01
5.60222685e-01 2.19093695e-01 7.79158056e-01 -1.47843409e+00
-4.95301992e-01 4.44180995e-01 -2.57401109e-01 -3.87174897e-02
-9.19104367e-02 5.28730638e-02 1.18378794e+00 -8.88256192e-01
4.83471006e-01 1.25151145e+00 7.08224237e-01 8.06864262e-01
-1.47651517e+00 -4.78992701e-01 3.33361298e-01 -5.51881678e-02
-9.98026669e-01 -5.54902017e-01 7.20022857e-01 -4.31033671e-01
7.42415965e-01 -2.44979467e-02 6.23399317e-02 1.29326296e+00
-2.24905759e-02 1.01931787e+00 1.31946707e+00 -6.46808803e-01
-6.37244806e-02 5.44453502e-01 1.64400339e-01 7.03281403e-01
-1.84969574e-01 1.24720976e-01 -2.70012498e-01 -9.71421376e-02
2.71348894e-01 -2.76940763e-01 -1.06353931e-01 -4.63784188e-01
-1.24683070e+00 7.67475784e-01 5.23723900e-01 8.18893909e-02
-2.75817484e-01 2.60482222e-01 5.63339353e-01 5.73118627e-01
9.04889941e-01 8.12614501e-01 -9.00173604e-01 2.23095968e-01
-6.71191990e-01 3.10138054e-03 6.67601883e-01 9.60893691e-01
1.11128366e+00 -2.47846752e-01 -3.18486691e-01 9.28735256e-01
3.14074010e-01 2.68836260e-01 4.45637792e-01 -1.22440195e+00
4.92763937e-01 5.34399033e-01 -2.22221948e-02 -2.62023240e-01
-6.67826176e-01 -7.10711420e-01 -8.74239802e-01 3.43780577e-01
4.16377544e-01 -3.77011120e-01 -1.00688076e+00 2.08199787e+00
2.48991907e-01 -6.32124394e-02 -4.55599464e-02 4.76695597e-01
6.16776705e-01 5.30642331e-01 2.42367208e-01 -4.07375097e-02
1.03486609e+00 -1.52562881e+00 -4.09463644e-01 -4.66751426e-01
1.15756643e+00 -7.92172432e-01 1.13633680e+00 2.76035219e-01
-1.13411903e+00 -7.45060265e-01 -6.78452730e-01 -2.49904618e-01
-4.87416387e-01 2.49081567e-01 5.52717149e-01 3.13362271e-01
-1.57304108e+00 5.71576536e-01 -7.59494126e-01 -1.96234375e-01
5.56115091e-01 5.47956765e-01 -4.83020812e-01 7.84894228e-02
-1.04492915e+00 1.23242986e+00 6.49003327e-01 -8.67147073e-02
-1.02034414e+00 -5.62505841e-01 -9.86985207e-01 -1.23551063e-01
4.25168186e-01 -8.50954294e-01 1.65932167e+00 -1.30609632e+00
-1.24893880e+00 1.20823801e+00 -5.64088412e-02 -1.95300952e-01
4.47545916e-01 -1.53891400e-01 4.24359031e-02 -6.91514686e-02
3.68151039e-01 1.20507634e+00 9.16191161e-01 -1.48702538e+00
-9.92818356e-01 -2.12753743e-01 2.14117631e-01 4.17624801e-01
-8.31643417e-02 -2.11351082e-01 -2.92805344e-01 -6.95889533e-01
-2.94844747e-01 -1.23495924e+00 -4.63600099e-01 -1.09215192e-01
-4.15741473e-01 -6.06001198e-01 7.49493480e-01 -5.92690945e-01
9.14057255e-01 -2.03718185e+00 4.37319010e-01 -5.39708361e-02
1.65527567e-01 5.01609445e-01 -7.51682639e-01 2.24389806e-01
-4.03700471e-01 2.07991228e-01 -4.93097782e-01 -8.74721706e-01
-6.12378009e-02 1.52259320e-01 6.11720234e-03 2.70875216e-01
4.61875945e-01 1.20327997e+00 -1.10793018e+00 -3.54126364e-01
4.17717807e-02 4.12547737e-01 -4.89861965e-01 3.07653725e-01
-3.83147985e-01 8.01097333e-01 -2.27046758e-01 3.78525883e-01
1.79692000e-01 -4.99800682e-01 -2.03543127e-01 -2.71335170e-02
2.80909717e-01 5.99258721e-01 -8.18324924e-01 1.75988340e+00
-8.64353359e-01 4.31307763e-01 1.97939768e-01 -1.44586384e+00
7.56598651e-01 4.25896227e-01 2.55623251e-01 -5.50039828e-01
-6.11085705e-02 3.46869200e-01 -9.47883502e-02 -2.75085896e-01
9.02417526e-02 -2.51832396e-01 -1.33123979e-01 9.71309304e-01
4.97764021e-01 -2.97263592e-01 2.54439294e-01 3.48581731e-01
9.19568777e-01 5.48036575e-01 2.35591337e-01 -2.07750365e-01
6.99440062e-01 -1.71813607e-01 3.37826848e-01 8.58851075e-01
-2.82283244e-03 1.91852853e-01 3.81226331e-01 -5.73451638e-01
-1.09445357e+00 -7.00131118e-01 1.15351574e-02 1.79977202e+00
-2.88793802e-01 -2.86648840e-01 -6.07214093e-01 -1.33942795e+00
-1.48866385e-01 7.26883650e-01 -7.40807831e-01 -2.77931720e-01
-5.38032472e-01 -5.85748434e-01 3.91071081e-01 4.61751431e-01
4.37668473e-01 -1.32421410e+00 -1.98923096e-01 8.13234448e-02
-3.58965904e-01 -9.06850338e-01 -6.23906076e-01 8.43378663e-01
-8.73031795e-01 -8.52630854e-01 -6.33597732e-01 -9.70030487e-01
9.01511788e-01 -6.30360097e-02 1.40322256e+00 2.97727048e-01
1.28204912e-01 4.03077990e-01 -3.83154213e-01 -3.64631802e-01
-8.71366978e-01 7.76121318e-01 -3.08753371e-01 -1.34903729e-01
1.75081074e-01 -6.14667594e-01 -3.55387449e-01 2.87844568e-01
-8.57282221e-01 2.84126908e-01 8.24089170e-01 1.09828150e+00
4.92773473e-01 -3.51295173e-01 9.15545642e-01 -1.30924213e+00
5.01293898e-01 -4.02935594e-01 -4.36074555e-01 2.94097811e-01
-7.06897795e-01 6.48787975e-01 5.52304327e-01 -2.93954670e-01
-1.08447444e+00 4.86388773e-01 -3.66204828e-01 9.73893180e-02
-4.28136796e-01 3.60319525e-01 -1.90497175e-01 -1.41810745e-01
8.27337861e-01 -6.26104977e-03 -5.53692318e-02 -6.30552173e-01
7.12279141e-01 8.15374434e-01 4.51597482e-01 -5.74650109e-01
8.79141927e-01 2.13973969e-01 1.65544778e-01 -3.69215786e-01
-1.66163778e+00 -6.66832864e-01 -1.12026560e+00 -1.97007209e-02
5.96819222e-01 -7.69070446e-01 -1.99138701e-01 5.00230610e-01
-1.17897308e+00 -1.06543946e+00 -4.94747937e-01 1.90871209e-01
-8.08245242e-01 2.19844028e-01 -4.30215985e-01 -3.88020635e-01
-2.29347110e-01 -1.17123473e+00 1.30642259e+00 -2.11481363e-01
-2.45857120e-01 -1.56873178e+00 1.18468245e-02 5.10913074e-01
3.94275635e-01 5.21467850e-02 9.06285048e-01 -1.08328390e+00
-3.61581951e-01 1.47557333e-02 -1.84439763e-01 8.56376410e-01
3.39306861e-01 -6.41507208e-01 -1.14196229e+00 -5.64581275e-01
1.56001687e-01 -1.07758522e+00 1.17882061e+00 1.80932552e-01
1.02001536e+00 -4.96165246e-01 -2.84631878e-01 7.61848032e-01
1.06745529e+00 -1.64444193e-01 2.98958182e-01 5.96127570e-01
9.37763512e-01 7.43529320e-01 6.15371764e-01 -5.47059327e-02
5.30578554e-01 7.25766540e-01 3.71932149e-01 -4.26314741e-01
-3.29301387e-01 3.28226238e-02 3.27861518e-01 6.19886160e-01
-1.08598452e-02 -1.79585621e-01 -9.29336429e-01 5.28046727e-01
-1.89979851e+00 -6.41221285e-01 1.09547824e-02 2.14081764e+00
1.20123613e+00 1.97006539e-01 1.90258339e-01 7.17202276e-02
5.86781085e-01 7.24115297e-02 -5.77828169e-01 -4.90881920e-01
9.86458585e-02 2.56800354e-01 3.72261912e-01 8.74914646e-01
-1.52221382e+00 1.08466780e+00 5.91240215e+00 6.83451235e-01
-9.24428344e-01 5.21669626e-01 8.14843178e-01 1.37232229e-01
-2.23491758e-01 -2.31809486e-02 -9.93685782e-01 2.20040753e-01
9.42199826e-01 1.47067472e-01 3.51773441e-01 7.55735040e-01
-1.09584905e-01 1.88030619e-02 -1.48314297e+00 6.28928959e-01
2.16335699e-01 -9.31672335e-01 1.42410681e-01 4.77513822e-04
9.95691836e-01 2.63040572e-01 5.55572957e-02 4.85168636e-01
6.66043818e-01 -8.40164125e-01 5.06642878e-01 2.48387411e-01
7.91663229e-01 -6.31294847e-01 8.30865860e-01 6.70250297e-01
-8.48904431e-01 1.59964375e-02 3.53683578e-03 -1.19911171e-01
1.65791482e-01 3.75213027e-01 -7.89516091e-01 2.96803445e-01
2.50548065e-01 6.53683126e-01 -9.34667408e-01 8.81174445e-01
-8.00593615e-01 4.92567003e-01 -2.02955961e-01 3.65438432e-01
5.43396771e-01 1.66209236e-01 4.51602131e-01 1.39540577e+00
-5.08082844e-02 -2.75860935e-01 5.79084635e-01 4.81641829e-01
-3.66692454e-01 -1.09108584e-02 -6.06841683e-01 9.38070565e-02
1.63271800e-01 1.52075458e+00 -6.42765880e-01 -3.90692562e-01
-4.78300422e-01 1.17910838e+00 7.16856062e-01 3.62491161e-01
-4.50196862e-01 -3.21470976e-01 -1.90480426e-02 -2.54093390e-02
2.71949340e-02 1.29401043e-01 -1.15996227e-01 -9.63964283e-01
-8.67994949e-02 -9.01467323e-01 6.65409982e-01 -6.64420485e-01
-1.29095948e+00 7.11767673e-01 -1.36743141e-02 -9.06760275e-01
-5.51990926e-01 -6.77388310e-01 -3.24567616e-01 1.10924935e+00
-1.72382927e+00 -1.50539470e+00 -9.38502327e-02 5.99930525e-01
7.12865949e-01 -2.37257048e-01 9.27767694e-01 3.16230059e-02
-1.38739973e-01 7.56884038e-01 -1.22388944e-01 1.48942797e-02
1.05658948e+00 -1.60556030e+00 6.10911250e-01 7.02939868e-01
2.02095374e-01 1.68750018e-01 3.70085269e-01 -3.95898163e-01
-5.26896477e-01 -1.44022655e+00 1.17170501e+00 -8.97212505e-01
5.23166001e-01 -5.24089932e-01 -8.90563309e-01 1.07851958e+00
3.29530329e-01 -2.74316352e-02 6.59595966e-01 3.55121046e-01
-4.20996755e-01 6.16675876e-02 -8.96283925e-01 3.51212889e-01
9.12701428e-01 -5.37742138e-01 -6.26051545e-01 7.67170370e-01
9.16930914e-01 -1.92872375e-01 -6.75490737e-01 5.14055431e-01
5.11715934e-02 -6.19104087e-01 8.46695662e-01 -6.65844142e-01
4.42528427e-01 -9.62104276e-02 2.23154724e-01 -1.75260222e+00
-5.41348398e-01 -6.48834825e-01 -2.27175113e-02 1.18955839e+00
9.40691710e-01 -8.01760554e-01 5.78599215e-01 4.85643893e-01
-3.74193132e-01 -8.11349630e-01 -5.75729966e-01 -8.95191610e-01
1.90689787e-01 -2.30608806e-01 4.81131822e-02 1.00792098e+00
-2.36538514e-01 1.11018288e+00 -3.17331165e-01 -8.54393169e-02
6.50714934e-01 -2.34664917e-01 7.36597240e-01 -1.29904199e+00
-3.37974012e-01 -4.08591002e-01 -2.91677918e-02 -9.97028649e-01
5.68772197e-01 -1.42315686e+00 2.07344919e-01 -1.63919437e+00
3.10127467e-01 -8.09818983e-01 -4.54668999e-01 1.13797021e+00
-5.65737367e-01 5.04796863e-01 1.17536694e-01 2.36561418e-01
-8.81645083e-01 4.45822358e-01 1.29908597e+00 -1.65723398e-01
-2.45023131e-01 4.85595495e-01 -1.07833576e+00 8.77242327e-01
1.00069571e+00 -8.04576695e-01 -4.55273896e-01 -5.51089823e-01
4.36617017e-01 -3.31222981e-01 2.75456816e-01 -7.00203896e-01
2.08038777e-01 2.78638691e-01 1.38021931e-01 -2.62179881e-01
2.09445074e-01 -6.54002964e-01 -2.94183493e-01 3.75521958e-01
-7.98087716e-01 -1.05739450e-02 2.30250180e-01 2.54986495e-01
-4.56244126e-02 -6.53106630e-01 9.63290036e-01 -2.51274914e-01
-4.88565773e-01 3.11805129e-01 -1.27932087e-01 3.74984533e-01
7.02965975e-01 1.43911198e-01 -2.48732671e-01 -4.19883639e-01
-1.06776333e+00 5.02291083e-01 3.89892578e-01 3.83782834e-01
6.87614903e-02 -1.28978813e+00 -9.75518048e-01 -5.03775962e-02
9.00333226e-02 3.31635356e-01 -2.26038724e-01 8.47913563e-01
4.80815172e-02 4.64675695e-01 -2.91284174e-02 -6.31463945e-01
-1.14880002e+00 7.69257903e-01 2.55253941e-01 -7.24289715e-01
-3.54440719e-01 1.08148682e+00 4.22693908e-01 -1.01093590e+00
4.02302504e-01 6.34187311e-02 -1.38111606e-01 1.81622818e-01
3.21890771e-01 2.04699188e-01 1.38613060e-01 -6.41650379e-01
-6.55614510e-02 4.62093621e-01 -3.31043363e-01 -2.92213380e-01
1.48801851e+00 -2.56328613e-01 -1.85362265e-01 5.85453331e-01
1.34148264e+00 -3.77536446e-01 -1.38161838e+00 -6.19549513e-01
3.16461086e-01 -2.49735098e-02 -1.10483922e-01 -1.02549887e+00
-9.44688380e-01 9.58377838e-01 1.98503658e-01 9.54984054e-02
1.31717038e+00 2.80994892e-01 4.32934910e-01 7.82820225e-01
2.09943175e-01 -1.01594293e+00 4.20775115e-01 7.11591005e-01
9.32534993e-01 -1.62245691e+00 -7.75619820e-02 -2.18769029e-01
-7.96109140e-01 7.07720637e-01 7.04476118e-01 -1.42010981e-02
4.84990656e-01 8.32402408e-02 2.90780485e-01 -1.98648795e-01
-9.56163883e-01 -3.26745212e-01 4.55793202e-01 7.57413328e-01
4.63156551e-01 -2.45095849e-01 1.93820726e-02 1.99440956e-01
-8.61933157e-02 -2.33631462e-01 2.30871290e-01 7.87629783e-01
-3.60671431e-01 -1.64748001e+00 -7.00276578e-03 4.92972016e-01
-5.38724661e-01 -3.62291962e-01 -4.57025707e-01 7.09235966e-01
2.92659402e-01 7.47130573e-01 -2.08604723e-01 3.37991677e-02
1.48363963e-01 4.84769970e-01 3.38472217e-01 -1.20203292e+00
-6.51641369e-01 7.17121065e-02 2.09940672e-01 -4.72046524e-01
-7.58353829e-01 -4.44297552e-01 -6.40435755e-01 2.45667383e-01
-4.74828899e-01 3.66138875e-01 5.72715402e-01 1.07432222e+00
1.20936081e-01 4.96112049e-01 7.27406740e-01 -1.17736077e+00
-5.41374266e-01 -1.19819081e+00 -2.10154444e-01 5.53599477e-01
6.15285814e-01 -5.40478468e-01 -6.24428272e-01 3.60023737e-01] | [9.473678588867188, 3.8527956008911133] |
8b612000-dae7-4158-834f-5f5fc9192644 | a-data-bootstrapping-recipe-for-low-resource-1 | null | null | https://aclanthology.org/2021.conll-1.45 | https://aclanthology.org/2021.conll-1.45.pdf | A Data Bootstrapping Recipe for Low-Resource Multilingual Relation Classification | Relation classification (sometimes called ‘extraction’) requires trustworthy datasets for fine-tuning large language models, as well as for evaluation. Data collection is challenging for Indian languages, because they are syntactically and morphologically diverse, as well as different from resource-rich languages like English. Despite recent interest in deep generative models for Indian languages, relation classification is still not well-served by public data sets. In response, we present IndoRE, a dataset with 39K entity- and relation-tagged gold sentences in three Indian languages, plus English. We start with a multilingual BERT (mBERT) based system that captures entity span positions and type information and provides competitive monolingual relation classification. Using this system, we explore and compare transfer mechanisms between languages. In particular, we study the accuracy-efficiency tradeoff between expensive gold instances vs. translated and aligned ‘silver’ instances. | ['Soumen Chakrabarti', 'Niloy Ganguly', 'Animesh Mukherjee', 'Bidisha Samanta', 'Arijit Nag'] | null | null | null | null | conll-emnlp-2021-11 | ['relation-classification'] | ['natural-language-processing'] | [-2.13697568e-01 1.83024257e-01 -4.97147799e-01 -4.39126700e-01
-1.43213093e+00 -9.63700175e-01 5.19214988e-01 4.02215213e-01
-5.99243224e-01 1.23317051e+00 3.96223515e-01 -6.11851454e-01
-1.18811410e-02 -9.08654153e-01 -6.11850560e-01 -3.03592175e-01
1.57745443e-02 1.27755928e+00 -5.30588403e-02 -4.14349824e-01
-3.33210289e-01 1.31896764e-01 -6.76564872e-01 3.28780144e-01
1.06327176e+00 5.86274803e-01 -3.54096144e-02 4.94328946e-01
-1.56244889e-01 8.42940331e-01 -5.20920575e-01 -1.22942603e+00
4.12680246e-02 -3.90535206e-01 -1.27652287e+00 -3.28391165e-01
3.43178511e-01 1.31552503e-01 -3.85375112e-01 8.30913901e-01
5.77162385e-01 -2.75602013e-01 6.50950432e-01 -9.08949077e-01
-1.14236999e+00 1.40230417e+00 -5.73431432e-01 4.14304525e-01
3.34993362e-01 -4.69340906e-02 1.44415200e+00 -6.87495708e-01
1.33827221e+00 1.04312301e+00 6.84354603e-01 4.73132640e-01
-1.38720155e+00 -6.56182766e-01 -9.84351560e-02 2.03879878e-01
-1.52808666e+00 -8.12093019e-01 3.49814624e-01 -4.20065194e-01
1.59116066e+00 4.06443268e-01 4.46896523e-01 1.22434628e+00
1.03292227e-01 7.03461170e-01 1.03606856e+00 -5.56170881e-01
-3.77279192e-01 1.40914887e-01 2.14051902e-01 4.26798880e-01
5.16977310e-01 -1.37237400e-01 -6.27081215e-01 -5.04320338e-02
2.61016160e-01 -6.49672449e-01 -1.79682374e-01 2.77904794e-03
-1.19675839e+00 7.60327876e-01 2.66667008e-01 4.42238808e-01
-2.03882530e-01 -3.24593574e-01 4.25496429e-01 4.62668955e-01
8.50181103e-01 4.60259378e-01 -1.04476070e+00 -3.27735364e-01
-6.62474453e-01 1.95213899e-01 1.09374678e+00 1.30083847e+00
6.59989059e-01 -1.83034480e-01 8.34665149e-02 1.20703650e+00
4.58730236e-02 4.89686608e-01 2.87849903e-01 -4.03287590e-01
1.07237327e+00 5.77147663e-01 -4.63642716e-01 -6.74714863e-01
-6.44782543e-01 -4.77168381e-01 -9.18132365e-01 -6.02408528e-01
5.75728238e-01 -2.70707309e-01 -4.35307741e-01 1.82740116e+00
2.94013083e-01 -5.67696691e-01 4.95515913e-01 5.03632307e-01
1.12531567e+00 3.81915510e-01 3.07499558e-01 -3.43599796e-01
1.71770108e+00 -6.45099998e-01 -4.31700379e-01 -4.76357102e-01
9.99878764e-01 -8.70262384e-01 8.38503063e-01 -4.04166281e-02
-1.20682383e+00 -1.09561346e-01 -6.52125537e-01 -5.19045591e-01
-5.83904862e-01 1.19520113e-01 1.13393855e+00 5.42133808e-01
-9.16874707e-01 2.41759092e-01 -9.01772082e-01 -7.48626530e-01
4.35457379e-01 2.39593849e-01 -7.74084210e-01 6.88419491e-02
-1.41833770e+00 1.31442094e+00 6.66717291e-01 1.59190536e-01
-3.33002299e-01 -4.71425116e-01 -1.13376772e+00 -2.07828075e-01
3.55034232e-01 -4.34283465e-01 1.30304468e+00 -4.06413913e-01
-8.98686469e-01 1.47577524e+00 -1.13553442e-01 -5.22847593e-01
1.89016834e-01 -2.63364941e-01 -6.94434285e-01 -3.68300825e-01
3.46000791e-01 3.93017858e-01 -1.18776023e-01 -9.35871124e-01
-7.16587901e-01 -5.06042242e-01 -9.49710235e-03 3.10253948e-01
-3.23050469e-02 5.38049877e-01 -3.11439484e-01 -6.03047073e-01
1.06441833e-01 -9.28223670e-01 -1.02960989e-01 -8.39521348e-01
-8.07610750e-01 -3.47664773e-01 1.20634481e-01 -1.08368397e+00
1.33986056e+00 -1.86396575e+00 -1.32460948e-02 -1.74531492e-03
2.47184664e-01 1.42925888e-01 -5.96476533e-03 6.38963103e-01
-1.97100826e-02 3.95463586e-01 -8.59663337e-02 -5.55824898e-02
-5.95069006e-02 5.17964959e-01 -1.04543446e-02 3.56441289e-01
6.87017560e-01 1.28945565e+00 -9.62997794e-01 -7.64861822e-01
-3.86275977e-01 2.87321091e-01 -4.46154237e-01 -7.46914670e-02
-1.30716115e-01 2.99794823e-01 -2.04568654e-01 1.06769311e+00
3.34043682e-01 -2.02520728e-01 8.05937171e-01 -1.11355603e-01
-3.11788078e-02 9.89642978e-01 -7.71344543e-01 1.46543252e+00
-5.85500240e-01 5.99849999e-01 -8.77989978e-02 -6.45117223e-01
7.47979999e-01 3.11938316e-01 1.83199376e-01 -7.48129368e-01
9.58926752e-02 6.74934089e-01 3.82911325e-01 -2.53904343e-01
8.61594319e-01 -8.90236348e-02 -4.35018688e-01 3.13911706e-01
3.44422519e-01 -3.16154212e-02 5.31008303e-01 2.79082209e-01
1.04657447e+00 2.09291860e-01 6.62745178e-01 -2.94207931e-01
1.69134080e-01 2.74228990e-01 8.33300769e-01 2.89506674e-01
8.84071738e-02 2.69248784e-01 6.24766350e-01 -3.84306192e-01
-9.17226493e-01 -9.26767409e-01 -5.31033933e-01 1.19982302e+00
-2.13083819e-01 -6.03718996e-01 -5.29774964e-01 -7.74615645e-01
-1.36057302e-01 5.37546933e-01 -3.73422951e-01 5.50800003e-02
-9.21998203e-01 -1.22421777e+00 9.06486571e-01 6.01658702e-01
1.91648126e-01 -1.08211076e+00 5.05733043e-02 4.02316034e-01
-3.88078511e-01 -1.48079741e+00 -3.18164557e-01 6.94491446e-01
-3.51889461e-01 -1.13483942e+00 -3.50640506e-01 -8.86977196e-01
1.64240226e-01 -4.09852445e-01 2.00637555e+00 -3.04314822e-01
2.94906814e-02 -2.31162295e-01 -3.76506180e-01 -3.90344739e-01
-7.14586079e-01 9.04917836e-01 1.13197356e-01 -5.72230339e-01
8.27323258e-01 -4.02113914e-01 2.09299698e-02 -4.84012254e-02
-4.36499268e-01 9.31878611e-02 6.42799914e-01 7.50640333e-01
6.99570596e-01 -2.44506270e-01 5.68178535e-01 -1.46524775e+00
4.89307553e-01 -6.73726857e-01 -5.18182695e-01 6.83711588e-01
-5.87814271e-01 1.06484771e-01 3.28449488e-01 -1.47900254e-01
-9.73967195e-01 -1.71967581e-01 -3.34712088e-01 5.40018439e-01
-2.05976758e-02 6.23475730e-01 -5.65660775e-01 2.93885797e-01
8.05183172e-01 -1.49845526e-01 -5.74803293e-01 -5.94032705e-01
5.03211558e-01 8.75787854e-01 7.65870690e-01 -9.65920329e-01
5.80523252e-01 -1.47123754e-01 -3.00187588e-01 -7.39321351e-01
-9.72859859e-01 -1.70286343e-01 -9.82717931e-01 4.78715450e-01
9.80681598e-01 -1.10036123e+00 -5.16734004e-01 3.53294998e-01
-1.20742810e+00 -4.83994573e-01 -2.54621595e-01 4.52857405e-01
-2.10027799e-01 -2.06588849e-01 -1.32112956e+00 -4.60547239e-01
-5.23854256e-01 -9.63418067e-01 1.09131467e+00 4.45747375e-02
-5.31458616e-01 -1.08553934e+00 3.62789810e-01 5.10746181e-01
6.46812245e-02 1.70272559e-01 1.06746936e+00 -1.00036454e+00
-5.50738811e-01 -6.43351600e-02 -3.64969641e-01 -1.01128101e-01
1.29539549e-01 5.07613458e-02 -7.94541717e-01 4.72736321e-02
-6.78368926e-01 -7.14468300e-01 6.43926084e-01 1.08621910e-01
4.18412507e-01 -4.37966049e-01 -4.37382370e-01 7.07384884e-01
1.13060236e+00 2.51818784e-02 3.63398075e-01 5.02159834e-01
9.38323140e-01 6.20446086e-01 4.74820226e-01 -1.22894742e-01
1.08299053e+00 6.53554142e-01 -3.20522428e-01 -1.30252615e-01
-2.04493895e-01 -3.24080139e-01 6.09383881e-02 1.31868625e+00
-2.12615952e-01 -3.70745510e-01 -1.41912293e+00 6.54191673e-01
-1.48659575e+00 -7.78335154e-01 -3.50933433e-01 1.88143158e+00
1.70305109e+00 1.44957721e-01 8.78204703e-02 -1.97507977e-01
3.77788961e-01 3.58049050e-02 -1.91093981e-01 -3.94992381e-01
-8.02932739e-01 5.44655144e-01 5.16571641e-01 5.87457240e-01
-1.18904841e+00 1.50227523e+00 6.28831100e+00 9.31691289e-01
-8.73006761e-01 3.03147137e-01 7.25724876e-01 2.03394920e-01
-4.83460724e-01 2.46550933e-01 -1.16501117e+00 1.63499087e-01
1.20245743e+00 -1.77423328e-01 3.02427292e-01 6.11618578e-01
-3.82765204e-01 -2.75126100e-02 -1.27166092e+00 7.85109460e-01
-8.66259038e-02 -1.29133427e+00 -1.38370425e-01 2.53042132e-01
6.72432780e-01 4.97997493e-01 -3.10082853e-01 5.36738575e-01
9.62850869e-01 -1.14451742e+00 7.49083757e-01 6.06113896e-02
1.35259891e+00 -7.46307790e-01 8.12866747e-01 1.83242649e-01
-1.22172463e+00 4.70121950e-01 -3.41035753e-01 1.13397576e-01
1.48263231e-01 6.14855468e-01 -7.61596262e-01 7.08955824e-01
6.53093338e-01 6.21518314e-01 -7.73634255e-01 4.06839401e-01
-4.23313677e-01 8.57747853e-01 -3.95744890e-01 7.33819604e-02
4.78116274e-02 -1.06548093e-01 3.12996179e-01 1.52998435e+00
-1.84174441e-02 1.46433398e-01 2.38031060e-01 4.55906063e-01
-4.74710703e-01 4.19161350e-01 -6.50715649e-01 -2.92621017e-01
6.39957607e-01 1.35994232e+00 -7.65872777e-01 -2.33969778e-01
-5.07957697e-01 7.25404978e-01 1.00023913e+00 2.89961308e-01
-4.18925583e-01 -9.57251899e-03 5.78441978e-01 -2.49979086e-02
-1.36237577e-01 -3.55102718e-01 -2.43263289e-01 -1.50233042e+00
-1.22873299e-02 -1.11238635e+00 6.52946889e-01 -3.82479399e-01
-1.51351607e+00 9.35579300e-01 -2.30963200e-01 -6.53950691e-01
-5.57389259e-01 -4.79703188e-01 1.67002797e-01 1.04135656e+00
-1.21625614e+00 -1.77084780e+00 2.56558686e-01 3.42888147e-01
3.31525028e-01 -1.97894707e-01 1.06441331e+00 6.99768841e-01
-8.14514339e-01 9.79052007e-01 -2.95613846e-03 7.92905927e-01
6.78867579e-01 -1.48088491e+00 8.87643218e-01 7.70901620e-01
9.19876635e-01 7.30013311e-01 4.09508735e-01 -7.67945707e-01
-1.16275787e+00 -1.03135896e+00 1.72488797e+00 -9.90974009e-01
9.68351364e-01 -6.51686847e-01 -8.56416345e-01 1.09093153e+00
3.76504183e-01 -1.07035473e-01 1.03463018e+00 1.00664175e+00
-5.08768737e-01 -7.68625513e-02 -8.23168576e-01 5.39319038e-01
1.22232795e+00 -8.10487151e-01 -4.30108637e-01 5.04364371e-01
6.58962131e-01 -6.87042236e-01 -1.20521104e+00 4.09206748e-01
3.82292330e-01 -5.44478953e-01 7.37037957e-01 -9.59121466e-01
5.63968956e-01 1.89156145e-01 -3.83817971e-01 -1.40109050e+00
-3.72217566e-01 -6.04453921e-01 -1.58689320e-01 1.71503663e+00
1.16505957e+00 -5.98585427e-01 5.84015608e-01 5.23810327e-01
4.12933156e-03 -9.25570667e-01 -7.09562063e-01 -8.21970701e-01
6.00382209e-01 -5.44563591e-01 6.81485653e-01 1.23411632e+00
-2.24032812e-02 1.11672473e+00 -1.00835830e-01 -2.32283119e-02
2.70176917e-01 3.25246274e-01 6.09358907e-01 -1.17831147e+00
-3.33364755e-01 -4.05948371e-01 -3.72861564e-01 -3.95450652e-01
4.13384229e-01 -1.30894387e+00 -1.07038982e-01 -1.59862506e+00
4.21636075e-01 -9.32502627e-01 7.60612562e-02 8.29665720e-01
-4.08187836e-01 3.84423256e-01 -1.40957460e-01 2.92158842e-01
-5.69175243e-01 1.73864961e-01 7.81166732e-01 8.04182515e-02
-6.54373616e-02 -1.43347874e-01 -9.61433113e-01 5.93845785e-01
6.28285348e-01 -5.27570069e-01 -1.37749240e-02 -7.84454167e-01
8.12552512e-01 4.60773446e-02 -1.23235926e-01 -5.24478674e-01
-2.44155172e-02 -2.07724079e-01 2.91025162e-01 -5.68525732e-01
-1.55849224e-02 -3.90999585e-01 1.10183433e-01 7.77839683e-03
-2.88959563e-01 2.88776726e-01 4.13581729e-02 1.28859013e-01
-2.61001915e-01 1.62096828e-01 3.86952817e-01 -2.05572695e-01
-5.74507833e-01 3.81186932e-01 -7.65893888e-03 8.63874376e-01
7.26944208e-01 1.65851846e-01 -6.94053531e-01 -1.42385080e-01
-5.90551078e-01 1.73133016e-01 4.29997772e-01 4.10206676e-01
-9.73319113e-02 -1.25867772e+00 -1.28089273e+00 8.78713056e-02
4.54848826e-01 1.16996638e-01 -6.22787364e-02 7.34335005e-01
-5.23271024e-01 4.11829799e-01 2.09661886e-01 -1.76928073e-01
-1.07390440e+00 4.29043442e-01 4.03681137e-02 -8.20337653e-01
-2.82352090e-01 1.08461761e+00 -1.60894930e-01 -9.56078112e-01
-2.19338283e-01 -2.66707182e-01 -1.80379912e-01 1.93036526e-01
7.87727311e-02 -5.34079075e-02 4.18399602e-01 -9.89500940e-01
-5.77731550e-01 8.83238390e-02 -2.06575081e-01 -1.30170912e-01
1.32404518e+00 -3.91601883e-02 -3.59528601e-01 5.57407200e-01
8.97970080e-01 5.91382504e-01 -3.50795627e-01 -3.98163646e-01
6.69095397e-01 -1.61066577e-02 -4.62242007e-01 -8.87665749e-01
-9.78205979e-01 4.96042669e-01 -1.25426337e-01 1.85595244e-01
7.13977337e-01 5.90175152e-01 1.09905648e+00 3.27501059e-01
5.23754716e-01 -9.97558773e-01 -7.32436180e-01 9.32359099e-01
5.61214924e-01 -1.30488372e+00 -9.24082324e-02 -5.58411777e-01
-6.42991543e-01 3.47974658e-01 6.57414079e-01 3.77300620e-01
5.60271859e-01 6.85767651e-01 3.35868150e-01 -2.65272707e-01
-9.08966780e-01 -5.32851875e-01 3.84964377e-01 6.34310007e-01
1.15824211e+00 5.37629902e-01 -5.11164427e-01 9.25920546e-01
-9.92779374e-01 -4.80151236e-01 9.35823023e-02 6.88055158e-01
1.46053091e-01 -1.45666027e+00 4.22534347e-02 7.16649532e-01
-8.80913675e-01 -7.79142380e-01 -5.80830634e-01 9.16922212e-01
3.16124797e-01 9.09866095e-01 9.88910720e-02 -3.28912944e-01
2.95751363e-01 1.47828877e-01 5.69343925e-01 -1.04754770e+00
-7.61509538e-01 -7.22629279e-02 8.84313405e-01 -2.48903841e-01
-2.72986323e-01 -8.06570947e-01 -1.03554952e+00 -3.98461998e-01
-5.01134932e-01 3.20479274e-01 3.03651184e-01 9.77174819e-01
1.96931735e-01 3.10335517e-01 9.46087465e-02 -2.26270646e-01
-1.76693246e-01 -1.28239238e+00 -5.53300679e-01 1.91550583e-01
-1.44935936e-01 -3.31629723e-01 9.09736753e-02 -1.47428932e-02] | [9.965147018432617, 9.282919883728027] |
83677a2d-c5e6-439d-bb99-c5832063f430 | adnet-leveraging-error-bias-towards-normal | 2109.05721 | null | https://arxiv.org/abs/2109.05721v2 | https://arxiv.org/pdf/2109.05721v2.pdf | ADNet: Leveraging Error-Bias Towards Normal Direction in Face Alignment | The recent progress of CNN has dramatically improved face alignment performance. However, few works have paid attention to the error-bias with respect to error distribution of facial landmarks. In this paper, we investigate the error-bias issue in face alignment, where the distributions of landmark errors tend to spread along the tangent line to landmark curves. This error-bias is not trivial since it is closely connected to the ambiguous landmark labeling task. Inspired by this observation, we seek a way to leverage the error-bias property for better convergence of CNN model. To this end, we propose anisotropic direction loss (ADL) and anisotropic attention module (AAM) for coordinate and heatmap regression, respectively. ADL imposes strong binding force in normal direction for each landmark point on facial boundaries. On the other hand, AAM is an attention module which can get anisotropic attention mask focusing on the region of point and its local edge connected by adjacent points, it has a stronger response in tangent than in normal, which means relaxed constraints in the tangent. These two methods work in a complementary manner to learn both facial structures and texture details. Finally, we integrate them into an optimized end-to-end training pipeline named ADNet. Our ADNet achieves state-of-the-art results on 300W, WFLW and COFW datasets, which demonstrates the effectiveness and robustness. | ['Fangyun Wei', 'Jongyoo Kim', 'Chong Li', 'Hao Yang', 'Yangyu Huang'] | 2021-09-13 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Huang_ADNet_Leveraging_Error-Bias_Towards_Normal_Direction_in_Face_Alignment_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Huang_ADNet_Leveraging_Error-Bias_Towards_Normal_Direction_in_Face_Alignment_ICCV_2021_paper.pdf | iccv-2021-1 | ['face-alignment'] | ['computer-vision'] | [-2.88867116e-01 2.72642933e-02 -9.55213830e-02 -7.20358372e-01
-2.31004983e-01 1.13618053e-01 4.46372509e-01 -3.63404751e-01
-3.30248922e-01 3.73576373e-01 4.59705174e-01 2.12923095e-01
-6.48421096e-03 -7.18195677e-01 -6.28167987e-01 -7.58494556e-01
2.72757828e-01 3.57744604e-01 5.94457276e-02 -3.54842484e-01
2.16669008e-01 6.44696832e-01 -9.51003909e-01 5.03902556e-03
6.97261512e-01 1.03538573e+00 -1.77976236e-01 -8.73319432e-02
-1.84655190e-01 2.21446678e-01 -1.96850687e-01 -5.11722088e-01
1.49060547e-01 -3.26374531e-01 -6.39097214e-01 -6.08559605e-03
8.40787768e-01 -3.55252236e-01 -1.96501032e-01 1.28661728e+00
8.30871165e-01 -1.01908416e-01 7.61670649e-01 -1.30324996e+00
-8.39265704e-01 1.62737638e-01 -1.22516131e+00 1.97184071e-01
-3.30793522e-02 1.24228902e-01 9.10230696e-01 -1.15063500e+00
4.59914505e-01 1.37125063e+00 8.48035514e-01 6.03021085e-01
-9.98404622e-01 -8.26275110e-01 4.74614501e-01 1.93695903e-01
-1.43156517e+00 -4.48107958e-01 1.19192135e+00 -2.91228741e-01
5.34370840e-01 2.01882925e-02 4.77901548e-01 9.77151453e-01
9.94977877e-02 6.33716643e-01 9.33570504e-01 -1.56034023e-01
-1.52699307e-01 -1.68917134e-01 -1.10540256e-01 8.03777099e-01
3.55639495e-02 -5.85237779e-02 -4.61443722e-01 1.32048681e-01
9.86917198e-01 1.25190601e-01 -4.19960648e-01 -2.47940093e-01
-8.53153050e-01 5.91090798e-01 8.64709914e-01 3.26544642e-01
-2.57741660e-01 1.01076566e-01 1.99938327e-01 5.30175641e-02
7.24694014e-01 4.58075218e-02 -5.14315426e-01 1.87312797e-01
-7.04359710e-01 2.12338120e-01 2.41321504e-01 8.64086747e-01
7.75597155e-01 -6.55576065e-02 -3.07917118e-01 1.07900345e+00
7.69447029e-01 4.37019795e-01 3.80111933e-01 -6.63469493e-01
5.72925389e-01 6.49239957e-01 -3.38532537e-01 -1.50067246e+00
-6.32958531e-01 -4.93271768e-01 -1.08629322e+00 2.68841952e-01
4.54080522e-01 7.31764287e-02 -9.31461871e-01 1.92338884e+00
6.21000826e-01 3.66785258e-01 -3.67918253e-01 1.28895938e+00
9.02997077e-01 4.60576981e-01 1.17028199e-01 -1.64882801e-02
1.48522413e+00 -1.07696736e+00 -7.40121901e-01 -1.23024791e-01
3.47651124e-01 -1.09109616e+00 1.26510656e+00 -4.51492891e-02
-9.64823127e-01 -6.24597132e-01 -8.25863242e-01 -3.12536597e-01
-7.09294826e-02 2.08401456e-01 3.94051701e-01 2.21343249e-01
-9.13160503e-01 4.47560579e-01 -8.50453436e-01 -2.47493386e-01
7.63337553e-01 3.50603640e-01 -4.80452329e-01 -1.87634137e-02
-8.83984327e-01 7.23844647e-01 -3.22194725e-01 5.97920895e-01
-3.69472861e-01 -8.10724974e-01 -7.62049198e-01 -2.93003656e-02
2.09947571e-01 -4.14177209e-01 8.35882902e-01 -9.87402260e-01
-1.63867569e+00 9.85017776e-01 -2.41187096e-01 1.41227171e-01
6.25057340e-01 -3.61153603e-01 -3.29838991e-01 -2.76016027e-01
3.22278701e-02 6.88069105e-01 7.11887240e-01 -1.07969832e+00
-2.88015187e-01 -8.47765982e-01 -3.23531598e-01 1.49900615e-01
-5.00951290e-01 6.90196967e-03 -6.06527627e-01 -9.26679432e-01
4.38853770e-01 -7.85491586e-01 2.82500181e-02 3.55059385e-01
-4.85652566e-01 -5.19692540e-01 8.76225054e-01 -5.48568726e-01
1.24151552e+00 -2.16381431e+00 1.78263679e-01 4.53950047e-01
2.01170683e-01 3.34015995e-01 -3.03185403e-01 -9.78704449e-03
-1.36533514e-01 2.22007811e-01 -2.12173581e-01 -4.22993779e-01
-7.37429876e-03 2.52018198e-02 -2.88528115e-01 6.03093445e-01
5.09439468e-01 8.90866458e-01 -6.16211593e-01 -5.98899364e-01
-1.43476769e-01 6.99901581e-01 -8.50315094e-01 1.74379885e-01
2.78635267e-02 5.53270340e-01 -6.56909466e-01 7.17051625e-01
1.22926438e+00 -1.51256509e-02 -1.96716934e-01 -6.56468034e-01
-1.45349413e-01 -1.56084690e-02 -1.08027196e+00 1.77595198e+00
-3.33255082e-01 3.25361669e-01 3.81056517e-02 -7.74246156e-01
1.12926126e+00 1.01479217e-01 2.87795961e-01 -7.81345069e-01
2.23203823e-01 2.96837687e-01 6.03459850e-02 -5.80488205e-01
-1.88218094e-02 5.91004156e-02 5.90382636e-01 3.18641007e-01
9.93844420e-02 1.73305258e-01 -3.96136284e-01 -4.24995989e-01
4.47162569e-01 3.98650229e-01 -3.75266746e-02 -4.60462362e-01
7.34327793e-01 -5.80438972e-01 8.58875275e-01 -4.85786051e-02
-2.54724950e-01 8.39798629e-01 5.90796411e-01 -7.50589132e-01
-9.95020390e-01 -6.71223342e-01 -4.33154076e-01 9.71565962e-01
2.75840700e-01 -3.22111696e-01 -9.34509039e-01 -8.00683558e-01
-9.63688865e-02 -1.29783535e-02 -8.83062899e-01 -1.71624720e-01
-1.03436267e+00 -7.67003953e-01 3.42349976e-01 6.56817317e-01
1.06137395e+00 -1.06055307e+00 -1.00866601e-01 -5.35079129e-02
5.68115823e-02 -8.88368249e-01 -1.00782609e+00 -4.32583004e-01
-7.94721782e-01 -1.00041366e+00 -9.71898437e-01 -8.94109011e-01
1.12924767e+00 -2.85529401e-02 1.01461887e+00 3.62777382e-01
-1.24857850e-01 -6.57740980e-02 -8.12712312e-02 -3.90256464e-01
4.93599117e-01 2.67274350e-01 -1.16994075e-01 3.57142806e-01
5.66237211e-01 -6.39678121e-01 -9.64510083e-01 7.54522324e-01
-6.36785388e-01 -9.55709070e-02 4.68275040e-01 8.78538370e-01
5.82067430e-01 -3.56440574e-01 1.39200121e-01 -6.20847404e-01
4.22319591e-01 -3.44589055e-01 -5.21748126e-01 2.15098917e-01
-5.46857297e-01 8.02573636e-02 4.88199383e-01 -3.18719417e-01
-9.21539128e-01 -4.66497652e-02 -5.52747190e-01 -3.10445458e-01
-4.65334691e-02 2.66643614e-01 -4.79769349e-01 -2.66264558e-01
2.46332064e-01 -8.95438865e-02 1.27769917e-01 -5.89441121e-01
7.85090178e-02 3.77454013e-01 3.93405169e-01 -7.61656046e-01
7.12392807e-01 6.05699420e-01 3.05168897e-01 -6.26476407e-01
-9.05656576e-01 -2.65400093e-02 -6.69235766e-01 -1.42455861e-01
8.78650308e-01 -5.24328470e-01 -8.01299274e-01 6.49996459e-01
-1.34227192e+00 -2.45512858e-01 1.74815550e-01 4.01204288e-01
-2.83773214e-01 2.35200703e-01 -4.29093450e-01 -4.90586936e-01
-4.07886624e-01 -1.27568042e+00 1.25512767e+00 4.83099163e-01
5.03548346e-02 -9.50557411e-01 8.54015797e-02 6.74839988e-02
5.87780118e-01 2.41958171e-01 8.53740573e-01 -4.29780662e-01
-3.87350857e-01 -7.64184520e-02 -5.29143751e-01 2.69731253e-01
1.40917571e-02 2.20282033e-01 -1.00972354e+00 -1.84037253e-01
7.44043589e-02 -1.35746717e-01 8.78684342e-01 4.09527421e-01
1.34279001e+00 -3.19836557e-01 -1.36204407e-01 1.16140234e+00
1.17252827e+00 -6.65391907e-02 8.03054869e-01 4.05617267e-01
8.98220897e-01 7.41451621e-01 5.61663747e-01 1.50167480e-01
2.96855509e-01 1.06528664e+00 4.31040734e-01 -5.01048326e-01
-3.04851741e-01 -3.15520525e-01 2.15672236e-02 6.71411932e-01
-5.95270574e-01 1.95903614e-01 -8.39006305e-01 1.20053187e-01
-1.80406272e+00 -6.89729929e-01 -5.66640571e-02 2.09486556e+00
9.14932787e-01 -5.83453067e-02 -1.56009961e-02 -1.39767393e-01
6.61052406e-01 3.39188039e-01 -4.10337627e-01 -1.30801825e-02
-1.58134997e-01 1.05818927e-01 2.30416343e-01 5.89377880e-01
-9.29262698e-01 1.07956326e+00 5.51304722e+00 1.05808401e+00
-1.44663393e+00 1.25318497e-01 1.06615245e+00 1.04257099e-01
-3.13344985e-01 -2.04850495e-01 -1.11922967e+00 7.11065829e-01
5.61481789e-02 4.66640919e-01 2.93377727e-01 7.20058262e-01
1.71135396e-01 3.32009733e-01 -9.19187367e-01 1.03441918e+00
1.03334390e-01 -1.04369330e+00 7.57802576e-02 1.46413878e-01
6.56126618e-01 -1.58631742e-01 3.46999943e-01 6.48132041e-02
-2.94202805e-01 -1.25512695e+00 5.51060915e-01 5.90416610e-01
8.31449211e-01 -7.98643708e-01 9.18814421e-01 -3.54001187e-02
-1.24553823e+00 2.01754823e-01 -7.32486725e-01 1.60877898e-01
-4.18902636e-02 5.47602296e-01 -3.79694253e-01 3.39228511e-01
7.87035227e-01 6.85549676e-01 -5.35654783e-01 8.08883607e-01
-4.08345878e-01 4.24944758e-01 -3.40746671e-01 2.22882539e-01
3.08025151e-01 -5.51206231e-01 3.00360352e-01 1.09027922e+00
3.02699298e-01 1.55357301e-01 -8.33136067e-02 1.00044942e+00
-2.11507082e-01 2.38353565e-01 -3.59516174e-01 4.75047469e-01
3.45771462e-01 1.43919671e+00 -5.59287906e-01 1.01214588e-01
-3.91271114e-01 8.47398996e-01 5.56712329e-01 5.01250684e-01
-1.00287879e+00 -3.32020432e-01 7.60300577e-01 3.47332686e-01
-2.57530771e-02 -1.21652171e-01 -3.98819685e-01 -8.79842699e-01
3.41506153e-01 -7.05063343e-01 1.34997323e-01 -5.22050381e-01
-1.40394807e+00 7.23143935e-01 -3.59632909e-01 -9.80094194e-01
4.07063544e-01 -6.61845624e-01 -9.82419848e-01 1.07710481e+00
-1.88606203e+00 -1.34730899e+00 -7.43116379e-01 7.01614380e-01
4.10278589e-01 -2.34221026e-01 5.08982539e-01 5.83894670e-01
-8.60336721e-01 9.72106040e-01 -3.57407779e-01 4.89320844e-01
9.50416505e-01 -9.28224504e-01 3.17071617e-01 5.77120066e-01
-6.86732829e-02 8.95824373e-01 2.51628250e-01 -5.45830190e-01
-9.79518235e-01 -1.14295924e+00 8.85460794e-01 -2.69576311e-01
3.83438975e-01 -1.74660042e-01 -1.07931435e+00 6.33629858e-01
1.13431461e-01 5.08759797e-01 4.32793826e-01 2.65311003e-01
-6.05241120e-01 -4.72686976e-01 -1.01782322e+00 7.22496331e-01
1.21932840e+00 -3.32818478e-01 -3.43870133e-01 2.58162796e-01
4.88266557e-01 -5.17746627e-01 -7.44206965e-01 7.75035381e-01
5.98289609e-01 -1.03544903e+00 9.83358324e-01 -7.15761960e-01
6.56084418e-01 -4.94910419e-01 1.06342852e-01 -1.17556119e+00
-4.69450235e-01 -5.34895539e-01 1.65548429e-01 1.54819119e+00
3.10488611e-01 -6.67625427e-01 7.40373790e-01 4.26865280e-01
-2.47942150e-01 -1.44276261e+00 -1.03975141e+00 -3.77749771e-01
1.18694752e-01 3.34373885e-03 1.01361728e+00 9.56568122e-01
-4.87462848e-01 1.96910277e-01 -3.10495496e-01 1.18974783e-01
3.46803367e-01 -9.69225764e-02 8.11056972e-01 -9.93342280e-01
1.82021424e-01 -9.02710497e-01 -4.30673242e-01 -1.27204657e+00
1.95955545e-01 -7.76087880e-01 -1.03566490e-01 -1.10945630e+00
5.30341454e-02 -7.47618914e-01 -2.33836874e-01 5.86157143e-01
-3.85022461e-01 6.05367899e-01 1.89317744e-02 3.01076651e-01
-2.97439694e-01 8.86136234e-01 1.49821126e+00 -3.69719677e-02
-5.85768260e-02 -1.65625647e-01 -5.12467444e-01 1.08110011e+00
7.10615516e-01 -2.43439049e-01 -9.37052369e-02 -8.88071775e-01
2.54583716e-01 -6.43782675e-01 2.73750573e-01 -7.29843557e-01
2.87030578e-01 -1.67981759e-01 6.29291236e-01 -3.78520429e-01
1.33155286e-01 -9.46456134e-01 -2.42494658e-01 1.29730254e-01
-2.02003375e-01 2.72565216e-01 -4.17382177e-03 2.54663944e-01
-3.65213603e-01 4.22385968e-02 1.05725658e+00 8.87051597e-02
-3.18630457e-01 8.25997651e-01 4.29618895e-01 1.60381153e-01
8.44450653e-01 -2.05812141e-01 -2.63173431e-01 1.34568242e-02
-4.20172483e-01 3.95965368e-01 3.45411748e-01 4.47310120e-01
5.96463442e-01 -1.71711862e+00 -8.46943557e-01 6.33545578e-01
-2.73009911e-02 3.18383485e-01 2.51009434e-01 1.27813101e+00
-6.03462577e-01 1.54913396e-01 -3.78407955e-01 -7.22562313e-01
-1.14837492e+00 1.90607280e-01 6.12917483e-01 1.32520422e-02
-4.68932450e-01 1.24518931e+00 3.52973670e-01 -4.26606953e-01
3.79050583e-01 -1.85705870e-01 -4.13929760e-01 -5.96031547e-02
5.12932539e-01 3.42452109e-01 -2.37992350e-02 -7.90288329e-01
-4.30161178e-01 1.49076390e+00 -1.13170028e-01 2.24658236e-01
1.18794107e+00 7.97437802e-02 -4.31188285e-01 1.95339392e-03
1.35228968e+00 2.27616236e-01 -1.36399758e+00 -2.26974398e-01
-1.07421070e-01 -6.75288379e-01 -3.01379245e-02 -4.69216466e-01
-1.60370314e+00 1.09086394e+00 7.29634643e-01 -2.27183372e-01
1.05694556e+00 -1.01574987e-01 7.87476182e-01 -1.81176320e-01
-9.13338736e-03 -9.99271452e-01 2.39826560e-01 5.35445631e-01
1.18623030e+00 -1.36191177e+00 3.10993791e-02 -6.63609624e-01
-4.84526813e-01 1.26082885e+00 1.10608840e+00 -1.82734191e-01
9.07514036e-01 1.74385787e-03 1.76288828e-01 -3.44424725e-01
-1.67762250e-01 -5.01101688e-02 6.07881188e-01 4.87437457e-01
7.92368770e-01 -3.88356984e-01 -4.62364554e-01 5.21755934e-01
-1.57675549e-01 -1.46198511e-01 -3.22216630e-01 4.44702864e-01
-2.51999110e-01 -9.32778418e-01 -2.99692929e-01 2.39529125e-02
-5.03348529e-01 -6.71472400e-04 -1.83986858e-01 7.37956941e-01
4.70478266e-01 4.49439168e-01 4.02896196e-01 -2.38537505e-01
4.33204710e-01 -1.56864300e-01 3.46548766e-01 -2.72755355e-01
-2.84308463e-01 3.29333246e-01 -4.14527625e-01 -7.78847396e-01
-2.81839758e-01 -3.05089355e-01 -1.24016368e+00 -3.25575948e-01
-2.36998230e-01 4.32083756e-02 5.46783388e-01 7.74283111e-01
4.35130328e-01 4.06890363e-01 7.12076008e-01 -8.05018961e-01
-4.67837423e-01 -1.17746437e+00 -3.76732200e-01 5.72599649e-01
2.19515815e-01 -8.06997001e-01 -2.28628695e-01 -3.05526525e-01] | [13.44289493560791, 0.499544233083725] |
34ca9a92-5c4f-4552-acae-7f6a7f1fb2f4 | pointgrid-a-deep-network-for-3d-shape | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Le_PointGrid_A_Deep_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Le_PointGrid_A_Deep_CVPR_2018_paper.pdf | PointGrid: A Deep Network for 3D Shape Understanding | This paper presents a new deep learning architecture called PointGrid that is designed for 3D model recognition from unorganized point clouds. The new architecture embeds the input point cloud into a 3D grid by a simple, yet effective, sampling strategy and directly learns transformations and features from their raw coordinates. The proposed method is an integration of point and grid, a hybrid model, that leverages the simplicity of grid-based approaches such as VoxelNet while avoid its information loss. PointGrid learns better global information compared with PointNet and is much simpler than PointNet++, Kd-Net, Oct-Net and O-CNN, yet provides comparable recognition accuracy. With experiments on popular shape recognition benchmarks, PointGrid demonstrates competitive performance over existing deep learning methods on both classification and segmentation. | ['Truc Le', 'Ye Duan'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['3d-part-segmentation'] | ['computer-vision'] | [-4.88473833e-01 -5.01356982e-02 -1.53412640e-01 -3.90853345e-01
-8.14873815e-01 -3.66927773e-01 6.65154159e-01 1.39351368e-01
-1.49169415e-01 3.83005857e-01 -3.87605488e-01 -3.21940988e-01
5.16972356e-02 -1.21235025e+00 -1.16735280e+00 -4.71974134e-01
-1.15925185e-01 1.11787617e+00 2.44680673e-01 8.20591599e-02
3.12098235e-01 1.34271491e+00 -1.12457812e+00 -1.79580525e-01
7.69388735e-01 1.34807909e+00 -6.67412207e-03 2.78944135e-01
-5.33708334e-01 2.56175846e-01 -2.83960551e-01 -4.04315675e-03
6.02670610e-01 4.55588043e-01 -6.96812510e-01 1.30947039e-01
1.06919837e+00 -4.68449026e-01 -2.74003029e-01 8.69115710e-01
3.95750493e-01 2.31589358e-02 7.47201085e-01 -1.17857611e+00
-7.21874595e-01 -5.50689697e-02 -5.90296805e-01 -1.57825172e-01
1.36339471e-01 1.65956140e-01 7.20730186e-01 -1.33110714e+00
6.28437638e-01 1.43795025e+00 1.25382864e+00 1.49788767e-01
-1.34651387e+00 -6.27383351e-01 -1.99219342e-02 -2.91172653e-01
-1.41168904e+00 2.35194489e-01 9.15044785e-01 -4.79078740e-01
1.39437127e+00 7.53415599e-02 1.04958773e+00 6.78651273e-01
1.46057740e-01 7.23879397e-01 8.66382241e-01 -9.39297378e-02
3.94052565e-01 -3.05866957e-01 2.21475922e-02 8.63346040e-01
1.31799147e-01 4.08863455e-01 -1.97686762e-01 -4.45272416e-01
1.63249540e+00 2.75063246e-01 -2.36551370e-02 -1.05720556e+00
-1.27242148e+00 7.67090201e-01 1.08822358e+00 6.54849857e-02
-4.29410875e-01 4.49841946e-01 1.96954116e-01 2.21053019e-01
7.50288606e-01 2.50075132e-01 -5.65244675e-01 3.43435220e-02
-1.15410066e+00 5.63648760e-01 6.27655387e-01 1.25203991e+00
1.28736067e+00 3.45742762e-01 2.19936281e-01 6.33602798e-01
4.48387921e-01 6.76170230e-01 1.69570714e-01 -8.80160749e-01
3.29798251e-01 1.05314040e+00 7.96200335e-03 -1.11018634e+00
-4.10793215e-01 -4.04481828e-01 -1.03751576e+00 7.22078681e-01
-4.45756055e-02 2.83520043e-01 -1.42023265e+00 9.89004910e-01
6.52162910e-01 8.60142946e-01 -3.32647860e-01 8.52589190e-01
1.15382242e+00 6.55617535e-01 -1.76265359e-01 6.66355252e-01
8.11803520e-01 -7.71775961e-01 8.54059458e-02 -8.42168648e-03
2.97796726e-01 -4.21612710e-01 9.90075886e-01 1.60009667e-01
-1.32377541e+00 -7.54250526e-01 -1.14786434e+00 -4.81820017e-01
-4.89233077e-01 -5.39275073e-02 7.83120930e-01 1.91615596e-01
-1.39238906e+00 9.58489358e-01 -1.24549329e+00 -3.59327495e-01
9.61355090e-01 6.40238762e-01 -4.01947349e-01 1.04609780e-01
-3.15908432e-01 8.26503038e-01 9.95213017e-02 -2.67203540e-01
-8.36540103e-01 -1.28556132e+00 -1.08054757e+00 -2.89845606e-03
-2.18161419e-01 -1.02454102e+00 1.23052263e+00 -1.05521277e-01
-1.61606991e+00 1.09140241e+00 -7.54291713e-02 -6.79798782e-01
5.29041231e-01 -2.92392224e-01 3.48396242e-01 -9.50146187e-03
2.08387837e-01 1.10531998e+00 8.13842654e-01 -1.49784040e+00
-5.33763885e-01 -6.99583590e-01 -2.53043354e-01 1.03117414e-01
4.06443834e-01 -5.54478586e-01 -3.38599086e-01 -3.64883482e-01
6.53871000e-01 -7.24655747e-01 -5.47203660e-01 4.46160048e-01
-3.50790203e-01 -6.00602329e-01 1.24238384e+00 -1.01498783e-01
1.34958059e-01 -1.93736982e+00 -2.10733533e-01 4.25698102e-01
5.77449679e-01 2.33051851e-01 -1.78905874e-01 2.83720940e-01
2.86823530e-02 2.12658212e-01 -2.45250419e-01 -4.96351808e-01
1.36095271e-01 3.83528262e-01 -5.61560333e-01 6.33646488e-01
3.23392093e-01 1.44798636e+00 -7.12279975e-01 -3.26713085e-01
7.68995821e-01 6.86780393e-01 -5.70023239e-01 6.81901500e-02
-4.31601614e-01 3.34382594e-01 -4.36267704e-01 9.79654133e-01
1.14935565e+00 -4.18366045e-01 -6.16541207e-01 -2.17598692e-01
-1.91497386e-01 3.33843708e-01 -9.89252031e-01 2.05733514e+00
-4.78054613e-01 2.99946398e-01 3.69545184e-02 -8.90752673e-01
1.51537788e+00 5.42632304e-02 7.41387844e-01 -3.65396559e-01
4.04034704e-02 3.62888366e-01 -5.92034400e-01 1.98303416e-01
3.14804912e-01 1.25826076e-01 1.48947805e-01 1.11298800e-01
2.70565122e-01 -1.06897461e+00 -5.63196957e-01 -2.07178667e-01
9.40554738e-01 3.43701243e-01 3.29513162e-01 -3.13043773e-01
1.72159195e-01 3.69784594e-01 3.30889881e-01 7.14613736e-01
1.06316000e-01 8.47542465e-01 -7.08650157e-04 -9.95810211e-01
-1.23112559e+00 -1.39279652e+00 -2.97591269e-01 2.74694532e-01
2.72597075e-01 -2.18313381e-01 -3.61654937e-01 -5.38842380e-01
8.47184181e-01 3.12959045e-01 -4.63186055e-01 3.02904129e-01
-6.93456471e-01 -5.32685891e-02 4.79632020e-01 8.85906756e-01
8.42499375e-01 -9.55814123e-01 -3.16062987e-01 2.48447508e-01
5.78755915e-01 -9.48193908e-01 -2.44195327e-01 1.81349576e-01
-1.50085878e+00 -9.71085846e-01 -3.78933042e-01 -9.47363734e-01
7.71403849e-01 2.67080665e-01 1.52763498e+00 8.38716775e-02
-1.58993796e-01 4.38692749e-01 -2.53774133e-02 -6.19064987e-01
9.25350189e-02 2.82500476e-01 -7.96080753e-02 -4.64506835e-01
7.64219940e-01 -9.68660891e-01 -4.84783471e-01 9.96561646e-02
-5.41431904e-01 -1.80580258e-01 3.94121706e-01 9.56479311e-01
1.29141676e+00 -3.10593754e-01 8.84591267e-02 -5.65804482e-01
4.42717582e-01 -2.04732582e-01 -9.75465417e-01 -3.83822531e-01
-4.04000640e-01 -3.84547472e-01 4.16156054e-01 -2.24241391e-02
-3.89841557e-01 3.06399614e-01 -5.53376019e-01 -1.16564929e+00
-5.60884237e-01 1.86163932e-01 -1.75755564e-02 -8.22666705e-01
4.62739915e-01 3.89279991e-01 1.16003700e-01 -8.38134289e-01
4.47596043e-01 3.01477879e-01 6.16859794e-01 -6.94835126e-01
1.21413708e+00 1.01535463e+00 1.18748978e-01 -9.02098656e-01
-4.30768758e-01 -5.04673123e-01 -1.30798650e+00 2.90557921e-01
6.74169302e-01 -9.58958507e-01 -7.13767648e-01 5.15521586e-01
-1.47622049e+00 -4.01061326e-01 -7.97500312e-01 2.18895182e-01
-8.56812656e-01 1.30742684e-01 -5.76801717e-01 -2.97911286e-01
-5.87087870e-01 -8.86396468e-01 1.69887269e+00 2.76881028e-02
2.20594294e-02 -1.20262182e+00 1.97635621e-01 -1.88777998e-01
4.08971280e-01 7.03127623e-01 8.94339859e-01 -6.27639353e-01
-1.24964046e+00 -3.69592994e-01 -3.72741789e-01 3.21731806e-01
1.46114737e-01 -2.08328199e-02 -6.80010676e-01 -4.46409404e-01
-2.74984948e-02 -2.82358289e-01 7.37455249e-01 5.32136023e-01
1.41817641e+00 -2.38759533e-01 -6.29593015e-01 1.27025783e+00
1.74012077e+00 -7.64681352e-03 5.46731710e-01 3.47775489e-01
9.53133881e-01 -2.23405376e-01 2.07439661e-01 2.09072307e-01
6.49551392e-01 4.26077247e-01 7.71294236e-01 -5.77232718e-01
1.59989726e-02 -6.08268976e-01 -3.42515081e-01 8.38631451e-01
-1.11584485e-01 4.29799914e-01 -1.27608287e+00 5.72725713e-01
-1.71353328e+00 -6.14655614e-01 2.82543823e-02 1.88881791e+00
4.67263430e-01 1.83724891e-02 -7.11483806e-02 -3.99028137e-02
1.66433752e-01 1.42086744e-01 -8.16319704e-01 -3.22004646e-01
-3.25744376e-02 7.74110019e-01 6.10794425e-01 3.76734376e-01
-1.27581859e+00 1.25893128e+00 7.14186430e+00 7.04165041e-01
-1.36948442e+00 8.91220570e-03 5.09320915e-01 2.19434887e-01
-1.16668507e-01 -3.72719496e-01 -8.34023178e-01 1.12530299e-01
3.17910790e-01 -1.29986092e-01 6.19109310e-02 1.43566358e+00
-1.25257755e-02 5.82327306e-01 -1.48036969e+00 1.28338194e+00
-1.86906770e-01 -1.92317486e+00 3.46613765e-01 2.87713200e-01
7.02972829e-01 9.05681849e-01 1.04468152e-01 3.06314290e-01
7.15961277e-01 -1.36990535e+00 5.39394736e-01 3.16433132e-01
9.07042086e-01 -6.00173056e-01 5.60310841e-01 4.26352829e-01
-1.21040392e+00 4.28712249e-01 -9.68394518e-01 -8.96000713e-02
-3.98171507e-02 6.02915943e-01 -1.12881660e+00 4.87974435e-01
8.73951495e-01 1.11411536e+00 -3.32604855e-01 1.40530860e+00
1.20367028e-01 3.61850500e-01 -8.35807502e-01 2.29228586e-01
6.43023610e-01 -4.34487253e-01 6.97501838e-01 1.09841597e+00
4.59481359e-01 -4.96014096e-02 6.03545249e-01 1.47999907e+00
-1.77697852e-01 -9.79299378e-03 -1.08961427e+00 5.07680297e-01
7.29436576e-01 1.18245220e+00 -5.28703451e-01 -3.30012739e-01
-5.18777549e-01 5.24877727e-01 5.84766686e-01 2.61664450e-01
-3.78246337e-01 -1.86541215e-01 1.08584774e+00 3.35783273e-01
6.64477408e-01 -7.03190267e-01 -7.96472013e-01 -8.84066224e-01
-1.04386635e-01 -4.22329515e-01 -1.13189854e-01 -7.63268292e-01
-1.62212205e+00 6.72757983e-01 -4.33449559e-02 -1.46790373e+00
-7.58216456e-02 -8.34352255e-01 -8.17316949e-01 8.48832309e-01
-1.84175241e+00 -1.49344456e+00 -6.00824893e-01 7.22200990e-01
4.70250875e-01 -1.81155294e-01 7.46188998e-01 -3.53047699e-02
2.04661146e-01 3.63959432e-01 1.56639531e-01 3.54535192e-01
-8.23052898e-02 -1.65688086e+00 1.25122094e+00 1.30333647e-01
3.13668430e-01 5.14726162e-01 -6.71507493e-02 -6.59882665e-01
-1.49224222e+00 -1.58420372e+00 4.45205778e-01 -5.77752888e-01
3.89396101e-01 -4.28559035e-01 -1.11645532e+00 8.09512615e-01
-2.17031082e-03 3.78139347e-01 2.54594117e-01 -7.95102641e-02
-3.55699450e-01 -1.63869396e-01 -1.51951814e+00 2.87638396e-01
1.20428634e+00 -4.42876399e-01 -8.35920334e-01 4.14266378e-01
8.85640323e-01 -1.03855026e+00 -1.15497458e+00 7.31966019e-01
4.51911986e-02 -8.23004007e-01 1.39616787e+00 -5.67680240e-01
8.50374401e-02 -2.52510160e-01 -1.61665156e-01 -1.61957586e+00
-7.10869074e-01 -3.45146716e-01 -1.74734741e-01 7.85420179e-01
3.85349691e-02 -8.04257333e-01 1.45850313e+00 2.53342599e-01
-5.03215730e-01 -1.04579639e+00 -1.32812440e+00 -9.64066327e-01
6.18987322e-01 -4.03510153e-01 1.40803874e+00 9.91296411e-01
-6.19565189e-01 -6.98750988e-02 3.57278138e-01 3.92598122e-01
1.02469766e+00 4.08772618e-01 1.13627219e+00 -1.63809431e+00
3.75551403e-01 -5.21292508e-01 -9.95928884e-01 -1.60297620e+00
2.24721447e-01 -1.11914885e+00 -2.81539083e-01 -1.83468246e+00
-4.92768168e-01 -8.53421926e-01 -1.02072835e-01 6.45689249e-01
5.56886613e-01 4.54809755e-01 1.55607402e-01 4.35630471e-01
-2.86781073e-01 7.25920677e-01 1.53857255e+00 -4.29842353e-01
-3.40931803e-01 3.54317948e-02 -1.29627749e-01 9.09614861e-01
6.20102048e-01 -1.06546707e-01 -4.96958308e-02 -8.67139995e-01
-2.25432217e-01 -2.07153946e-01 7.11741745e-01 -1.17906165e+00
4.26277876e-01 -1.17745614e-02 7.26458669e-01 -1.47436023e+00
5.50239086e-01 -1.13313508e+00 7.86373690e-02 1.81057021e-01
2.00027391e-01 2.93489277e-01 4.82363582e-01 4.65419620e-01
-3.63879621e-01 3.28958035e-01 7.47170091e-01 -5.44680178e-01
-5.23393631e-01 1.04269528e+00 2.88679332e-01 -3.84886920e-01
9.11649883e-01 -7.29013741e-01 -3.80637906e-02 2.03659814e-02
-8.03281546e-01 3.22900712e-01 7.78953910e-01 3.63015115e-01
1.11192858e+00 -1.83595908e+00 -6.40478194e-01 5.66778660e-01
-1.62836924e-01 1.04268408e+00 -2.56942153e-01 3.90969396e-01
-1.09362578e+00 6.65592432e-01 -1.86757103e-01 -1.42807508e+00
-8.05590868e-01 3.09570611e-01 6.85704291e-01 -1.18628897e-01
-1.17987001e+00 8.61863554e-01 1.64538249e-01 -1.20645428e+00
2.69532979e-01 -9.81850803e-01 2.71300048e-01 -5.61153412e-01
1.19089410e-02 3.25863034e-01 3.71683806e-01 -4.81167614e-01
-3.43178511e-01 1.05523956e+00 4.87985685e-02 4.20036227e-01
1.68987262e+00 3.46870989e-01 -3.13947618e-01 3.34480613e-01
1.26607156e+00 -4.70805049e-01 -1.37879944e+00 -4.70644116e-01
2.97991242e-02 -7.58802593e-01 1.92840457e-01 -5.44373333e-01
-1.12132227e+00 1.02910900e+00 4.90224332e-01 1.90766037e-01
6.12803102e-01 1.06676027e-01 1.00516641e+00 4.34954762e-01
7.94236362e-01 -5.50098658e-01 -1.04656868e-01 8.58487189e-01
1.09441733e+00 -1.21441627e+00 -8.82766545e-02 -5.25984645e-01
7.93184862e-02 1.03780496e+00 8.45697701e-01 -1.00345337e+00
1.17779982e+00 4.20441926e-01 7.77162090e-02 -6.48780763e-01
-6.00064456e-01 -1.49220368e-02 2.83303529e-01 1.05008495e+00
-5.42944521e-02 6.28748983e-02 4.81058627e-01 3.18128556e-01
-5.42354822e-01 1.67426094e-01 -1.08108059e-01 7.30382204e-01
-4.27946150e-01 -8.22818637e-01 -4.33266371e-01 6.38663948e-01
1.35387212e-01 1.02021724e-01 -4.55412358e-01 1.05798125e+00
1.02048270e-01 -2.18736529e-02 6.92111254e-01 -2.79557854e-01
5.95038235e-01 -5.37378229e-02 4.34497535e-01 -7.53522396e-01
-5.61857164e-01 -1.01883493e-01 -5.98787308e-01 -8.44033957e-01
-2.77837008e-01 -5.11654258e-01 -1.32424951e+00 -5.34943223e-01
-1.49443388e-01 -2.14583687e-02 8.28937769e-01 6.01390779e-01
7.64282703e-01 1.71401918e-01 6.08393252e-01 -1.67878604e+00
-6.16554379e-01 -7.44193733e-01 -7.07039654e-01 2.43159965e-01
4.94232416e-01 -8.27186286e-01 -1.90491393e-01 -3.19642514e-01] | [7.981328010559082, -3.651167154312134] |
6eb41f3a-dc27-416e-ba89-a6342939af80 | towards-trustworthy-explanation-on-causal | 2306.14115 | null | https://arxiv.org/abs/2306.14115v1 | https://arxiv.org/pdf/2306.14115v1.pdf | Towards Trustworthy Explanation: On Causal Rationalization | With recent advances in natural language processing, rationalization becomes an essential self-explaining diagram to disentangle the black box by selecting a subset of input texts to account for the major variation in prediction. Yet, existing association-based approaches on rationalization cannot identify true rationales when two or more snippets are highly inter-correlated and thus provide a similar contribution to prediction accuracy, so-called spuriousness. To address this limitation, we novelly leverage two causal desiderata, non-spuriousness and efficiency, into rationalization from the causal inference perspective. We formally define a series of probabilities of causation based on a newly proposed structural causal model of rationalization, with its theoretical identification established as the main component of learning necessary and sufficient rationales. The superior performance of the proposed causal rationalization is demonstrated on real-world review and medical datasets with extensive experiments compared to state-of-the-art methods. | ['Hengrui Cai', 'Yong Cai', 'Yunlong Wang', 'Tong Wu', 'Wenbo Zhang'] | 2023-06-25 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [ 5.59152305e-01 4.76743579e-01 -1.02160871e+00 -4.63385969e-01
-4.04869199e-01 -2.06508636e-01 7.03425467e-01 5.68466306e-01
-4.53503877e-02 9.05968964e-01 5.78008711e-01 -7.19324589e-01
-6.73305929e-01 -6.56603813e-01 -6.47649109e-01 -3.60415578e-01
-7.30615780e-02 4.07210857e-01 -1.70802802e-01 1.24000795e-01
6.38114452e-01 -4.07533869e-02 -1.38422191e+00 2.82001406e-01
1.35887754e+00 4.73896712e-01 -1.77027836e-01 2.74828464e-01
4.36458178e-02 1.07468462e+00 -2.02046886e-01 -5.05915940e-01
-1.31843716e-01 -6.31014705e-01 -7.29748905e-01 -4.72542882e-01
-5.21672554e-02 -2.10340619e-01 1.79710230e-04 9.54626918e-01
8.47181976e-02 -2.22581878e-01 9.61334705e-01 -1.22935164e+00
-8.39478254e-01 1.49330544e+00 -7.59150028e-01 3.59069377e-01
5.04852176e-01 -1.76742673e-01 1.78026402e+00 -7.59213686e-01
3.23787481e-01 1.38547015e+00 4.84806299e-01 4.49340075e-01
-1.21411312e+00 -8.75672758e-01 4.26731706e-01 5.67148328e-01
-1.16082335e+00 -1.50085717e-01 7.27108300e-01 -3.69760782e-01
8.31448793e-01 4.01248395e-01 5.71360946e-01 1.08432412e+00
5.08117914e-01 8.65988016e-01 1.00382888e+00 -5.10138512e-01
2.21546337e-01 -1.51824638e-01 4.44216073e-01 7.44083822e-01
9.18738604e-01 3.25534970e-01 -8.48715067e-01 -6.28462315e-01
5.52532434e-01 5.93641773e-02 -2.62048513e-01 -8.22018832e-02
-1.27672100e+00 1.12909389e+00 1.05972230e-01 -8.03556070e-02
-6.01384282e-01 1.91015258e-01 -4.29148749e-02 -1.86421603e-01
2.07803413e-01 7.43155718e-01 -7.64857173e-01 -2.01420113e-02
-5.93726039e-01 3.86409730e-01 5.42087674e-01 6.17757499e-01
4.13105428e-01 -2.32370347e-01 -3.26260537e-01 3.61815274e-01
6.12013102e-01 4.74759191e-01 4.46158022e-01 -4.08026725e-01
2.59024560e-01 9.31450486e-01 2.88918793e-01 -1.09456837e+00
-6.56391203e-01 -4.79652852e-01 -9.29858446e-01 -4.02052373e-01
3.19691300e-01 -6.72127232e-02 -7.20806241e-01 1.90722084e+00
3.70600164e-01 4.84498709e-01 5.44699803e-02 7.61569023e-01
6.33920848e-01 2.12898687e-01 6.11960948e-01 -7.79988170e-01
1.54083061e+00 -5.72567284e-01 -9.23158526e-01 -1.82503685e-01
4.28463489e-01 -7.65553057e-01 1.20791137e+00 5.25150418e-01
-6.32108212e-01 -9.34532285e-02 -1.26528585e+00 2.83475220e-01
1.72776043e-01 6.57814071e-02 1.22545147e+00 5.63827217e-01
-1.03867881e-01 7.23020852e-01 -5.50395906e-01 -5.85960150e-02
3.81676137e-01 1.89861700e-01 7.16754049e-02 2.65957117e-01
-1.74325025e+00 6.06445134e-01 3.75856489e-01 -1.62808776e-01
-6.29197478e-01 -1.25360239e+00 -5.09970427e-01 2.04091102e-01
8.52644265e-01 -1.32689989e+00 1.26248813e+00 -4.86820072e-01
-1.23414421e+00 4.98059928e-01 -3.54112387e-01 -5.39162040e-01
5.52334607e-01 -6.39673471e-01 -4.25446302e-01 -2.01853663e-01
3.91600549e-01 -5.69832372e-03 8.57748091e-01 -1.02887654e+00
-7.61904538e-01 -3.50404292e-01 -1.46436423e-01 2.29499228e-02
-1.58221766e-01 -2.90970266e-01 4.77267802e-02 -5.70019186e-01
2.54777133e-01 -6.51146650e-01 -3.28141570e-01 -8.56847107e-01
-9.77671087e-01 -5.94332278e-01 1.33355975e-01 -1.60976425e-01
1.85584211e+00 -1.50431812e+00 4.10802551e-02 2.76530713e-01
7.58957326e-01 -1.07705101e-01 3.50802004e-01 3.27885389e-01
-6.04523540e-01 6.45174146e-01 -3.33694965e-02 6.96869940e-02
-1.13540016e-01 8.23785365e-03 -7.26170182e-01 3.13864768e-01
2.10789546e-01 9.12590563e-01 -1.15562034e+00 -5.29265642e-01
8.61594081e-03 -2.02628031e-01 -7.80471981e-01 2.84542203e-01
-3.14369053e-01 1.70065179e-01 -9.53946829e-01 3.62978607e-01
3.24977875e-01 -6.95484817e-01 5.73825955e-01 -3.54999602e-02
-1.42747071e-02 7.72831380e-01 -1.06187713e+00 1.06548989e+00
1.68428896e-03 -5.94316311e-02 -8.89581680e-01 -9.12291408e-01
5.96329331e-01 4.49209929e-01 5.36069512e-01 -1.86958909e-01
8.00903663e-02 1.82136610e-01 2.90626884e-01 -5.79307437e-01
3.51191640e-01 -4.85745966e-01 -2.06655025e-01 7.04343736e-01
-4.15751725e-01 3.95947218e-01 -7.89153203e-02 3.77216518e-01
1.25016916e+00 -3.01674187e-01 1.32814837e+00 -3.01369250e-01
3.98183763e-01 2.52334446e-01 8.78090382e-01 9.59712863e-01
-1.18202046e-01 3.45594198e-01 8.99792135e-01 -3.66046876e-01
-8.23570788e-01 -1.00508571e+00 9.85784754e-02 6.94108903e-01
1.52684644e-01 -6.11502230e-01 -3.20971668e-01 -8.65480661e-01
2.50801891e-01 1.23212826e+00 -1.03211451e+00 -3.09265643e-01
-3.46363693e-01 -1.24844372e+00 5.31825364e-01 4.19961482e-01
8.48292112e-02 -6.02779031e-01 -6.49318874e-01 8.50606263e-02
-3.65753800e-01 -7.34242499e-01 -2.29506910e-01 1.97917879e-01
-8.81398141e-01 -1.66311789e+00 3.42714451e-02 2.49837250e-01
4.41425025e-01 2.88139492e-01 1.19273996e+00 1.08909741e-01
2.75381953e-02 -5.97419329e-02 -6.76949173e-02 -6.43103957e-01
-5.54051101e-01 4.79011536e-02 3.01698565e-01 -9.31289420e-02
6.03205800e-01 -6.12644613e-01 -7.53893793e-01 9.39851031e-02
-4.96210188e-01 4.52623248e-01 8.36446047e-01 9.60607588e-01
5.89235842e-01 2.14813352e-01 9.59555864e-01 -1.49174488e+00
9.97996509e-01 -6.72347784e-01 -2.57413596e-01 4.66328532e-01
-1.48241532e+00 5.32935798e-01 6.45457685e-01 -3.22718173e-01
-1.34395540e+00 -2.43410140e-01 3.00337404e-01 -8.29895884e-02
-3.69395278e-02 9.04569805e-01 -2.38885850e-01 8.99790883e-01
7.54118085e-01 -1.61828429e-01 -3.45541447e-01 -1.64340734e-01
7.54976749e-01 1.73522711e-01 3.81330848e-01 -5.57954550e-01
6.68813169e-01 3.44388306e-01 1.35715783e-01 -2.11875305e-01
-1.22547591e+00 -4.77081865e-01 -3.60763520e-01 9.01357457e-02
6.80215478e-01 -6.53806269e-01 -1.20946658e+00 -2.58597344e-01
-1.38281775e+00 2.45081976e-01 -1.12443417e-01 7.34457076e-01
-4.72258061e-01 1.45161182e-01 -6.21161759e-02 -9.55412984e-01
-3.16705465e-01 -6.45171463e-01 6.42257929e-01 1.53441489e-01
-8.68029773e-01 -9.42407131e-01 2.33544275e-01 3.27209324e-01
-2.58058667e-01 9.37048942e-02 1.81555736e+00 -9.58224952e-01
-4.63437259e-01 -1.32411793e-01 -2.80748814e-01 -3.66295874e-01
2.01328218e-01 2.28346661e-02 -7.21343338e-01 6.36291981e-01
-2.58916132e-02 7.93040022e-02 9.45643127e-01 5.56001127e-01
8.84395659e-01 -4.76770818e-01 -7.13602781e-01 3.58731090e-03
1.20289242e+00 1.84708416e-01 4.34170574e-01 8.43205452e-02
7.01090753e-01 6.35970354e-01 8.92826855e-01 8.00309181e-01
5.99029720e-01 3.66504699e-01 4.53751415e-01 8.79729763e-02
1.64631099e-01 -1.01136661e+00 -4.72239591e-02 4.26439494e-01
-2.15989217e-01 -2.06687361e-01 -9.99647260e-01 2.54489392e-01
-2.35742784e+00 -1.07991457e+00 -6.92530572e-01 2.13244963e+00
1.09822059e+00 2.53693104e-01 -3.62443887e-02 3.24034840e-01
7.88372695e-01 -9.14390907e-02 -6.92237794e-01 -5.53796031e-02
-5.49371028e-03 -1.60450295e-01 5.44474900e-01 5.45904458e-01
-9.33504462e-01 9.58901823e-01 6.93224907e+00 8.06978345e-01
-6.76975489e-01 9.76517349e-02 6.80213809e-01 2.17343211e-01
-1.00834978e+00 2.50469536e-01 -8.10126245e-01 5.46759851e-02
7.29334474e-01 -5.86341083e-01 5.36778793e-02 7.72336900e-01
8.34433317e-01 -1.69330299e-01 -1.52475381e+00 7.27061212e-01
-1.96702436e-01 -1.45835507e+00 3.87088507e-01 -1.30184337e-01
6.75933599e-01 -7.12565124e-01 4.50760201e-02 -1.06620163e-01
7.11582065e-01 -1.25622916e+00 7.77722716e-01 3.66906971e-01
4.85506564e-01 -4.27261025e-01 6.93997324e-01 3.55346441e-01
-7.08230972e-01 -2.16757298e-01 -1.92240462e-01 -4.12308484e-01
9.61310565e-02 9.78834033e-01 -1.32017350e+00 7.24964499e-01
1.55097306e-01 8.96377504e-01 -3.07247639e-01 3.83492082e-01
-1.02489781e+00 1.16198850e+00 9.18452218e-02 -2.55570859e-01
-1.98518127e-01 3.33226770e-01 5.04918575e-01 7.90482640e-01
-5.92161976e-02 5.10656297e-01 -3.68577063e-01 1.18200910e+00
1.47734936e-02 2.86804974e-01 -3.88226241e-01 -2.21597806e-01
5.76512754e-01 8.12128901e-01 -5.97378790e-01 -3.12730014e-01
-1.69712216e-01 3.17273885e-01 2.69765198e-01 2.35995486e-01
-1.02683568e+00 2.27933973e-01 5.13300955e-01 1.06342286e-01
-2.18141586e-01 1.37771934e-01 -1.19286394e+00 -1.28784323e+00
-4.30783063e-01 -7.84566164e-01 7.83957422e-01 -4.99106735e-01
-1.40917671e+00 1.36418536e-01 3.44887346e-01 -9.57100749e-01
-3.24805081e-01 -2.99298346e-01 -5.75522006e-01 5.98130763e-01
-1.34600174e+00 -9.58034933e-01 -1.99393928e-03 3.99578243e-01
4.15664196e-01 -9.62442234e-02 8.25927317e-01 6.46245107e-02
-6.93987668e-01 5.76838493e-01 -4.35465693e-01 -4.58145142e-01
8.12500954e-01 -1.35443485e+00 -4.86998586e-03 8.93373311e-01
1.72402877e-02 1.14764214e+00 1.16064310e+00 -1.03029346e+00
-1.38957560e+00 -7.56960809e-01 1.31445050e+00 -6.37310088e-01
9.18016970e-01 5.63683733e-02 -6.78005040e-01 5.17172515e-01
-1.02078542e-01 -7.70660102e-01 1.07074845e+00 9.22420800e-01
-7.12114930e-01 1.78342968e-01 -6.72834933e-01 9.01339829e-01
1.19715166e+00 8.32526535e-02 -1.14909661e+00 8.83232132e-02
7.96539366e-01 1.57443970e-01 -3.48266393e-01 6.14311039e-01
8.40666115e-01 -9.26173508e-01 9.13546622e-01 -1.02666771e+00
1.19654059e+00 -1.09730043e-01 2.00948231e-02 -1.16398013e+00
-4.05202776e-01 -8.65678489e-01 -1.13070291e-02 9.08285439e-01
7.43848026e-01 -5.54073215e-01 7.75795937e-01 6.82557523e-01
2.30848268e-01 -8.04952681e-01 -6.76061213e-01 -2.76420623e-01
-7.83792660e-02 -7.89149463e-01 8.23474228e-01 1.09921587e+00
5.32844722e-01 1.04027367e+00 -6.09273434e-01 2.76726544e-01
7.76552260e-01 4.28166837e-01 7.20976472e-01 -1.46049821e+00
-3.40432078e-01 -4.09250766e-01 1.08918481e-01 -9.76188481e-01
1.64790913e-01 -7.79654920e-01 -1.40210733e-01 -1.37738752e+00
6.70939744e-01 -4.72406894e-01 -4.32229936e-01 5.64603329e-01
-8.38567555e-01 -4.86521602e-01 -2.84049779e-01 5.43016970e-01
-4.20660913e-01 6.34739697e-01 1.08939636e+00 1.16702951e-02
-4.35382873e-01 1.45845070e-01 -1.48017168e+00 1.12355590e+00
6.34516656e-01 -8.06355357e-01 -9.71314192e-01 -1.05487676e-02
8.79874170e-01 2.98174649e-01 3.15115452e-01 -2.02899743e-02
1.17135666e-01 -8.13503981e-01 -6.11804090e-02 -5.56304753e-01
-4.60737437e-01 -4.51254696e-01 1.56532228e-01 5.24625123e-01
-8.63597929e-01 -6.54410869e-02 -1.17829151e-01 1.08676374e+00
9.17186812e-02 3.66904996e-02 3.50880086e-01 1.67448640e-01
-5.01507103e-01 -1.28981799e-01 -1.64677069e-01 9.81074348e-02
7.72176385e-01 1.97434977e-01 -5.76164424e-01 -3.51612359e-01
-4.29240942e-01 3.23326550e-02 -2.44843975e-01 5.31223893e-01
7.82313049e-01 -1.03610444e+00 -7.60825753e-01 -6.47650436e-02
1.74255222e-01 -4.59625572e-01 3.41677129e-01 1.03112674e+00
8.77706781e-02 8.63846660e-01 3.21798086e-01 -3.94036651e-01
-1.11069393e+00 7.28066325e-01 -7.81950802e-02 -8.07296455e-01
-3.97689342e-01 6.53958917e-01 7.35159576e-01 1.11061379e-01
-1.81129560e-01 -4.55728948e-01 -3.64787698e-01 -5.64114228e-02
4.21141177e-01 3.00465405e-01 -2.86221176e-01 4.38960679e-02
-6.30392551e-01 1.96364373e-01 -8.03611577e-02 -9.96599421e-02
1.13510323e+00 -7.18416050e-02 -1.39701918e-01 5.77721119e-01
3.84647548e-01 2.34543562e-01 -7.01617181e-01 -2.00261042e-01
3.12218517e-01 -4.45483863e-01 -1.32807598e-01 -1.00916600e+00
-3.29250813e-01 6.64670169e-01 -7.74834491e-03 2.16173053e-01
8.18161488e-01 1.19008079e-01 2.67529994e-01 1.98249206e-01
2.63619702e-02 -7.20587015e-01 5.96837625e-02 1.26140967e-01
9.01749253e-01 -1.25666106e+00 5.21109760e-01 -9.65192854e-01
-7.32652485e-01 9.38038468e-01 3.18367809e-01 1.28075659e-01
7.78836966e-01 2.48005632e-02 -1.95435464e-01 -4.18777436e-01
-1.19808829e+00 1.32600635e-01 4.63459522e-01 1.20135993e-01
6.61876261e-01 6.10800862e-01 -1.10543334e+00 1.54053295e+00
-3.28958720e-01 -2.79771909e-03 5.48148334e-01 2.17341125e-01
-2.91374981e-01 -9.85625267e-01 -3.14298838e-01 4.98239726e-01
-6.54972553e-01 -4.98847663e-01 -4.51878667e-01 8.14184487e-01
-2.80730743e-02 1.25318503e+00 -5.03419340e-01 -6.90259159e-01
2.55006343e-01 4.36098501e-02 1.64491281e-01 -6.51822031e-01
-3.01502377e-01 1.27280086e-01 2.24644467e-01 -4.80261236e-01
-5.87006509e-01 -7.53381431e-01 -1.54784262e+00 -4.50352550e-01
-4.18150991e-01 2.93529183e-01 1.97795913e-01 1.28543878e+00
4.30234164e-01 7.41031229e-01 3.74831349e-01 1.13139994e-01
-8.84180665e-01 -8.23830426e-01 -4.06955242e-01 2.76707053e-01
1.60615426e-02 -1.05492282e+00 -4.20848012e-01 1.35816038e-01] | [8.283187866210938, 5.6162519454956055] |
8ce66774-464e-4b82-828f-684161afcba3 | multi-resolution-location-based-training-for | 2301.06458 | null | https://arxiv.org/abs/2301.06458v1 | https://arxiv.org/pdf/2301.06458v1.pdf | Multi-resolution location-based training for multi-channel continuous speech separation | The performance of automatic speech recognition (ASR) systems severely degrades when multi-talker speech overlap occurs. In meeting environments, speech separation is typically performed to improve the robustness of ASR systems. Recently, location-based training (LBT) was proposed as a new training criterion for multi-channel talker-independent speaker separation. Assuming fixed array geometry, LBT outperforms widely-used permutation-invariant training in fully overlapped utterances and matched reverberant conditions. This paper extends LBT to conversational multi-channel speaker separation. We introduce multi-resolution LBT to estimate the complex spectrograms from low to high time and frequency resolutions. With multi-resolution LBT, convolutional kernels are assigned consistently based on speaker locations in physical space. Evaluation results show that multi-resolution LBT consistently outperforms other competitive methods on the recorded LibriCSS corpus. | ['DeLiang Wang', 'Hassan Taherian'] | 2023-01-16 | null | null | null | null | ['speech-separation', 'speaker-separation'] | ['speech', 'speech'] | [ 8.45182016e-02 -6.99341238e-01 3.50947291e-01 -4.15880620e-01
-1.66540742e+00 -7.12707758e-01 3.07161570e-01 -2.83350050e-01
-3.26181263e-01 4.66156721e-01 4.72554475e-01 -3.35533261e-01
-4.17975903e-01 1.49455026e-01 -4.16591853e-01 -1.02122891e+00
-3.20461541e-01 2.00642511e-01 -2.06313565e-01 -5.55315353e-02
-1.15318321e-01 5.62550247e-01 -1.39977086e+00 6.40956283e-01
6.64660752e-01 6.51063919e-01 5.09912848e-01 1.25749576e+00
-5.37397563e-02 3.67095083e-01 -1.15004563e+00 3.11095268e-01
3.16033103e-02 -2.68939704e-01 -4.17720705e-01 1.49113145e-02
3.86832565e-01 1.50219336e-01 -4.04484630e-01 6.59684181e-01
1.19217563e+00 3.82800490e-01 4.72036481e-01 -7.78696418e-01
-1.28039479e-01 1.05263126e+00 -5.05088568e-01 5.92926979e-01
5.58694243e-01 -3.23060751e-01 6.75199509e-01 -1.17957139e+00
-3.93826813e-01 1.44246328e+00 8.56288135e-01 2.90609062e-01
-1.25772250e+00 -8.94427598e-01 2.07569435e-01 1.80971310e-01
-1.81942785e+00 -1.18089318e+00 7.38211811e-01 -7.67856389e-02
1.25985134e+00 8.30316484e-01 7.70849735e-02 1.19241619e+00
-1.93114772e-01 7.34542847e-01 9.57096159e-01 -6.47960603e-01
1.49387136e-01 -9.56577212e-02 2.03672588e-01 1.23259142e-01
-1.05628893e-01 -1.60196245e-01 -1.10217416e+00 -3.65134805e-01
5.42690814e-01 -6.44521236e-01 -6.81030810e-01 1.97785869e-01
-1.30086207e+00 2.09757686e-01 -1.24324560e-01 6.33450329e-01
-3.21290672e-01 -1.76275611e-01 3.91698629e-01 3.94560009e-01
6.04074955e-01 5.40605545e-01 -3.87008280e-01 -2.55406410e-01
-9.90087509e-01 -1.35399476e-01 9.52506244e-01 9.31675732e-01
4.32615727e-02 6.11754417e-01 -1.95587993e-01 1.57148862e+00
1.91826701e-01 8.40799809e-01 6.29271507e-01 -5.71792543e-01
6.65738881e-01 -3.98090065e-01 1.85710847e-01 -4.74578828e-01
-5.08407116e-01 -8.66822362e-01 -7.76877582e-01 -3.67689133e-01
2.55646706e-01 -3.44838232e-01 -8.69129062e-01 1.43854201e+00
1.77273437e-01 4.40480977e-01 5.21105886e-01 1.16428328e+00
8.40900838e-01 8.75425339e-01 -5.46153724e-01 -6.14609063e-01
1.05478585e+00 -8.22490752e-01 -1.11062109e+00 -3.06545079e-01
1.83236852e-01 -1.06962800e+00 6.83909118e-01 6.52394593e-01
-9.21840250e-01 -5.90556502e-01 -1.06005383e+00 6.06433630e-01
-7.90775791e-02 4.30751920e-01 2.23601595e-01 1.32920015e+00
-1.08145893e+00 -1.34116277e-01 -7.16133475e-01 -9.43352282e-02
-2.20618635e-01 2.48280779e-01 -3.73781651e-01 -1.28747746e-01
-1.27242506e+00 4.97726262e-01 -2.49451116e-01 4.79565859e-01
-7.59276271e-01 -6.98919177e-01 -8.45327735e-01 1.72475144e-01
8.34523439e-02 -7.68219605e-02 1.53768146e+00 -6.54220998e-01
-1.89847910e+00 2.93242961e-01 -5.12990296e-01 -4.11691338e-01
7.23004863e-02 -3.83954018e-01 -1.07863998e+00 9.29420292e-02
-2.07575411e-01 -3.04862857e-01 1.14457512e+00 -1.32903337e+00
-1.45968378e-01 -1.53730720e-01 -6.24430716e-01 5.64599991e-01
-2.69686490e-01 3.30192268e-01 -2.63341933e-01 -6.54117346e-01
5.86381078e-01 -6.25891507e-01 1.39996231e-01 -9.72155154e-01
-4.75698233e-01 3.48427966e-02 7.08338380e-01 -9.96919274e-01
1.21344078e+00 -2.57560420e+00 -2.65626353e-03 2.46782959e-01
-1.91596851e-01 4.04494196e-01 -1.90027982e-01 4.56718177e-01
-4.05037463e-01 -4.29240257e-01 4.13084254e-02 -6.41880393e-01
9.41145793e-02 -1.97669938e-01 -2.71499038e-01 6.12742543e-01
-2.25283191e-01 2.67742544e-01 -4.35944229e-01 -1.77064314e-01
3.18839490e-01 7.13870525e-01 -1.84231117e-01 3.14629138e-01
6.14472210e-01 5.30407548e-01 -4.30598222e-02 5.89711964e-01
8.66988420e-01 2.23202065e-01 1.69746965e-01 -3.02072465e-01
-3.09940428e-01 6.95377052e-01 -1.31979203e+00 1.57119715e+00
-8.22427750e-01 1.00016224e+00 9.33401167e-01 -1.02144873e+00
1.00942683e+00 8.86607349e-01 2.84700990e-01 -4.18412656e-01
-4.81136218e-02 3.23461980e-01 2.73762941e-01 -4.45663631e-01
4.68066156e-01 -4.63702865e-02 -1.40088126e-01 8.39973018e-02
4.25341502e-02 -1.00865155e-01 -4.39701915e-01 -6.74238726e-02
1.13949239e+00 -6.73235416e-01 2.65943050e-01 -3.15763682e-01
6.25109613e-01 -6.90244377e-01 4.39919472e-01 8.98808956e-01
-1.25632167e-01 8.47975135e-01 -3.84349465e-01 1.28198713e-01
-5.41276455e-01 -1.32155800e+00 -2.23647416e-01 1.30108595e+00
-1.50168270e-01 -1.24420322e-01 -7.63903141e-01 -6.66496754e-02
-1.31724611e-01 8.75923753e-01 -7.28852721e-03 2.55649567e-01
-8.13678145e-01 -9.47619557e-01 9.58835125e-01 4.29732919e-01
3.52539510e-01 -5.02854168e-01 7.32378848e-03 3.84548396e-01
-5.46010792e-01 -1.44591033e+00 -8.48249793e-01 5.88906407e-01
-1.73656881e-01 -3.78149539e-01 -9.20396149e-01 -7.85654545e-01
1.05585210e-01 7.96817482e-01 5.38876057e-01 -5.75313151e-01
-3.36390972e-01 7.14357555e-01 -3.25964183e-01 -3.40074152e-01
-3.98490906e-01 -8.01075697e-02 7.73684561e-01 3.99910599e-01
2.88167953e-01 -7.10919559e-01 -2.63471127e-01 7.67500401e-01
-3.12918335e-01 -4.00792629e-01 5.39670825e-01 7.92545736e-01
9.45329964e-02 3.18740845e-01 9.80232179e-01 1.39667196e-02
9.07852888e-01 -8.87846425e-02 -3.24303716e-01 3.56198072e-01
5.75581864e-02 -1.59765765e-01 5.19335508e-01 -7.98710465e-01
-1.51121747e+00 -2.20930800e-01 -2.13202074e-01 -5.81245482e-01
-4.29834992e-01 3.31982762e-01 -5.11279106e-01 -3.20168957e-02
7.34199226e-01 6.32406712e-01 -3.07712793e-01 -5.65821111e-01
9.58724394e-02 1.46543813e+00 5.17705381e-01 -4.28267688e-01
4.48966980e-01 1.20446965e-01 -7.26861715e-01 -1.73485291e+00
-3.04126590e-01 -1.07393634e+00 -3.27550143e-01 -1.86379552e-01
6.14621639e-01 -1.16667783e+00 -6.14436209e-01 5.97294092e-01
-1.00572670e+00 -3.68958145e-01 3.25142980e-01 1.22155750e+00
-4.13990945e-01 2.49719471e-01 -5.15770137e-01 -1.54757011e+00
-3.51093262e-01 -1.11286318e+00 1.08265758e+00 -4.18047570e-02
-3.23261380e-01 -7.14709699e-01 -1.45097878e-02 5.61919570e-01
5.41064978e-01 -6.89719319e-01 2.27569357e-01 -8.14012110e-01
-2.29778603e-01 -1.06631726e-01 2.90386081e-01 4.03733432e-01
5.66833913e-01 -5.02276659e-01 -1.53789437e+00 -6.11249804e-01
1.94392458e-01 5.23708872e-02 6.74431801e-01 9.47693765e-01
1.00697148e+00 -2.41136417e-01 -5.32019496e-01 5.76975167e-01
6.18999183e-01 5.23634613e-01 2.99504161e-01 -4.57000881e-02
5.64798892e-01 5.16827822e-01 4.07533050e-01 3.64825457e-01
-5.82275279e-02 8.10322702e-01 -3.37499768e-01 -5.16284145e-02
-1.34700537e-01 3.29120457e-01 6.56364679e-01 1.36487424e+00
2.56214768e-01 -5.85546732e-01 -8.81181359e-01 5.44609129e-01
-1.49666834e+00 -1.05127835e+00 6.39834777e-02 2.31392360e+00
5.55419624e-01 -1.57044426e-01 1.38145415e-02 3.31519753e-01
9.98957157e-01 1.27146572e-01 -3.52102518e-01 -3.73519123e-01
-4.75797504e-01 3.12221739e-02 6.06280208e-01 7.51186848e-01
-1.05234087e+00 4.69157517e-01 6.67471933e+00 1.09615159e+00
-1.22501361e+00 2.85981983e-01 1.56714290e-01 -4.68808234e-01
5.95465638e-02 -8.60959530e-01 -9.03364360e-01 8.21638852e-02
1.21685934e+00 5.12503162e-02 5.62443376e-01 5.00336170e-01
4.09644544e-01 4.10936363e-02 -1.02691686e+00 1.70562637e+00
3.64108205e-01 -8.21241856e-01 -5.97386956e-01 -1.29063621e-01
3.57523412e-01 1.43636659e-01 3.71781021e-01 2.07126975e-01
1.25986159e-01 -1.09719777e+00 6.33653939e-01 5.35932072e-02
8.11954439e-01 -7.24434078e-01 4.32672799e-01 2.95709550e-01
-1.54727876e+00 -2.20612720e-01 -5.32179326e-03 2.88434535e-01
1.77741066e-01 5.26992202e-01 -1.44847596e+00 6.92584455e-01
8.25454354e-01 1.65890172e-01 1.07300416e-01 1.06763697e+00
1.73657849e-01 1.09966075e+00 -4.97260273e-01 6.20464459e-02
1.65299028e-02 2.14069888e-01 1.08677983e+00 1.78006434e+00
5.35292685e-01 2.30492905e-01 2.95805167e-02 2.90983975e-01
2.79914647e-01 -1.50625229e-01 -3.02356899e-01 9.51106548e-02
9.08981085e-01 1.05425394e+00 -4.77012187e-01 5.83164860e-03
-1.32871613e-01 1.05977309e+00 -1.14791542e-01 8.62285733e-01
-6.43221617e-01 -3.88947845e-01 1.01186788e+00 -1.58624426e-01
5.13126612e-01 -6.58459961e-01 5.65813929e-02 -9.41276312e-01
1.44348748e-03 -1.02220440e+00 1.27418205e-01 -6.83372080e-01
-1.17690933e+00 8.09527755e-01 -7.09500238e-02 -1.27288115e+00
-2.70221174e-01 -5.21994293e-01 -4.49688166e-01 1.26847601e+00
-1.35114598e+00 -8.68555546e-01 1.79551437e-01 6.02379560e-01
1.10766625e+00 -5.07263005e-01 1.23229825e+00 5.03056645e-01
-6.06589735e-01 1.12033582e+00 5.04941285e-01 1.00984201e-01
9.46567953e-01 -1.08269441e+00 4.45032895e-01 7.98849702e-01
2.42984205e-01 7.34679639e-01 8.58554780e-01 -2.36851051e-01
-1.43568695e+00 -7.75583625e-01 4.84015107e-01 -2.30080396e-01
3.96343023e-01 -8.63261580e-01 -1.16246426e+00 4.35874015e-01
2.81532675e-01 -2.63758302e-01 1.12503636e+00 3.79351109e-01
-5.08344471e-01 -4.10498261e-01 -7.96134710e-01 2.16584548e-01
6.76984906e-01 -7.99542904e-01 -5.85837305e-01 2.00365081e-01
6.63703322e-01 -6.00808620e-01 -6.58485055e-01 1.94170430e-01
6.13744855e-01 -6.28893197e-01 1.25824118e+00 -3.20849717e-02
-6.57909453e-01 -3.43361586e-01 -5.67797899e-01 -1.65252686e+00
-3.16084445e-01 -1.16104376e+00 2.10062593e-01 1.12632394e+00
5.88345349e-01 -7.36333311e-01 2.30413109e-01 1.11636810e-01
-6.05115592e-01 3.34964357e-02 -1.28411269e+00 -1.13386428e+00
-3.47278863e-01 -6.98018134e-01 5.32614946e-01 8.98404002e-01
2.09400043e-01 4.45093870e-01 -4.25980926e-01 1.04186380e+00
6.17091298e-01 -5.38174920e-02 4.97079313e-01 -8.51922333e-01
-5.19863546e-01 -2.96593636e-01 -1.19191334e-01 -1.37685466e+00
2.37799793e-01 -4.80715692e-01 5.27237236e-01 -1.29197478e+00
-3.34850162e-01 -4.35877830e-01 -4.78149861e-01 -1.91816073e-02
-5.63850850e-02 -1.34022266e-01 -3.00606102e-01 -6.87884763e-02
-3.84058714e-01 4.43611234e-01 6.96924210e-01 -2.60716796e-01
-4.87908542e-01 3.85413945e-01 -2.20473021e-01 4.59057599e-01
8.85898888e-01 -8.42816383e-02 -3.84282380e-01 -3.69034678e-01
-4.39820558e-01 4.13595349e-01 3.51919346e-02 -1.19027531e+00
3.74004573e-01 2.26851240e-01 2.95964420e-01 -8.48899662e-01
9.56565261e-01 -5.93267143e-01 1.37970909e-01 -1.93453521e-01
-5.90137899e-01 -4.15669858e-01 5.59325337e-01 5.47406554e-01
-3.04296345e-01 2.43016228e-01 6.35798752e-01 2.01754272e-01
-3.68880689e-01 -2.78189242e-01 -8.08954775e-01 -3.03939104e-01
4.10477459e-01 -2.30316445e-01 8.17539617e-02 -5.80238521e-01
-7.40357220e-01 -9.07348841e-02 -4.87580121e-01 5.53400040e-01
7.47064710e-01 -1.09667158e+00 -8.94289911e-01 3.70071441e-01
-5.26890643e-02 -1.85016945e-01 7.46544302e-01 9.35463548e-01
1.22117646e-01 7.70651221e-01 3.56723070e-01 -8.18460226e-01
-1.91098666e+00 1.37598142e-01 6.36780381e-01 2.72181571e-01
-4.35999691e-01 1.34876275e+00 4.31771159e-01 -5.41837513e-01
5.62216640e-01 -2.73866624e-01 -1.01825409e-01 -1.57086611e-01
9.46743488e-01 3.17916334e-01 5.73299468e-01 -7.84784079e-01
-6.07838511e-01 2.93155402e-01 -1.16730899e-01 -6.61335230e-01
1.03892434e+00 -5.42094290e-01 2.21998870e-01 9.08900738e-01
1.13907719e+00 4.68350977e-01 -9.06924605e-01 -3.69389236e-01
-3.18342656e-01 -3.58253330e-01 2.69424915e-01 -7.75506973e-01
-3.96725893e-01 8.32841814e-01 6.16014898e-01 4.11884874e-01
1.21081030e+00 -2.25217827e-02 3.65697592e-01 5.83644986e-01
2.91094780e-01 -9.66952085e-01 1.58699527e-01 6.08525753e-01
1.13233161e+00 -9.68918443e-01 -4.44865704e-01 -4.98876899e-01
-5.45293510e-01 1.06469369e+00 4.06275779e-01 5.34907520e-01
6.66977465e-01 7.60307133e-01 4.81480390e-01 2.56287634e-01
-4.44879562e-01 -6.72630221e-02 3.54363024e-01 8.64526272e-01
5.97063899e-01 4.01508808e-01 3.66358370e-01 7.81902194e-01
-5.07291973e-01 -8.58890653e-01 1.17098227e-01 5.69434345e-01
-5.43938994e-01 -7.51815438e-01 -1.21446931e+00 1.28583491e-01
-3.21128428e-01 -3.00315291e-01 -2.41861388e-01 2.31957182e-01
-5.07272959e-01 1.52721918e+00 1.12646498e-01 -5.14440894e-01
3.83557022e-01 3.94823790e-01 3.70180190e-01 -4.40366685e-01
-6.21087551e-01 8.87848616e-01 4.14409071e-01 -2.90727347e-01
-3.32412332e-01 -7.59009004e-01 -1.11560202e+00 9.65685956e-03
-6.19037986e-01 5.98289490e-01 1.01758361e+00 8.11887383e-01
3.74504209e-01 1.12877965e+00 7.30794609e-01 -9.46453333e-01
-5.58003128e-01 -1.34216177e+00 -8.54824901e-01 -2.01053396e-01
9.21093643e-01 -4.53544527e-01 -6.17289245e-01 -1.52622417e-01] | [14.866063117980957, 5.9532084465026855] |
a8b8c7ad-7157-4a63-9786-dcb6e6c6e7ba | temp-taxonomy-expansion-with-dynamic-margin | null | null | https://aclanthology.org/2021.emnlp-main.313 | https://aclanthology.org/2021.emnlp-main.313.pdf | TEMP: Taxonomy Expansion with Dynamic Margin Loss through Taxonomy-Paths | As an essential form of knowledge representation, taxonomies are widely used in various downstream natural language processing tasks. However, with the continuously rising of new concepts, many existing taxonomies are unable to maintain coverage by manual expansion. In this paper, we propose TEMP, a self-supervised taxonomy expansion method, which predicts the position of new concepts by ranking the generated taxonomy-paths. For the first time, TEMP employs pre-trained contextual encoders in taxonomy construction and hypernym detection problems. Experiments prove that pre-trained contextual embeddings are able to capture hypernym-hyponym relations. To learn more detailed differences between taxonomy-paths, we train the model with dynamic margin loss by a novel dynamic margin function. Extensive evaluations exhibit that TEMP outperforms prior state-of-the-art taxonomy expansion approaches by 14.3% in accuracy and 15.8% in mean reciprocal rank on three public benchmarks. | ['Xiaojie Yuan', 'Haiying Wu', 'Ning Jiang', 'Yanlong Wen', 'Hongyuan Xu', 'Zichen Liu'] | null | null | null | null | emnlp-2021-11 | ['taxonomy-expansion'] | ['natural-language-processing'] | [ 2.55224198e-01 2.32877418e-01 -4.52440500e-01 -3.42747360e-01
-1.53949112e-01 -6.89696133e-01 6.39370799e-01 7.46380210e-01
-6.67307794e-01 6.90576971e-01 4.51667964e-01 -3.27988684e-01
-2.95836002e-01 -1.10587156e+00 -3.40027183e-01 -2.15833440e-01
-1.76118910e-01 9.01978016e-01 2.31277913e-01 -4.85629261e-01
2.53934473e-01 1.61194354e-01 -1.75576830e+00 2.06077993e-01
1.01410246e+00 8.62440050e-01 -2.25436073e-02 3.62153351e-02
-5.16346991e-01 3.84961337e-01 -4.21103448e-01 -7.93681383e-01
3.32109362e-01 7.55429342e-02 -1.01648676e+00 -3.95837843e-01
3.12405050e-01 -4.99416441e-02 -4.63828623e-01 1.04092419e+00
7.41505548e-02 2.85596937e-01 5.97782075e-01 -1.18301666e+00
-7.20319927e-01 9.45031285e-01 -3.97892833e-01 2.06749991e-01
5.32361686e-01 -3.13915163e-01 1.74016869e+00 -1.05842149e+00
8.70644987e-01 1.13495767e+00 6.06492639e-01 4.02146608e-01
-1.16711164e+00 -9.26187515e-01 2.19398499e-01 5.34833133e-01
-1.49825454e+00 1.22907430e-01 4.49743986e-01 -4.38610613e-01
1.31439137e+00 8.32549855e-02 6.40264571e-01 8.51885080e-01
-7.03515336e-02 2.13904887e-01 7.33724535e-01 -3.90155703e-01
1.36075869e-01 1.79928154e-01 3.68453652e-01 6.04969621e-01
6.24512434e-01 3.93209904e-02 -5.16256332e-01 -1.96248725e-01
3.88558179e-01 -5.78895845e-02 -1.52296618e-01 -4.42913473e-01
-9.31645393e-01 8.73332202e-01 6.75003648e-01 3.30196351e-01
-1.84224442e-01 -1.33828178e-01 5.37949324e-01 3.52533519e-01
4.17860270e-01 1.08026135e+00 -4.68796670e-01 -2.79543009e-02
-5.72066247e-01 2.45406240e-01 6.19375885e-01 1.13192272e+00
6.78746283e-01 -4.96501356e-01 1.49145573e-01 8.12910676e-01
-5.19795204e-03 -1.81684509e-01 9.71376657e-01 -5.32770932e-01
3.97820741e-01 1.36278963e+00 -2.18445897e-01 -7.50682890e-01
-5.32295465e-01 -6.84069574e-01 -7.16853380e-01 -2.97091693e-01
-3.84608954e-02 3.67612928e-01 -6.97726190e-01 1.62228930e+00
5.73004305e-01 2.69327402e-01 7.87845403e-02 5.48167706e-01
8.04866195e-01 4.32627112e-01 2.07611620e-01 -8.21870565e-02
1.53353083e+00 -7.70459652e-01 -4.26523119e-01 -2.26592407e-01
7.32524931e-01 -5.43762445e-01 1.25481367e+00 2.58366138e-01
-5.66093385e-01 -3.77773911e-01 -1.09022224e+00 -1.54427022e-01
-8.38874400e-01 -2.92373657e-01 9.93648410e-01 5.83851933e-01
-4.71321881e-01 6.36653960e-01 -4.69419986e-01 -6.68951631e-01
4.18152273e-01 2.87373453e-01 -5.11348128e-01 -2.41534844e-01
-1.51249433e+00 1.01201642e+00 1.03203046e+00 -4.78125513e-01
-6.04623139e-01 -1.02798259e+00 -8.25739384e-01 4.35571522e-01
5.27029157e-01 -7.37672925e-01 1.07415438e+00 -1.44965172e-01
-1.08408058e+00 8.86630654e-01 3.69901061e-02 -7.01640904e-01
-7.55136237e-02 -3.42399150e-01 -6.46193385e-01 -8.43310431e-02
1.97195649e-01 7.13948548e-01 2.54675001e-01 -1.06461036e+00
-9.82855916e-01 -3.07971776e-01 2.95467138e-01 4.11354482e-01
-9.15003955e-01 -2.18210727e-01 -2.27517113e-01 -4.85028088e-01
2.27412030e-01 -6.26015842e-01 -2.32606500e-01 -2.68920243e-01
-2.66910702e-01 -7.55766332e-01 4.16011125e-01 -1.10712565e-01
1.62384260e+00 -1.78554869e+00 1.40354991e-01 1.68212667e-01
4.99068916e-01 3.02878827e-01 -9.35328305e-02 5.95602930e-01
-1.81589887e-01 1.55870646e-01 -4.30191547e-01 -1.82707280e-01
1.09142609e-01 4.12584543e-01 -5.92466474e-01 -1.00971058e-01
6.09108657e-02 9.31703687e-01 -1.27374005e+00 -3.82373810e-01
5.98125868e-02 1.70547798e-01 -7.55109370e-01 1.36465222e-01
-3.62819582e-01 -1.77942336e-01 1.52401090e-01 5.60479283e-01
3.32427025e-01 -1.94575489e-01 5.41359842e-01 -7.45891482e-02
2.60342598e-01 7.27912664e-01 -8.49395096e-01 1.96403193e+00
-6.40532315e-01 2.65449286e-01 -8.12401891e-01 -1.00730360e+00
9.12270188e-01 1.61475673e-01 5.33433795e-01 -4.96289819e-01
1.63873523e-01 4.43887442e-01 1.40351340e-01 -2.30877906e-01
8.17957580e-01 -2.59555787e-01 -1.48248777e-01 3.67468059e-01
2.37172082e-01 -7.13588595e-02 4.92935956e-01 4.71063554e-01
1.30899584e+00 -1.25283599e-01 6.60823762e-01 -6.76796809e-02
4.68320370e-01 8.13596100e-02 6.15779817e-01 3.71359348e-01
1.85516521e-01 1.93636149e-01 2.68592149e-01 -5.10416269e-01
-8.97924423e-01 -1.12432134e+00 -3.39008152e-01 1.26099956e+00
2.32910767e-01 -9.98589933e-01 -2.46400669e-01 -8.30170572e-01
3.03231567e-01 8.16439986e-01 -5.19724488e-01 -5.19112766e-01
-3.78793210e-01 -5.86656153e-01 6.30551040e-01 6.07724249e-01
2.06288964e-01 -1.06735563e+00 -6.52800262e-01 5.00136912e-01
-2.52170265e-01 -1.26220679e+00 -2.63873100e-01 2.72888631e-01
-8.85067999e-01 -1.20991504e+00 -1.52118832e-01 -7.45848238e-01
4.73600149e-01 1.38966098e-01 1.37507868e+00 8.43564346e-02
-4.00323361e-01 -6.54666051e-02 -7.33309031e-01 -1.87681571e-01
-1.47162333e-01 6.48138165e-01 2.62121528e-01 -4.87915456e-01
9.81896818e-01 -1.03957653e+00 -3.21260124e-01 1.11128010e-01
-9.51725364e-01 -2.08153710e-01 4.72634882e-01 1.08821642e+00
4.47774589e-01 3.40319812e-01 6.45301104e-01 -1.14840317e+00
8.98595154e-01 -6.34148479e-01 -5.82201600e-01 3.40700984e-01
-1.22807038e+00 3.23725641e-01 6.09978318e-01 -4.19992477e-01
-8.92570734e-01 -5.64952800e-03 -1.47488937e-01 -1.71324179e-01
-1.75860357e-02 8.84284377e-01 -1.35790288e-01 2.92963505e-01
9.31751549e-01 -2.08120607e-02 -5.60499609e-01 -6.04424000e-01
8.50503564e-01 7.81535506e-01 8.35790575e-01 -5.69970191e-01
9.74626124e-01 2.85336465e-01 4.80290987e-02 -2.70683706e-01
-1.18952584e+00 -9.89863873e-01 -7.50516474e-01 3.10998917e-01
4.06518012e-01 -7.59737015e-01 -5.99823475e-01 -2.64048576e-01
-1.14719832e+00 1.77849188e-01 -3.70468467e-01 2.61745989e-01
-1.66099295e-01 4.22243506e-01 -3.00049335e-01 -4.18968618e-01
-6.48598850e-01 -6.01628423e-01 9.17266846e-01 3.66521142e-02
-4.24565136e-01 -7.72420883e-01 3.09567660e-01 2.20350280e-01
1.70902431e-01 -8.11044723e-02 1.45020223e+00 -9.09143567e-01
-4.13306415e-01 -3.38468283e-01 -3.09446752e-01 2.32692808e-02
3.12567919e-01 -3.89918387e-01 -8.23154926e-01 -1.26784340e-01
-4.57105160e-01 -3.19962233e-01 1.13369799e+00 -2.70993680e-01
1.14154088e+00 -3.11201692e-01 -5.92822433e-01 7.05999374e-01
1.34008026e+00 1.14847638e-01 4.84661669e-01 5.10059834e-01
6.68302894e-01 6.72130466e-01 7.48479724e-01 4.81197268e-01
4.25099105e-01 6.26090288e-01 5.78550041e-01 5.79421580e-01
3.60061601e-02 -6.68246508e-01 -6.95930570e-02 7.31511950e-01
2.41029024e-01 -2.12933615e-01 -1.07072878e+00 8.00383389e-01
-1.78266823e+00 -8.18826914e-01 2.20431805e-01 2.27698469e+00
1.28628445e+00 3.78801316e-01 -5.43098412e-02 3.60118508e-01
4.33230817e-01 -2.70108998e-01 -5.82196951e-01 -2.87337512e-01
2.29538120e-02 5.50859034e-01 3.84390295e-01 4.26940143e-01
-1.04042888e+00 1.52758968e+00 5.47260189e+00 6.36008859e-01
-5.04631579e-01 6.79784268e-02 -1.62875950e-01 -1.29907904e-02
-4.95428920e-01 2.65337378e-01 -8.23294461e-01 2.06940889e-01
6.19472861e-01 -6.03593051e-01 2.29982033e-01 1.04857886e+00
-7.25249529e-01 2.13728949e-01 -1.26111984e+00 9.66007769e-01
-9.29196551e-02 -1.25583184e+00 2.84507573e-01 1.06593125e-01
6.37335181e-01 -7.25797042e-02 -1.60222203e-01 7.15854526e-01
6.57071769e-01 -1.11054134e+00 1.32914200e-01 1.09160384e-02
1.13758814e+00 -7.94654429e-01 8.48950624e-01 5.94586395e-02
-1.52645862e+00 -3.58438790e-01 -7.89891958e-01 -2.23238111e-01
1.58775896e-01 8.01863968e-01 -1.18377912e+00 7.35910058e-01
4.74864274e-01 7.77275145e-01 -5.90303957e-01 1.18609476e+00
-6.67891204e-01 5.07765710e-01 -3.70216608e-01 5.25168441e-02
9.32373777e-02 1.08945303e-01 4.82923925e-01 1.20006418e+00
1.49681941e-01 1.55962467e-01 8.64420757e-02 6.81769967e-01
-4.94744837e-01 2.62782484e-01 -5.27818024e-01 -1.89152837e-01
1.01285923e+00 1.29253662e+00 -4.21489358e-01 -3.47905546e-01
-1.57567069e-01 8.29839826e-01 5.71218729e-01 -1.85841359e-02
-6.17811263e-01 -6.42399311e-01 1.06898332e+00 4.55195121e-02
-1.08543612e-01 -4.80762357e-03 -3.90891671e-01 -1.27208233e+00
-5.24172839e-03 -6.53259993e-01 7.42363572e-01 -5.02562463e-01
-1.57203150e+00 7.10334420e-01 1.84176132e-01 -1.06399786e+00
-3.01144421e-01 -6.40434861e-01 -2.68744111e-01 5.17560303e-01
-1.42225480e+00 -1.10179603e+00 -3.75306159e-01 1.61017910e-01
4.06257540e-01 -4.29210156e-01 1.20437217e+00 3.94496739e-01
-3.12289566e-01 8.30018580e-01 3.10034258e-03 6.37367517e-02
6.76965773e-01 -1.43824553e+00 7.86778748e-01 7.11913586e-01
5.22064686e-01 1.01557779e+00 5.80667317e-01 -6.87138200e-01
-7.25775957e-01 -1.03463650e+00 1.26857698e+00 -3.74114990e-01
8.38910341e-01 -5.21371961e-01 -1.15720868e+00 5.66217899e-01
4.93927710e-02 -3.83603036e-01 8.54413092e-01 7.90051341e-01
-1.23475218e+00 -1.93611905e-01 -9.58461821e-01 5.58505654e-01
1.65510964e+00 -4.11679417e-01 -1.09477723e+00 2.67920613e-01
1.29174006e+00 -1.02074675e-01 -1.01877582e+00 5.89062512e-01
7.02956021e-01 -5.31109452e-01 1.05035496e+00 -8.80818129e-01
5.83602905e-01 -1.18023820e-01 -1.48485228e-01 -1.46890950e+00
-4.44260150e-01 -4.26589787e-01 -3.67242575e-01 1.13253164e+00
3.74729693e-01 -6.52842343e-01 1.00528753e+00 1.25661090e-01
-1.29395828e-01 -8.96935344e-01 -1.00226212e+00 -1.08475387e+00
3.36381674e-01 -3.05641443e-01 1.00799716e+00 1.10017729e+00
7.52166748e-01 7.73985147e-01 1.12658247e-01 -2.45920066e-02
4.80129898e-01 2.21795872e-01 4.79927480e-01 -1.84459329e+00
-7.23832324e-02 -7.83764720e-01 -7.54519880e-01 -9.31294441e-01
3.92443419e-01 -1.21360672e+00 -2.52932101e-01 -1.69458342e+00
2.65512615e-01 -5.62266290e-01 -5.72632074e-01 6.83326781e-01
-3.25389653e-01 -5.96182123e-02 -1.13652959e-01 1.28266707e-01
-4.53705490e-01 8.05255830e-01 6.45563364e-01 -2.18973204e-01
-9.48470309e-02 -3.51902574e-01 -8.32709372e-01 4.60694909e-01
8.53231370e-01 -6.97303832e-01 -7.08350301e-01 -3.61573756e-01
6.68626666e-01 -6.86151087e-01 -2.61152275e-02 -1.02066541e+00
3.64592105e-01 -8.10461640e-02 -2.28292808e-01 -4.89598840e-01
3.37770939e-01 -8.64646196e-01 -1.84510842e-01 2.85365343e-01
-4.61924970e-01 6.29183128e-02 1.30559027e-01 5.95898330e-01
-2.57871926e-01 -3.32287401e-01 4.75865483e-01 2.42867600e-02
-8.60629320e-01 2.97587365e-01 -4.86407056e-02 2.53219157e-01
7.07498193e-01 -1.19244173e-01 -2.92013675e-01 5.94855398e-02
-3.35915357e-01 3.82683784e-01 2.68281341e-01 7.41676390e-01
7.27401912e-01 -1.34231734e+00 -4.51173693e-01 -3.57603505e-02
8.28816712e-01 1.59887359e-01 -1.63295239e-01 1.98168799e-01
-1.46602035e-01 6.82889640e-01 -1.71611577e-01 -8.32628980e-02
-1.23011577e+00 7.29974866e-01 2.12837502e-01 -6.10747159e-01
-5.84551632e-01 1.10283077e+00 8.76757950e-02 -8.34973752e-01
5.02354681e-01 -4.44730580e-01 -3.30793470e-01 1.32351846e-01
4.80026215e-01 1.50721595e-01 2.59298146e-01 -3.06114495e-01
-3.56109738e-01 3.16616297e-01 -3.69527578e-01 -6.86971024e-02
1.26639116e+00 1.39770180e-01 -2.39618137e-01 2.42789283e-01
1.06114161e+00 -3.06698918e-01 -4.95159805e-01 -6.96325004e-01
6.34623349e-01 -6.01845384e-01 -2.71740079e-01 -9.75529730e-01
-5.99590957e-01 8.74399662e-01 3.80633771e-01 8.83618090e-03
1.17850423e+00 -2.46838648e-02 9.06633437e-01 6.32686555e-01
6.12533987e-01 -7.75521398e-01 1.12445213e-01 7.53491640e-01
6.73887432e-01 -8.96205127e-01 1.00075632e-01 -6.78919375e-01
-3.92487735e-01 8.49348843e-01 9.13791299e-01 2.16956377e-01
5.94220698e-01 6.57142885e-03 -2.82719165e-01 -3.02782476e-01
-9.57867682e-01 -4.94708836e-01 3.21619928e-01 7.27740765e-01
6.00173712e-01 9.36829150e-02 -5.37948012e-01 6.98370874e-01
-7.51267135e-01 -1.21362053e-01 3.33390206e-01 5.77475011e-01
-5.99650323e-01 -1.54718101e+00 1.71307370e-01 6.28351748e-01
-1.42300390e-02 -5.24025679e-01 -5.18464267e-01 4.89344686e-01
4.29876029e-01 8.15196514e-01 2.22422276e-02 -6.54326379e-01
3.55708390e-01 3.99941981e-01 2.93619663e-01 -1.18086123e+00
-5.01312137e-01 -6.12419128e-01 4.69124131e-02 -2.41019264e-01
2.71646772e-02 -4.06311691e-01 -1.41100633e+00 -2.36902222e-01
-5.29794216e-01 2.65187413e-01 3.57232690e-01 8.05654049e-01
4.48264569e-01 5.01291335e-01 1.85224697e-01 -5.25076278e-02
-7.74964392e-01 -1.21914804e+00 -4.06286925e-01 6.66817069e-01
-3.23988259e-01 -1.05235195e+00 -1.45862028e-01 -2.76683390e-01] | [9.250040054321289, 8.062782287597656] |
0c845c82-588c-452c-8bce-e3911a99ebae | beyond-convolutions-a-novel-deep-learning | 2102.13631 | null | https://arxiv.org/abs/2102.13631v1 | https://arxiv.org/pdf/2102.13631v1.pdf | Beyond Convolutions: A Novel Deep Learning Approach for Raw Seismic Data Ingestion | Traditional seismic processing workflows (SPW) are expensive, requiring over a year of human and computational effort. Deep learning (DL) based data-driven seismic workflows (DSPW) hold the potential to reduce these timelines to a few minutes. Raw seismic data (terabytes) and required subsurface prediction (gigabytes) are enormous. This large-scale, spatially irregular time-series data poses seismic data ingestion (SDI) as an unconventional yet fundamental problem in DSPW. Current DL research is limited to small-scale simplified synthetic datasets as they treat seismic data like images and process them with convolution networks. Real seismic data, however, is at least 5D. Applying 5D convolutions to this scale is computationally prohibitive. Moreover, raw seismic data is highly unstructured and hence inherently non-image like. We propose a fundamental shift to move away from convolutions and introduce SESDI: Set Embedding based SDI approach. SESDI first breaks down the mammoth task of large-scale prediction into an efficient compact auxiliary task. SESDI gracefully incorporates irregularities in data with its novel model architecture. We believe SESDI is the first successful demonstration of end-to-end learning on real seismic data. SESDI achieves SSIM of over 0.8 on velocity inversion task on real proprietary data from the Gulf of Mexico and outperforms the state-of-the-art U-Net model on synthetic datasets. | ['Anshumali Shrivastava', 'Antoine Vial-Aussavy', 'Anu Chandran', 'Menal Gupta', 'Aditya Desai', 'Zhaozhuo Xu'] | 2021-02-26 | null | null | null | null | ['irregular-time-series'] | ['time-series'] | [ 1.60633013e-01 7.81979263e-02 6.54911518e-01 -2.39837706e-01
-9.76681232e-01 -2.72769868e-01 6.73638880e-01 8.80907774e-02
-8.36709082e-01 3.50704074e-01 6.36354089e-01 -4.81814295e-01
-3.05781841e-01 -1.22111738e+00 -9.57765222e-01 -7.78279185e-01
-9.18584943e-01 7.53236353e-01 6.71821117e-01 -5.91823161e-01
2.12306231e-01 7.00433731e-01 -1.30008149e+00 4.74850178e-01
1.78998619e-01 1.07277381e+00 1.77939892e-01 1.01528788e+00
-2.06996482e-02 8.14346313e-01 -3.09682459e-01 8.97909254e-02
5.55898905e-01 4.23601478e-01 -8.43904495e-01 -4.40260381e-01
4.71674472e-01 -5.06634951e-01 -7.22743690e-01 4.34835047e-01
1.00605023e+00 4.04469728e-01 3.85358214e-01 -1.10760760e+00
-3.74347448e-01 6.87823176e-01 -5.99661529e-01 5.22186220e-01
-5.81635498e-02 4.84968513e-01 6.73162282e-01 -1.28258884e+00
5.51381350e-01 1.20889032e+00 1.41852391e+00 3.04979771e-01
-1.32424176e+00 -5.22113204e-01 -4.31816310e-01 -7.53836557e-02
-1.15966833e+00 -4.41568047e-01 7.37205744e-01 -7.62906790e-01
1.54847682e+00 1.20907873e-01 6.78372920e-01 1.04631460e+00
2.57573366e-01 6.94454789e-01 5.04939497e-01 1.01564221e-01
4.35662150e-01 -9.09027517e-01 4.32721302e-02 4.10542518e-01
3.61384660e-01 2.56419599e-01 -1.04902923e+00 -3.52649689e-02
8.66543174e-01 3.85144465e-02 -1.72332183e-01 2.41098896e-01
-1.68330383e+00 8.07947636e-01 3.76864374e-01 -2.10089207e-01
-5.80485523e-01 7.02331305e-01 9.11805749e-01 5.07760882e-01
6.30056679e-01 4.28228766e-01 -5.34617066e-01 -5.08536339e-01
-1.37948740e+00 5.49854815e-01 7.49247789e-01 5.72474360e-01
6.79419398e-01 8.58128369e-01 3.46749276e-01 4.47700918e-01
1.44699886e-01 6.80188537e-01 3.85233432e-01 -9.61196661e-01
8.09894741e-01 7.55084306e-02 -7.27060437e-02 -1.14175320e+00
-8.04505825e-01 3.37369144e-02 -1.30624342e+00 5.28746784e-01
4.22411889e-01 -4.88629639e-01 -9.45429742e-01 1.12920749e+00
1.66180685e-01 5.21682858e-01 2.87564844e-01 8.66483271e-01
8.84989917e-01 1.04180491e+00 8.53230953e-02 5.74769616e-01
1.07976091e+00 -7.15247929e-01 -2.80084908e-01 -3.80593866e-01
7.36265302e-01 -5.47740161e-01 1.13322675e+00 2.53604203e-01
-1.08406818e+00 -3.12224180e-01 -9.63330925e-01 -4.39055301e-02
-2.84001857e-01 -4.69033033e-01 8.83440852e-01 1.75196618e-01
-1.26099110e+00 8.73441279e-01 -1.44419634e+00 1.24285191e-01
6.68157399e-01 3.72759193e-01 -6.31610930e-01 -4.90625314e-02
-1.38750899e+00 5.79020798e-01 4.17252362e-01 3.91904503e-01
-1.13847721e+00 -1.40190184e+00 -1.17517710e+00 2.42569342e-01
1.71193630e-01 -3.67274612e-01 1.09591806e+00 -1.29264504e-01
-1.03084600e+00 3.34297866e-01 3.36564720e-01 -8.24806631e-01
6.47501826e-01 -7.53252506e-02 -2.83016086e-01 2.26884767e-01
8.89000073e-02 4.64769244e-01 5.66352427e-01 -1.02050078e+00
-4.77387428e-01 -3.49013321e-02 -3.00853640e-01 -2.10501418e-01
-3.19899946e-01 -2.10761558e-02 -2.72505246e-02 -8.52191687e-01
2.63952136e-01 -5.35992146e-01 -5.16816139e-01 1.92678168e-01
-9.86703932e-02 4.06053010e-03 9.91340876e-01 -1.00067687e+00
9.11921680e-01 -2.13718820e+00 -2.07515761e-01 1.19946390e-01
3.92792791e-01 2.79707640e-01 -2.62174070e-01 7.82966733e-01
-1.89576849e-01 -2.21817084e-02 -7.36484766e-01 -3.61978889e-01
5.65805025e-02 4.31578875e-01 -4.99236435e-01 3.35986346e-01
4.05911028e-01 7.72585630e-01 -9.54566121e-01 -3.05083454e-01
2.26494055e-02 6.21770442e-01 -7.71294296e-01 2.18277082e-01
-1.12812351e-02 1.92436069e-01 -1.90833569e-01 8.31869066e-01
9.82975900e-01 2.01066732e-02 -3.60511839e-01 -4.29713368e-01
-5.59267521e-01 -5.00822924e-02 -1.37619603e+00 1.62797523e+00
-4.71404761e-01 8.52787077e-01 -1.20527022e-01 -1.21364331e+00
7.96990395e-01 2.79967904e-01 6.72248721e-01 -7.06200182e-01
-3.61878693e-01 4.77360606e-01 -4.07872051e-02 -8.17285120e-01
6.62679791e-01 -2.14017600e-01 -1.83395311e-01 5.39919913e-01
6.38659671e-02 -2.94554412e-01 1.95316970e-01 1.96130514e-01
1.46807146e+00 1.93641886e-01 -3.77834648e-01 -5.60252547e-01
-2.10103571e-01 3.87414008e-01 3.74723405e-01 6.89478755e-01
2.29737133e-01 9.78916347e-01 5.33995211e-01 -1.21164155e+00
-1.61194026e+00 -1.03364384e+00 9.63831097e-02 9.38309371e-01
-4.75736529e-01 5.05829677e-02 -2.96884567e-01 -1.41469715e-02
2.51810491e-01 3.08830798e-01 -6.98368192e-01 1.33258343e-01
-1.12449932e+00 -1.16032875e+00 1.20669734e+00 9.88283992e-01
4.54909533e-01 -1.11329794e+00 -1.03206456e+00 7.15947092e-01
9.87411365e-02 -8.64960849e-01 -2.14755431e-01 5.07903576e-01
-9.25403655e-01 -6.59873247e-01 -6.57469749e-01 -6.14809930e-01
3.43438774e-01 1.12213902e-01 1.23551178e+00 -2.96732962e-01
-3.60549033e-01 -8.34917873e-02 -2.02206492e-01 -4.47268337e-01
-5.78188486e-02 -1.47901058e-01 6.40094057e-02 -1.97635099e-01
4.17441651e-02 -8.48673582e-01 -7.48249352e-01 -1.74170420e-01
-1.42847931e+00 1.67026415e-01 3.69192243e-01 1.02420533e+00
4.77888405e-01 -2.22777277e-02 5.39473355e-01 -6.46569490e-01
4.80606347e-01 -9.28646147e-01 -5.97601593e-01 -1.77358150e-01
-1.33751124e-01 1.62884519e-01 6.50743902e-01 -3.27553421e-01
-1.19965005e+00 -3.77220823e-03 -6.14046454e-01 -3.80090475e-01
7.12275207e-02 1.18472123e+00 5.07642031e-01 -7.29037374e-02
8.84954154e-01 2.41135642e-01 8.90954360e-02 -5.07599533e-01
-1.87965706e-01 5.66544116e-01 8.74614835e-01 -6.18219972e-01
8.70787859e-01 9.05565858e-01 3.07340950e-01 -1.16256189e+00
-5.77003062e-01 -3.34316373e-01 -6.59509838e-01 -2.82315642e-01
7.29924321e-01 -1.05810368e+00 -7.12404370e-01 7.39378512e-01
-1.15974641e+00 -9.71866131e-01 -4.69402045e-01 4.23198849e-01
-3.50017995e-01 3.37778479e-01 -8.07333410e-01 -5.58656335e-01
-6.74899697e-01 -8.86704564e-01 1.28081703e+00 -2.70783007e-01
-6.47421554e-02 -1.05196512e+00 9.18182656e-02 -1.09858789e-01
8.46006393e-01 7.48610854e-01 5.75891554e-01 -4.85980302e-01
-5.91262817e-01 -4.12335306e-01 -3.73724014e-01 -7.97834713e-04
-3.16656768e-01 1.38496682e-02 -9.47957158e-01 -1.56777188e-01
-4.57963310e-02 -4.51162368e-01 1.18456221e+00 3.69954109e-01
1.19892204e+00 -8.20095763e-02 1.93916976e-01 1.00998306e+00
1.58932459e+00 -5.00181429e-02 7.39830852e-01 6.27772510e-01
9.51024950e-01 5.22334337e-01 5.01511157e-01 8.49387944e-01
2.89965421e-01 -1.12606334e-02 6.02284133e-01 -3.72433037e-01
-1.65290669e-01 5.64064384e-02 1.54660851e-01 8.98150921e-01
-5.13114989e-01 -1.55215949e-01 -1.88742340e+00 9.11503017e-01
-1.84421158e+00 -1.26332808e+00 -4.85003650e-01 1.88886750e+00
6.63006306e-01 2.29671001e-01 -3.60498756e-01 3.83573562e-01
-6.20145798e-02 4.46476966e-01 -2.41384640e-01 -1.24892101e-01
-2.62956083e-01 4.30585653e-01 9.57928777e-01 3.89431089e-01
-1.26413274e+00 8.01089048e-01 5.81628513e+00 6.41584218e-01
-1.38693047e+00 2.18017682e-01 3.95928442e-01 4.33974192e-02
-4.17640954e-01 -4.87698555e-01 -5.86838663e-01 4.29108769e-01
1.43538153e+00 1.10493889e-02 1.24136835e-01 7.19247103e-01
5.08081079e-01 -8.39372501e-02 -9.10515547e-01 8.14032972e-01
-2.34341472e-01 -2.22582388e+00 5.40051237e-02 -8.32413584e-02
6.34696007e-01 7.82732368e-01 -2.25807920e-01 9.86735895e-02
5.31041682e-01 -1.07844162e+00 7.44276524e-01 5.92540205e-01
9.26312923e-01 -8.03179383e-01 9.83564198e-01 1.15852386e-01
-1.29876077e+00 -5.56673743e-02 -5.41372240e-01 -2.04607293e-01
6.13006473e-01 6.15384758e-01 -6.95145607e-01 3.75018895e-01
1.10075331e+00 7.28906214e-01 -1.80757672e-01 9.69173253e-01
4.67354178e-01 8.81869674e-01 -7.44484186e-01 4.33718830e-01
9.59673762e-01 1.89127147e-01 3.16696912e-01 1.62881362e+00
5.56284130e-01 1.57190442e-01 4.69579957e-02 2.38442585e-01
7.17692450e-02 -4.55897599e-01 -8.48762035e-01 7.39603937e-02
2.94577718e-01 7.60447741e-01 -7.05115736e-01 -4.13546622e-01
-5.75277507e-01 6.07538044e-01 6.30561784e-02 2.72139102e-01
-5.05818069e-01 -4.25612301e-01 8.48110974e-01 2.60543555e-01
3.83819461e-01 -7.21503079e-01 -4.44404811e-01 -8.75373423e-01
-1.27059385e-01 -4.81832385e-01 3.42985183e-01 -5.79824448e-01
-1.28975391e+00 6.46600068e-01 -6.71118274e-02 -1.56894398e+00
-3.90766934e-02 -6.84467554e-01 -7.20550716e-01 8.36013734e-01
-1.79743075e+00 -1.32915008e+00 -6.03939295e-01 4.76438493e-01
8.33801389e-01 -1.71022251e-01 7.76671290e-01 7.11645782e-01
-1.72149390e-01 1.57789558e-01 5.06142020e-01 4.68866110e-01
3.38142723e-01 -1.09469962e+00 1.25589895e+00 9.42625701e-01
-4.68828708e-01 1.11792825e-01 8.67800176e-01 -7.73866236e-01
-1.82441020e+00 -1.47095907e+00 8.17732215e-01 -2.09610015e-01
1.29657531e+00 -2.35799372e-01 -1.37469029e+00 7.23336637e-01
-5.99696487e-02 5.45746684e-01 5.52527547e-01 -5.44045746e-01
-2.01472014e-01 -8.51714388e-02 -9.45965707e-01 4.98307467e-01
9.73569274e-01 -2.82674313e-01 -4.45628792e-01 5.09661615e-01
8.77139986e-01 -4.59362090e-01 -1.27652109e+00 3.94578010e-01
4.57690448e-01 -6.50041997e-01 1.37984753e+00 -6.36070371e-01
9.88062620e-01 -2.93264657e-01 -2.78939247e-01 -1.31669950e+00
-8.51079226e-02 -5.51124334e-01 -8.73707682e-02 5.55038273e-01
3.02774340e-01 -5.10235727e-01 9.37708199e-01 5.81436276e-01
-8.65985513e-01 -6.24741971e-01 -1.05845416e+00 -8.45248580e-01
1.27709016e-01 -9.21996951e-01 6.76364243e-01 1.03552723e+00
-2.78027058e-01 -2.41885662e-01 -5.05044699e-01 5.19948423e-01
1.00700676e+00 -2.12706909e-01 6.04549646e-01 -1.30249667e+00
-9.13614482e-02 -8.52191001e-02 -5.40568411e-01 -5.78508675e-01
-2.65507281e-01 -5.63616872e-01 1.07126042e-01 -1.49591506e+00
-5.38429022e-01 -4.94389147e-01 9.07213762e-02 5.85380316e-01
2.95490324e-01 4.93484914e-01 -1.58885583e-01 2.22747535e-01
1.59090936e-01 5.55315197e-01 1.03121161e+00 -2.88879037e-01
1.09428786e-01 -2.67113507e-01 -2.25876570e-02 6.87173486e-01
8.65951478e-01 -5.19755125e-01 -2.67933726e-01 -1.28915262e+00
5.66761851e-01 3.34661812e-01 4.16369140e-01 -1.28366446e+00
5.04464924e-01 -2.11840466e-01 7.92333037e-02 -8.43007863e-01
2.72625387e-01 -7.51163542e-01 2.87452877e-01 6.29821479e-01
-5.85222952e-02 3.21672410e-01 4.18389052e-01 4.52923805e-01
-4.95453924e-01 -2.75801364e-02 6.38914049e-01 -2.91371673e-01
-1.19281566e+00 5.85480154e-01 -3.99319947e-01 2.89400220e-02
7.50847995e-01 -4.01789457e-01 -1.46675169e-01 8.90168175e-03
-7.36970603e-01 2.82195896e-01 -7.69615695e-02 -4.26836461e-02
9.49534714e-01 -9.98335719e-01 -1.06355166e+00 4.29888546e-01
-1.35916203e-01 8.86167347e-01 7.46864557e-01 6.93102062e-01
-1.57523310e+00 -3.60765569e-02 -3.86411250e-01 -4.63915169e-01
-8.39190125e-01 2.96902619e-02 1.01088054e-01 -1.77682102e-01
-1.26546729e+00 1.27659333e+00 -1.53966144e-01 -5.50875604e-01
-6.33358732e-02 -4.94583547e-01 4.56439555e-02 3.90353084e-01
7.85636365e-01 5.46724796e-01 3.57007474e-01 -3.13372344e-01
2.17947760e-03 2.72879988e-01 3.95931989e-01 -1.20432191e-01
2.42127728e+00 4.31867063e-01 -2.64467876e-02 3.28360349e-01
1.22896504e+00 -5.37310958e-01 -1.76317143e+00 -2.59614497e-01
1.89701170e-01 -4.85016704e-01 1.87648296e-01 -1.86024055e-01
-1.14137316e+00 1.34395790e+00 3.05246592e-01 6.24547414e-02
8.92769814e-01 -2.73828834e-01 1.21171153e+00 8.60130429e-01
2.88732678e-01 -1.08314323e+00 1.83116019e-01 9.58635330e-01
1.14448369e+00 -1.31573153e+00 -8.83411914e-02 -3.48228924e-02
-6.26944959e-01 1.41800809e+00 3.80092293e-01 -3.19892615e-01
9.43607271e-01 1.10650051e+00 3.54115590e-02 -6.16376817e-01
-6.64814532e-01 2.51732796e-01 -4.52586383e-01 6.60092950e-01
1.88145339e-01 -6.16356060e-02 4.00422513e-01 2.43465587e-01
-2.97981977e-01 7.12540075e-02 7.58866847e-01 1.37537098e+00
-5.60894251e-01 -4.87024754e-01 -5.27137637e-01 6.75494909e-01
-2.59151340e-01 -4.21685785e-01 8.33956659e-01 7.11149335e-01
-1.98560521e-01 1.30497709e-01 5.17595708e-01 -3.01356047e-01
3.91384482e-01 -1.15886398e-01 -3.30625117e-01 -4.41674471e-01
-4.89876628e-01 -1.23311371e-01 1.32564127e-01 -5.35276830e-01
-2.10325614e-01 -8.67568374e-01 -1.52027297e+00 -7.80252755e-01
4.37377244e-01 -2.61574626e-01 6.92300856e-01 6.88338995e-01
1.97112232e-01 6.82640016e-01 3.31076741e-01 -1.82492375e+00
-5.54380834e-01 -1.01543212e+00 -5.45534849e-01 2.59180725e-01
6.42620265e-01 -4.71637189e-01 -2.99737632e-01 4.47023213e-01] | [6.855991363525391, 2.5121665000915527] |
5779444a-8d0c-4ee7-9ec7-8262f82affe9 | meta-reinforced-synthetic-data-for-one-shot-1 | 1911.07164 | null | https://arxiv.org/abs/1911.07164v1 | https://arxiv.org/pdf/1911.07164v1.pdf | Meta-Reinforced Synthetic Data for One-Shot Fine-Grained Visual Recognition | One-shot fine-grained visual recognition often suffers from the problem of training data scarcity for new fine-grained classes. To alleviate this problem, an off-the-shelf image generator can be applied to synthesize additional training images, but these synthesized images are often not helpful for actually improving the accuracy of one-shot fine-grained recognition. This paper proposes a meta-learning framework to combine generated images with original images, so that the resulting ``hybrid'' training images can improve one-shot learning. Specifically, the generic image generator is updated by a few training instances of novel classes, and a Meta Image Reinforcing Network (MetaIRNet) is proposed to conduct one-shot fine-grained recognition as well as image reinforcement. The model is trained in an end-to-end manner, and our experiments demonstrate consistent improvement over baselines on one-shot fine-grained image classification benchmarks. | ['Yanwei Fu', 'Satoshi Tsutsui', 'David Crandall'] | 2019-11-17 | meta-reinforced-synthetic-data-for-one-shot | http://papers.nips.cc/paper/8570-meta-reinforced-synthetic-data-for-one-shot-fine-grained-visual-recognition | http://papers.nips.cc/paper/8570-meta-reinforced-synthetic-data-for-one-shot-fine-grained-visual-recognition.pdf | neurips-2019-12 | ['fine-grained-visual-recognition'] | ['computer-vision'] | [ 5.02107203e-01 1.97825104e-01 -4.41280454e-01 -6.35413408e-01
-1.03314269e+00 -4.43202615e-01 7.55010128e-01 -4.69353825e-01
-2.58943528e-01 7.44281888e-01 2.13420928e-01 1.11166582e-01
1.50489658e-01 -8.49121928e-01 -1.02848816e+00 -6.73569679e-01
5.85627079e-01 4.19311166e-01 1.91733446e-02 -2.55315136e-02
1.89762041e-01 2.64023244e-01 -1.81488693e+00 6.76579952e-01
8.59795570e-01 1.04407084e+00 4.17755604e-01 7.69214511e-01
-2.87426054e-01 9.24930334e-01 -8.28409135e-01 -4.41520065e-01
4.13554013e-01 -5.17535269e-01 -5.99694550e-01 6.80229247e-01
7.13000953e-01 -6.37537897e-01 -3.46249223e-01 1.02957761e+00
3.83191764e-01 2.85183400e-01 5.78241110e-01 -1.30141675e+00
-1.07446742e+00 5.46945632e-01 -5.27891397e-01 1.22358896e-01
-6.02965020e-02 7.75792658e-01 7.54714251e-01 -1.11609685e+00
4.85042870e-01 1.28965676e+00 3.83025885e-01 9.04340386e-01
-1.17128277e+00 -7.84358382e-01 4.04355615e-01 -5.77271096e-02
-1.20256221e+00 -5.88081479e-01 5.63260198e-01 -2.31891572e-01
1.00455320e+00 -8.41746926e-02 4.36906546e-01 1.23467314e+00
-1.82241410e-01 7.22354174e-01 1.20123255e+00 -4.60419923e-01
2.51240671e-01 9.17407647e-02 6.79460689e-02 8.26996684e-01
1.15268096e-01 5.57727396e-01 -4.58507925e-01 1.24442428e-01
7.52433240e-01 3.40484649e-01 8.07858855e-02 -2.30758876e-01
-1.05076838e+00 7.72430599e-01 7.46323049e-01 1.41783580e-01
-4.38690931e-01 2.09334373e-01 1.90813690e-01 4.22596157e-01
5.09995162e-01 6.71885848e-01 -2.97296792e-01 -1.50729567e-01
-8.82867634e-01 1.53823227e-01 4.82289851e-01 7.40288377e-01
1.08461297e+00 3.59373331e-01 -6.37863338e-01 1.18382013e+00
-1.12314634e-02 4.20474619e-01 7.56151319e-01 -9.31426942e-01
2.47608021e-01 5.02405822e-01 3.02464254e-02 -3.93117934e-01
2.32573971e-01 -6.52807176e-01 -8.67544711e-01 3.17223400e-01
8.97632316e-02 -1.30471334e-01 -1.62746429e+00 1.64792097e+00
8.89009982e-02 4.44062144e-01 1.21791236e-01 8.67833734e-01
8.97645175e-01 8.36802185e-01 2.64582008e-01 1.47948498e-02
1.13009763e+00 -1.66532290e+00 -1.87809825e-01 -4.98183757e-01
2.12262854e-01 -4.54459667e-01 1.48014331e+00 3.75499316e-02
-9.64375556e-01 -9.52259302e-01 -9.85095263e-01 3.42697382e-01
-4.48704600e-01 -1.28399059e-01 4.55472171e-01 6.39048040e-01
-6.74761355e-01 4.01557386e-01 -4.84146446e-01 2.67937835e-02
8.75126302e-01 -1.48141859e-02 -2.89881527e-01 -4.55591917e-01
-8.91114354e-01 5.06341100e-01 4.58457321e-01 -2.36462072e-01
-1.48790753e+00 -1.03093672e+00 -7.98928678e-01 1.94584623e-01
2.73631096e-01 -8.51843953e-01 1.31128228e+00 -1.31524384e+00
-1.36708045e+00 8.15833092e-01 9.39759836e-02 -3.87562633e-01
2.65499532e-01 1.67902067e-01 -2.75895596e-01 -3.33252177e-03
4.55377370e-01 1.07567751e+00 1.46586800e+00 -1.27975440e+00
-8.84946704e-01 -1.25345364e-01 1.72763065e-01 2.20969453e-01
-3.17884803e-01 -3.14046353e-01 -4.18020070e-01 -1.10426164e+00
-6.12235963e-01 -8.40298235e-01 -4.43545550e-01 -2.23657653e-01
-2.54824072e-01 -3.80735733e-02 6.76174819e-01 -1.21200390e-01
7.54112959e-01 -2.26003480e+00 -1.57083392e-01 -1.63577244e-01
1.23365328e-01 2.85545588e-01 -8.42850983e-01 -1.00929253e-01
-1.93360224e-01 4.93477844e-02 -1.20774783e-01 -2.12964877e-01
-1.89223774e-02 1.99375540e-01 -6.63074255e-01 -1.32364064e-01
5.27851880e-01 1.34783053e+00 -8.82547796e-01 5.79798557e-02
1.82077989e-01 1.02686264e-01 -5.14585078e-01 4.89267677e-01
-6.13191724e-01 2.99623847e-01 -2.07266495e-01 6.79457784e-01
5.10729194e-01 -8.08554769e-01 -2.27874428e-01 -3.45340818e-01
2.85573184e-01 -2.86908627e-01 -7.93717146e-01 1.61532545e+00
-5.99878311e-01 1.12140343e-01 -6.69824362e-01 -6.69603229e-01
8.64694536e-01 -9.15443003e-02 -1.25368088e-01 -1.14605236e+00
-1.03987560e-01 -6.25215545e-02 -2.63065249e-02 -2.01626450e-01
4.33817089e-01 -5.04621148e-01 -3.25934172e-01 8.82034123e-01
2.67540962e-01 -1.38359442e-01 1.15030505e-01 1.86645046e-01
9.48313355e-01 -2.61220634e-02 2.31737882e-01 1.83115512e-01
5.46875186e-02 1.63366407e-01 4.37232763e-01 1.16659915e+00
-1.17678434e-01 6.56685770e-01 -2.91302036e-02 -5.76646626e-01
-1.36187136e+00 -1.05202436e+00 3.73387754e-01 1.60714865e+00
1.43516853e-01 -6.02710657e-02 -9.35479283e-01 -8.57785225e-01
1.05915561e-01 7.83994973e-01 -8.73491108e-01 -6.73228860e-01
-1.26382589e-01 -6.28190577e-01 5.63144803e-01 7.67146289e-01
7.68686712e-01 -1.36192620e+00 -5.93095303e-01 3.87093216e-01
-2.24927068e-02 -8.63541067e-01 -8.38241756e-01 3.10450613e-01
-6.79426074e-01 -8.90661657e-01 -1.03398240e+00 -8.28010917e-01
9.13497329e-01 7.46096015e-01 1.19570518e+00 1.65117890e-01
-5.67602277e-01 1.69288456e-01 -4.37308818e-01 -3.95049676e-02
-3.72148305e-01 2.42184065e-02 -1.85319081e-01 2.87832648e-01
2.10787192e-01 -1.95418909e-01 -6.17438674e-01 2.61968940e-01
-9.29375648e-01 1.71166450e-01 8.08663785e-01 1.58083606e+00
7.52687514e-01 5.75741753e-02 9.10619318e-01 -1.08773243e+00
7.84568012e-01 -3.06296438e-01 -4.61529851e-01 4.60583419e-01
-6.18131578e-01 3.05285662e-01 8.91207933e-01 -6.91202343e-01
-1.30910230e+00 -3.04403212e-02 -3.52922734e-03 -1.04508138e+00
-4.24165696e-01 2.35002637e-01 9.31166857e-03 -2.33425319e-01
8.29299212e-01 5.46108127e-01 -1.42704561e-01 -4.38358128e-01
8.60842288e-01 6.79629266e-01 7.32883692e-01 -5.43636918e-01
7.51636028e-01 1.41751513e-01 -6.20318353e-01 -5.06625593e-01
-1.36859536e+00 -1.47866026e-01 -2.45476425e-01 -1.87949333e-02
6.72873735e-01 -1.10335231e+00 -2.11703360e-01 7.54639149e-01
-6.51846111e-01 -8.25098038e-01 -7.34519780e-01 -4.79078442e-02
-6.23549342e-01 -2.06487894e-01 -5.07033885e-01 -3.08200866e-01
-4.02798861e-01 -1.19830596e+00 1.29341578e+00 5.73295772e-01
8.68311524e-02 -6.08150899e-01 -2.65789535e-02 4.54237014e-01
5.32166660e-01 -2.32123345e-01 7.99241185e-01 -2.17101440e-01
-6.28555298e-01 -1.18428357e-01 -4.95232821e-01 3.50654274e-01
2.74647772e-01 -2.77172834e-01 -1.05108106e+00 -4.65214968e-01
-3.80255967e-01 -1.00116420e+00 1.11188722e+00 1.23854816e-01
1.41557336e+00 -4.60238874e-01 -2.11986437e-01 6.57912910e-01
1.34460282e+00 2.63966292e-01 6.42872095e-01 1.26089409e-01
6.95151746e-01 1.24136075e-01 8.15840960e-01 5.17008305e-01
2.90526032e-01 4.91733313e-01 1.48716405e-01 -6.58958778e-02
-7.13642120e-01 -6.36016011e-01 -6.09256094e-03 2.24929601e-01
3.14300478e-01 -1.73907295e-01 -4.82484370e-01 4.76886600e-01
-1.55318153e+00 -1.23954284e+00 8.76236022e-01 1.87426329e+00
8.58895838e-01 8.84867981e-02 1.61071986e-01 -2.07298651e-01
7.62189090e-01 3.56237441e-01 -1.09412026e+00 -1.36034280e-01
-1.08494438e-01 4.15752053e-01 2.73645133e-01 2.74888039e-01
-8.98327887e-01 1.46414125e+00 6.04608917e+00 9.94315803e-01
-1.45625985e+00 1.98401183e-01 1.20316100e+00 -1.97454855e-01
-2.56321251e-01 -1.09071732e-01 -7.95643508e-01 6.13615096e-01
8.36882591e-01 -1.86120972e-01 6.80814028e-01 1.17788327e+00
-2.79924363e-01 3.06805730e-01 -7.59010434e-01 1.09146965e+00
2.19828799e-01 -1.83103120e+00 3.29766572e-01 -1.88655928e-01
1.16343081e+00 2.17021704e-02 2.43595749e-01 5.77251077e-01
5.93338370e-01 -1.02605665e+00 7.16636121e-01 5.95960975e-01
1.37993276e+00 -8.31704915e-01 2.34904245e-01 4.39646512e-01
-1.08563268e+00 -3.08280528e-01 -5.77111483e-01 1.37281874e-02
5.75102575e-04 4.46668893e-01 -8.07716906e-01 -5.70281669e-02
6.43039525e-01 6.32312894e-01 -7.95112669e-01 8.58089685e-01
-1.64548457e-01 4.37060714e-01 1.86554566e-01 -1.34354718e-02
5.08668959e-01 3.31049770e-01 4.39434871e-02 9.16216195e-01
1.67124361e-01 3.08711529e-01 3.34562182e-01 8.50073338e-01
-5.86689770e-01 -5.02011895e-01 -5.14624476e-01 -4.31221932e-01
5.93428612e-01 1.30643690e+00 -6.66726053e-01 -9.39833701e-01
-3.84221375e-01 1.30341363e+00 5.51923156e-01 3.38594764e-01
-6.82814479e-01 -4.53458458e-01 6.55997217e-01 -1.27475083e-01
6.50028884e-01 3.82359862e-01 -1.86612964e-01 -1.41903377e+00
-2.22968221e-01 -1.25852013e+00 4.44086164e-01 -1.00523984e+00
-1.78752160e+00 7.66359568e-01 -4.38106358e-01 -1.22723615e+00
-4.42725003e-01 -2.43199944e-01 -5.56019962e-01 7.32279241e-01
-1.39404786e+00 -1.15792036e+00 -6.63079083e-01 7.37067223e-01
9.27966654e-01 -4.31262881e-01 9.11508977e-01 1.58491552e-01
-3.73837143e-01 1.02381432e+00 -3.02168608e-01 4.04798165e-02
6.20806217e-01 -1.05415154e+00 7.91297793e-01 8.13790381e-01
1.35392904e-01 3.30117285e-01 3.79476279e-01 -7.57815301e-01
-1.36486733e+00 -1.80544019e+00 4.70118880e-01 -1.48076832e-01
4.29941922e-01 -2.18815684e-01 -9.26208198e-01 5.26271522e-01
5.54005206e-02 5.11190355e-01 5.82637548e-01 -1.40339568e-01
-9.19525981e-01 -2.97236115e-01 -1.27833056e+00 6.88212156e-01
1.20636177e+00 -5.90114832e-01 -6.13684237e-01 3.47780168e-01
9.68317032e-01 -2.12795317e-01 -6.56703174e-01 3.51359576e-01
4.69885916e-01 -7.47759819e-01 1.00877416e+00 -8.85435522e-01
6.16088629e-01 -2.31551960e-01 -2.43038327e-01 -1.67419302e+00
-7.14071155e-01 -1.93544894e-01 -8.69828612e-02 8.87293279e-01
3.65867049e-01 -5.75695038e-01 1.03529227e+00 4.59334135e-01
-2.44783580e-01 -5.51725686e-01 -4.41005647e-01 -8.56805503e-01
-1.98116582e-02 -7.04622045e-02 9.31396008e-01 6.59371376e-01
-4.20188874e-01 5.89623094e-01 -6.62894011e-01 -1.07223310e-01
8.12909484e-01 6.15751922e-01 1.04672861e+00 -7.23158240e-01
-7.66912460e-01 -5.06320119e-01 -2.67964333e-01 -1.03646529e+00
1.85435474e-01 -7.40009546e-01 4.44016516e-01 -1.20498729e+00
5.80569506e-01 -6.40339792e-01 -4.47718859e-01 7.75474370e-01
-4.72754240e-01 8.32613051e-01 4.44339037e-01 2.36280188e-01
-7.94351697e-01 6.15963519e-01 1.58269703e+00 -5.34318328e-01
1.01225235e-01 -2.32682779e-01 -1.17778003e+00 2.93641567e-01
5.54023504e-01 -3.66086423e-01 -5.76365173e-01 -3.02159160e-01
-3.39592993e-01 2.98032552e-01 4.06000733e-01 -8.50386977e-01
1.49129227e-01 -3.64499420e-01 6.93454981e-01 -2.44389415e-01
3.19772124e-01 -3.89166325e-01 1.82771713e-01 5.19293547e-01
-5.44060588e-01 -3.69290859e-01 9.44025964e-02 5.80089986e-01
-3.07162851e-01 -1.28257051e-01 1.26355898e+00 -5.56337655e-01
-1.16702366e+00 7.34719276e-01 5.13913073e-02 2.34515831e-01
1.06314766e+00 -3.11831623e-01 -6.45358682e-01 -2.06033751e-01
-6.49779916e-01 -1.14531867e-01 6.92360640e-01 6.86680734e-01
7.56588042e-01 -1.57540774e+00 -5.48187852e-01 5.20452559e-01
5.73442936e-01 -2.58213282e-01 6.08493924e-01 6.92704767e-02
5.56699894e-02 2.34065950e-01 -4.80294496e-01 -4.62485492e-01
-8.33960295e-01 9.10327852e-01 3.98224801e-01 -2.51945555e-01
-5.55806637e-01 1.13246584e+00 5.82739294e-01 -4.57473338e-01
2.18936250e-01 1.52233630e-01 2.58385897e-01 -1.66267172e-01
9.46718931e-01 -7.24563189e-03 -5.47751673e-02 -3.02247614e-01
-4.38811406e-02 4.13804650e-01 -6.59838319e-01 4.46257368e-02
1.29892111e+00 -4.54228669e-02 2.90931314e-01 3.17115158e-01
1.11853492e+00 -4.44164515e-01 -1.86455047e+00 -3.87862027e-01
-4.27814782e-01 -7.51578987e-01 -2.18640774e-01 -1.32549477e+00
-1.21747589e+00 6.77978218e-01 6.01336539e-01 -1.09762326e-01
1.30901182e+00 6.49718568e-02 8.91228974e-01 4.24801111e-01
5.28462410e-01 -7.34096527e-01 6.99176073e-01 5.05941868e-01
7.32694089e-01 -1.49224591e+00 -4.96816009e-01 -3.87845933e-02
-8.32904816e-01 8.42507184e-01 9.58482385e-01 -4.40950930e-01
5.24293482e-01 3.35408628e-01 1.67396590e-01 4.43854630e-02
-1.14623165e+00 -2.08130598e-01 1.86049134e-01 7.31279314e-01
-8.58866330e-03 2.96303749e-01 4.32003558e-01 7.08089113e-01
-9.77851525e-02 3.52116942e-01 2.48796508e-01 7.46537328e-01
-6.87954187e-01 -9.79583561e-01 -7.52849132e-02 9.32115018e-01
1.82789303e-02 -2.68611610e-01 -2.86285549e-01 2.95941889e-01
2.38240823e-01 7.21491575e-01 3.12326699e-01 -4.89859432e-01
2.63822258e-01 -1.16592206e-01 6.95047975e-01 -9.43511486e-01
-5.70760369e-01 -5.13200909e-02 -9.60980356e-02 -6.36163235e-01
-8.98450986e-02 -1.74076855e-01 -1.00894165e+00 -1.98525593e-01
-1.18640877e-01 -3.88970822e-02 4.02578413e-01 1.05736411e+00
6.40135884e-01 7.47696996e-01 9.71327245e-01 -8.85704339e-01
-6.86365068e-01 -8.26448560e-01 -4.50468898e-01 6.17261767e-01
2.81655967e-01 -6.61277235e-01 -1.67762443e-01 1.70803204e-01] | [9.962484359741211, 2.368959426879883] |
250fba80-3f93-452c-b561-1d6324b17351 | integrative-semantic-dependency-parsing-via | 1401.6050 | null | http://arxiv.org/abs/1401.6050v1 | http://arxiv.org/pdf/1401.6050v1.pdf | Integrative Semantic Dependency Parsing via Efficient Large-scale Feature Selection | Semantic parsing, i.e., the automatic derivation of meaning representation
such as an instantiated predicate-argument structure for a sentence, plays a
critical role in deep processing of natural language. Unlike all other top
systems of semantic dependency parsing that have to rely on a pipeline
framework to chain up a series of submodels each specialized for a specific
subtask, the one presented in this article integrates everything into one
model, in hopes of achieving desirable integrity and practicality for real
applications while maintaining a competitive performance. This integrative
approach tackles semantic parsing as a word pair classification problem using a
maximum entropy classifier. We leverage adaptive pruning of argument candidates
and large-scale feature selection engineering to allow the largest feature
space ever in use so far in this field, it achieves a state-of-the-art
performance on the evaluation data set for CoNLL-2008 shared task, on top of
all but one top pipeline system, confirming its feasibility and effectiveness. | ['Chunyu Kit', 'Xiaotian Zhang', 'Hai Zhao'] | 2014-01-23 | null | null | null | null | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [ 5.37832141e-01 6.71081603e-01 -1.12389266e-01 -6.50719404e-01
-1.09454107e+00 -7.60398507e-01 5.15555799e-01 5.67125797e-01
-6.37642264e-01 5.40693343e-01 2.68620610e-01 -6.58379853e-01
-1.60828844e-01 -7.27248251e-01 -5.09398043e-01 -3.91414016e-01
9.16801542e-02 5.68268478e-01 5.43670654e-01 -3.56473863e-01
2.21529961e-01 7.12262392e-02 -1.57289040e+00 5.71261227e-01
5.31096935e-01 8.52043509e-01 2.29029164e-01 4.87767577e-01
-5.39940298e-01 4.94624943e-01 -5.62369525e-01 -7.85469890e-01
3.36237289e-02 -2.23407060e-01 -1.23994744e+00 -5.56052029e-01
2.79490262e-01 2.12946221e-01 3.55995178e-01 8.51341128e-01
4.03998137e-01 -9.35767964e-02 1.76236525e-01 -8.75644565e-01
-2.17544004e-01 1.04271412e+00 -6.54009506e-02 2.49891728e-01
6.39430046e-01 -2.11502150e-01 1.69523346e+00 -5.92422605e-01
7.48295784e-01 1.33544767e+00 6.32651329e-01 6.71521664e-01
-1.25917172e+00 -3.85298580e-01 2.34735683e-01 5.70513085e-02
-6.46768034e-01 -2.87119329e-01 5.56254983e-01 -1.17394574e-01
1.82208872e+00 1.21373095e-01 5.39387703e-01 9.86227810e-01
1.89573213e-01 7.56123781e-01 8.80943239e-01 -7.34733284e-01
3.50782841e-01 -3.30394238e-01 8.19247127e-01 5.91014445e-01
4.29594040e-01 -3.56168360e-01 -5.99795640e-01 -2.41294265e-01
-7.64526203e-02 -5.98679960e-01 1.62916735e-01 -9.24948677e-02
-9.19526100e-01 1.00295568e+00 -7.92788193e-02 4.95370448e-01
-1.32192984e-01 8.44100267e-02 6.41982079e-01 3.75101268e-01
4.31322098e-01 4.47355032e-01 -9.93048966e-01 -3.37794483e-01
-6.36194110e-01 4.26785469e-01 1.19600677e+00 7.48343050e-01
3.62488478e-01 -4.48878944e-01 -9.10106152e-02 7.33607948e-01
1.84488431e-01 1.68433666e-01 4.64862794e-01 -8.30121696e-01
8.24162543e-01 1.01012325e+00 -2.82433540e-01 -1.95036709e-01
-7.99966931e-01 -3.40721518e-01 -1.62090193e-02 5.85637614e-02
6.07549846e-01 -2.75330156e-01 -6.87687457e-01 2.06722212e+00
6.17844522e-01 -2.69194424e-01 4.31277037e-01 5.43283165e-01
8.70247900e-01 3.53026867e-01 6.92538261e-01 1.31055981e-01
2.10742736e+00 -6.03413701e-01 -4.15482759e-01 -7.01463163e-01
9.32849944e-01 -6.42893195e-01 1.00412023e+00 3.11735213e-01
-1.17865729e+00 -2.23653287e-01 -1.16028905e+00 -4.65245783e-01
-4.07788932e-01 -2.14336932e-01 9.33299184e-01 6.78916514e-01
-9.49687719e-01 6.07307255e-01 -7.16054976e-01 -4.11433399e-01
2.38594234e-01 3.94251496e-01 -7.25657701e-01 4.57703471e-02
-1.42507410e+00 9.05652046e-01 7.84965336e-01 -5.22937290e-02
-2.19942734e-01 -7.47261941e-01 -1.13312542e+00 1.90655768e-01
5.10713518e-01 -9.68859732e-01 1.48735905e+00 -6.71632349e-01
-1.33462918e+00 1.16777289e+00 -3.00913811e-01 -8.17198753e-01
2.13936448e-01 -7.03180969e-01 -2.38941208e-01 8.84993002e-02
2.32288823e-01 5.74102998e-01 5.89405417e-01 -6.73659980e-01
-7.51786292e-01 -5.27518690e-01 2.19752461e-01 1.70888796e-01
-1.24402970e-01 5.34858167e-01 -2.14463189e-01 -2.71770805e-01
3.26600045e-01 -7.23361611e-01 -1.66574270e-01 -8.57427180e-01
-2.27554396e-01 -5.77760935e-01 4.13652182e-01 -7.42486537e-01
1.19345677e+00 -2.01683617e+00 1.38745084e-01 -1.96483955e-01
-8.11626837e-02 3.96667480e-01 -5.08083291e-02 5.77377617e-01
-2.55438030e-01 2.53616661e-01 -5.12740314e-01 -6.20714664e-01
1.52980149e-01 3.64097953e-01 -2.18883947e-01 1.50693491e-01
7.28745401e-01 1.09471893e+00 -9.73233402e-01 -3.75503182e-01
1.76077653e-02 2.80134141e-01 -5.50019681e-01 1.70811430e-01
-3.78918886e-01 8.69123265e-02 -6.69033885e-01 3.74033868e-01
5.53192437e-01 -4.05296497e-02 5.07672191e-01 1.25759497e-01
6.71666162e-03 9.96440828e-01 -9.97759581e-01 2.08279276e+00
-6.34785593e-01 1.87517241e-01 1.10462725e-01 -1.00496602e+00
8.96126211e-01 3.23113680e-01 2.21954748e-01 -6.14848256e-01
1.24730580e-01 4.23704237e-01 1.20928288e-01 -2.91383535e-01
4.24348414e-01 -2.66138613e-01 -6.06300771e-01 2.89410412e-01
4.57422137e-01 1.61361787e-02 3.56776118e-01 2.50035971e-01
1.39112210e+00 4.81595218e-01 5.43343842e-01 -5.69591999e-01
6.93681002e-01 2.75078386e-01 6.39779329e-01 5.70683241e-01
2.64255069e-02 5.10461807e-01 9.23898637e-01 -5.10052621e-01
-8.91415596e-01 -8.57025921e-01 -2.21686289e-01 1.30819476e+00
-1.08833700e-01 -5.87324202e-01 -1.05528235e+00 -1.00851190e+00
-2.21422732e-01 1.08282232e+00 -5.16962528e-01 1.16490692e-01
-9.65578377e-01 -9.02529240e-01 6.82709992e-01 5.10615885e-01
2.88741976e-01 -1.43619418e+00 -1.20645785e+00 5.00205636e-01
-1.14419371e-01 -1.54492342e+00 3.21580410e-01 8.09591472e-01
-7.00969160e-01 -1.36485744e+00 2.74555087e-02 -8.22165549e-01
2.87101954e-01 -3.05477947e-01 1.62451661e+00 1.48206055e-01
-1.65053695e-01 1.31373674e-01 -6.95164502e-01 -4.30363715e-01
-6.29649043e-01 3.91822308e-01 -6.52104974e-01 -4.28438485e-01
7.14985967e-01 -3.93550813e-01 -2.95187682e-01 -3.85945618e-01
-9.09955740e-01 5.31151630e-02 4.62642044e-01 8.20713043e-01
3.60266030e-01 -3.00451458e-01 5.62689900e-01 -1.24341726e+00
6.52384996e-01 -4.05172646e-01 -4.93713468e-01 3.49302739e-01
-4.86674130e-01 3.41009110e-01 7.72729695e-01 3.28606576e-01
-1.28049195e+00 1.99634776e-01 -6.81951344e-01 6.78144515e-01
-4.80715305e-01 3.57023120e-01 -4.55316663e-01 5.97656250e-01
4.92636174e-01 -8.77879262e-02 -1.37264058e-01 -6.26559854e-01
4.91366953e-01 3.57959956e-01 5.61720908e-01 -8.67089629e-01
4.43145484e-01 2.89915353e-01 2.01497719e-01 -4.97895151e-01
-1.21374011e+00 -4.69222546e-01 -8.21687043e-01 5.87128341e-01
1.18718886e+00 -7.83916116e-01 -4.04835790e-01 1.70554951e-01
-1.39940679e+00 -1.16112411e-01 -3.41163665e-01 2.02213332e-01
-4.88201648e-01 4.41809744e-01 -6.33145213e-01 -6.15332067e-01
-7.09309340e-01 -9.64479506e-01 1.34885991e+00 8.27609599e-02
-5.25699258e-01 -1.06602156e+00 -4.83711623e-02 4.62543517e-01
1.60523042e-01 2.14695811e-01 1.28922975e+00 -1.40776289e+00
-5.56677431e-02 -2.33526245e-01 -5.45769632e-02 3.40155810e-01
-2.32552171e-01 -2.18156159e-01 -1.08011246e+00 6.40595406e-02
2.03305617e-01 -2.58608609e-01 9.09860253e-01 1.30620256e-01
6.74896479e-01 1.78697839e-01 -5.35410494e-02 3.84611934e-01
1.36640322e+00 5.41914105e-02 3.41107935e-01 6.13812685e-01
2.28144377e-01 9.05197859e-01 7.62162268e-01 1.12814769e-01
4.59439993e-01 3.83432031e-01 5.82744062e-01 2.44197786e-01
-1.31375194e-01 -1.84559554e-01 3.54641229e-01 3.11832637e-01
3.90102923e-01 -2.66590208e-01 -1.13703394e+00 5.64506948e-01
-1.96965599e+00 -6.16229177e-01 -4.07599181e-01 2.01457500e+00
6.73740029e-01 5.97688198e-01 -1.05814427e-01 2.57240206e-01
4.80219543e-01 1.85528398e-01 -2.04592973e-01 -9.01819110e-01
-1.15336604e-01 7.38141775e-01 2.99335510e-01 5.19433320e-01
-1.17562604e+00 1.35450089e+00 6.04120398e+00 5.78710079e-01
-5.91286361e-01 2.90413857e-01 5.30533254e-01 1.92647800e-01
-4.05102670e-01 3.63650411e-01 -1.22595644e+00 1.77754194e-01
1.39086616e+00 9.64170247e-02 1.81808874e-01 7.88623691e-01
-1.39995113e-01 -3.25892240e-01 -1.10521960e+00 4.08914924e-01
-3.01935375e-01 -1.23888755e+00 -1.47537500e-01 -3.53298694e-01
-9.85775664e-02 3.07734251e-01 -4.77961570e-01 3.93600464e-01
3.99240673e-01 -9.43957567e-01 8.08608830e-01 -1.69013575e-01
4.00172561e-01 -5.76596737e-01 9.29611683e-01 3.85616481e-01
-1.09043753e+00 -2.06032947e-01 -1.10336542e-01 -2.48716727e-01
5.70538461e-01 8.04926991e-01 -4.45084870e-01 8.41124117e-01
6.12235665e-01 3.99446636e-01 -5.88448346e-01 4.81396347e-01
-9.23345149e-01 7.45438814e-01 -4.37665075e-01 2.26675160e-03
3.34569544e-01 1.67519838e-01 6.15620136e-01 1.60244453e+00
7.77157582e-03 7.72643983e-02 -3.04625016e-02 5.03312111e-01
1.04651593e-01 2.84910262e-01 -3.55351418e-01 6.47717044e-02
5.67710638e-01 1.42314661e+00 -8.09676945e-01 -3.39962900e-01
-5.94743431e-01 6.72611356e-01 5.66718400e-01 -2.87190735e-01
-8.34517181e-01 -3.06911558e-01 5.85796833e-01 -2.38987684e-01
4.47664797e-01 -2.15335488e-01 -5.52198470e-01 -9.19559956e-01
3.35978210e-01 -8.34205091e-01 7.87571251e-01 -2.97118783e-01
-1.03254807e+00 8.37679386e-01 2.19536617e-01 -6.06871486e-01
-5.19361496e-01 -8.39430332e-01 -6.90868795e-01 1.01833582e+00
-1.73654091e+00 -1.29407418e+00 2.76883761e-03 2.77102590e-01
5.82371593e-01 4.26652282e-02 1.19157732e+00 -5.69574209e-03
-4.96550381e-01 4.75828260e-01 -5.68121731e-01 -3.38710216e-03
3.05230856e-01 -1.36174083e+00 1.16231954e+00 1.06689298e+00
2.54688859e-01 6.65903687e-01 5.11815786e-01 -5.10107040e-01
-1.50332654e+00 -6.93150580e-01 1.39761722e+00 -6.50794387e-01
8.74623299e-01 -6.25401616e-01 -9.22941923e-01 5.90188205e-01
1.68073848e-01 -6.74491096e-03 7.66066015e-01 5.16347706e-01
-6.10894859e-01 2.75593221e-01 -1.10242176e+00 1.70163363e-01
1.32103765e+00 -3.10609043e-01 -1.24611795e+00 5.13926595e-02
1.11280954e+00 -4.91881728e-01 -8.26602340e-01 3.59204769e-01
4.81032401e-01 -1.18498874e+00 8.12703192e-01 -9.45839465e-01
5.38199902e-01 8.10236484e-03 -2.43724570e-01 -9.15414453e-01
-1.68433115e-01 -8.04996312e-01 -4.86062765e-02 1.50557756e+00
7.68020332e-01 -7.83548176e-01 7.54920661e-01 7.95710742e-01
-4.13544267e-01 -8.39238524e-01 -1.19379425e+00 -6.54924333e-01
3.33686084e-01 -7.88238049e-01 6.47422612e-01 3.99289906e-01
3.98895964e-02 8.91017675e-01 3.93450826e-01 5.30455410e-02
5.60814202e-01 1.72230303e-01 4.52071428e-01 -1.36612821e+00
-4.12863016e-01 -4.05011982e-01 -3.50980848e-01 -6.64174438e-01
7.33978570e-01 -1.23944068e+00 -6.97184578e-02 -1.72476888e+00
7.75884539e-02 -4.65707809e-01 -3.19075048e-01 7.90432811e-01
-3.99168730e-01 -7.35775530e-02 2.96503693e-01 -1.43102556e-01
-5.46407819e-01 5.01480997e-02 5.45930862e-01 2.71004230e-01
-9.50470790e-02 -3.06655735e-01 -1.14970565e+00 8.37970614e-01
9.14337575e-01 -7.19859004e-01 -3.69625270e-01 -6.75295174e-01
4.50874686e-01 -1.35040015e-01 1.28581405e-01 -8.29552233e-01
-1.10962704e-01 1.25868365e-01 -1.74593136e-01 -1.75784767e-01
2.39326879e-01 -6.49650931e-01 -1.34718493e-01 4.01520103e-01
-3.05435151e-01 1.95771411e-01 2.60707885e-01 3.56012225e-01
-3.22331846e-01 -6.44479752e-01 6.02097154e-01 -2.23912969e-01
-1.03247416e+00 -2.01488957e-01 4.15542349e-02 5.27733028e-01
9.51176763e-01 -5.27496338e-02 -4.32898879e-01 3.32660347e-01
-6.18874848e-01 1.68630496e-01 1.10639639e-01 6.31220222e-01
1.44024029e-01 -5.12765288e-01 -8.22586119e-01 2.59475440e-01
1.32483035e-01 1.79720387e-01 -2.77155619e-02 5.40202856e-01
-3.15205306e-01 4.97883052e-01 1.25223652e-01 -3.85633677e-01
-1.36563253e+00 2.52642334e-01 -1.94333196e-02 -8.71284783e-01
-8.90937746e-01 1.02239370e+00 -8.16765279e-02 -3.56629997e-01
-1.74861893e-01 -3.93825799e-01 -2.85808921e-01 4.64602597e-02
4.78984296e-01 -1.16306618e-01 6.64352179e-01 -5.69951177e-01
-6.85574770e-01 4.36634928e-01 4.62090336e-02 -4.09821272e-02
1.49600065e+00 8.75474438e-02 -3.24581891e-01 9.79068950e-02
1.05623794e+00 1.63456090e-02 -1.02311194e+00 -1.44724576e-02
7.73687899e-01 9.83779505e-03 -7.82780647e-02 -1.02448010e+00
-5.85780978e-01 8.54813993e-01 2.12776996e-02 3.02994728e-01
9.22386348e-01 4.15420085e-01 1.00528967e+00 3.54130179e-01
3.67320806e-01 -1.03775430e+00 -5.13843298e-01 8.84431958e-01
4.62276757e-01 -1.10932791e+00 -3.51470499e-03 -7.41740823e-01
-5.81161022e-01 1.24759603e+00 2.22898096e-01 -1.87811479e-01
4.62667584e-01 5.32574594e-01 -3.42155434e-02 -2.33167857e-01
-1.03195536e+00 -3.37260008e-01 4.72862273e-02 4.04412031e-01
8.26030374e-01 1.57762513e-01 -8.60697687e-01 1.12335551e+00
-4.35387731e-01 -4.30370033e-01 6.97547793e-02 1.16206193e+00
-5.89831591e-01 -1.77538621e+00 6.16985969e-02 1.11719951e-01
-9.38094556e-01 -4.72867191e-01 -3.32045943e-01 9.09451962e-01
6.04960434e-02 1.12283134e+00 -4.52129468e-02 -9.42450911e-02
5.42997420e-01 6.81039810e-01 5.53909421e-01 -9.69838381e-01
-1.31363201e+00 -4.01109159e-01 7.65301108e-01 -7.65371323e-01
-2.35573292e-01 -7.75949657e-01 -1.80420804e+00 2.24891260e-01
-9.38087627e-02 2.43438765e-01 1.04142606e+00 1.21558106e+00
4.61561382e-01 3.89229804e-01 4.96250018e-02 -4.27750945e-01
-8.00354540e-01 -6.71310365e-01 -2.66908497e-01 6.38065875e-01
-1.19802229e-01 -4.51901376e-01 -2.04640359e-01 -2.08028644e-01] | [10.292357444763184, 9.440902709960938] |
2504ceb5-8069-4ad8-8355-6f4cc39ca7c1 | automatic-code-summarization-via-multi | 2006.05405 | null | https://arxiv.org/abs/2006.05405v5 | https://arxiv.org/pdf/2006.05405v5.pdf | Retrieval-Augmented Generation for Code Summarization via Hybrid GNN | Source code summarization aims to generate natural language summaries from structured code snippets for better understanding code functionalities. However, automatic code summarization is challenging due to the complexity of the source code and the language gap between the source code and natural language summaries. Most previous approaches either rely on retrieval-based (which can take advantage of similar examples seen from the retrieval database, but have low generalization performance) or generation-based methods (which have better generalization performance, but cannot take advantage of similar examples). This paper proposes a novel retrieval-augmented mechanism to combine the benefits of both worlds. Furthermore, to mitigate the limitation of Graph Neural Networks (GNNs) on capturing global graph structure information of source code, we propose a novel attention-based dynamic graph to complement the static graph representation of the source code, and design a hybrid message passing GNN for capturing both the local and global structural information. To evaluate the proposed approach, we release a new challenging benchmark, crawled from diversified large-scale open-source C projects (total 95k+ unique functions in the dataset). Our method achieves the state-of-the-art performance, improving existing methods by 1.42, 2.44 and 1.29 in terms of BLEU-4, ROUGE-L and METEOR. | ['Yang Liu', 'JingKai Siow', 'Xiaofei Xie', 'Yu Chen', 'Shangqing Liu'] | 2020-06-09 | retrieval-augmented-generation-for-code | https://openreview.net/forum?id=zv-typ1gPxA | https://openreview.net/pdf?id=zv-typ1gPxA | iclr-2021-1 | ['code-summarization'] | ['computer-code'] | [-4.00641002e-02 -3.38389054e-02 -2.92637378e-01 -1.34390462e-02
-1.09706032e+00 -5.30822039e-01 2.94121325e-01 7.31279612e-01
9.70541537e-02 4.55419332e-01 6.44838333e-01 -2.41617456e-01
-1.18848823e-01 -7.35580146e-01 -5.87733448e-01 -2.37395659e-01
-2.90645510e-01 -1.37669191e-01 5.06261170e-01 -3.33865911e-01
7.99474299e-01 -1.24367632e-01 -1.58421481e+00 4.70542014e-01
1.50144279e+00 3.26431155e-01 5.75013816e-01 8.12193453e-01
-7.04652309e-01 9.74560618e-01 -8.32340240e-01 -3.51936996e-01
-8.46549347e-02 -7.07379103e-01 -7.38692939e-01 -3.24904233e-01
3.30494344e-01 -6.39157295e-02 -4.34081316e-01 1.38221562e+00
4.94442433e-01 2.40037888e-02 3.23308051e-01 -1.14742327e+00
-9.61291730e-01 1.02058554e+00 -7.97313154e-01 2.18128771e-01
6.71374083e-01 1.70641346e-03 1.13866711e+00 -4.85605448e-01
7.70837069e-01 1.00473666e+00 7.11787343e-01 4.46976691e-01
-9.97043610e-01 -4.00119215e-01 1.66133508e-01 5.96222021e-02
-9.88925159e-01 -2.25334570e-01 8.02861154e-01 -5.33007622e-01
1.31567264e+00 1.53679505e-01 4.28791016e-01 6.85211778e-01
4.22500193e-01 6.73284829e-01 3.78047347e-01 -2.46173114e-01
8.72792155e-02 -1.30664170e-01 5.88016689e-01 9.68236744e-01
5.59980810e-01 -6.09531283e-01 -1.89466268e-01 -6.41445637e-01
1.34274125e-01 3.26813430e-01 -5.85643828e-01 -2.63018012e-01
-1.20700693e+00 6.66815639e-01 6.51628196e-01 4.27580446e-01
-2.95118600e-01 3.69251817e-01 8.27546835e-01 3.82403284e-01
3.22555840e-01 5.68902791e-01 -2.51055449e-01 -3.99066240e-01
-9.27864015e-01 3.50133747e-01 9.34148550e-01 1.29323101e+00
9.83882070e-01 -4.99882847e-02 -5.03875375e-01 8.57341409e-01
4.82745379e-01 4.34628636e-01 1.03535533e+00 -3.91012996e-01
8.44983637e-01 1.39344335e+00 -3.21081400e-01 -1.37610233e+00
-2.80447632e-01 -6.02051020e-01 -6.57542646e-01 -2.67112464e-01
-1.72873586e-01 1.39627546e-01 -7.52938747e-01 1.48464763e+00
-4.09466922e-02 -9.20975506e-02 2.06828877e-01 3.79942566e-01
1.29732645e+00 8.35012317e-01 -2.40280554e-01 -7.84885138e-02
1.17227972e+00 -1.27783597e+00 -4.87910062e-01 -3.25222284e-01
9.73905742e-01 -7.27014482e-01 1.19742036e+00 6.61441609e-02
-7.28924632e-01 -4.03300941e-01 -1.13268209e+00 -5.83742224e-02
-3.07625353e-01 2.83719003e-01 4.20969367e-01 5.26843309e-01
-1.19800174e+00 5.53359568e-01 -6.33229613e-01 -4.89674509e-01
2.30345979e-01 -4.38427031e-02 -2.13592902e-01 -2.53571510e-01
-6.14735544e-01 2.67972469e-01 5.56766272e-01 -3.40987056e-01
-8.22832584e-01 -7.27771342e-01 -8.85987580e-01 4.92169797e-01
5.00659227e-01 -4.73896176e-01 1.09442806e+00 -7.56231606e-01
-9.93825674e-01 3.43409240e-01 -6.12634560e-03 -3.14714968e-01
9.85355228e-02 -1.48977354e-01 -2.16706708e-01 6.99595883e-02
1.47376925e-01 7.53887445e-02 3.90000552e-01 -1.15943038e+00
-4.52054143e-01 -2.03832179e-01 3.55002463e-01 -7.86567628e-02
-5.73728561e-01 -4.15820219e-02 -6.04801416e-01 -6.64577842e-01
-2.15897888e-01 -7.64967322e-01 -2.02327684e-01 -5.43238044e-01
-4.74197775e-01 -3.08533579e-01 5.80579102e-01 -9.68262792e-01
1.94751585e+00 -2.14835525e+00 1.56776696e-01 -1.36937007e-01
5.33640325e-01 4.59859133e-01 -5.51868737e-01 9.72954810e-01
6.03920408e-02 3.40648055e-01 -4.90949154e-01 -1.02601811e-01
-1.28631920e-01 -2.43024409e-01 -2.33213857e-01 1.00406237e-01
-7.65157044e-02 8.84102345e-01 -1.18955779e+00 -3.66266608e-01
-4.44209874e-01 6.11228645e-02 -8.77020299e-01 2.30273813e-01
-4.14675623e-01 -1.37029305e-01 -6.37248099e-01 3.66188198e-01
5.75682700e-01 -4.28881496e-01 1.18826985e-01 1.94657981e-01
6.05607070e-02 3.67284745e-01 -8.33793104e-01 2.05409622e+00
-6.99871659e-01 4.94135290e-01 -2.55619437e-01 -7.98722744e-01
1.05230391e+00 6.48754537e-02 1.59444362e-01 -4.87681985e-01
-2.53985316e-01 3.08332115e-01 2.01920769e-03 -6.53864980e-01
6.31997228e-01 5.59329331e-01 -2.78167009e-01 7.19320774e-01
1.23952786e-02 7.22405612e-02 6.71496868e-01 8.19932163e-01
1.83932328e+00 2.85638701e-02 4.45690483e-01 -3.23525101e-01
7.88520873e-01 1.10380426e-01 4.68381166e-01 8.45322132e-01
1.98343456e-01 5.53957701e-01 1.06486690e+00 -2.37228259e-01
-7.51024723e-01 -4.74421263e-01 5.26428342e-01 1.04171395e+00
2.06834212e-01 -1.19588315e+00 -9.29316640e-01 -1.11631405e+00
-4.04536426e-02 6.73673213e-01 -4.95905071e-01 -5.28177917e-01
-6.37274325e-01 -6.27348483e-01 6.62222922e-01 3.74477953e-01
5.34385264e-01 -1.00958765e+00 -4.51706648e-01 2.91568518e-01
-5.21520138e-01 -6.10396385e-01 -7.81188190e-01 -4.01804984e-01
-8.92925501e-01 -1.22824216e+00 -6.53291166e-01 -6.30420625e-01
6.82940841e-01 6.03834927e-01 1.31216049e+00 8.38691056e-01
-1.07148059e-01 2.99515009e-01 -7.37854958e-01 -1.42197117e-01
-8.95609677e-01 5.86644173e-01 -6.82957172e-01 -3.77488732e-01
9.57360342e-02 -4.97001052e-01 -3.63418192e-01 -4.31139134e-02
-1.07076859e+00 -4.33544535e-03 7.78632224e-01 8.32511187e-01
1.14716135e-01 -1.03339970e-01 8.40036750e-01 -9.82816398e-01
9.79338467e-01 -9.71548975e-01 -5.90804636e-01 5.27403533e-01
-7.10387945e-01 3.65389079e-01 8.46287370e-01 -2.44189531e-01
-1.03997707e+00 -2.09700987e-01 -7.46219456e-02 -1.90516841e-02
2.65848845e-01 1.06165802e+00 1.40139535e-01 -8.74244666e-04
9.23913300e-01 5.39329171e-01 -8.67109299e-02 -6.56150579e-01
2.83521354e-01 8.92405927e-01 3.81726414e-01 -5.33921599e-01
6.47692382e-01 3.27506997e-02 -4.76397783e-01 -4.54336226e-01
-4.16764587e-01 -6.83855474e-01 -2.15388477e-01 -1.04752583e-02
5.69574952e-01 -8.20382118e-01 -2.91700274e-01 3.52153122e-01
-1.41569912e+00 5.98818585e-02 4.51525338e-02 1.78262860e-01
-2.55045742e-01 8.83902788e-01 -5.85556149e-01 -4.75337803e-01
-7.03221679e-01 -1.32504117e+00 1.07423043e+00 3.11497331e-01
7.74283186e-02 -7.04402030e-01 3.63646775e-01 5.48436120e-02
7.38756895e-01 4.58209127e-01 1.33955038e+00 -8.67327154e-01
-7.72671759e-01 -4.40134197e-01 -4.04006422e-01 1.42540619e-01
3.13924521e-01 1.11031532e-01 -5.56813657e-01 -5.48141122e-01
-1.72946543e-01 -8.81452411e-02 9.02608216e-01 -1.75487772e-01
9.34747875e-01 -5.64607203e-01 -3.83341342e-01 2.86710888e-01
1.83001065e+00 9.56461579e-02 6.97122037e-01 2.61357069e-01
8.72846603e-01 5.01544774e-01 2.54559427e-01 5.67564070e-01
6.00059152e-01 4.10500824e-01 7.15965688e-01 4.90229040e-01
-3.35802883e-01 -3.84301484e-01 6.12834632e-01 1.38535929e+00
2.67518729e-01 -3.85887355e-01 -1.19753826e+00 9.03035343e-01
-2.00929332e+00 -7.07763195e-01 -4.34106916e-01 2.15342259e+00
7.26273179e-01 1.08228751e-01 1.09302506e-01 -2.43122071e-01
9.05804157e-01 1.39051765e-01 -3.35389227e-01 -2.05629125e-01
2.91502386e-01 -3.26183885e-01 2.06789151e-01 1.60905436e-01
-5.78990877e-01 5.60976326e-01 5.01302195e+00 1.08147514e+00
-8.34526122e-01 3.35200168e-02 -5.09174354e-03 1.70061529e-01
-5.74661553e-01 2.84781754e-01 -5.40626884e-01 7.28176713e-01
1.09043908e+00 -8.80841374e-01 4.97441500e-01 1.25884914e+00
-2.88688153e-01 -6.22508898e-02 -1.02568114e+00 7.79758215e-01
3.63375485e-01 -1.43784165e+00 2.75641322e-01 -1.19725376e-01
9.08475637e-01 2.89741695e-01 -5.10026276e-01 6.73698843e-01
3.38206768e-01 -5.24649262e-01 4.82890487e-01 4.76416767e-01
5.41233897e-01 -5.92433333e-01 1.04522872e+00 4.75676626e-01
-1.52219748e+00 -2.94264406e-01 -4.62515563e-01 1.00263663e-01
-3.18369091e-01 6.65314674e-01 -6.95679188e-01 1.09072590e+00
5.91410935e-01 9.78048027e-01 -1.19048882e+00 1.44663131e+00
4.07814793e-03 4.97260898e-01 7.38553628e-02 -5.00835419e-01
2.29236066e-01 1.32328600e-01 7.61578798e-01 1.43530309e+00
5.16628623e-01 -3.05294186e-01 3.45659167e-01 1.09466326e+00
-3.52658808e-01 4.20559555e-01 -8.21867824e-01 -3.16530406e-01
3.40294927e-01 1.34624696e+00 -8.12561393e-01 -5.97786665e-01
-6.47986710e-01 5.82607031e-01 4.37007576e-01 1.33553073e-01
-6.66859448e-01 -1.16666567e+00 2.74317533e-01 -4.80896421e-02
2.68657207e-01 -7.25485012e-02 1.30403891e-01 -1.38744140e+00
4.64024961e-01 -1.06099010e+00 3.63451779e-01 -6.55111551e-01
-9.39587176e-01 8.48749518e-01 6.18258007e-02 -1.31953526e+00
-2.34217197e-01 -8.56219232e-02 -8.99066389e-01 6.23092473e-01
-1.33252025e+00 -8.63897264e-01 -6.82945132e-01 9.11919251e-02
7.00252175e-01 -3.43180805e-01 5.04787982e-01 4.11641091e-01
-4.58037883e-01 5.51056385e-01 3.56560081e-01 9.33322459e-02
5.44307649e-01 -1.44254339e+00 8.52066219e-01 1.18991768e+00
-8.00459534e-02 1.07594538e+00 4.45514172e-01 -9.56887364e-01
-1.73717213e+00 -1.21868777e+00 7.16163456e-01 -3.42743933e-01
5.86948335e-01 -3.01161498e-01 -1.29599643e+00 3.64183158e-01
2.62592077e-01 -2.28089750e-01 3.80172610e-01 -8.23053271e-02
-4.81031299e-01 -3.00084706e-02 -9.60797369e-01 4.70970839e-01
1.06395280e+00 -4.24799711e-01 -6.36150956e-01 2.72590309e-01
1.06097853e+00 -2.15677425e-01 -8.22506845e-01 7.23360479e-02
2.42571503e-01 -1.04103494e+00 4.74806756e-01 -4.39589649e-01
9.11293030e-01 -3.80893469e-01 1.94077045e-02 -1.41600919e+00
-2.47486413e-01 -6.96570277e-01 -1.29287884e-01 1.55930424e+00
3.66217434e-01 -5.83169937e-01 4.59971368e-01 9.30374786e-02
-4.94340092e-01 -6.81677401e-01 -3.52274150e-01 -7.54225969e-01
-1.91335037e-01 -1.30033977e-02 9.13229823e-01 7.42210627e-01
3.86335969e-01 3.27935398e-01 -1.16585456e-01 -1.25276625e-01
3.48949701e-01 3.98248047e-01 9.94625330e-01 -1.20658314e+00
-3.76197308e-01 -5.60091674e-01 -5.06037414e-01 -9.37957048e-01
1.93252012e-01 -1.27493572e+00 1.25077993e-01 -1.94920671e+00
6.62243187e-01 -5.93029447e-02 -2.48076990e-01 4.93520707e-01
-4.96742159e-01 -2.60181874e-01 2.73040950e-01 3.63125980e-01
-8.99670541e-01 5.04729629e-01 8.30724418e-01 -3.26157153e-01
-2.82897383e-01 -2.87008971e-01 -9.54619348e-01 4.95410293e-01
6.33494079e-01 -8.27522337e-01 -6.13519013e-01 -5.95403433e-01
5.76662362e-01 3.97526413e-01 8.99684951e-02 -1.08494961e+00
4.65582967e-01 1.34277388e-01 -4.45157021e-01 -5.04477024e-01
-6.33855760e-01 -4.07653481e-01 3.13115567e-02 7.14607000e-01
-3.57844889e-01 4.46392298e-01 2.79666215e-01 8.94364595e-01
-3.58217835e-01 -6.67014599e-01 4.97473180e-01 -3.22120041e-01
-7.07414150e-01 1.29228532e-01 -2.17281237e-01 2.77652174e-01
7.03630567e-01 5.82665317e-02 -9.99745071e-01 -2.25422814e-01
2.19520941e-01 3.16347301e-01 7.65875518e-01 7.63360441e-01
6.40240312e-01 -1.27280378e+00 -8.81475389e-01 7.29875863e-02
6.56276166e-01 -2.77513325e-01 2.78679043e-01 7.38023520e-01
-8.33293498e-01 2.32596397e-01 -1.87531188e-01 -2.80367702e-01
-1.17783928e+00 6.52882874e-01 7.77155682e-02 -7.29654312e-01
-6.85658336e-01 4.64139104e-01 1.50615275e-01 -5.05420387e-01
-8.17637444e-02 -4.46090966e-01 -3.90625596e-01 -1.15409791e-01
6.45716369e-01 5.66588700e-01 3.82657319e-01 -3.15223545e-01
-3.36326748e-01 5.37194788e-01 -3.51041138e-01 4.22999740e-01
1.33918774e+00 7.99852833e-02 -6.75464213e-01 2.56354392e-01
1.48476660e+00 3.72163266e-01 -6.71144128e-01 -3.02169800e-01
3.43222976e-01 -5.47271252e-01 -2.37978622e-01 -6.05296791e-01
-1.00105357e+00 6.50358856e-01 2.06506640e-01 5.30489862e-01
1.03524888e+00 -7.68408105e-02 8.56966674e-01 6.93303943e-01
4.44620758e-01 -6.41802192e-01 3.18932831e-01 5.30938685e-01
1.00226939e+00 -9.86605346e-01 8.06132257e-02 -2.70063639e-01
-3.03646713e-01 1.21211243e+00 6.80467367e-01 -2.35735983e-01
2.34180480e-01 -7.47042010e-03 -3.18351477e-01 -4.08300251e-01
-9.57866549e-01 1.47474721e-01 3.47864896e-01 3.40333551e-01
5.84165156e-01 -3.94926131e-01 -4.69589800e-01 5.64780354e-01
6.23172559e-02 -2.21475258e-01 1.03856242e+00 1.11734164e+00
-6.37693584e-01 -1.04031897e+00 -1.49494866e-02 9.47753787e-01
-6.08032703e-01 -3.11827958e-01 -3.45817238e-01 5.71409166e-01
-4.06698316e-01 9.58079338e-01 -4.00987983e-01 -4.45275813e-01
4.10885304e-01 -7.20736086e-02 -4.69708443e-02 -9.71876919e-01
-8.57386112e-01 -2.35691890e-01 -8.05050060e-02 -6.15382195e-01
-1.51294217e-01 -3.40715677e-01 -1.23235834e+00 -1.97606787e-01
-5.99361539e-01 3.89555842e-01 7.14485109e-01 4.39395100e-01
8.23399067e-01 8.67204785e-01 5.26996970e-01 -5.40276766e-01
-6.63342416e-01 -1.02948666e+00 -2.38437206e-01 3.95013183e-01
4.75413144e-01 -4.12034720e-01 -5.08829653e-01 6.30754158e-02] | [7.554312229156494, 7.96572208404541] |
289377a6-da7e-4539-be2b-119514dd688e | j-net-randomly-weighted-u-net-for-audio | 1911.12926 | null | https://arxiv.org/abs/1911.12926v1 | https://arxiv.org/pdf/1911.12926v1.pdf | J-Net: Randomly weighted U-Net for audio source separation | Several results in the computer vision literature have shown the potential of randomly weighted neural networks. While they perform fairly well as feature extractors for discriminative tasks, a positive correlation exists between their performance and their fully trained counterparts. According to these discoveries, we pose two questions: what is the value of randomly weighted networks in difficult generative audio tasks such as audio source separation and does such positive correlation still exist when it comes to large random networks and their trained counterparts? In this paper, we demonstrate that the positive correlation still exists. Based on this discovery, we can try out different architecture designs or tricks without training the whole model. Meanwhile, we find a surprising result that in comparison to the non-trained encoder (down-sample path) in Wave-U-Net, fixing the decoder (up-sample path) to random weights results in better performance, almost comparable to the fully trained model. | ['Hung-Yi Lee', 'Yen-Min Hsu', 'Bo-Wen Chen'] | 2019-11-29 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 4.65316266e-01 2.70137787e-01 -1.18168011e-01 -1.82098895e-01
-7.24270821e-01 -4.19614255e-01 4.10568625e-01 -3.41435254e-01
-4.40769196e-01 5.25246143e-01 3.76908451e-01 -2.60965884e-01
-4.75989670e-01 -6.43960357e-01 -6.29239917e-01 -9.62572753e-01
-3.37721556e-01 3.79082203e-01 3.42317194e-01 -2.95162201e-01
-1.01256303e-01 2.86528558e-01 -1.45825601e+00 1.17096201e-01
2.83460796e-01 1.02924180e+00 1.92599058e-01 7.20203996e-01
2.41020307e-01 9.46686327e-01 -4.65086788e-01 -3.90658140e-01
3.24772060e-01 -3.91808152e-01 -5.80364823e-01 -2.37438709e-01
4.50678676e-01 -2.10004047e-01 -6.01354539e-01 1.05924058e+00
8.16522837e-01 -3.72586101e-02 5.08607030e-01 -1.11738658e+00
-5.73101282e-01 1.31899548e+00 -4.50918525e-01 3.91789407e-01
4.01157998e-02 1.64864764e-01 1.62494171e+00 -7.92149186e-01
4.20765519e-01 1.09741569e+00 9.02308226e-01 4.61283356e-01
-1.26558137e+00 -7.75026977e-01 6.59519583e-02 2.60484993e-01
-1.30676818e+00 -8.31974864e-01 9.02796090e-01 -3.90021265e-01
9.30695117e-01 1.95280179e-01 6.71986401e-01 1.43908262e+00
-1.74915522e-01 7.25905180e-01 7.73604929e-01 -4.13561881e-01
3.16788554e-02 8.80124122e-02 1.79688945e-01 6.39578521e-01
2.64703840e-01 2.60616779e-01 -7.70857394e-01 -1.00891329e-02
4.95831370e-01 -3.49771589e-01 -5.91898620e-01 -2.77890831e-01
-1.20374727e+00 6.37975633e-01 1.90295279e-01 7.99073339e-01
-1.25853807e-01 4.94661450e-01 3.12517703e-01 5.13303816e-01
3.41019452e-01 8.80742192e-01 -6.65102422e-01 -1.70106158e-01
-1.27524638e+00 1.82939768e-01 4.93866295e-01 5.76878667e-01
5.49127817e-01 8.34309280e-01 -1.85997605e-01 8.21630001e-01
1.63048148e-01 4.31226343e-01 7.89523423e-01 -8.18685234e-01
2.49109402e-01 1.72631130e-01 -4.31058615e-01 -9.97728229e-01
-6.35552049e-01 -1.05519521e+00 -8.90962541e-01 1.29789412e-01
6.88060403e-01 -5.62384188e-01 -7.24788249e-01 2.21663785e+00
-1.53033063e-01 4.85912919e-01 -1.68446481e-01 7.50301003e-01
6.93474472e-01 4.81230170e-01 -2.53963888e-01 -3.97750326e-02
1.11768508e+00 -6.22387350e-01 -3.67785901e-01 -4.45062160e-01
2.45694906e-01 -6.49928689e-01 9.53452110e-01 5.08906066e-01
-9.52707231e-01 -6.63711965e-01 -1.34435999e+00 1.83792263e-01
-1.03689171e-01 4.05564792e-02 7.17716277e-01 6.87604129e-01
-1.04611719e+00 9.70793128e-01 -5.74371696e-01 -1.41721115e-01
4.25007880e-01 5.46264112e-01 -2.26842746e-01 1.34107143e-01
-1.23027694e+00 9.15214777e-01 3.44591975e-01 9.77881625e-02
-1.07935703e+00 -6.73985541e-01 -6.34651124e-01 3.46202880e-01
3.28340054e-01 -7.12942064e-01 1.17973399e+00 -1.24987483e+00
-1.50652099e+00 5.42789519e-01 -7.12094754e-02 -7.29147792e-01
2.52157182e-01 -5.31111136e-02 -3.33433181e-01 8.25829208e-02
-4.02922094e-01 6.12688124e-01 1.03600097e+00 -1.05540144e+00
-5.24329782e-01 -2.00877205e-01 3.87459323e-02 -2.92543679e-01
-5.93299270e-01 1.82415675e-02 4.55327071e-02 -9.11936462e-01
7.78226182e-02 -9.19141471e-01 -1.35650218e-01 -2.30454341e-01
-4.56752241e-01 -2.10666955e-01 5.10647297e-01 -3.20846826e-01
1.19576120e+00 -2.29837012e+00 2.59621650e-01 2.36616984e-01
3.81168693e-01 3.90627980e-01 -5.34135282e-01 2.08119154e-01
-4.64650571e-01 6.90246075e-02 -5.15077338e-02 -8.02197754e-02
1.21463820e-01 6.40429780e-02 -4.49961513e-01 3.21413517e-01
4.52110589e-01 9.05563414e-01 -9.00304675e-01 -2.16958046e-01
-7.14099333e-02 5.84256053e-01 -5.71919858e-01 -1.27079263e-01
-2.11392506e-03 1.00548780e-02 -2.42368937e-01 3.63518089e-01
4.08295453e-01 -1.57589853e-01 3.40923548e-01 -2.36437738e-01
3.20620351e-02 3.99831891e-01 -1.07963753e+00 1.53435266e+00
-2.24754736e-01 1.20442712e+00 2.58404808e-03 -1.25857401e+00
5.70029020e-01 5.73363900e-01 5.37612975e-01 -5.10222554e-01
2.88277954e-01 1.39841035e-01 6.93605781e-01 -4.68675584e-01
3.20088118e-01 -5.29321969e-01 1.33004576e-01 3.80836666e-01
8.67225647e-01 1.96142383e-02 5.72290830e-02 4.10172530e-03
1.49784780e+00 -2.01061159e-01 8.45291689e-02 -1.73547387e-01
2.45663792e-01 -3.33311319e-01 3.92879903e-01 8.03115010e-01
-1.10331420e-02 9.15004730e-01 5.03687024e-01 -9.98739004e-02
-9.08089340e-01 -1.13408327e+00 -1.93171859e-01 1.19021773e+00
-2.80549973e-01 -5.06272733e-01 -5.95528185e-01 -4.10997063e-01
-1.76897362e-01 4.27027971e-01 -5.86220026e-01 -4.03568596e-01
-5.10484099e-01 -7.67619014e-01 8.59128237e-01 4.81625497e-01
1.55145660e-01 -8.96329224e-01 -5.83079100e-01 1.89349875e-01
4.63659279e-02 -1.07733357e+00 -8.79606754e-02 7.54919767e-01
-7.75794446e-01 -8.62255096e-01 -8.52112234e-01 -5.80917120e-01
2.32946783e-01 4.26288217e-01 1.17960489e+00 -1.06997766e-01
-7.60138556e-02 3.36603761e-01 -2.85053730e-01 -7.00515985e-01
-9.86937732e-02 4.43507373e-01 1.98066518e-01 -7.78168365e-02
1.66819513e-01 -1.04703367e+00 -4.65817183e-01 2.47470349e-01
-8.51747155e-01 -3.09356689e-01 6.81345463e-01 8.18362534e-01
5.29711135e-02 3.03688496e-01 7.95455396e-01 -6.64612174e-01
7.35545039e-01 -4.14566875e-01 -2.02598184e-01 1.00350775e-01
-5.66464543e-01 2.84950644e-01 5.00776172e-01 -7.13363886e-01
-6.63281143e-01 -4.70974622e-03 -5.07006049e-01 -2.55094528e-01
1.15925044e-01 3.29234093e-01 1.46833695e-02 2.83016954e-02
6.77064240e-01 2.77089253e-02 -1.23674244e-01 -4.28765118e-01
3.11933786e-01 4.35202122e-01 3.57670248e-01 -4.34996039e-01
1.19337630e+00 3.12237799e-01 -3.11134160e-01 -8.73458445e-01
-9.82988834e-01 -1.43382430e-01 -3.43826175e-01 -1.91816688e-01
7.95081437e-01 -7.94570923e-01 -5.26702821e-01 2.52396166e-01
-9.57331955e-01 -4.04187083e-01 -7.32038379e-01 6.38967812e-01
-3.81237656e-01 -7.87364244e-02 -3.85479301e-01 -7.81845629e-01
-1.27011642e-01 -1.03285980e+00 6.67208612e-01 7.01483116e-02
-5.22631943e-01 -8.66989493e-01 1.43056735e-01 7.09191263e-02
7.71286786e-01 -2.11894587e-01 1.00893879e+00 -8.04316521e-01
-2.86657453e-01 -2.20618442e-01 -1.32091179e-01 6.51550651e-01
-3.45401615e-02 1.43162861e-01 -1.43121588e+00 -1.94173548e-02
8.67632851e-02 -3.63311946e-01 1.33369219e+00 4.46832776e-01
1.03195369e+00 -1.35333389e-01 -3.29915690e-03 5.26876390e-01
1.10721815e+00 7.12403953e-02 5.68737626e-01 3.77073511e-02
4.64834541e-01 4.74816591e-01 -2.47502640e-01 2.91711777e-01
3.63655686e-02 6.72364116e-01 4.34154630e-01 4.21829484e-02
-4.44331437e-01 -3.23327720e-01 5.94513595e-01 9.72559214e-01
-3.98942143e-01 -2.46421859e-01 -6.83070123e-01 3.43548656e-01
-1.62928843e+00 -1.33544874e+00 3.21079530e-02 2.21285582e+00
8.15829039e-01 6.05668187e-01 7.20991045e-02 6.96714520e-01
3.78831565e-01 6.22899771e-01 -2.02893317e-01 -1.89600542e-01
-2.82438725e-01 6.24528050e-01 3.38991791e-01 2.65388966e-01
-9.55746651e-01 5.05967855e-01 7.38800907e+00 1.27629924e+00
-1.37000728e+00 2.59300500e-01 4.30622995e-01 -4.56717163e-01
-5.23465991e-01 -1.06666990e-01 -5.68991721e-01 3.75871271e-01
1.19911969e+00 3.97882499e-02 4.16685492e-01 5.74685752e-01
-1.21254446e-02 9.86476988e-02 -1.20750296e+00 8.06409299e-01
6.50353655e-02 -1.27029109e+00 -1.91180140e-01 1.40722826e-01
5.22780478e-01 1.65021434e-01 3.41438174e-01 5.14005899e-01
4.12313551e-01 -1.16678190e+00 9.07953024e-01 3.37656438e-01
5.46726286e-01 -5.19077063e-01 5.92405260e-01 3.28305840e-01
-8.33778918e-01 -1.63599044e-01 -4.78552639e-01 -1.34028330e-01
1.72829255e-01 9.63169336e-01 -7.05207646e-01 2.45913193e-01
3.29447865e-01 4.35356259e-01 -5.13093829e-01 1.11869049e+00
-2.30438292e-01 1.04488444e+00 -2.21890152e-01 8.05919841e-02
2.49727145e-01 9.85907540e-02 6.18930638e-01 1.23649681e+00
4.68058407e-01 -4.14284140e-01 -5.01760781e-01 5.58998764e-01
-2.22789153e-01 -2.71479666e-01 -8.24554622e-01 -2.69231170e-01
6.68257326e-02 1.16312110e+00 -5.90494215e-01 -8.90821368e-02
-2.93854177e-01 6.45890474e-01 3.25292856e-01 2.79168099e-01
-9.10663664e-01 -3.78058374e-01 6.66793823e-01 3.09753507e-01
4.07922953e-01 -3.01425964e-01 -1.32953569e-01 -9.67156231e-01
-7.21801072e-02 -9.23962593e-01 -3.71298678e-02 -6.94208622e-01
-1.16262472e+00 6.96950793e-01 -3.14701915e-01 -1.15736783e+00
-2.34859809e-01 -5.57372093e-01 -5.99716246e-01 3.55119973e-01
-1.44079828e+00 -7.41931140e-01 7.56536797e-02 4.91135567e-01
5.00948310e-01 -3.74901414e-01 9.02198374e-01 6.26095772e-01
-5.42816103e-01 8.10840547e-01 -4.07830663e-02 2.44891480e-01
6.09284043e-01 -1.04793119e+00 -1.77085325e-02 6.47888839e-01
9.65297282e-01 6.58044100e-01 8.97888541e-01 1.80105604e-02
-1.27870500e+00 -6.91445470e-01 8.13503504e-01 -3.64281535e-01
8.86210740e-01 -4.46539760e-01 -6.69931352e-01 4.29209471e-01
4.63419169e-01 1.39543344e-03 6.94288671e-01 4.33345467e-01
-7.01401353e-01 -2.37058043e-01 -7.84145951e-01 5.02950907e-01
1.16231179e+00 -4.51379508e-01 -5.32903969e-01 1.60422266e-01
7.58040547e-01 1.48418348e-03 -3.68339032e-01 5.17602623e-01
9.16542649e-01 -1.20188630e+00 1.19122875e+00 -5.87914109e-01
6.70857012e-01 1.13143370e-01 -1.96390554e-01 -1.58066916e+00
-3.02995950e-01 -5.20759225e-01 -9.71836522e-02 1.23530686e+00
8.04837108e-01 -5.16142488e-01 8.12095463e-01 1.86222419e-01
-1.45525232e-01 -9.13476229e-01 -1.19277215e+00 -9.01606500e-01
1.88493386e-01 -1.02854073e+00 1.76641256e-01 8.11956942e-01
-4.86776829e-02 5.96239030e-01 -4.73489881e-01 -1.43503428e-01
3.76996279e-01 -2.64737546e-01 3.59717995e-01 -1.42377317e+00
-7.36130774e-01 -8.05800021e-01 -5.41400313e-01 -9.51356351e-01
4.16997582e-01 -9.72894609e-01 -1.07372478e-01 -1.17456543e+00
4.26869281e-02 -3.63330156e-01 -6.52731061e-01 3.61281514e-01
9.92667750e-02 5.42012095e-01 3.57546657e-01 -6.34017214e-02
-4.68523413e-01 2.57815421e-01 9.89230037e-01 -4.11195248e-01
3.59846056e-02 3.22556406e-01 -1.13082802e+00 1.05042613e+00
7.93337822e-01 -6.29020333e-01 -4.46078658e-01 -5.48694730e-01
6.24821126e-01 -1.12373330e-01 3.25860590e-01 -1.44991350e+00
8.37861001e-02 1.90383106e-01 2.35799447e-01 -4.01733108e-02
5.54892600e-01 -8.91668975e-01 -1.31489649e-01 2.06184253e-01
-4.55642313e-01 -1.69703290e-01 -7.12983236e-02 4.64452773e-01
-3.98018301e-01 -6.99904084e-01 6.40027821e-01 -1.03205152e-01
-3.76960129e-01 1.13601880e-02 -7.03314602e-01 2.22206205e-01
6.21357322e-01 -2.30127528e-01 -1.67171821e-01 -5.97061217e-01
-7.38903224e-01 -1.72705188e-01 -2.44537115e-01 3.55082273e-01
4.83937234e-01 -1.28370106e+00 -5.64816177e-01 1.44740820e-01
-9.07560140e-02 -2.68579841e-01 1.62631273e-01 9.32636261e-01
1.50323734e-01 3.57869714e-01 -9.03495587e-03 -3.30711365e-01
-1.10630107e+00 3.05656344e-01 2.66264081e-01 -3.10795844e-01
-2.84040540e-01 1.24739385e+00 -1.53756831e-02 -1.36048809e-01
4.68408972e-01 -2.98304796e-01 -2.46279854e-02 5.18213034e-01
3.92714143e-01 3.25279891e-01 1.33789524e-01 -3.58251095e-01
-3.01538199e-01 6.29763305e-01 2.87514657e-01 -2.24899977e-01
1.65346539e+00 2.98781812e-01 2.13612512e-01 6.25237882e-01
1.13528919e+00 3.32168996e-01 -1.06767762e+00 -3.27090442e-01
8.05285648e-02 -1.33588359e-01 6.88968748e-02 -4.73859400e-01
-1.58115351e+00 1.26405394e+00 6.54079854e-01 4.21056151e-01
9.37515438e-01 1.48429662e-01 5.40171742e-01 4.99956220e-01
2.93543279e-01 -9.65941846e-01 2.30225712e-01 6.76360369e-01
7.06065774e-01 -1.06773138e+00 2.36745086e-03 9.12568197e-02
-6.09390497e-01 1.07054555e+00 2.88794875e-01 -3.86868358e-01
8.11563432e-01 5.72456717e-01 -1.17383271e-01 -1.50475964e-01
-8.65141153e-01 -5.78445494e-01 2.42969617e-01 6.83671713e-01
6.85496032e-01 -5.20233884e-02 6.59391508e-02 6.45892799e-01
-4.93286520e-01 -6.63773268e-02 3.28243554e-01 4.41252053e-01
-7.77919531e-01 -1.11379480e+00 -7.91570693e-02 5.69520593e-01
-5.85339606e-01 -3.83177876e-01 -3.46014738e-01 6.37702346e-01
3.76079828e-01 1.01299071e+00 2.29027886e-02 -8.96127641e-01
2.26076424e-01 4.40619409e-01 5.57412446e-01 -6.46471381e-01
-8.53831649e-01 -8.19910243e-02 2.71398634e-01 -3.07973742e-01
-4.67186570e-01 -4.18204337e-01 -8.60274434e-01 2.75188349e-02
-5.62063456e-01 -1.92848891e-02 4.22086000e-01 9.13064420e-01
6.66807741e-02 8.84272575e-01 4.71350849e-01 -1.00206470e+00
-7.64899671e-01 -1.12820411e+00 -6.65892184e-01 -6.64618274e-04
4.83827800e-01 -6.07053101e-01 -5.79905450e-01 -1.83873445e-01] | [15.376206398010254, 5.405327796936035] |
974e32c3-ff6f-47d6-b7da-ae66d6753465 | subdimensional-expansion-for-multi-objective | 2102.01353 | null | https://arxiv.org/abs/2102.01353v2 | https://arxiv.org/pdf/2102.01353v2.pdf | Subdimensional Expansion for Multi-objective Multi-agent Path Finding | Conventional multi-agent path planners typically determine a path that optimizes a single objective, such as path length. Many applications, however, may require multiple objectives, say time-to-completion and fuel use, to be simultaneously optimized in the planning process. Often, these criteria may not be readily compared and sometimes lie in competition with each other. Simply applying standard multi-objective search algorithms to multi-agent path finding may prove to be inefficient because the size of the space of possible solutions, i.e., the Pareto-optimal set, can grow exponentially with the number of agents (the dimension of the search space). This paper presents an approach that bypasses this so-called curse of dimensionality by leveraging our prior multi-agent work with a framework called subdimensional expansion. One example of subdimensional expansion, when applied to A*, is called M* and M* was limited to a single objective function. We combine principles of dominance and subdimensional expansion to create a new algorithm named multi-objective M* (MOM*), which dynamically couples agents for planning only when those agents have to "interact" with each other. MOM* computes the complete Pareto-optimal set for multiple agents efficiently and naturally trades off sub-optimal approximations of the Pareto-optimal set and computational efficiency. Our approach is able to find the complete Pareto-optimal set for problem instances with hundreds of solutions which the standard multi-objective A* algorithms could not find within a bounded time. | ['Howie Choset', 'Sivakumar Rathinam', 'Zhongqiang Ren'] | 2021-02-02 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-8.16070214e-02 1.23174943e-01 -2.35145718e-01 2.44115561e-01
-6.93956614e-01 -9.84658122e-01 2.76577830e-01 4.63122785e-01
-5.73802948e-01 1.22746146e+00 -9.85668674e-02 -3.33234012e-01
-9.63475704e-01 -1.06233752e+00 -2.66828328e-01 -6.73126817e-01
-4.67477292e-01 1.17852879e+00 2.30713069e-01 -4.02117610e-01
4.54178602e-01 5.46932876e-01 -1.36888778e+00 -2.60147542e-01
1.01228309e+00 5.62718034e-01 4.08780605e-01 7.60171175e-01
-1.56809956e-01 2.69403514e-02 -7.31539965e-01 1.66874811e-01
3.93753856e-01 -3.31104308e-01 -9.80165720e-01 5.50121740e-02
-6.25854671e-01 -2.58555800e-01 2.32081681e-01 7.81729698e-01
3.93609971e-01 4.78796542e-01 6.46765530e-01 -1.85792994e+00
-7.06757903e-02 5.19199610e-01 -6.13581717e-01 4.49143201e-02
4.93182242e-01 1.43626466e-01 9.69436169e-01 -1.07930608e-01
6.20654821e-01 1.30700707e+00 3.22533995e-01 3.27532321e-01
-1.33908606e+00 -8.37596040e-03 1.57710597e-01 3.10856141e-02
-1.29613161e+00 -3.75778824e-02 3.41769427e-01 -2.11461931e-01
1.16300070e+00 6.98216081e-01 7.38723397e-01 2.24221483e-01
4.54517692e-01 4.61135507e-01 9.25072372e-01 -4.17511374e-01
5.51519454e-01 -9.01708007e-02 -2.59372473e-01 4.98342961e-01
4.74136770e-01 4.31665242e-01 1.16468072e-02 -4.29059774e-01
4.85072672e-01 -3.25815886e-01 -1.76711664e-01 -6.01944149e-01
-1.29203916e+00 1.01695526e+00 6.31188899e-02 1.31624818e-01
-5.35616159e-01 2.24710986e-01 1.60401583e-01 3.64582628e-01
2.46269442e-02 9.86124218e-01 -4.20326114e-01 -2.99552679e-01
-4.72130567e-01 7.43026137e-01 9.88616943e-01 7.46551752e-01
7.43545890e-01 -2.70771086e-01 6.71927258e-02 3.54213595e-01
3.53802353e-01 1.37032807e-01 -1.64754651e-02 -1.30808020e+00
4.86015171e-01 7.24390686e-01 7.54006565e-01 -7.32629657e-01
-8.53703797e-01 -2.52342016e-01 -1.78876251e-01 1.05366290e+00
6.42558813e-01 -4.68749493e-01 -5.12256145e-01 1.65305102e+00
5.66891909e-01 -5.02413034e-01 3.38135958e-01 9.11212921e-01
-3.26248035e-02 7.76914358e-01 -9.53527763e-02 -6.98368788e-01
1.02246237e+00 -1.00571144e+00 -4.18299705e-01 -1.58189923e-01
7.48696446e-01 -7.39520371e-01 8.08951557e-01 2.62679040e-01
-1.46931303e+00 2.40552783e-01 -1.21300614e+00 5.99593997e-01
-3.92689049e-01 -6.71315670e-01 6.14401817e-01 7.18787909e-01
-1.05848122e+00 5.92971265e-01 -5.86107790e-01 -3.60940874e-01
-1.13872677e-01 7.80047596e-01 6.64299503e-02 3.23350588e-03
-8.09308231e-01 1.27620173e+00 6.92210495e-01 -3.10244679e-01
-7.84452856e-01 -4.51043338e-01 -4.14537638e-01 2.07241103e-01
1.09150684e+00 -8.80364001e-01 1.14193034e+00 -6.72281563e-01
-1.54914224e+00 1.14029072e-01 8.84831697e-02 2.97251325e-02
5.58770478e-01 4.46577877e-01 -7.29108155e-02 -6.58711269e-02
2.43524566e-01 4.79630530e-01 3.29079568e-01 -1.46603477e+00
-9.22579944e-01 -1.60065904e-01 6.64152205e-01 6.56173706e-01
-7.55447224e-02 1.67832866e-01 -3.30072604e-02 -5.27298972e-02
-8.54572654e-02 -1.07802093e+00 -9.63533998e-01 -3.84838849e-01
-2.61688799e-01 -3.43123108e-01 4.91564453e-01 2.16796659e-02
1.28210282e+00 -1.43197775e+00 6.87838197e-01 5.48109651e-01
7.85605535e-02 -1.39316946e-01 -4.21915323e-01 9.67131376e-01
2.50294507e-01 1.70989782e-01 -2.66657650e-01 -6.43305704e-02
2.93508708e-01 5.20374954e-01 2.73932934e-01 5.77442169e-01
-2.33679712e-01 6.67624593e-01 -1.15280187e+00 -4.60644662e-01
-1.36501398e-02 -2.57094920e-01 -6.54494822e-01 -2.07157165e-01
-6.01034939e-01 1.33304551e-01 -7.33054042e-01 5.88465333e-01
5.21872342e-01 -3.25929299e-02 3.86136949e-01 4.67309803e-01
-8.15565288e-01 -1.18047319e-01 -1.60526204e+00 1.39687741e+00
-3.49848896e-01 1.73823729e-01 4.55066293e-01 -8.92614901e-01
6.44682109e-01 2.41415307e-01 1.16701829e+00 -6.12531006e-01
1.01274513e-01 5.17184734e-01 2.09287927e-01 -1.87401697e-01
7.93076396e-01 -1.95811063e-01 -2.11869255e-01 9.21184421e-01
-5.60734034e-01 -1.77316383e-01 7.25839138e-01 -1.20775856e-01
1.22080278e+00 -1.02844790e-01 4.65912461e-01 -6.08530939e-01
4.06078607e-01 7.16733694e-01 4.95029539e-01 5.23446202e-01
-1.68181896e-01 -1.70044322e-02 8.06187332e-01 -3.49459738e-01
-1.03592813e+00 -9.21760082e-01 2.77714193e-01 9.66929853e-01
5.59446931e-01 -3.54516506e-01 -5.35246074e-01 -5.26387393e-01
3.18239518e-02 6.84634566e-01 -4.24767852e-01 3.30378413e-01
-7.49550164e-01 -9.28250730e-01 1.87716067e-01 -4.77428101e-02
-1.37594774e-01 -7.87449002e-01 -1.33152461e+00 7.83047736e-01
3.72451767e-02 -5.94700754e-01 -3.68083537e-01 1.72607988e-01
-6.75042152e-01 -1.21322715e+00 -6.29879951e-01 -3.34036469e-01
7.73108780e-01 2.33147651e-01 8.90225232e-01 1.97207689e-01
-1.06600374e-01 3.79031360e-01 -2.83335716e-01 -5.45455098e-01
-5.24915695e-01 8.44062492e-02 1.46994427e-01 -4.97048676e-01
-2.77608365e-01 -6.29635215e-01 -5.24791241e-01 5.30262530e-01
-8.72758687e-01 -2.28185534e-01 4.26311910e-01 4.34859365e-01
5.20507038e-01 6.87588096e-01 7.75187731e-01 -5.31730764e-02
1.11084378e+00 -5.94315648e-01 -9.92293477e-01 3.81473780e-01
-7.37286687e-01 3.05830002e-01 6.66884363e-01 -3.92481178e-01
-5.32453418e-01 -2.18342487e-02 3.88229787e-01 1.16945710e-02
1.63994640e-01 7.80008674e-01 -1.41938552e-01 -2.86289543e-01
3.76873583e-01 -1.47245407e-01 1.43363044e-01 -2.06900865e-01
3.15644175e-01 2.18969181e-01 1.83240697e-01 -8.10399115e-01
6.93926692e-01 3.04960877e-01 6.48377776e-01 -4.45949525e-01
-3.15593109e-02 -3.64658564e-01 -3.51814181e-01 -4.03612196e-01
4.77230549e-01 -4.30765599e-02 -1.34053600e+00 -1.05124511e-01
-1.31947315e+00 -2.63397127e-01 -4.73868251e-01 4.80903834e-01
-1.02056766e+00 2.69160092e-01 1.75164551e-01 -1.04786527e+00
1.41545475e-01 -1.33285737e+00 4.69414622e-01 2.12079078e-01
-4.17080492e-01 -9.49938357e-01 4.25646305e-01 4.05674949e-02
4.48945403e-01 6.67410254e-01 1.16673410e+00 -3.33068520e-01
-6.96106195e-01 1.32813044e-02 8.12035426e-02 -6.22112930e-01
7.76982233e-02 -1.40425684e-02 1.16314813e-02 -4.49404389e-01
-3.61118436e-01 8.10095295e-02 1.39227554e-01 5.61559677e-01
4.13367867e-01 -6.80915654e-01 -6.72694743e-01 9.51361805e-02
1.84990084e+00 7.75111914e-01 4.31815654e-01 1.00291765e+00
5.88168725e-02 9.23800409e-01 9.41507399e-01 8.48069727e-01
6.03257418e-01 1.01361728e+00 9.00045514e-01 2.61574924e-01
3.53069335e-01 3.22526634e-01 2.21159190e-01 1.54880360e-01
-4.93341655e-01 -7.41547823e-01 -1.09403169e+00 7.62696326e-01
-2.10685706e+00 -1.00972080e+00 -1.48880500e-02 2.20352817e+00
4.80099797e-01 5.40815704e-02 6.97914481e-01 2.55246282e-01
5.60639918e-01 -1.29237175e-01 -6.26280546e-01 -9.09654737e-01
1.17257215e-01 -1.92156881e-01 8.72268856e-01 9.04422164e-01
-7.20481694e-01 4.62563783e-01 6.77475500e+00 5.60270846e-01
-6.19634449e-01 -3.15982029e-02 2.14107558e-01 -4.27732050e-01
-5.14723003e-01 2.08512351e-01 -4.20080870e-01 3.67537349e-01
1.02670944e+00 -8.76808822e-01 9.32627320e-01 5.54597318e-01
7.08308935e-01 -5.50502777e-01 -1.07443523e+00 4.60573912e-01
-4.53805834e-01 -1.29086602e+00 -4.17267084e-01 7.12217033e-01
8.51935565e-01 -2.53553689e-01 -1.79530665e-01 -8.41437951e-02
6.64701879e-01 -1.24189389e+00 7.55629182e-01 1.28794134e-01
3.16421241e-01 -1.30386913e+00 4.53404903e-01 5.84021986e-01
-1.32881176e+00 -4.42731947e-01 -9.39867049e-02 -2.70382464e-01
8.72396708e-01 2.63495632e-02 -6.99084342e-01 7.86954582e-01
3.96821886e-01 -1.97974429e-01 2.70931840e-01 1.40725148e+00
3.31145078e-01 -4.87337828e-01 -6.59137130e-01 -3.79637271e-01
8.56013834e-01 -4.86190706e-01 1.15530276e+00 6.16968632e-01
5.42879522e-01 3.54123145e-01 5.71350873e-01 8.19882751e-01
6.63496435e-01 6.67370856e-02 -4.26389635e-01 -1.85696900e-01
5.03728390e-01 9.24866915e-01 -1.10096180e+00 9.41108763e-02
-9.26987603e-02 2.95419693e-01 -2.33269689e-04 2.79253781e-01
-9.67373371e-01 -3.96135688e-01 9.25515294e-01 -8.06477368e-02
-4.48250920e-02 -6.79269671e-01 -4.52687204e-01 -2.65765846e-01
-1.16330661e-01 -7.38392472e-01 3.92573684e-01 -2.79198736e-01
-6.82729125e-01 4.98854846e-01 4.85466570e-01 -1.17881262e+00
-6.04591668e-01 -3.95741791e-01 -6.58992589e-01 8.80395114e-01
-1.53027093e+00 -6.32881224e-01 1.02481186e-01 4.42653179e-01
4.42709923e-01 -5.21858633e-02 8.31995368e-01 1.23278536e-01
-3.69954735e-01 1.00536168e-01 3.63779634e-01 -8.71348202e-01
3.78381088e-02 -1.16962218e+00 -2.02405483e-01 6.88236415e-01
-5.96038640e-01 1.55874342e-01 1.04483593e+00 -5.61262548e-01
-1.62454128e+00 -5.51511049e-01 6.61338747e-01 -1.36707366e-01
7.75080502e-01 3.51296186e-01 -1.66182101e-01 3.83290231e-01
2.29415432e-01 -6.70323610e-01 4.35720623e-01 -1.03768259e-01
5.02225637e-01 9.72669721e-02 -1.25262845e+00 9.05903816e-01
9.14574742e-01 2.99050242e-01 -1.30080268e-01 4.40150172e-01
7.16787875e-01 -2.43484959e-01 -8.34216058e-01 3.40984672e-01
4.40956682e-01 -7.66497195e-01 1.01494265e+00 -6.69468939e-01
6.33708984e-02 -4.50838864e-01 -3.07204336e-01 -1.70577657e+00
-4.36867207e-01 -1.03772068e+00 1.26009118e-02 8.42431366e-01
4.96586084e-01 -8.02182972e-01 6.72268927e-01 8.06407511e-01
-3.54858845e-01 -9.65931475e-01 -1.32131994e+00 -1.35063612e+00
2.58251429e-01 -1.71316430e-01 1.05724204e+00 6.95445776e-01
3.68967563e-01 1.39569528e-02 -5.50801605e-02 3.26468408e-01
8.44749570e-01 3.41131896e-01 5.01525283e-01 -1.18751216e+00
-4.14248347e-01 -9.00429308e-01 5.72457910e-03 -5.48725128e-01
-7.69005641e-02 -5.17398834e-01 5.43583324e-03 -1.90568149e+00
-1.37059465e-01 -1.07588577e+00 1.23800347e-02 4.36296463e-01
4.15968895e-01 -4.34896618e-01 5.16937077e-01 2.23205745e-01
-5.15661478e-01 3.59298140e-01 1.49839878e+00 -1.53685659e-01
-8.10836613e-01 5.57442978e-02 -6.85077846e-01 5.78570247e-01
9.40964043e-01 -6.27738118e-01 -5.62166154e-01 -3.85761380e-01
3.63582879e-01 6.88323677e-01 -7.19172880e-02 -7.15766668e-01
2.89678752e-01 -1.25165439e+00 -3.79059345e-01 -5.36316514e-01
4.55410123e-01 -9.69648063e-01 6.07330084e-01 7.00504422e-01
-6.81174919e-02 4.91107821e-01 1.43503219e-01 4.50100273e-01
1.17618032e-01 -6.67569876e-01 4.77749646e-01 -3.38743389e-01
-6.90813482e-01 1.19208276e-01 -6.47832692e-01 -3.04448515e-01
1.73145342e+00 -5.01286209e-01 -5.30347884e-01 -1.80549279e-01
-5.02043128e-01 7.89084554e-01 6.09297395e-01 1.88615888e-01
4.70721126e-01 -1.14816630e+00 -7.29797542e-01 -5.87923825e-01
-3.05939883e-01 -1.21897161e-01 1.10103421e-01 8.23816359e-01
-5.59798717e-01 6.37940407e-01 -4.80409801e-01 -1.69908166e-01
-1.19928765e+00 7.51499057e-01 3.98492247e-01 -7.10495710e-01
-3.21185887e-01 4.18112963e-01 -1.45336613e-01 -1.89291328e-01
-1.50241062e-01 -5.67227677e-02 -1.69276595e-01 3.39663327e-01
3.43810081e-01 9.79502976e-01 -3.29476625e-01 -4.28464264e-01
-7.13095665e-01 6.06535375e-01 4.22593355e-01 -6.93707347e-01
1.47243619e+00 -3.15567106e-01 -3.58070314e-01 -1.85476139e-01
7.66035140e-01 -3.16821709e-02 -1.03057981e+00 3.38506639e-01
1.06409691e-01 -5.44564545e-01 -4.02105451e-02 -9.22960877e-01
-7.88267076e-01 5.95240556e-02 1.55014694e-01 6.07848644e-01
1.16319275e+00 -1.09319977e-01 6.02692008e-01 4.31830287e-01
6.40106857e-01 -1.29439175e+00 -7.79143302e-03 5.25245488e-01
8.83431971e-01 -7.87715554e-01 2.32375339e-01 -3.42967242e-01
-4.11808401e-01 1.24432337e+00 5.07408559e-01 5.67307137e-02
2.47713357e-01 5.03661990e-01 -1.98568955e-01 -1.21143945e-01
-8.36043477e-01 -4.28719819e-01 -1.40230805e-01 6.85614467e-01
-2.88122833e-01 4.10913140e-01 -7.78564095e-01 2.16425866e-01
-1.91552132e-01 -2.50780702e-01 8.76839638e-01 1.21989393e+00
-8.63089442e-01 -1.52176237e+00 -8.76045048e-01 1.04401477e-01
-3.51696871e-02 3.98022830e-01 -1.51571617e-01 1.02096403e+00
8.73406380e-02 1.34259176e+00 4.66033816e-03 -9.01053324e-02
3.85079741e-01 -4.26880300e-01 5.76631308e-01 -4.19250727e-01
-4.34543639e-01 -6.17592409e-02 4.91431445e-01 -4.80856597e-01
-4.80425060e-01 -6.72506928e-01 -1.55539000e+00 -7.83369839e-01
-8.14592838e-02 3.80342931e-01 7.26828277e-01 1.05261588e+00
5.12899011e-02 5.65893471e-01 5.91180861e-01 -1.02941799e+00
-5.88790596e-01 -2.25865319e-01 -4.86856520e-01 -3.80316734e-01
3.56455922e-01 -8.29881787e-01 -2.00021733e-03 -6.60559773e-01] | [4.957076549530029, 1.9174559116363525] |
4ef05f8e-1c5b-4377-90b8-41e494d21398 | visual-semantic-information-pursuit-a-survey | 1903.05434 | null | http://arxiv.org/abs/1903.05434v1 | http://arxiv.org/pdf/1903.05434v1.pdf | Visual Semantic Information Pursuit: A Survey | Visual semantic information comprises two important parts: the meaning of
each visual semantic unit and the coherent visual semantic relation conveyed by
these visual semantic units. Essentially, the former one is a visual perception
task while the latter one corresponds to visual context reasoning. Remarkable
advances in visual perception have been achieved due to the success of deep
learning. In contrast, visual semantic information pursuit, a visual scene
semantic interpretation task combining visual perception and visual context
reasoning, is still in its early stage. It is the core task of many different
computer vision applications, such as object detection, visual semantic
segmentation, visual relationship detection or scene graph generation. Since it
helps to enhance the accuracy and the consistency of the resulting
interpretation, visual context reasoning is often incorporated with visual
perception in current deep end-to-end visual semantic information pursuit
methods. However, a comprehensive review for this exciting area is still
lacking. In this survey, we present a unified theoretical paradigm for all
these methods, followed by an overview of the major developments and the future
trends in each potential direction. The common benchmark datasets, the
evaluation metrics and the comparisons of the corresponding methods are also
introduced. | ['Daqi Liu', 'Josef Kittler', 'Miroslaw Bober'] | 2019-03-13 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 5.77900171e-01 2.51277909e-02 -2.46356130e-01 -3.69373411e-01
-6.36715665e-02 -4.90302891e-01 6.18064225e-01 4.13788438e-01
-1.15619905e-01 3.65576476e-01 5.65067232e-02 -8.43879357e-02
-1.15253188e-01 -5.09725988e-01 -5.58504045e-01 -7.82430649e-01
4.68967795e-01 2.12402031e-01 4.39740717e-01 -2.31352486e-02
4.93223935e-01 1.98714510e-01 -1.70520508e+00 4.74288046e-01
7.01963305e-01 1.19642019e+00 7.45294631e-01 3.49393934e-01
-6.25146627e-01 8.26054215e-01 -3.53860885e-01 -1.85331434e-01
-1.16869435e-01 -5.79574645e-01 -9.42044318e-01 6.69254243e-01
4.94772732e-01 2.93399483e-01 -1.52102277e-01 1.44932044e+00
1.28706843e-01 3.21246237e-01 5.48611045e-01 -1.56041181e+00
-7.58984268e-01 2.16718614e-01 -5.64337075e-01 2.28279382e-01
2.63128787e-01 1.10994879e-04 1.32133937e+00 -9.35454965e-01
7.32930601e-01 1.16246235e+00 2.60606647e-01 2.89332062e-01
-1.20417452e+00 -6.67652255e-03 5.21798790e-01 7.97906280e-01
-1.01706374e+00 -1.58220768e-01 1.24596846e+00 -6.86807632e-01
7.68943131e-01 2.66757786e-01 1.05851090e+00 1.00555623e+00
-7.67552555e-02 1.01403344e+00 1.17335224e+00 -4.62971747e-01
3.40365112e-01 8.24966282e-02 4.55244660e-01 5.81691682e-01
1.63802758e-01 -7.24352524e-02 -3.74387980e-01 2.33478591e-01
6.36711597e-01 1.67854279e-01 -1.41025022e-01 -7.95176446e-01
-9.99905348e-01 7.74648964e-01 8.30926776e-01 4.15836692e-01
-3.97775948e-01 2.80448347e-01 4.13784951e-01 -1.35367975e-01
1.40362844e-01 1.34060651e-01 1.46100121e-02 4.11666512e-01
-8.57174397e-01 3.08124185e-01 4.16571498e-01 8.41653287e-01
6.89031005e-01 2.36740038e-01 -2.91476965e-01 9.21133816e-01
5.54123998e-01 3.05654615e-01 2.05614164e-01 -1.09508395e+00
8.98065418e-02 8.49290907e-01 -7.30087459e-02 -1.27947807e+00
-4.65929747e-01 -2.95014799e-01 -6.80358469e-01 3.17550868e-01
3.93309474e-01 2.55346447e-01 -1.08096814e+00 1.51942050e+00
3.67388010e-01 2.01248065e-01 4.07313835e-03 1.24110603e+00
1.26271069e+00 5.41121066e-01 5.11255741e-01 -1.59126550e-01
1.82487750e+00 -9.87940073e-01 -8.67235243e-01 -9.10296202e-01
-2.80886471e-01 -9.35039520e-01 9.30592299e-01 -4.40115891e-02
-9.50707853e-01 -6.61150157e-01 -1.11310124e+00 -5.79614460e-01
-5.44890940e-01 -1.41774900e-02 8.54766011e-01 1.66664332e-01
-1.01880288e+00 1.45698652e-01 -5.03304183e-01 -5.98851025e-01
6.86182082e-01 -3.44443060e-02 -2.96438485e-01 -2.40980238e-01
-1.08373523e+00 7.03005075e-01 6.49217069e-01 1.66824386e-01
-7.62925565e-01 -3.32436174e-01 -9.90017235e-01 1.89362183e-01
5.74332058e-01 -9.66912627e-01 9.75918770e-01 -1.22633529e+00
-9.33935881e-01 1.29904974e+00 -2.80392349e-01 -3.24720532e-01
2.94429660e-01 1.04872316e-01 -4.24715951e-02 3.70528430e-01
2.76640564e-01 7.34248161e-01 8.93018425e-01 -1.73485780e+00
-7.89329052e-01 -5.03858447e-01 -1.18014358e-01 4.10489857e-01
2.71566659e-01 -1.85590982e-02 -9.06440556e-01 -6.41376495e-01
3.52513462e-01 -7.47915447e-01 -3.40138912e-01 9.75715220e-02
-5.26129186e-01 -4.62620884e-01 9.13435876e-01 -4.07847792e-01
7.45738804e-01 -2.06718946e+00 3.75247151e-01 1.01947635e-01
4.13439035e-01 3.04539949e-01 6.42892122e-02 3.43162566e-01
-2.32341543e-01 -3.24383348e-01 -5.78752637e-01 -4.08374727e-01
-1.41559288e-01 2.42010921e-01 -4.47865099e-01 4.08814281e-01
6.36988431e-02 1.17873287e+00 -1.08662474e+00 -5.54845572e-01
6.68518364e-01 4.52568948e-01 -1.73569515e-01 1.32225946e-01
-6.45815253e-01 5.97985506e-01 -3.97888124e-01 6.48322284e-01
3.95470411e-01 -6.57144845e-01 4.45329815e-01 -3.13661069e-01
-3.84212099e-02 -5.32978028e-02 -9.27670121e-01 1.87096655e+00
-8.15104172e-02 1.09010744e+00 -4.72749211e-02 -1.49720943e+00
9.39112186e-01 3.27500291e-02 4.57025260e-01 -8.53620529e-01
1.56323552e-01 -3.94346491e-02 -2.63767093e-01 -6.75611377e-01
4.65973765e-01 -1.68349385e-01 1.91508606e-01 -8.27199593e-02
-1.01954430e-01 -1.61543176e-01 1.82267115e-01 6.88368976e-02
2.24002600e-01 1.30773738e-01 7.46304512e-01 -9.72701237e-02
6.48786724e-01 3.37211668e-01 5.12373209e-01 3.96877170e-01
-4.88941997e-01 5.53395987e-01 5.92397392e-01 -4.01742369e-01
-8.64425063e-01 -1.18544197e+00 -9.73559991e-02 8.79632235e-01
8.52184832e-01 -3.95282835e-01 -6.44619286e-01 -5.61958075e-01
-3.52197699e-02 7.69103050e-01 -7.28500724e-01 -7.80494288e-02
-3.24787766e-01 -6.16563439e-01 -1.22321680e-01 7.64983714e-01
7.11209059e-01 -1.38852179e+00 -7.24081278e-01 -6.94213510e-02
-3.29807043e-01 -1.37671006e+00 -7.20512196e-02 -6.08122572e-02
-8.32958460e-01 -1.25848639e+00 -6.16709292e-01 -1.15606964e+00
5.57568252e-01 7.21389472e-01 1.10931492e+00 4.27936278e-02
-4.94419128e-01 5.43347538e-01 -2.29844153e-01 -3.79421294e-01
-5.63934930e-02 -5.00758350e-01 -4.63123620e-01 1.84208468e-01
4.70154852e-01 -4.45425630e-01 -6.99821949e-01 1.06534421e-01
-8.79372776e-01 3.46923351e-01 4.53795880e-01 7.70813882e-01
9.91349399e-01 -1.36925831e-01 2.09627822e-01 -8.23194623e-01
2.79436588e-01 -3.46040338e-01 -5.73588610e-01 4.36550051e-01
-4.94672477e-01 -1.96276177e-02 3.42572778e-01 9.40871239e-02
-1.07339561e+00 2.43133724e-01 -4.80602719e-02 -4.35183883e-01
-4.43970859e-01 4.51420516e-01 -2.85035282e-01 1.25171110e-01
3.81882668e-01 4.16934520e-01 2.02818867e-02 -4.33470607e-01
7.81643033e-01 2.51730859e-01 6.92396104e-01 -1.57136023e-01
3.21331203e-01 7.06141174e-01 2.51494110e-01 -1.23751068e+00
-1.17315435e+00 -7.98593283e-01 -6.06039584e-01 -4.12261933e-01
1.43734658e+00 -5.05045414e-01 -6.28852487e-01 3.47749949e-01
-1.38480628e+00 -8.72431025e-02 -2.33318239e-01 1.61824331e-01
-8.75813484e-01 7.11930931e-01 -1.26341432e-01 -6.06940329e-01
-2.25718901e-01 -1.47082520e+00 1.24637175e+00 4.51777518e-01
-3.70477885e-02 -1.30504823e+00 -1.53073177e-01 7.31247008e-01
-6.00642823e-02 4.16113853e-01 1.08829713e+00 -4.08613354e-01
-6.47825539e-01 1.03385821e-01 -8.41296434e-01 2.35259533e-01
-4.25135382e-02 -1.52147800e-01 -1.11315632e+00 7.38958418e-02
-1.53926313e-02 -4.77099121e-02 1.08989394e+00 6.90761268e-01
1.07226217e+00 5.78124896e-02 -5.76337278e-01 5.07371306e-01
1.67378962e+00 2.20251486e-01 5.36340117e-01 9.55248177e-02
1.00083196e+00 1.03732359e+00 7.12754667e-01 -6.28689304e-02
4.13994372e-01 9.45325315e-01 8.10990214e-01 -2.42033362e-01
-5.94936967e-01 -1.81576297e-01 8.16562846e-02 4.97523099e-01
-1.31534755e-01 -4.62952964e-02 -1.07415438e+00 4.73480642e-01
-2.16270137e+00 -9.86213565e-01 -3.77836496e-01 1.93525267e+00
2.84056276e-01 -3.96938920e-02 5.36787994e-02 1.16469897e-01
9.25008655e-01 4.38667327e-01 -5.97598195e-01 -1.35391355e-01
-2.37297997e-01 -3.53768677e-01 9.55673754e-02 4.63467032e-01
-1.20897007e+00 1.35409629e+00 6.05573273e+00 6.81150138e-01
-1.01238823e+00 1.60148498e-02 5.75371683e-01 3.88120800e-01
-6.94770366e-02 1.85946226e-01 -4.84376490e-01 2.64267713e-01
5.04534505e-02 9.15517956e-02 3.41158301e-01 1.09908736e+00
1.52232066e-01 -3.58357131e-01 -9.61218059e-01 1.42624366e+00
3.19165528e-01 -1.53429556e+00 3.55030566e-01 -1.74021453e-01
5.88672936e-01 -1.59866646e-01 1.11082189e-01 -2.13842720e-01
-1.28058642e-01 -8.95426095e-01 7.50507355e-01 5.13841450e-01
6.47096813e-01 -6.38323009e-01 4.51434344e-01 -8.96405578e-02
-1.47001112e+00 -8.11949298e-02 -4.54383224e-01 1.49021313e-01
3.23575586e-01 5.56306422e-01 -5.92829347e-01 5.30759692e-01
7.04904974e-01 1.16017818e+00 -5.47169924e-01 1.17743206e+00
-6.02807403e-01 3.17674220e-01 2.36367106e-01 5.78482337e-02
3.96142215e-01 -3.65542650e-01 8.70317340e-01 1.11983597e+00
-3.36196750e-01 -2.17636138e-01 4.53216255e-01 1.18849790e+00
1.25167757e-01 1.14374124e-01 -6.10647380e-01 -7.89556801e-02
1.38965249e-01 1.27535164e+00 -1.16648471e+00 -4.37633902e-01
-5.42526960e-01 1.07765198e+00 2.78925210e-01 5.46433866e-01
-6.72529340e-01 -2.33362749e-01 9.50894415e-01 -1.04661547e-01
1.22001834e-01 -3.39892268e-01 -7.11188793e-01 -9.29661334e-01
-8.23010355e-02 -4.12936956e-01 4.72467244e-01 -9.58856940e-01
-1.12713957e+00 3.49993467e-01 7.18369335e-02 -1.10214055e+00
-1.08647786e-01 -8.92160535e-01 -5.34283698e-01 6.86949492e-01
-1.52594638e+00 -1.09166050e+00 -6.96251631e-01 5.58638036e-01
9.42523837e-01 1.72202513e-01 4.33937639e-01 -1.27506807e-01
-4.06408101e-01 -4.09872644e-02 -3.22295688e-02 2.50940602e-02
1.73718780e-01 -1.38964200e+00 1.84873685e-01 9.44886863e-01
4.43978220e-01 3.19947660e-01 7.59315968e-01 -5.60915053e-01
-1.09056842e+00 -9.57610786e-01 8.62669826e-01 -3.12769353e-01
5.24803281e-01 -3.65591913e-01 -1.02035725e+00 4.70229477e-01
3.09604079e-01 -2.78217997e-03 5.57624996e-01 -2.01865658e-01
-1.62875444e-01 1.72395119e-03 -6.09919250e-01 8.54142666e-01
1.04970622e+00 -5.71877778e-01 -8.36102128e-01 3.81049216e-01
7.31509566e-01 -1.48538768e-01 -3.24017376e-01 2.22873732e-01
2.87776738e-01 -1.05888224e+00 1.39186907e+00 -7.07762182e-01
3.63365680e-01 -6.15941942e-01 -2.22213179e-01 -9.93485212e-01
-3.31147373e-01 -1.87221393e-01 1.39666989e-01 1.02171123e+00
-9.20344964e-02 -1.51280075e-01 6.17106318e-01 2.73007572e-01
-1.96649656e-01 -4.95751470e-01 -5.96586227e-01 -3.64382088e-01
-3.32631528e-01 -7.69068003e-01 4.55120672e-03 9.14648533e-01
-1.16489232e-01 8.08398128e-01 -1.58575952e-01 -5.29625118e-02
9.62827981e-01 7.40453839e-01 3.86843473e-01 -1.18502736e+00
-1.61795989e-01 -8.99892569e-01 -7.77356684e-01 -1.04954195e+00
1.83129326e-01 -1.22067034e+00 6.13330957e-03 -2.10501456e+00
3.75270188e-01 7.35357702e-02 -1.97410882e-01 3.62682045e-01
-4.74198192e-01 2.64186054e-01 4.93426025e-01 1.60051808e-01
-7.17897713e-01 5.41579843e-01 1.40209603e+00 -2.38577574e-01
-1.93350852e-01 -6.14269264e-02 -7.32147336e-01 8.71506751e-01
6.78540051e-01 1.56600438e-02 -6.75322831e-01 -2.85772294e-01
7.52483904e-02 -9.23788249e-02 7.88993120e-01 -7.34211504e-01
2.55026788e-01 -2.99251735e-01 3.30181509e-01 -4.89810020e-01
3.49918664e-01 -7.87271082e-01 -9.48509201e-02 3.21309865e-01
-3.08805078e-01 -1.07383594e-01 7.87060857e-02 9.35747206e-01
-5.39042473e-01 -1.59313247e-01 1.06343317e+00 -2.61753470e-01
-1.62323427e+00 1.18315868e-01 -3.07556897e-01 1.59187570e-01
1.22794187e+00 -4.42114890e-01 -1.98210761e-01 -8.99542570e-02
-1.07227004e+00 1.81186020e-01 2.62671918e-01 6.62142456e-01
9.71637487e-01 -1.19609857e+00 -3.70633364e-01 -7.57741705e-02
4.14818048e-01 5.75284548e-02 5.94634473e-01 9.73094106e-01
-2.83171892e-01 5.23467302e-01 -3.21420312e-01 -1.05256283e+00
-1.42873931e+00 1.02956092e+00 2.05685377e-01 1.18653476e-01
-7.88052976e-01 6.20680809e-01 8.69839966e-01 3.49347681e-01
2.84577549e-01 -2.39323959e-01 -7.52564132e-01 -2.42091101e-02
5.81282735e-01 3.03677559e-01 -2.00115457e-01 -8.52155089e-01
-4.76382881e-01 8.60805809e-01 3.38656038e-01 2.04650372e-01
9.45736945e-01 -5.05237520e-01 -2.75073320e-01 5.69668829e-01
1.06825006e+00 -3.42091560e-01 -1.31759226e+00 -3.17327976e-01
2.21286461e-01 -2.71692485e-01 1.03808410e-01 -7.01499522e-01
-1.28864157e+00 1.24760187e+00 4.90122795e-01 2.36098304e-01
1.30047929e+00 4.77709651e-01 4.44520593e-01 -7.06339404e-02
9.02586356e-02 -9.53220963e-01 2.61608750e-01 3.39904726e-01
1.01166856e+00 -1.50058711e+00 -5.17830588e-02 -9.43760693e-01
-1.00667906e+00 1.05185461e+00 4.54104573e-01 1.54886460e-02
6.20145559e-01 -2.96126723e-01 2.98013557e-02 -5.15927315e-01
-2.84108162e-01 -8.44217539e-01 8.01152706e-01 7.69599497e-01
3.35337579e-01 5.28050736e-02 -2.77978778e-01 5.04139423e-01
1.90400228e-01 -3.25133681e-01 5.69011793e-02 5.90558469e-01
-5.83289862e-01 -8.43369782e-01 -2.25566044e-01 -2.25598402e-02
-2.41012037e-01 -7.64973983e-02 -6.34267569e-01 6.63635910e-01
1.60849601e-01 9.85422790e-01 1.71032310e-01 -1.07325122e-01
2.03850403e-01 -7.46472832e-03 4.79581177e-01 -6.14399850e-01
-1.64317325e-01 1.69305757e-01 -1.23610742e-01 -7.72906244e-01
-7.04332530e-01 -5.32682836e-01 -1.55649328e+00 1.33924335e-01
2.45189786e-01 -2.64035106e-01 7.91049242e-01 1.02599502e+00
1.27283603e-01 6.10958695e-01 1.43346950e-01 -7.36338735e-01
1.86447069e-01 -4.14651334e-01 -5.01191318e-01 7.82010138e-01
1.60921335e-01 -9.91520882e-01 2.80288216e-02 2.73001224e-01] | [10.268915176391602, 1.4763625860214233] |
34576ea7-4a92-48f4-9c41-8557a02f0eb8 | avoiding-reasoning-shortcuts-adversarial | 1906.07132 | null | https://arxiv.org/abs/1906.07132v1 | https://arxiv.org/pdf/1906.07132v1.pdf | Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QA | Multi-hop question answering requires a model to connect multiple pieces of evidence scattered in a long context to answer the question. In this paper, we show that in the multi-hop HotpotQA (Yang et al., 2018) dataset, the examples often contain reasoning shortcuts through which models can directly locate the answer by word-matching the question with a sentence in the context. We demonstrate this issue by constructing adversarial documents that create contradicting answers to the shortcut but do not affect the validity of the original answer. The performance of strong baseline models drops significantly on our adversarial evaluation, indicating that they are indeed exploiting the shortcuts rather than performing multi-hop reasoning. After adversarial training, the baseline's performance improves but is still limited on the adversarial evaluation. Hence, we use a control unit that dynamically attends to the question at different reasoning hops to guide the model's multi-hop reasoning. We show that this 2-hop model trained on the regular data is more robust to the adversaries than the baseline model. After adversarial training, this 2-hop model not only achieves improvements over its counterpart trained on regular data, but also outperforms the adversarially-trained 1-hop baseline. We hope that these insights and initial improvements will motivate the development of new models that combine explicit compositional reasoning with adversarial training. | ['Yichen Jiang', 'Mohit Bansal'] | 2019-06-17 | avoiding-reasoning-shortcuts-adversarial-1 | https://aclanthology.org/P19-1262 | https://aclanthology.org/P19-1262.pdf | acl-2019-7 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 2.03505471e-01 5.48584640e-01 9.38753858e-02 -3.49963903e-01
-1.60160601e+00 -1.33222806e+00 5.67281544e-01 1.25797898e-01
-3.68918002e-01 5.85210502e-01 4.19737011e-01 -7.91803598e-01
8.53744000e-02 -1.00175011e+00 -1.11062634e+00 -2.31973335e-01
4.48113948e-01 5.97631991e-01 8.60638916e-01 -6.57782257e-01
6.87821582e-02 1.06114924e-01 -8.67166877e-01 8.27535629e-01
7.53157318e-01 4.64961201e-01 -3.54922801e-01 1.11678720e+00
-3.22021633e-01 1.25083637e+00 -9.16844249e-01 -1.18183756e+00
4.35263187e-01 -5.35515189e-01 -1.29127765e+00 -6.04550660e-01
1.08143222e+00 -4.93206382e-01 -4.14170235e-01 1.02838910e+00
2.19703645e-01 6.13272749e-02 2.95235813e-01 -1.17977452e+00
-9.68902469e-01 9.09480512e-01 -1.81736559e-01 3.81109744e-01
5.58103263e-01 3.09258252e-01 1.64549458e+00 -7.33611643e-01
8.03950250e-01 1.28893483e+00 8.35308611e-01 8.13086808e-01
-1.08928978e+00 -4.40289229e-01 2.93977439e-01 4.07033771e-01
-7.88255274e-01 -2.63887167e-01 7.55175173e-01 -1.25551075e-01
9.58466351e-01 6.51512861e-01 1.00340448e-01 1.23701930e+00
9.60369259e-02 7.95020163e-01 8.99716496e-01 -3.72901082e-01
9.82056023e-04 -3.03304717e-02 3.61049473e-01 8.42673957e-01
-2.21351460e-01 2.73243394e-02 -1.84979022e-01 -4.88604486e-01
-5.26177585e-02 -3.52921069e-01 -1.37096286e-01 -4.30468395e-02
-8.80725324e-01 1.14127541e+00 8.02433670e-01 3.43844980e-01
2.23870110e-02 3.49367112e-01 4.58970666e-01 6.98821843e-01
3.55315208e-02 9.68113184e-01 -6.71963692e-01 1.73949748e-01
-7.80485570e-01 7.06867754e-01 1.06881392e+00 7.19214380e-01
6.11677170e-01 -4.21176195e-01 -5.92306197e-01 5.27135730e-01
6.74225837e-02 4.44647849e-01 1.77095160e-01 -1.44839227e+00
8.80340338e-01 6.44607425e-01 5.07313684e-02 -9.41514134e-01
-3.88508663e-02 -1.67963400e-01 -1.78506956e-01 2.83611827e-02
8.23399186e-01 -2.65852541e-01 -6.60677195e-01 1.97449899e+00
3.98593545e-01 -5.44957407e-02 3.13812762e-01 8.97847295e-01
7.53597021e-01 6.91951156e-01 1.81499511e-01 3.61773312e-01
1.25667667e+00 -1.43502653e+00 -5.23126662e-01 -4.65535343e-01
7.50907362e-01 -6.93916500e-01 1.46712196e+00 1.06219597e-01
-1.27231634e+00 -3.50251943e-01 -1.09764421e+00 -5.29334009e-01
-4.86348510e-01 -6.95456803e-01 3.60271454e-01 2.33622640e-01
-1.01876831e+00 3.59323949e-01 -3.20406854e-01 -1.59753352e-01
2.19376996e-01 -1.33279800e-01 -1.29795641e-01 -4.64460343e-01
-1.83847916e+00 1.11779130e+00 4.32969742e-02 -1.76535517e-01
-9.06913936e-01 -7.97807276e-01 -7.61813045e-01 -5.24527766e-02
6.73735499e-01 -8.68665695e-01 1.73669505e+00 -8.82441819e-01
-1.15172184e+00 7.90212095e-01 -2.12005883e-01 -6.55876875e-01
6.76107347e-01 -2.45487973e-01 -4.09228027e-01 5.66043437e-01
4.69913304e-01 5.64561665e-01 8.04960847e-01 -1.34229970e+00
-4.33397532e-01 -2.52603978e-01 9.51197445e-01 -9.07145590e-02
6.00333102e-02 9.21947584e-02 -3.91050249e-01 -6.15333617e-01
-2.03038678e-01 -8.71930301e-01 -1.31615788e-01 4.61170562e-02
-5.10359704e-01 -3.39548200e-01 9.09733236e-01 -7.66936183e-01
1.00792861e+00 -1.89116836e+00 5.31238876e-02 6.91062585e-02
1.40824735e-01 5.49985826e-01 -5.48560858e-01 5.13712943e-01
7.57424235e-02 4.36937928e-01 -6.13957763e-01 -9.24574584e-02
2.82979816e-01 5.74218571e-01 -1.01640427e+00 -5.33997035e-03
4.90065247e-01 1.37265241e+00 -1.10762656e+00 -4.95619953e-01
-2.43887126e-01 -2.91360319e-02 -8.23176444e-01 3.87518972e-01
-8.59183788e-01 2.00238943e-01 -3.66339177e-01 3.68219852e-01
5.40274084e-01 -8.23244601e-02 -2.32349001e-02 -1.19740278e-01
4.30860102e-01 4.90806609e-01 -7.46774912e-01 1.55716002e+00
-5.22407413e-01 5.53012490e-01 1.80724502e-01 -6.92023814e-01
3.78263056e-01 2.85700083e-01 -2.01754704e-01 -7.02241004e-01
-2.39715964e-01 2.60505438e-01 9.23983604e-02 -7.76158750e-01
3.53020728e-01 -4.28579897e-01 -2.86178559e-01 6.41752183e-01
-7.55206719e-02 -3.51710260e-01 3.09946001e-01 7.44158030e-01
1.65432668e+00 6.59223646e-02 -3.45747232e-01 1.48492560e-01
8.74820888e-01 3.31232935e-01 2.98039854e-01 1.10737586e+00
-3.16092402e-01 5.90840697e-01 7.51385987e-01 -3.15651894e-01
-9.02086020e-01 -1.31367218e+00 1.21267945e-01 1.47274387e+00
3.04463897e-02 -4.36193585e-01 -8.61542404e-01 -1.53538978e+00
2.04096392e-01 1.17275131e+00 -7.54150271e-01 -2.85874277e-01
-9.81773376e-01 -1.60291102e-02 1.20807850e+00 5.70224226e-01
5.10647416e-01 -1.17626846e+00 -2.25152284e-01 1.54005647e-01
-5.23524165e-01 -1.11037028e+00 -5.49332559e-01 2.47575156e-02
-5.11784673e-01 -1.45071661e+00 -3.56494784e-01 -6.16871774e-01
5.54536223e-01 2.08174577e-03 1.45535779e+00 5.93468487e-01
1.58674449e-01 5.55742681e-01 -3.65659922e-01 -2.20344543e-01
-8.56681585e-01 2.58893549e-01 -6.84123695e-01 -4.09737319e-01
2.92151809e-01 -3.97717923e-01 -5.06341636e-01 2.74967223e-01
-1.27079999e+00 -4.18436706e-01 2.53196150e-01 7.10797608e-01
3.87983859e-01 -2.07711473e-01 7.59278297e-01 -1.33207750e+00
7.27464020e-01 -6.71138108e-01 -4.31125075e-01 6.03642404e-01
-2.41609648e-01 3.04743022e-01 1.14971423e+00 -3.27618927e-01
-9.98007119e-01 -6.28587663e-01 -5.51871777e-01 -2.63835192e-01
5.13084829e-02 3.10904264e-01 -1.60213023e-01 1.28267393e-01
9.26292658e-01 -3.14914942e-01 -3.85496289e-01 -1.92322671e-01
8.41561615e-01 1.87850371e-01 7.92471766e-01 -9.66821790e-01
1.33829975e+00 3.25244546e-01 -2.45463789e-01 8.88916384e-03
-1.56083286e+00 -9.38253179e-02 -2.52121776e-01 1.50628969e-01
9.69706833e-01 -4.06644434e-01 -5.07550299e-01 -1.16614290e-01
-1.58937204e+00 -5.54351032e-01 -4.95025039e-01 -2.25013644e-01
-3.58644426e-01 3.65353107e-01 -8.95366371e-01 -3.58950466e-01
-2.74443030e-01 -1.04734349e+00 8.63429070e-01 -1.68731287e-01
-5.38533270e-01 -1.12695253e+00 2.52483487e-01 8.70670378e-01
2.89524138e-01 2.23118365e-01 1.35739875e+00 -1.06463265e+00
-7.16754794e-01 -3.56707364e-01 -1.37930036e-01 6.19361877e-01
-1.66190356e-01 -4.76347283e-02 -1.03366196e+00 7.82823190e-02
2.18514174e-01 -7.31590986e-01 8.75343204e-01 -4.41461623e-01
1.04467261e+00 -6.36937320e-01 9.40846875e-02 2.27898210e-01
1.25882900e+00 -3.50362599e-01 5.54712415e-01 3.95081460e-01
6.16660118e-01 7.72915423e-01 3.69343907e-01 -6.53752744e-01
4.22776163e-01 2.54862368e-01 6.33740902e-01 7.04394728e-02
-3.44788015e-01 -5.54235518e-01 4.36795503e-01 6.13217354e-01
6.78766251e-01 -2.85873801e-01 -9.12430763e-01 6.88793719e-01
-1.82635081e+00 -1.23477185e+00 -3.54830474e-01 1.62618959e+00
1.13803303e+00 3.97804797e-01 -1.80918872e-01 1.78725589e-02
4.71606970e-01 3.19221646e-01 -6.95808828e-01 -1.01557183e+00
-2.41190314e-01 4.12759990e-01 1.27241313e-01 1.22896373e+00
-8.72501433e-01 1.04911065e+00 6.60153484e+00 4.70324248e-01
-5.40010095e-01 3.46050054e-01 2.10603386e-01 -6.74730390e-02
-1.03201103e+00 1.57577321e-01 -4.59653020e-01 2.77133048e-01
9.12424445e-01 9.75161605e-03 6.15945876e-01 4.98479635e-01
-3.58774304e-01 3.99797931e-02 -1.38299632e+00 -7.97426403e-02
1.86670125e-01 -1.27859914e+00 1.92986637e-01 -3.90610099e-01
7.77933061e-01 -5.90334646e-02 1.60030439e-01 6.94893181e-01
8.62226963e-01 -9.50957060e-01 7.36175299e-01 2.66304046e-01
3.12778711e-01 -6.92537487e-01 8.05596173e-01 5.04475296e-01
-6.83452845e-01 -2.73561388e-01 -9.45428386e-02 5.44512235e-02
2.40703270e-01 2.89051473e-01 -6.54264569e-01 5.30428648e-01
3.92215043e-01 2.37545632e-02 -7.49400437e-01 3.50914419e-01
-1.07320237e+00 8.84445488e-01 -1.00581720e-01 6.88162446e-02
7.08356202e-01 2.08560854e-01 6.48309052e-01 1.14954257e+00
-1.95096314e-01 2.48390108e-01 -1.83454249e-02 9.96041179e-01
-5.63680768e-01 -1.61265299e-01 -5.71589530e-01 6.68159826e-03
2.62148529e-01 1.05254424e+00 -7.16941953e-02 -5.32853425e-01
-6.53274834e-01 1.02627838e+00 7.59311438e-01 5.39826274e-01
-1.09569025e+00 -4.13485140e-01 5.12675166e-01 5.32536432e-02
3.69138271e-01 8.25096443e-02 -2.26247549e-01 -1.01057816e+00
3.31217617e-01 -1.35657620e+00 7.81784654e-01 -9.98795688e-01
-1.72103810e+00 5.53560555e-01 -1.68313831e-01 -3.80013049e-01
-3.59923333e-01 -4.84888405e-01 -8.93876195e-01 1.03856993e+00
-1.64652288e+00 -1.12068152e+00 1.18988700e-01 7.04328537e-01
4.33304220e-01 2.32547358e-01 9.60845828e-01 5.47951050e-02
-4.11410481e-01 8.48695815e-01 -3.03496301e-01 4.42931741e-01
9.33169067e-01 -1.48443735e+00 5.15042484e-01 1.10588539e+00
3.96792263e-01 8.67070615e-01 8.43669057e-01 -5.53391755e-01
-1.23932934e+00 -1.12065494e+00 1.08960867e+00 -1.23756957e+00
1.27265501e+00 -1.64043963e-01 -1.36989689e+00 1.07038546e+00
5.68985820e-01 -7.32505992e-02 7.32449591e-01 1.14722073e-01
-1.12664068e+00 -3.52355130e-02 -1.39540243e+00 8.87475193e-01
8.52368057e-01 -1.00044751e+00 -1.51255465e+00 4.89051133e-01
1.29169691e+00 -3.82923454e-01 -5.16572773e-01 2.79492974e-01
2.58412302e-01 -8.90990138e-01 9.03427720e-01 -1.26409876e+00
9.19729888e-01 -2.50918835e-01 -3.76079738e-01 -1.04483664e+00
-3.26601090e-03 -5.48549712e-01 -4.12185252e-01 1.23942876e+00
7.73765922e-01 -5.85667551e-01 5.26156425e-01 9.41276908e-01
-1.36615440e-01 -6.69911504e-01 -8.87583852e-01 -5.49051166e-01
7.31541574e-01 -6.65385664e-01 6.48510516e-01 7.45954156e-01
-1.23659357e-01 5.41072130e-01 2.26439655e-01 4.99588847e-01
4.56917286e-01 2.19314963e-01 6.87085211e-01 -6.20855570e-01
-5.38496554e-01 -3.13604414e-01 1.69598803e-01 -1.00432932e+00
6.48795843e-01 -1.19143307e+00 1.95816800e-01 -1.68310368e+00
-1.32045895e-01 -1.90364748e-01 -2.28277028e-01 5.37488580e-01
-7.13184536e-01 2.41469890e-01 2.92601109e-01 -3.09719499e-02
-7.29173362e-01 1.29398614e-01 1.35326397e+00 -4.44340020e-01
3.72054905e-01 -1.00105003e-01 -1.08588839e+00 8.59157622e-01
7.90071666e-01 -6.84702694e-01 -4.94751245e-01 -9.33355331e-01
5.89457631e-01 -1.15493111e-01 7.02084541e-01 -6.36936247e-01
4.68236208e-01 3.16457711e-02 5.76865301e-02 -2.56292462e-01
2.31372550e-01 -7.79336929e-01 -5.01275003e-01 4.17643964e-01
-1.00212336e+00 2.63550580e-01 3.21317315e-01 5.41998625e-01
-2.03186512e-01 -6.29279375e-01 5.75567007e-01 -3.25140715e-01
-2.91029245e-01 2.76278909e-02 -1.61597207e-01 8.44315588e-01
7.90044069e-01 3.08577508e-01 -6.72354519e-01 -4.39009786e-01
-7.76838362e-01 5.50918937e-01 3.24354887e-01 2.37381443e-01
4.23621595e-01 -1.20468795e+00 -7.84621119e-01 -2.27147296e-01
4.72354218e-02 9.20164064e-02 5.83542436e-02 3.63003880e-01
-3.44145447e-01 2.54390925e-01 3.85595322e-01 -3.46067660e-02
-1.06719530e+00 8.64873230e-01 5.17603993e-01 -6.35166764e-01
-3.32497209e-01 9.50877964e-01 8.10503401e-03 -9.34093297e-01
1.96480706e-01 -2.36565486e-01 2.21642211e-01 -1.32633224e-01
5.41397035e-01 2.85461754e-01 4.45651114e-02 -3.08933735e-01
-2.60617733e-01 4.53903079e-01 -2.11899742e-01 -4.25211400e-01
8.84230018e-01 1.09590113e-01 -3.28520358e-01 3.67566615e-01
1.36268389e+00 5.44012308e-01 -9.11226928e-01 -3.34966451e-01
3.90762389e-02 -3.52575094e-01 -5.89887142e-01 -1.17817414e+00
-8.49966168e-01 1.01140916e+00 -1.51547670e-01 5.98941743e-01
8.02537143e-01 3.54104280e-01 1.25450683e+00 7.14248359e-01
-1.43214734e-02 -7.75432169e-01 4.04858083e-01 8.10976207e-01
1.32483709e+00 -9.91427124e-01 -4.02042568e-01 -2.96315670e-01
-5.61547995e-01 8.72545362e-01 8.84769619e-01 -2.78905302e-01
2.04439372e-01 3.49906832e-01 5.14026999e-01 -2.64427096e-01
-8.91912282e-01 -5.44150025e-02 2.43994758e-01 4.10097569e-01
4.23797518e-02 -3.07385921e-01 2.84413490e-02 6.27565444e-01
-5.08019328e-01 -3.87697488e-01 2.44610503e-01 7.61352777e-01
-2.79500246e-01 -1.23612583e+00 -4.27211195e-01 1.31438240e-01
-6.59543753e-01 -3.68052840e-01 -9.05203044e-01 7.90084600e-01
1.46894738e-01 1.31241167e+00 -2.95005232e-01 -1.82985410e-01
7.01839864e-01 5.24454415e-01 4.67723846e-01 -5.90539038e-01
-1.14639676e+00 -7.74523556e-01 3.52605343e-01 -6.58346832e-01
-8.32025185e-02 -2.77209431e-01 -1.46867001e+00 -4.93641555e-01
-2.84456331e-02 4.04899955e-01 2.89414406e-01 1.18725502e+00
1.62595779e-01 5.58418095e-01 5.28297424e-01 1.05387904e-01
-1.01921833e+00 -6.98040485e-01 2.56751508e-01 6.77139044e-01
6.02203548e-01 1.31640267e-02 -5.79612017e-01 -1.74825624e-01] | [11.09365463256836, 7.993206977844238] |
7bdaca2e-511c-4f47-ac83-d3442f6a3ce3 | learning-rich-representation-of-keyphrases | null | null | https://openreview.net/forum?id=mSw7ck7b7L | https://openreview.net/pdf?id=mSw7ck7b7L | Learning Rich Representation of Keyphrases from Text | In this work, we explore how to learn task-specific language models aimed towards learning rich representation of keyphrases from text documents. We experiment with different masking strategies for training transformer language models (LMs) in discriminative as well as generative settings. In the discriminative setting, we introduce a new pre-training objective - Keyphrase Boundary Infilling with Replacement (KBIR), showing large gains in performance (upto 9.26 points in F1) over SOTA, when LM pre-trained using KBIR is fine-tuned for the task of keyphrase extraction. In the generative setting, we introduce a new training setup for BART - KeyBART, that reproduces the keyphrases related to the input text in the CatSeq format, instead of the denoised original input. This also led to gains in performance (upto 4.33 points in F1@M) over SOTA for keyphrase generation. Additionally, we also fine-tune the pre-trained language models on named entity recognition (NER), question answering (QA), relation extraction (RE), abstractive summarization and achieve comparable performance with that of the SOTA, showing that learning rich representation of keyphrases is indeed beneficial for many other fundamental NLP tasks. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['keyphrase-generation', 'keyphrase-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.74776840e-01 5.04552424e-01 2.56674495e-02 1.71278164e-01
-1.65040994e+00 -7.42043197e-01 9.79574203e-01 5.58660090e-01
-7.25561619e-01 1.07023275e+00 7.61641979e-01 -2.41605490e-01
-2.39838362e-01 -7.97259271e-01 -9.63195562e-01 -5.17035723e-01
-3.70182768e-02 4.98810321e-01 1.35345072e-01 -5.68201840e-01
2.14438558e-01 3.95219028e-01 -1.13778937e+00 6.70571387e-01
9.28558469e-01 7.15985894e-01 1.14866771e-01 1.17564881e+00
-5.30713260e-01 1.07448399e+00 -1.10994041e+00 -5.67158937e-01
3.71624008e-02 -4.28216755e-01 -1.17347884e+00 -1.79520816e-01
8.08944345e-01 -3.24862152e-02 -4.30846274e-01 6.94304705e-01
6.15607440e-01 1.76802278e-01 8.13635409e-01 -7.55913496e-01
-5.61958730e-01 1.18658888e+00 -3.44832569e-01 4.87251461e-01
4.83835608e-01 -2.76074499e-01 1.37139380e+00 -9.06744957e-01
6.85609221e-01 1.19677401e+00 4.92505282e-01 3.18788260e-01
-1.25943613e+00 -3.65898997e-01 -9.19954032e-02 2.13908434e-01
-1.25972664e+00 -4.84318316e-01 3.68966281e-01 -1.49025526e-02
1.39966261e+00 5.67096233e-01 2.35134572e-01 1.17761493e+00
3.23334038e-01 1.09504080e+00 9.14226353e-01 -7.33017743e-01
-8.24845135e-02 2.33956337e-01 2.30816782e-01 3.44680607e-01
2.19677269e-01 -3.59366119e-01 -6.16387248e-01 -2.56845295e-01
3.67272675e-01 -5.36052585e-01 -2.71290839e-01 2.62670130e-01
-1.45945656e+00 7.57287681e-01 8.83394554e-02 6.62385523e-01
-5.84227383e-01 1.41582236e-01 5.84164202e-01 6.72304213e-01
4.57367390e-01 9.51018453e-01 -8.57375443e-01 -6.15390055e-02
-1.27775061e+00 5.42296767e-01 9.82182801e-01 9.99384582e-01
6.63312674e-01 8.38839188e-02 -1.09681475e+00 8.38176012e-01
-3.44389558e-01 5.70970893e-01 7.19731867e-01 -5.72085619e-01
9.29898083e-01 6.07754111e-01 -2.86453199e-02 -5.79292834e-01
-3.23488861e-01 -7.24007726e-01 -1.18030822e+00 -6.49854243e-01
5.28698117e-02 -3.28698874e-01 -1.20786023e+00 1.42994654e+00
2.69309916e-02 -1.37421235e-01 5.21872759e-01 3.81417051e-02
1.15366244e+00 1.13939989e+00 -1.40513137e-01 -3.21884036e-01
1.63997817e+00 -1.02478778e+00 -8.56418669e-01 -2.53226787e-01
8.12727511e-01 -1.00611770e+00 1.05042684e+00 4.39210832e-01
-1.08984721e+00 -5.67954302e-01 -7.83847153e-01 -3.39384258e-01
-5.75398624e-01 5.92639983e-01 1.69774100e-01 2.45344236e-01
-1.14081681e+00 5.64096808e-01 -3.30218345e-01 -4.85944599e-01
2.02832729e-01 8.83979797e-02 -4.39237654e-01 1.42396241e-01
-1.33773685e+00 1.02487063e+00 8.43547642e-01 -2.29202956e-01
-7.90889084e-01 -9.94070172e-01 -8.43154490e-01 1.47990406e-01
7.36154199e-01 -8.93904507e-01 1.31178689e+00 -1.92787364e-01
-1.42661834e+00 8.86178672e-01 -1.25500098e-01 -9.61603582e-01
3.38687479e-01 -7.63987124e-01 -3.92201930e-01 1.44852057e-01
1.66533187e-01 6.16609931e-01 9.75127935e-01 -9.52362120e-01
-6.19557142e-01 7.42425621e-02 -7.33229965e-02 1.33690447e-01
-3.62218022e-01 1.95122898e-01 -1.97829321e-01 -1.19790792e+00
-3.70606512e-01 -5.33933282e-01 -1.21650025e-01 -9.89794075e-01
-9.20989454e-01 -4.17614102e-01 4.55928087e-01 -1.13307118e+00
1.71049678e+00 -1.60366249e+00 2.52911896e-01 9.06440914e-02
-6.78499714e-02 4.82293844e-01 -1.48475409e-01 1.00591564e+00
-1.34522796e-01 2.95003206e-01 -3.96082878e-01 -2.28613630e-01
3.05166896e-02 2.42049828e-01 -1.02114797e+00 -1.84819385e-01
5.14357328e-01 1.36007178e+00 -7.16482699e-01 -5.74172199e-01
-1.24037564e-01 8.59759748e-02 -2.74453998e-01 4.29123580e-01
-3.78568381e-01 1.08467802e-01 -3.93337458e-01 1.33371815e-01
3.73505145e-01 -6.67962953e-02 -9.35558975e-02 -3.31592262e-01
7.49711022e-02 7.11683571e-01 -1.20953429e+00 1.58970594e+00
-7.89370894e-01 5.69653630e-01 -2.00061724e-01 -8.91336799e-01
7.20520914e-01 5.09557724e-01 8.74625593e-02 -5.54362118e-01
-2.76982654e-02 1.46163046e-01 -2.45381460e-01 -3.71264815e-01
9.69952345e-01 4.64097373e-02 -3.16046923e-01 4.74782050e-01
6.16568983e-01 -5.10681510e-01 5.32240331e-01 6.57821774e-01
1.27472746e+00 -1.76437497e-01 6.66639864e-01 -1.94742277e-01
8.52284849e-01 -6.36759726e-03 -1.50085568e-01 1.03375340e+00
6.82845294e-01 6.38461530e-01 5.93455434e-01 3.73236984e-02
-1.05018795e+00 -7.83562124e-01 2.08133757e-02 1.18495333e+00
-5.25319695e-01 -1.02049863e+00 -8.52609277e-01 -8.43033314e-01
-8.79484341e-02 1.16198444e+00 -5.53553641e-01 -4.08020943e-01
-1.08701074e+00 -7.25231171e-01 1.09071672e+00 4.14919943e-01
5.60405850e-01 -1.19503427e+00 -2.15051919e-01 3.47101539e-01
-3.17773908e-01 -1.35425365e+00 -5.39734721e-01 4.68374431e-01
-7.98828721e-01 -6.78623736e-01 -1.02285671e+00 -6.67024672e-01
3.51250887e-01 -1.87008038e-01 1.37484813e+00 -2.91479737e-01
-8.49762931e-02 5.03615439e-01 -5.10484934e-01 -2.52495229e-01
-7.99259961e-01 5.92645466e-01 -2.79775113e-01 -8.39648023e-02
-1.22449517e-01 -4.34432656e-01 -4.91679572e-02 -3.72739524e-01
-1.27051592e+00 3.15088183e-02 9.99759078e-01 8.12814236e-01
4.42847729e-01 9.57069471e-02 4.81611788e-01 -1.05054963e+00
1.03986239e+00 -2.66341835e-01 -1.86694562e-01 6.06885493e-01
-4.44142908e-01 7.89361358e-01 8.45898032e-01 -4.77746367e-01
-9.43108916e-01 -6.22155845e-01 -3.61815304e-01 2.16675550e-02
-1.78936377e-01 4.58178759e-01 -1.58733785e-01 4.58603024e-01
8.83735359e-01 5.41658998e-01 -6.68710828e-01 -8.64849806e-01
8.61124814e-01 6.20623410e-01 7.70871818e-01 -6.44420743e-01
1.08969235e+00 1.05215274e-01 -1.71196386e-01 -9.36399162e-01
-1.12282181e+00 -5.19856811e-01 -4.87420171e-01 5.18713117e-01
7.16619253e-01 -9.00983095e-01 -2.25929067e-01 2.51876056e-01
-1.35027015e+00 -1.44741803e-01 -6.79749250e-01 1.70157120e-01
-3.50278616e-01 4.64139789e-01 -6.24662995e-01 -5.76936841e-01
-1.20199621e+00 -5.42676270e-01 1.30056310e+00 6.10193163e-02
-4.46279079e-01 -9.28004801e-01 1.58626914e-01 2.87901968e-01
3.58104616e-01 1.34536803e-01 1.28980243e+00 -1.15958488e+00
-2.97714353e-01 -2.67374128e-01 -1.09973110e-01 6.83829129e-01
1.43520638e-01 -2.72652298e-01 -9.45010006e-01 -2.25984812e-01
-8.47315565e-02 -3.48863363e-01 1.35212791e+00 1.02642357e-01
7.86934674e-01 -8.00276279e-01 -8.26442242e-02 2.92843282e-01
1.04194868e+00 -2.92558163e-01 7.38278270e-01 3.04381967e-01
6.84830129e-01 5.30938864e-01 4.65179205e-01 1.55479029e-01
3.86925310e-01 6.35376751e-01 -1.80995256e-01 -1.36269629e-02
-3.51956904e-01 -5.49346447e-01 6.50113761e-01 9.68856931e-01
1.33334890e-01 -4.12505716e-01 -7.18733490e-01 6.75641060e-01
-1.61721051e+00 -8.14098001e-01 -1.52285159e-01 1.91293728e+00
1.42818797e+00 2.47068703e-01 -1.19057307e-02 2.15020090e-01
3.93103749e-01 3.41808289e-01 2.48080231e-02 -4.50216293e-01
-4.21619952e-01 8.86305749e-01 5.21080434e-01 5.12644589e-01
-1.14115083e+00 1.29768038e+00 5.64090300e+00 1.41514182e+00
-8.51589680e-01 1.01808064e-01 4.08651412e-01 2.18052462e-01
-2.82249063e-01 3.06216069e-02 -1.16496539e+00 7.30650201e-02
1.29914653e+00 -4.03812140e-01 1.80745974e-01 4.91876245e-01
-9.20279622e-02 -2.26943478e-01 -1.03432357e+00 9.34530616e-01
1.97331980e-01 -1.53587079e+00 4.90186304e-01 -2.23182186e-01
7.06946433e-01 -1.54946357e-01 -2.96427906e-01 6.72921300e-01
4.10255820e-01 -1.04410207e+00 6.85240328e-01 4.48405594e-01
7.17679381e-01 -6.77089274e-01 7.98472226e-01 5.44051230e-01
-9.83346581e-01 1.98030025e-01 -3.35166484e-01 2.92402238e-01
9.25106481e-02 7.39467859e-01 -1.24538708e+00 1.12928009e+00
4.65460420e-01 3.39056104e-01 -9.01461124e-01 6.84124231e-01
-4.86965328e-01 8.31457138e-01 -4.25695539e-01 1.06873244e-01
2.75379032e-01 1.83495060e-01 7.14742362e-01 1.74540865e+00
1.80513754e-01 -1.66150093e-01 -1.25538781e-01 5.83098888e-01
-4.89718437e-01 4.64232266e-01 -3.79053205e-01 -4.07031000e-01
2.49339908e-01 1.55193961e+00 -6.03009403e-01 -7.93003678e-01
1.79085478e-01 1.17511404e+00 2.45667800e-01 3.27510774e-01
-5.23330092e-01 -6.60751343e-01 1.99917525e-01 -5.20536900e-02
6.51236057e-01 -6.27029091e-02 8.24760273e-02 -1.34658635e+00
1.86123237e-01 -1.18792272e+00 6.23215318e-01 -6.36969864e-01
-1.11686492e+00 8.21952760e-01 2.65723795e-01 -7.74621844e-01
-6.56531751e-01 -2.62844056e-01 -6.21694267e-01 8.52639794e-01
-1.51233482e+00 -1.28502965e+00 1.34360805e-01 2.93179333e-01
6.76246226e-01 -1.41831532e-01 9.57330585e-01 1.41240567e-01
-3.83369327e-01 7.47364044e-01 1.88916594e-01 2.11121723e-01
8.12378407e-01 -1.63802063e+00 8.32462430e-01 9.24489915e-01
7.61790335e-01 7.47716546e-01 7.04392970e-01 -6.20547175e-01
-1.33535969e+00 -1.27725244e+00 1.23707438e+00 -7.33829796e-01
7.95948565e-01 -5.54177344e-01 -1.06339920e+00 6.30941153e-01
5.75497925e-01 -3.53827208e-01 3.10425729e-01 -2.01099679e-01
-3.29003811e-01 -1.28066763e-01 -8.03737640e-01 4.92812306e-01
6.32071197e-01 -6.98683202e-01 -1.10268891e+00 4.43856210e-01
1.07401443e+00 -4.51411098e-01 -8.89603734e-01 4.11326468e-01
-1.14537720e-02 -4.41903412e-01 1.09894252e+00 -8.61657321e-01
3.08711112e-01 -1.54588342e-01 -8.10882971e-02 -1.37815225e+00
9.68208164e-03 -1.14245760e+00 -5.10462940e-01 1.74965239e+00
5.85210145e-01 -4.22074705e-01 3.30828726e-01 -2.03707933e-01
-9.86006632e-02 -7.64316797e-01 -7.63746202e-01 -6.73532844e-01
3.07072312e-01 -3.91438484e-01 4.05642450e-01 5.06485820e-01
-4.08143610e-01 1.12526023e+00 -3.90159577e-01 -2.90607382e-02
1.20357454e-01 1.42308306e-02 9.43570793e-01 -8.52043152e-01
-3.81828576e-01 -2.40555629e-01 7.42167458e-02 -1.15546250e+00
4.22625728e-02 -9.90061820e-01 -2.88472444e-01 -1.78100145e+00
2.99237873e-02 3.18972021e-02 -7.73363486e-02 5.66040814e-01
-4.93108004e-01 2.04031486e-02 2.55007356e-01 1.50251746e-01
-4.86679584e-01 5.43806612e-01 1.01929891e+00 -1.38595179e-01
-2.16878936e-01 1.48097739e-01 -9.10317004e-01 4.33820218e-01
5.88151872e-01 -6.10593438e-01 -1.46846250e-01 -1.12157084e-01
4.52840447e-01 -8.88184309e-02 3.04093719e-01 -8.00785720e-01
1.53865144e-01 1.49142593e-01 2.27652475e-01 -9.05580044e-01
9.57596973e-02 -2.98287690e-01 -1.66118279e-01 2.24988490e-01
-5.71098268e-01 1.16507463e-01 5.13475597e-01 3.07689995e-01
-3.13775361e-01 -3.90595287e-01 3.46550822e-01 -1.73441753e-01
-2.63334543e-01 -1.17292568e-01 -3.30520838e-01 6.88282311e-01
3.32912832e-01 1.41557649e-01 -2.68317044e-01 -4.37063128e-01
-4.35795993e-01 9.27157402e-02 -1.21244602e-01 3.25844675e-01
3.86630684e-01 -1.01013339e+00 -1.10241961e+00 5.71516389e-03
9.43597686e-03 1.75625727e-01 1.11031227e-01 9.25393879e-01
-3.92579436e-01 9.89130020e-01 3.79567713e-01 -9.11861956e-02
-1.13394761e+00 3.50039542e-01 4.81499843e-02 -1.18845999e+00
-7.35532939e-01 9.03020561e-01 1.78499147e-03 -5.29709816e-01
2.07248330e-01 -1.01126480e+00 -4.30549502e-01 5.07870555e-01
5.28951406e-01 3.58450651e-01 7.23594189e-01 -2.51697123e-01
-1.29492074e-01 3.54092538e-01 -5.14699280e-01 -2.23929465e-01
1.28848183e+00 1.23824529e-01 -3.00408244e-01 3.80803972e-01
1.27939820e+00 5.44887722e-01 -4.47026253e-01 -5.33109426e-01
5.99217474e-01 3.73712569e-01 6.99145272e-02 -8.82020712e-01
-4.27844852e-01 6.65730894e-01 5.76817729e-02 2.33399779e-01
1.01964629e+00 2.40510985e-01 1.18615329e+00 8.55182171e-01
-4.46935743e-02 -1.02500987e+00 1.37781560e-01 8.17252994e-01
1.19950414e+00 -8.42199206e-01 1.64056346e-01 -1.49157912e-01
-7.68827081e-01 1.15827227e+00 6.83000311e-02 -1.75509885e-01
2.36875445e-01 2.19831884e-01 -7.63690323e-02 -2.01607645e-01
-9.06226158e-01 -3.70085180e-01 7.45341122e-01 2.15798259e-01
3.35796952e-01 -1.46466136e-01 -1.76285103e-01 6.89835191e-01
-7.86615014e-01 -4.07345235e-01 5.02688527e-01 8.16032767e-01
-3.29558790e-01 -1.18950355e+00 -4.15471852e-01 6.48631990e-01
-7.54566491e-01 -6.81797326e-01 -6.65019751e-01 7.93506145e-01
-1.73754677e-01 8.02957892e-01 -2.04794198e-01 -2.47031420e-01
4.99709517e-01 4.80995864e-01 4.57261264e-01 -9.49407697e-01
-1.06547177e+00 7.68740401e-02 4.45697367e-01 -1.09833241e-01
-9.30461884e-02 -2.00548083e-01 -1.11352491e+00 5.99009618e-02
-4.83095616e-01 5.47826231e-01 2.93110162e-01 1.12582815e+00
3.79850477e-01 7.39189327e-01 3.57435763e-01 -3.79804343e-01
-7.44530678e-01 -1.49154961e+00 -1.68781877e-01 2.19763950e-01
4.28532541e-01 -4.22054827e-02 -3.06575298e-01 2.69529581e-01] | [12.312111854553223, 9.020480155944824] |
6d2a8932-e6bd-4fd4-ae5e-f012811c46d9 | robustloc-robust-camera-pose-regression-in | 2211.11238 | null | https://arxiv.org/abs/2211.11238v4 | https://arxiv.org/pdf/2211.11238v4.pdf | RobustLoc: Robust Camera Pose Regression in Challenging Driving Environments | Camera relocalization has various applications in autonomous driving. Previous camera pose regression models consider only ideal scenarios where there is little environmental perturbation. To deal with challenging driving environments that may have changing seasons, weather, illumination, and the presence of unstable objects, we propose RobustLoc, which derives its robustness against perturbations from neural differential equations. Our model uses a convolutional neural network to extract feature maps from multi-view images, a robust neural differential equation diffusion block module to diffuse information interactively, and a branched pose decoder with multi-layer training to estimate the vehicle poses. Experiments demonstrate that RobustLoc surpasses current state-of-the-art camera pose regression models and achieves robust performance in various environments. Our code is released at: https://github.com/sijieaaa/RobustLoc | ['Diego Navarro Navarro', 'Andreas Hartmannsgruber', 'Wee Peng Tay', 'Rui She', 'Qiyu Kang', 'Sijie Wang'] | 2022-11-21 | null | null | null | null | ['camera-absolute-pose-regression', 'camera-localization', 'camera-relocalization', 'visual-localization'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-4.33415622e-01 -1.51931882e-01 -1.21026933e-01 -3.79559845e-01
-6.30583167e-01 -7.99167931e-01 5.21261573e-01 -7.04383969e-01
-3.33114505e-01 4.95904565e-01 -6.93360195e-02 -1.63180411e-01
3.76788855e-01 -3.97729486e-01 -1.20000196e+00 -7.25983560e-01
2.23305896e-01 2.83795744e-01 2.19097167e-01 -3.54588598e-01
2.24500030e-01 5.01610696e-01 -1.49682212e+00 -3.68534327e-01
7.31766343e-01 6.91832542e-01 1.56433776e-01 1.03557646e+00
5.05211353e-01 7.37756371e-01 -1.01539001e-01 -2.14450613e-01
4.80872273e-01 1.16024800e-01 -1.32909149e-01 5.82128856e-03
6.47383630e-01 -5.85584044e-01 -8.31257939e-01 1.11319411e+00
4.24945265e-01 2.16170456e-02 3.76652867e-01 -1.44500697e+00
-6.36451900e-01 1.22335389e-01 -7.84062207e-01 2.07033530e-01
3.05802673e-01 5.85639834e-01 4.26964819e-01 -9.49128509e-01
6.52460515e-01 1.28429794e+00 8.42983305e-01 5.11997461e-01
-9.44945276e-01 -9.47990417e-01 2.97635764e-01 2.51232505e-01
-1.46754682e+00 -6.63371205e-01 7.25468040e-01 -5.67160368e-01
8.39859903e-01 -6.87615052e-02 5.33587277e-01 1.26674747e+00
7.26501584e-01 7.66857147e-01 5.51209986e-01 2.78941244e-01
-9.24460739e-02 -1.52124772e-02 -2.06000686e-01 6.77437365e-01
2.93556571e-01 3.76189321e-01 -5.93577087e-01 1.79511473e-01
5.00347674e-01 -5.85398376e-02 -2.05304295e-01 -7.91605949e-01
-1.28629994e+00 7.33201921e-01 4.92478102e-01 -4.27869231e-01
-1.35343503e-02 5.97436547e-01 8.14288855e-02 2.00213060e-01
4.08413798e-01 3.46868098e-01 -2.77709186e-01 -2.03962713e-01
-6.47002339e-01 4.68539596e-01 6.38375461e-01 1.44275200e+00
8.83972883e-01 2.87114859e-01 4.59401786e-01 5.30052841e-01
4.74387944e-01 1.10828173e+00 1.83278784e-01 -1.53270161e+00
6.00323498e-01 1.93178698e-01 2.63266504e-01 -1.12390351e+00
-5.39719880e-01 -4.88214701e-01 -4.77023035e-01 2.41484165e-01
2.52200551e-02 -5.83865166e-01 -6.54272377e-01 1.66435778e+00
4.51752990e-01 4.86101449e-01 2.16920644e-01 1.10836995e+00
6.38236642e-01 6.43575490e-01 -5.69431365e-01 1.99389830e-01
7.13330567e-01 -1.14572608e+00 -5.67234993e-01 -7.52320230e-01
5.49864590e-01 -7.52831459e-01 5.08318603e-01 3.16743284e-01
-9.28116560e-01 -4.65476066e-01 -1.32990170e+00 -3.56565416e-01
-3.75682086e-01 2.00524047e-01 3.78739089e-01 2.45839775e-01
-1.19091177e+00 -6.71977177e-02 -1.06975174e+00 -3.71858627e-01
9.91824865e-02 3.27027380e-01 -5.35544872e-01 -6.98846132e-02
-9.21055615e-01 1.14687741e+00 -2.40779631e-02 4.56555516e-01
-1.11944509e+00 -7.08528340e-01 -1.32710493e+00 -4.21288908e-01
3.83834541e-01 -7.14770973e-01 1.36620867e+00 -7.00500369e-01
-1.58526075e+00 5.51719069e-01 -3.31392020e-01 -4.48367387e-01
7.21044838e-01 -6.67433739e-01 -2.19189331e-01 -1.49868175e-01
1.92357674e-01 1.01083410e+00 9.07479703e-01 -1.27802837e+00
-5.44572711e-01 -2.52087384e-01 -1.56205962e-03 5.03476202e-01
2.17905164e-01 -2.41044700e-01 -9.47944343e-01 -2.02907979e-01
1.33584931e-01 -1.45525432e+00 -3.11039239e-01 1.77701905e-01
-3.73121291e-01 4.39811796e-01 1.24749148e+00 -3.23918194e-01
6.08266294e-01 -2.20641804e+00 3.99845034e-01 -1.37439609e-01
3.08015257e-01 -1.05483249e-01 -2.33237624e-01 2.17591614e-01
1.00102574e-01 -2.74830550e-01 1.30635630e-02 -4.63538677e-01
-4.65233773e-02 9.69638601e-02 -2.45814860e-01 9.67484713e-01
2.14132831e-01 8.18425119e-01 -7.03837693e-01 -2.57533669e-01
4.55428153e-01 6.96246803e-01 -5.87269068e-01 8.70903879e-02
-3.87699418e-02 6.15875542e-01 -3.03585202e-01 6.34128451e-01
9.75409091e-01 1.73033193e-01 -3.46201807e-01 -6.99342415e-02
-3.19588244e-01 -1.20029598e-01 -1.26507246e+00 1.61737502e+00
-3.53312194e-01 1.13465607e+00 4.38605249e-01 -5.35545945e-01
7.52754867e-01 -1.09508447e-01 4.99783933e-01 -5.15823901e-01
3.59846860e-01 6.19552843e-02 -2.05245882e-01 -4.56790984e-01
7.38665223e-01 4.91943270e-01 -2.76122898e-01 5.80543913e-02
-1.13379307e-01 -6.48152888e-01 8.34831297e-02 2.67981142e-01
8.77322435e-01 4.58281100e-01 -5.63567802e-02 -1.32146209e-01
4.06670958e-01 2.06958782e-02 8.88741374e-01 3.58875394e-01
-4.55656290e-01 9.44544256e-01 4.84140873e-01 -4.51554596e-01
-1.08203781e+00 -9.73549008e-01 -1.35308681e-02 7.58745372e-01
9.10893440e-01 -1.19856045e-01 -6.26223028e-01 -1.93031743e-01
3.09551328e-01 5.64650774e-01 -5.20834565e-01 -4.64069307e-01
-6.48072898e-01 -4.77594584e-01 4.54463214e-01 5.65966308e-01
5.92643857e-01 -3.87579322e-01 -5.35015643e-01 -1.22957574e-02
-2.14418992e-01 -1.33953071e+00 -8.69264603e-01 7.41543248e-02
-3.85667920e-01 -1.04728961e+00 -2.48844475e-01 -5.71483433e-01
5.09160519e-01 8.16944003e-01 8.33676875e-01 -2.18937695e-01
-1.74183786e-01 4.74130064e-01 1.44092590e-01 -5.99339962e-01
-1.96910530e-01 1.89488679e-01 3.66852462e-01 -6.98149949e-02
3.51477295e-01 -3.70635062e-01 -5.74005127e-01 6.27476215e-01
-5.75096488e-01 2.01162428e-01 4.43956584e-01 5.14886439e-01
5.34210861e-01 -4.47985768e-01 7.16696009e-02 -4.40529227e-01
2.31165513e-01 -6.86912000e-01 -1.15049303e+00 -3.00299227e-01
-3.78113538e-01 -1.89910419e-02 3.15533638e-01 -3.76645297e-01
-9.67536271e-01 5.72430611e-01 1.77015901e-01 -7.85280883e-01
-6.24351427e-02 2.43571386e-01 -1.49122462e-01 -3.45671743e-01
6.42565250e-01 5.47919050e-02 2.46703684e-01 6.42305762e-02
5.49481869e-01 5.77628493e-01 7.33566999e-01 -3.36666316e-01
1.19095016e+00 6.80536389e-01 -1.09087519e-01 -7.62268305e-01
-6.40594304e-01 -5.82033932e-01 -8.03226531e-01 -5.88095605e-01
8.99008989e-01 -1.57860684e+00 -7.20295072e-01 6.77043617e-01
-1.14188635e+00 -5.39462268e-01 2.65794069e-01 7.17054188e-01
-6.49928570e-01 8.59149918e-02 -3.54029030e-01 -4.24253434e-01
8.67792591e-02 -1.51779771e+00 1.21548057e+00 4.34543461e-01
1.01353668e-01 -7.70586431e-01 1.35319278e-01 4.25735205e-01
3.54150206e-01 3.36678416e-01 -2.17292741e-01 1.66879162e-01
-9.58008707e-01 -2.61966169e-01 2.06527766e-02 1.25212625e-01
-1.19120985e-01 4.76233006e-01 -9.24104273e-01 -4.91620988e-01
-2.76243061e-01 -3.33801031e-01 9.02472317e-01 5.20906568e-01
8.66012394e-01 -1.88443586e-02 -5.94129443e-01 1.42998946e+00
1.23577273e+00 3.05186473e-02 4.27322477e-01 4.76758093e-01
1.05773509e+00 3.35735500e-01 5.12167871e-01 1.58066139e-01
9.75273013e-01 5.78281343e-01 8.63606989e-01 -6.90846816e-02
5.47040850e-02 -9.28094834e-02 7.28062093e-01 7.07673430e-01
2.45702446e-01 -2.45788872e-01 -9.74573731e-01 7.09726512e-01
-2.08955026e+00 -8.44644547e-01 -2.45598465e-01 1.68154168e+00
2.85245240e-01 1.90738261e-01 -1.89113095e-01 -7.31970727e-01
4.86710101e-01 4.29641396e-01 -1.14235711e+00 -4.37761426e-01
-1.82325780e-01 -8.35618615e-01 1.26478803e+00 7.54439294e-01
-1.22477818e+00 1.14382827e+00 6.47501183e+00 -1.64476223e-02
-1.43344748e+00 -1.69764720e-02 3.72499973e-01 -5.08008718e-01
-1.82202116e-01 -9.60013792e-02 -9.59560335e-01 2.86384761e-01
8.78566980e-01 -8.88552740e-02 5.15897572e-01 9.97278392e-01
3.63064915e-01 -2.37389639e-01 -9.30123031e-01 1.00955164e+00
4.19682890e-01 -1.19382751e+00 -5.34836709e-01 4.22868505e-02
8.73554826e-01 1.03063619e+00 3.74547511e-01 2.40093201e-01
7.73786426e-01 -8.13993156e-01 1.00411153e+00 6.90601230e-01
5.86189806e-01 -7.28712618e-01 7.18414128e-01 4.00622994e-01
-1.22079539e+00 -1.69532806e-01 -4.61314917e-01 3.23542356e-02
2.41941407e-01 2.37714678e-01 -6.04403257e-01 2.55277246e-01
8.60033691e-01 1.29985285e+00 -6.01285100e-01 9.73617196e-01
-1.52203754e-01 3.28171670e-01 -4.83712852e-01 2.79825956e-01
3.09893817e-01 -2.01116398e-01 8.37450147e-01 1.00319004e+00
4.51784432e-01 -1.56428859e-01 1.13847964e-01 7.49709427e-01
5.17097376e-02 -4.76808012e-01 -1.13621807e+00 5.41189969e-01
3.95409077e-01 1.29381037e+00 -1.53197467e-01 1.24330908e-01
-7.07562268e-01 8.90764177e-01 3.51262301e-01 6.16567910e-01
-1.17238712e+00 -2.07575142e-01 1.24169624e+00 -6.57972917e-02
3.70690644e-01 -8.45668018e-01 -2.82456964e-01 -1.34022081e+00
1.62535515e-02 -6.35155916e-01 -1.19278923e-01 -1.16318405e+00
-6.77015543e-01 3.02271187e-01 1.04421377e-01 -1.49569035e+00
-2.90571451e-01 -5.95098674e-01 -5.47754049e-01 6.07888579e-01
-1.57969391e+00 -1.27426922e+00 -5.32984674e-01 4.78944898e-01
6.62233710e-01 -1.93944409e-01 1.51542485e-01 1.31605156e-02
-9.08741951e-01 3.98546785e-01 2.74739355e-01 -5.07235527e-02
9.88217533e-01 -1.17356980e+00 6.43405914e-01 1.16297209e+00
-3.80409777e-01 4.41239357e-01 9.52864468e-01 -4.21742499e-01
-2.01113558e+00 -1.32254457e+00 3.27907145e-01 -8.21076274e-01
6.70264482e-01 -5.98866642e-01 -5.75402439e-01 1.13992083e+00
3.69216233e-01 3.47242802e-01 7.58872405e-02 -3.32741141e-01
-2.52430230e-01 -3.81398290e-01 -7.09629118e-01 7.67883897e-01
9.49300766e-01 -3.59539598e-01 -7.45672500e-03 2.15134680e-01
6.83973372e-01 -1.16925085e+00 -4.88855928e-01 2.36386225e-01
5.73960662e-01 -8.30211043e-01 8.94745052e-01 -8.09793845e-02
6.05239309e-02 -6.88568592e-01 -2.32217789e-01 -1.44457746e+00
-2.35160992e-01 -7.60212660e-01 -1.07376389e-01 6.88593984e-01
6.04327381e-01 -8.02979529e-01 6.04835272e-01 6.30648136e-01
-4.25542951e-01 -4.63303447e-01 -8.72702181e-01 -5.67562878e-01
1.39125824e-01 -4.89897519e-01 4.98259157e-01 6.27824724e-01
-2.99240828e-01 2.15354457e-01 -5.59926808e-01 7.65340567e-01
6.18153572e-01 -1.90402597e-01 1.38589740e+00 -8.98359179e-01
9.89214852e-02 -1.77935079e-01 -5.73987186e-01 -1.18483579e+00
5.19353032e-01 -4.98714685e-01 4.70604002e-01 -1.37311172e+00
1.06582135e-01 -1.31514817e-01 1.29723281e-01 2.13156670e-01
-4.50380556e-02 3.19066525e-01 7.53277019e-02 1.81132481e-01
-8.03444564e-01 6.52978003e-01 1.22533607e+00 -2.69133717e-01
-1.58736348e-01 -2.24215891e-02 -7.05332994e-01 7.79674828e-01
1.14712787e+00 -3.37366998e-01 -5.13040185e-01 -1.04434156e+00
5.06796479e-01 -6.92821965e-02 4.09907341e-01 -1.24940670e+00
6.05741918e-01 -4.07507151e-01 3.30891758e-01 -6.94840193e-01
5.79483449e-01 -8.30064356e-01 2.54172266e-01 3.08782011e-01
2.25914363e-02 5.12085676e-01 5.14066517e-01 6.82980061e-01
-1.29093394e-01 2.87497222e-01 8.11962247e-01 5.03367446e-02
-8.59183967e-01 5.50375760e-01 -5.85040748e-01 2.31656106e-03
1.24675870e+00 -3.63180399e-01 -5.92122078e-01 -7.32338548e-01
-3.23246926e-01 8.05275619e-01 7.94628441e-01 9.25059855e-01
5.26139617e-01 -1.32121444e+00 -7.51581371e-01 4.40428227e-01
2.71936893e-01 4.08465594e-01 7.94066489e-02 9.73327756e-01
-8.36487710e-01 3.18817914e-01 -9.38696265e-02 -9.97337461e-01
-1.08354998e+00 3.83940190e-01 7.62535989e-01 5.63981473e-01
-5.45842171e-01 8.11794400e-01 2.90717721e-01 -8.92053604e-01
4.33803014e-02 -4.21993911e-01 6.28121868e-02 -2.89569169e-01
3.47561032e-01 3.02178741e-01 -1.26425549e-01 -1.20642567e+00
-3.63662124e-01 9.52979565e-01 -1.17883444e-01 -8.02898258e-02
1.11379564e+00 -6.01555228e-01 1.50035650e-01 3.63657951e-01
1.29202735e+00 4.97757718e-02 -2.15273428e+00 -6.56366274e-02
-4.80413496e-01 -2.41561249e-01 3.31315666e-01 -2.50085354e-01
-1.21616709e+00 6.97202146e-01 5.44931710e-01 -4.50048387e-01
6.42621815e-01 -1.15259595e-01 6.86133444e-01 6.03534818e-01
1.53602481e-01 -1.23819005e+00 -1.06878921e-01 9.18999374e-01
1.01746929e+00 -1.55151367e+00 1.34915225e-02 -1.99502975e-01
-7.43763983e-01 1.02539170e+00 1.11122811e+00 -1.27669513e-01
8.78171623e-01 7.66795754e-01 5.45618773e-01 -3.80459167e-02
-9.38162029e-01 -4.22077021e-04 1.96624268e-02 6.00628316e-01
-5.45849167e-02 -5.45387268e-02 4.37070459e-01 8.92634317e-02
-4.73834246e-01 -2.97582090e-01 7.07834363e-01 6.98508501e-01
-4.09724593e-01 -4.23592985e-01 -4.08090830e-01 -1.80707239e-02
6.01196215e-02 2.27566272e-01 -3.76875997e-01 8.42901647e-01
1.67103648e-01 8.93658161e-01 1.81420103e-01 -5.58200121e-01
3.56837481e-01 -3.81290585e-01 7.02032074e-02 -1.86727002e-01
-6.73426986e-02 3.64774693e-04 -1.29570648e-01 -1.07610106e+00
-3.09529454e-01 -1.01419520e+00 -1.19170761e+00 -6.26909733e-01
-3.30497712e-01 -4.13402975e-01 8.59365284e-01 7.79224455e-01
6.17426574e-01 4.18281108e-01 7.02884078e-01 -1.29866529e+00
-1.84489995e-01 -7.12292492e-01 -1.40740290e-01 -1.32379740e-01
9.80154276e-01 -8.16921353e-01 -4.23731148e-01 2.07902879e-01] | [8.02891731262207, -2.0426535606384277] |
0a703d1c-cc10-4c98-abe9-63161ac9d0cb | disambiguating-prepositional-phrase | null | null | https://aclanthology.org/P14-3010 | https://aclanthology.org/P14-3010.pdf | Disambiguating prepositional phrase attachment sites with sense information captured in contextualized distributional data | null | ['Clayton Greenberg'] | 2014-06-01 | null | null | null | acl-2014-6 | ['prepositional-phrase-attachment'] | ['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] | [-7.21804141998291, 3.846081018447876] |
cd17f394-00f4-41c0-979c-93e7d9c55521 | margin-preserving-self-paced-contrastive | 2103.08454 | null | https://arxiv.org/abs/2103.08454v2 | https://arxiv.org/pdf/2103.08454v2.pdf | Margin Preserving Self-paced Contrastive Learning Towards Domain Adaptation for Medical Image Segmentation | To bridge the gap between the source and target domains in unsupervised domain adaptation (UDA), the most common strategy puts focus on matching the marginal distributions in the feature space through adversarial learning. However, such category-agnostic global alignment lacks of exploiting the class-level joint distributions, causing the aligned distribution less discriminative. To address this issue, we propose in this paper a novel margin preserving self-paced contrastive Learning (MPSCL) model for cross-modal medical image segmentation. Unlike the conventional construction of contrastive pairs in contrastive learning, the domain-adaptive category prototypes are utilized to constitute the positive and negative sample pairs. With the guidance of progressively refined semantic prototypes, a novel margin preserving contrastive loss is proposed to boost the discriminability of embedded representation space. To enhance the supervision for contrastive learning, more informative pseudo-labels are generated in target domain in a self-paced way, thus benefiting the category-aware distribution alignment for UDA. Furthermore, the domain-invariant representations are learned through joint contrastive learning between the two domains. Extensive experiments on cross-modal cardiac segmentation tasks demonstrate that MPSCL significantly improves semantic segmentation performance, and outperforms a wide variety of state-of-the-art methods by a large margin. | ['Yao Zhao', 'Jiayu Zhou', 'Yang Liu', 'Shuai Zheng', 'Zhenfeng Zhu', 'Zhizhe Liu'] | 2021-03-15 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 4.72427458e-01 7.30258003e-02 -3.25452685e-01 -5.80321491e-01
-9.11638856e-01 -5.47813654e-01 4.80943501e-01 7.38794357e-03
-4.08573925e-01 6.30415201e-01 1.60376169e-03 1.32855743e-01
-1.44916207e-01 -7.76707947e-01 -5.59636712e-01 -1.08691978e+00
3.59704494e-01 3.50211561e-01 2.27228731e-01 -6.44278228e-02
-1.56147331e-01 2.28962377e-01 -9.48730350e-01 7.55593181e-02
1.34810781e+00 8.79576206e-01 1.74109071e-01 -4.65675071e-03
-2.77357638e-01 1.80823073e-01 -5.93774676e-01 -2.88310230e-01
2.88234890e-01 -9.10624802e-01 -6.55203700e-01 2.92443722e-01
2.70091921e-01 3.10191717e-02 -2.63535708e-01 1.33162344e+00
5.87943554e-01 1.79908991e-01 8.99976730e-01 -1.21847343e+00
-6.69741154e-01 3.46626669e-01 -7.38783836e-01 2.12593347e-01
-9.60225016e-02 2.45754525e-01 8.88669372e-01 -7.08279014e-01
5.17047703e-01 1.21150899e+00 5.49459100e-01 6.99160635e-01
-1.48602223e+00 -9.76536512e-01 2.23785758e-01 6.65455163e-02
-1.32449162e+00 -4.65367399e-02 1.46744597e+00 -5.03230333e-01
3.94757539e-02 1.22228526e-01 5.35178423e-01 1.21294379e+00
4.91409451e-02 8.92043591e-01 1.33531570e+00 -3.94697309e-01
3.12087446e-01 2.80038118e-01 -6.08049147e-02 6.13060534e-01
-5.94235621e-02 1.49992168e-01 -2.17093140e-01 -2.22685665e-01
7.87660718e-01 1.81011558e-01 -2.29290038e-01 -1.10467947e+00
-1.21207142e+00 9.38305497e-01 6.65944099e-01 4.44713831e-01
-3.23103338e-01 -4.56380606e-01 7.05090225e-01 1.00522041e-02
4.99065220e-01 3.65809917e-01 -3.35181385e-01 3.98044646e-01
-7.42790699e-01 1.39139205e-01 2.10211396e-01 6.93209827e-01
7.71615028e-01 1.09599099e-01 -5.28447926e-01 1.17318273e+00
3.32712263e-01 5.48675895e-01 8.99171829e-01 -6.89573944e-01
4.35884476e-01 7.17696369e-01 -3.49466950e-01 -9.34212327e-01
-9.32296813e-02 -8.03178191e-01 -1.13346791e+00 2.93441594e-01
5.59752464e-01 3.97025980e-03 -1.11433840e+00 2.06013489e+00
6.47006035e-01 2.78402984e-01 2.42811829e-01 9.19625759e-01
5.57278752e-01 4.15639520e-01 5.80039024e-01 -2.23492026e-01
1.35764182e+00 -8.42169285e-01 -7.06536889e-01 -4.53741312e-01
2.55536258e-01 -5.39500356e-01 1.17755759e+00 -7.99073875e-02
-7.66451597e-01 -8.50983858e-01 -1.21772552e+00 2.00480551e-01
-1.01233497e-01 -2.03103364e-01 3.11297029e-01 5.65390944e-01
-4.02863085e-01 4.42764580e-01 -7.73344994e-01 9.38685089e-02
9.58674908e-01 1.57911815e-02 -2.97254980e-01 -1.84981316e-01
-1.40496731e+00 5.37532449e-01 5.38500071e-01 -2.67804861e-01
-6.71151519e-01 -8.61577868e-01 -9.79802847e-01 -2.36446351e-01
6.98918924e-02 -6.87189758e-01 8.01744282e-01 -1.35891128e+00
-1.28588223e+00 1.24822879e+00 1.95183903e-01 -2.74620652e-01
7.55142868e-01 2.33531058e-01 -4.29752946e-01 2.56588310e-01
5.73475122e-01 8.41723680e-01 1.29986048e+00 -1.40848029e+00
-4.93829370e-01 -5.87878168e-01 -3.32043111e-01 4.29942369e-01
-5.03044546e-01 -5.07405579e-01 -2.98063785e-01 -1.25561905e+00
1.80041298e-01 -8.41100931e-01 -3.25115442e-01 2.47557670e-01
-3.26805711e-01 -1.04245916e-01 8.10338259e-01 -7.45911956e-01
9.32077408e-01 -2.47504282e+00 2.99800485e-01 3.26342374e-01
2.00603291e-01 3.49782497e-01 -8.92331079e-02 -3.49304169e-01
-2.25768298e-01 -3.17622244e-01 -1.02133131e+00 -2.01562390e-01
-1.02550708e-01 3.26322198e-01 -2.24944323e-01 5.07554591e-01
3.76047373e-01 7.37869263e-01 -1.15018320e+00 -8.37135792e-01
2.11137265e-01 4.33818281e-01 -5.20298481e-01 3.87337059e-01
-6.63915575e-02 1.01153672e+00 -7.35461235e-01 5.40994763e-01
9.11578536e-01 -5.96913025e-02 1.78497061e-01 -2.95937747e-01
5.64306259e-01 -1.80454612e-01 -8.51551533e-01 2.02328873e+00
-4.10849780e-01 9.05436650e-02 -1.16912462e-01 -1.54213309e+00
1.22454870e+00 8.57339352e-02 5.21887541e-01 -9.68999028e-01
-4.55597155e-02 2.89388895e-01 -5.24571612e-02 -2.11956784e-01
-1.26039945e-02 -6.25654340e-01 -3.32252502e-01 2.25172639e-02
8.86424333e-02 -1.78688467e-01 -2.58691490e-01 -1.09602675e-01
5.37879109e-01 2.17909127e-01 3.76244992e-01 -3.50675553e-01
6.82847917e-01 -5.61906844e-02 1.04247212e+00 3.84299517e-01
-6.72580063e-01 8.12256932e-01 2.53554493e-01 -9.11888480e-02
-9.24340129e-01 -1.45197427e+00 -5.15574574e-01 7.97210455e-01
4.52160239e-01 3.59297067e-01 -9.57464337e-01 -1.24352109e+00
-1.48872957e-02 6.47636294e-01 -7.49738693e-01 -6.46069944e-01
-5.67594886e-01 -7.75489748e-01 4.76733804e-01 6.09106064e-01
8.24432731e-01 -1.07013607e+00 -2.89286494e-01 3.05880129e-01
-1.97965026e-01 -8.69874239e-01 -8.38571310e-01 1.35909438e-01
-1.14278364e+00 -9.73769784e-01 -1.27603996e+00 -1.11317217e+00
9.06747341e-01 -5.94777763e-02 1.01069176e+00 -4.80823517e-01
-2.65731186e-01 2.98566878e-01 -2.84005702e-01 -1.01623386e-01
-6.21373117e-01 -1.16143614e-01 -1.63852766e-01 3.22781622e-01
3.65448534e-01 -7.63606668e-01 -9.11696196e-01 3.16197991e-01
-9.04813707e-01 -5.33835292e-02 6.58325195e-01 1.25556469e+00
9.36254025e-01 -7.86955059e-02 1.00871086e+00 -1.13690841e+00
3.45761836e-01 -5.94240904e-01 -1.96586847e-01 1.11335337e-01
-7.06948221e-01 2.99365353e-02 8.65253508e-01 -7.27896214e-01
-1.20544064e+00 4.98848036e-02 -1.03317380e-01 -7.51760364e-01
-3.15648109e-01 1.07633390e-01 -5.27433693e-01 2.05841482e-01
7.82123983e-01 5.17699540e-01 4.21039462e-01 -1.49982914e-01
5.08978426e-01 4.53761518e-01 7.48138309e-01 -5.69223404e-01
9.81248617e-01 5.72961390e-01 -1.70156196e-01 -2.89808840e-01
-8.31200242e-01 -4.57435429e-01 -6.55353665e-01 5.12477607e-02
1.01346004e+00 -9.38464761e-01 1.84157029e-01 6.62504911e-01
-6.48891449e-01 -2.51134366e-01 -8.36294353e-01 4.34139013e-01
-7.43221283e-01 5.21746218e-01 -3.00509334e-01 -2.35273853e-01
-3.09714079e-01 -1.21812773e+00 8.80640626e-01 4.64055717e-01
-1.25785321e-01 -1.16977835e+00 1.80501491e-01 4.84174848e-01
1.29272789e-01 4.73648906e-01 1.03176475e+00 -8.70938659e-01
-1.52408481e-01 -1.37166172e-01 -1.87585875e-01 8.27381194e-01
6.06086314e-01 -6.38281941e-01 -9.14468527e-01 -4.78432745e-01
7.79509991e-02 -2.54993975e-01 7.54284620e-01 3.82600218e-01
1.31152785e+00 -1.57307744e-01 -4.03893501e-01 6.93816602e-01
1.27886939e+00 2.87815392e-01 5.12194753e-01 3.07257295e-01
8.99180889e-01 6.57646477e-01 9.37963367e-01 2.01185405e-01
1.50472060e-01 5.49921989e-01 2.03222170e-01 -3.81849736e-01
-3.27463388e-01 -3.85658890e-01 -8.48998129e-02 6.40347123e-01
4.48758066e-01 1.57815918e-01 -6.58589780e-01 7.44146287e-01
-1.58003521e+00 -6.48037195e-01 5.35636067e-01 2.12469840e+00
1.21930504e+00 1.59143284e-01 1.97425872e-01 1.09237455e-01
1.06221271e+00 2.87403286e-01 -9.88446832e-01 4.19119149e-02
-2.85765678e-02 1.75937295e-01 2.66173005e-01 1.87650472e-01
-1.38452661e+00 6.57240272e-01 4.82668924e+00 1.40079284e+00
-1.03937256e+00 3.54070067e-01 8.75321627e-01 3.45304906e-01
-4.43974227e-01 -2.40503684e-01 -2.54149824e-01 8.02249432e-01
3.10095787e-01 -1.09881852e-02 -3.83523991e-03 1.07391310e+00
-9.42418575e-02 3.47628176e-01 -8.40186417e-01 1.03507733e+00
2.23528817e-02 -1.16328049e+00 2.77792290e-02 -8.46367404e-02
9.12311196e-01 -4.55205739e-01 3.35498691e-01 4.98647630e-01
7.28861392e-02 -6.43605590e-01 5.45024931e-01 3.91623437e-01
1.06633162e+00 -9.35748577e-01 5.79663396e-01 7.85372704e-02
-9.85364318e-01 -4.39697243e-02 -2.30088472e-01 6.07375145e-01
3.07771731e-02 5.38477778e-01 -6.08125031e-01 5.98450303e-01
6.19845450e-01 8.06024194e-01 -4.75235939e-01 6.79493904e-01
-7.31197521e-02 5.49152970e-01 1.59682661e-01 3.90621930e-01
2.05157846e-01 -4.18423712e-01 8.04999590e-01 1.12122655e+00
-7.64753819e-02 -1.57828391e-01 3.40375632e-01 1.00336242e+00
-1.34824142e-01 8.29602331e-02 -3.94807100e-01 2.83144474e-01
4.94115591e-01 1.05414915e+00 -7.09182858e-01 -3.47880810e-01
-2.14266896e-01 1.25211334e+00 1.54080331e-01 3.07620376e-01
-8.48135829e-01 -3.86258513e-01 5.84424853e-01 1.09773383e-01
3.07964653e-01 2.84492612e-01 -4.30717766e-01 -8.95463526e-01
4.26106006e-02 -9.36833262e-01 6.61149919e-01 -2.79969037e-01
-1.89260399e+00 4.75512922e-01 1.05692118e-01 -1.71369648e+00
-1.00875020e-01 -1.54348761e-01 -6.96944237e-01 1.02990246e+00
-1.74076438e+00 -1.26262426e+00 -2.49209240e-01 7.68095493e-01
5.91772318e-01 -4.06159103e-01 7.21010804e-01 4.14772391e-01
-3.85281175e-01 9.64850068e-01 3.00189435e-01 3.84653270e-01
1.06790340e+00 -1.31224871e+00 8.32967013e-02 7.38234520e-01
-2.56204337e-01 4.05938506e-01 5.31250119e-01 -5.27629137e-01
-6.03236079e-01 -1.37203169e+00 1.21794805e-01 -7.55024999e-02
3.42981279e-01 -8.06442499e-02 -1.35490656e+00 3.34236354e-01
-3.05170147e-03 4.10939038e-01 7.13222682e-01 -1.83381394e-01
-6.13492012e-01 -4.20001626e-01 -1.58394384e+00 5.26240528e-01
8.92643154e-01 -6.57379150e-01 -8.57163846e-01 2.25176752e-01
7.98307061e-01 -2.95372009e-01 -1.02749395e+00 6.99901879e-01
2.28920847e-01 -6.75338209e-01 1.18693948e+00 -5.03537476e-01
3.19357753e-01 -3.50264996e-01 2.19113044e-02 -1.42701352e+00
-3.15319955e-01 -2.06926018e-01 1.56571671e-01 1.45994890e+00
2.21673399e-03 -8.68438840e-01 7.26943851e-01 1.87587574e-01
-3.19964439e-01 -7.18445182e-01 -1.09657180e+00 -7.66689658e-01
5.53782940e-01 6.68402463e-02 3.88685882e-01 1.35176551e+00
-2.99230039e-01 3.12202811e-01 1.46417338e-02 1.29697964e-01
9.38845456e-01 3.52317542e-01 3.50955158e-01 -1.04417551e+00
-4.32273448e-01 -4.80683953e-01 -4.78123248e-01 -8.67473006e-01
4.16395366e-01 -1.15836430e+00 1.51458114e-01 -1.16733539e+00
3.45834017e-01 -8.99102092e-01 -6.84883952e-01 4.01468575e-01
-5.72840691e-01 3.75973463e-01 1.22835683e-02 3.19739282e-01
-3.73412371e-01 8.28688681e-01 1.67235518e+00 -5.13785958e-01
-1.53214172e-01 9.12118852e-02 -8.80727053e-01 6.20125055e-01
8.36051941e-01 -5.36498904e-01 -6.57494545e-01 8.05791765e-02
-6.79776192e-01 -4.27733138e-02 4.63163614e-01 -8.29492688e-01
-1.26936585e-01 -8.33572373e-02 5.47971904e-01 -3.37996602e-01
4.97955680e-02 -6.59400702e-01 -2.61851430e-01 5.18356204e-01
-3.81262541e-01 -5.85237622e-01 1.99642420e-01 8.91299129e-01
-4.80077624e-01 -5.26470318e-02 1.46628737e+00 -7.60706216e-02
-8.08533072e-01 5.25019526e-01 9.24674571e-02 5.11333466e-01
1.06074154e+00 -4.23734039e-01 6.56613782e-02 2.08498016e-01
-8.32067847e-01 1.88542217e-01 4.06066537e-01 4.40993190e-01
5.51872969e-01 -1.44384980e+00 -6.49607420e-01 3.95295858e-01
3.19424808e-01 3.92940521e-01 7.80920029e-01 6.02059484e-01
-1.33090660e-01 -1.42180935e-01 -5.71888804e-01 -9.27070320e-01
-8.14214051e-01 6.53736115e-01 5.00113368e-01 -4.17328775e-01
-6.94983006e-01 7.75328636e-01 8.44117343e-01 -7.63708830e-01
4.42827120e-02 1.51119053e-01 -2.23472014e-01 -4.15701605e-03
3.43068779e-01 8.91154632e-02 -1.64954782e-01 -5.96011698e-01
-4.19445187e-01 6.91109538e-01 -1.92328990e-01 1.12529941e-01
9.74943817e-01 -1.99075386e-01 2.56677747e-01 3.15683246e-01
1.36843920e+00 -1.43964365e-02 -1.72052240e+00 -4.72215474e-01
-3.76176834e-01 -4.19866800e-01 -1.27106085e-01 -7.67089188e-01
-1.20861697e+00 7.80987918e-01 1.17853987e+00 -1.72402471e-01
1.42253363e+00 1.31525129e-01 9.15576220e-01 -4.05743718e-01
-1.50478957e-02 -1.02254725e+00 4.32219654e-01 -1.21619165e-01
6.96152151e-01 -1.34206057e+00 -3.14941287e-01 -4.60192502e-01
-9.31281924e-01 6.72745228e-01 6.94436193e-01 -3.72218132e-01
4.87269342e-01 -1.62775010e-01 2.09424108e-01 5.90286553e-02
1.70137733e-01 -3.25997360e-04 3.26962411e-01 1.06581414e+00
2.28074454e-02 2.47269928e-01 -2.87285805e-01 7.01230764e-01
2.06702143e-01 -4.47249264e-01 -8.88745412e-02 7.48852730e-01
-6.68284595e-02 -1.26752996e+00 -3.86435449e-01 2.41851032e-01
-3.43777001e-01 1.00945331e-01 1.61889195e-01 7.69513905e-01
2.44700760e-01 4.92659599e-01 1.74010485e-01 -1.60099864e-01
4.14908588e-01 1.90588105e-02 4.22155976e-01 -5.15178919e-01
-2.21490324e-01 1.49375230e-01 -4.11912382e-01 -2.16175526e-01
-4.03857350e-01 -7.70909190e-01 -1.43309176e+00 3.35430950e-01
-6.68794662e-02 2.62127444e-02 2.07190141e-01 8.92602801e-01
2.22502723e-01 7.86934853e-01 9.87475812e-01 -4.79600281e-01
-7.63947964e-01 -7.52102256e-01 -7.34029412e-01 9.53991711e-01
3.48759234e-01 -8.32921624e-01 -3.61968875e-01 1.13310285e-01] | [14.473763465881348, -1.871621012687683] |
06dd4ef2-1a87-45d4-9899-e620ba46c4cc | on-the-descriptive-power-of-lidar-intensity | 2108.01383 | null | https://arxiv.org/abs/2108.01383v1 | https://arxiv.org/pdf/2108.01383v1.pdf | On the descriptive power of LiDAR intensity images for segment-based loop closing in 3-D SLAM | We propose an extension to the segment-based global localization method for LiDAR SLAM using descriptors learned considering the visual context of the segments. A new architecture of the deep neural network is presented that learns the visual context acquired from synthetic LiDAR intensity images. This approach allows a single multi-beam LiDAR to produce rich and highly descriptive location signatures. The method is tested on two public datasets, demonstrating an improved descriptiveness of the new descriptors, and more reliable loop closure detection in SLAM. Attention analysis of the network is used to show the importance of focusing on the broader context rather than only on the 3-D segment. | ['Piotr Skrzypczyński', 'Jan Wietrzykowski'] | 2021-08-03 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [ 1.40943721e-01 -1.90708101e-01 -3.12166601e-01 -9.33548450e-01
-6.98716164e-01 -4.42449838e-01 7.78213263e-01 2.06276193e-01
-6.97661161e-01 7.52902687e-01 1.80107266e-01 -3.77040878e-02
-4.07713115e-01 -7.27801621e-01 -8.06537747e-01 -3.60217452e-01
-1.27103239e-01 6.91836059e-01 6.44228831e-02 -2.31283084e-01
3.50788713e-01 1.25026894e+00 -1.70078886e+00 -2.16971830e-01
5.30130744e-01 1.01045501e+00 2.84845024e-01 3.66367847e-01
-5.93194440e-02 4.00805295e-01 -5.08778334e-01 2.47709736e-01
4.42183912e-01 2.76858527e-02 -3.55104327e-01 -1.90930784e-01
1.32425678e+00 -4.52939510e-01 -3.12950432e-01 8.29580247e-01
4.25703496e-01 2.93223560e-01 5.33546448e-01 -1.20049238e+00
-5.14832377e-01 2.30774939e-01 -4.59621161e-01 6.15275726e-02
3.53455722e-01 1.82783663e-01 1.27915382e+00 -1.03333986e+00
1.02900934e+00 1.48656619e+00 1.03667653e+00 -1.31818950e-01
-1.53321099e+00 -6.79505110e-01 -5.06772995e-02 1.69403926e-01
-1.73120773e+00 -2.41746694e-01 7.40309000e-01 -3.10004264e-01
1.00886512e+00 -1.31792098e-01 7.94277251e-01 9.50790644e-01
4.88406479e-01 4.00147945e-01 1.07935095e+00 -1.63460419e-01
2.35838503e-01 -1.68492362e-01 -9.24327224e-02 7.76348531e-01
5.37390947e-01 6.38238549e-01 -8.74753892e-01 7.65860546e-04
5.74539483e-01 1.31458819e-01 5.57468273e-02 -1.05555761e+00
-1.18651295e+00 1.19306815e+00 1.30962849e+00 2.19636038e-01
-4.73250419e-01 7.35552073e-01 2.75616258e-01 -1.77556887e-01
2.09239438e-01 4.17612791e-01 -1.38228029e-01 5.19765541e-02
-1.29485095e+00 3.52133602e-01 3.10692310e-01 9.89231884e-01
1.52529562e+00 1.52488261e-01 1.17840379e-01 2.26102114e-01
4.73028004e-01 1.03225243e+00 -9.77900103e-02 -1.15574181e+00
1.07479721e-01 5.91684163e-01 -2.71290988e-02 -9.50701118e-01
-6.64958417e-01 -5.78872919e-01 -3.62328768e-01 5.44327140e-01
-3.90034355e-02 2.33236089e-01 -1.06599593e+00 1.74106014e+00
1.75140440e-01 1.16363876e-01 -1.67347327e-01 7.19704688e-01
5.22287488e-01 1.31742254e-01 -1.39239524e-02 5.02867043e-01
7.81016231e-01 -4.43973869e-01 -4.57897544e-01 -5.41878045e-01
5.25767803e-01 -3.13941270e-01 7.75120139e-01 -5.93356490e-02
-5.02565503e-01 -7.70128071e-01 -1.38652241e+00 -4.10456061e-01
-6.83466673e-01 1.37090027e-01 6.02191150e-01 1.98731497e-01
-1.20938015e+00 5.06881416e-01 -5.91102779e-01 -7.16754675e-01
5.62280476e-01 3.83401036e-01 -5.25558412e-01 -3.05926561e-01
-9.39562738e-01 1.12053871e+00 4.39331293e-01 5.71558289e-02
-9.66123462e-01 -6.36501729e-01 -1.40647411e+00 -2.32310668e-01
-2.26501673e-01 -5.55562973e-01 8.05907309e-01 -5.71250737e-01
-8.15658987e-01 9.06596720e-01 -2.37355381e-01 -8.11548412e-01
4.96642768e-01 -3.16264570e-01 1.06726162e-01 2.81621218e-01
5.62172890e-01 1.35290623e+00 5.67844331e-01 -1.50003088e+00
-8.09555173e-01 -3.91216397e-01 -7.21223950e-02 1.06004514e-01
5.94906628e-01 -5.87966502e-01 3.99706960e-02 -1.25856921e-01
3.76985043e-01 -1.12251544e+00 -1.58834681e-01 3.88120294e-01
-9.93223041e-02 -1.18929207e-01 1.13909733e+00 -2.09632248e-01
4.17881489e-01 -2.28368020e+00 -1.57630891e-01 2.67012537e-01
1.78187937e-02 -2.38846496e-01 -2.62611985e-01 4.72638786e-01
1.57109097e-01 -4.87077273e-02 -4.28159982e-01 -4.95421767e-01
1.75289437e-01 7.07826853e-01 -6.19077742e-01 8.36180151e-01
1.77889928e-01 1.20420778e+00 -9.96348500e-01 -3.54007185e-01
6.85176253e-01 6.22564077e-01 -4.30540502e-01 -3.32120433e-02
-7.70538226e-02 5.82098961e-01 3.27355564e-02 8.33095253e-01
9.34187353e-01 3.13721180e-01 -9.82336700e-02 -2.61382014e-01
-3.08332264e-01 3.02175254e-01 -8.79741073e-01 2.36456680e+00
-6.26707435e-01 1.12296605e+00 -1.06066361e-01 -5.68072617e-01
1.15989518e+00 -3.84096980e-01 3.79638672e-01 -9.82013464e-01
-4.84520830e-02 2.16232419e-01 -2.14349955e-01 -8.48156288e-02
7.44126737e-01 -1.80881873e-01 -3.25235784e-01 2.24377260e-01
2.21465394e-01 -6.99153125e-01 -8.28330219e-02 2.12822616e-01
7.75086999e-01 4.55672741e-01 2.91630059e-01 -4.16830003e-01
1.55401215e-01 2.18922585e-01 3.33470583e-01 1.01252151e+00
-3.96256834e-01 4.55139577e-01 -1.70266181e-01 -6.27889991e-01
-9.19242442e-01 -1.38290572e+00 -4.04589832e-01 9.19752181e-01
4.51629043e-01 -1.97462261e-01 -1.31401522e-02 -5.92523098e-01
7.79564500e-01 8.10945213e-01 -8.93943489e-01 -2.61080950e-01
-5.43660045e-01 4.78966050e-02 8.31904292e-01 6.17013574e-01
6.81223631e-01 -9.19060647e-01 -1.25716066e+00 1.22776777e-02
5.54110631e-02 -1.18855393e+00 1.61573887e-02 5.22062182e-01
-7.63481557e-01 -9.02411520e-01 1.79405972e-01 -5.20657420e-01
2.64976293e-01 3.46922547e-01 9.78144228e-01 -6.50482625e-02
-3.95139903e-01 7.20751882e-01 9.83405113e-03 -4.70308274e-01
-2.85629816e-02 1.37835860e-01 1.86701611e-01 -2.77380973e-01
4.50774997e-01 -6.05793357e-01 -3.52985412e-01 5.35603659e-03
-5.44237673e-01 -3.71061414e-01 7.42242217e-01 8.21699679e-01
7.48268485e-01 -7.07732916e-01 2.12309554e-01 -3.79397392e-01
2.86005765e-01 -3.61983269e-01 -5.87903082e-01 -1.32765666e-01
-7.44542241e-01 2.44569585e-01 -3.36792320e-02 -3.77560332e-02
-6.82300270e-01 4.11218554e-01 3.93624566e-02 -4.77149218e-01
-4.92035300e-01 5.76034665e-01 1.15289338e-01 -6.12809420e-01
7.64039040e-01 1.14980973e-01 -1.38266027e-01 -2.68563658e-01
7.48752713e-01 2.45341867e-01 6.91706181e-01 -5.10603666e-01
9.72984850e-01 9.63950574e-01 5.77598512e-01 -9.80503559e-01
-5.60885906e-01 -4.72857475e-01 -1.21281350e+00 -4.45747674e-02
6.96215689e-01 -1.05799258e+00 -6.56544745e-01 -6.50954917e-02
-1.17551327e+00 -2.20325336e-01 -7.09547222e-01 5.36092520e-01
-7.41300523e-01 2.76100338e-01 -3.91803607e-02 -5.98894060e-01
1.02511734e-01 -1.11073852e+00 1.60347795e+00 6.42068163e-02
-2.13099077e-01 -8.84266973e-01 4.82108295e-01 -1.06376238e-01
4.91734803e-01 6.21575296e-01 6.40513539e-01 -5.65180838e-01
-1.11138773e+00 -3.68079454e-01 -4.62695271e-01 3.18867415e-02
-1.11637041e-01 4.66978513e-02 -1.11131346e+00 -3.99013281e-01
-4.20587510e-01 -3.22047472e-01 1.10602260e+00 2.25553989e-01
5.58519959e-01 2.82781243e-01 -4.95725244e-01 1.04625297e+00
1.67605329e+00 -1.39502406e-01 3.83574724e-01 3.76112550e-01
7.19321311e-01 4.87609923e-01 7.69094944e-01 2.71962702e-01
5.00240982e-01 7.19217598e-01 1.03480494e+00 -9.34607387e-02
-2.17243165e-01 -6.33236945e-01 2.41223454e-01 1.63660273e-01
2.45894521e-01 2.78174669e-01 -1.04798782e+00 7.86648333e-01
-1.60930181e+00 -7.24283934e-01 1.62989777e-02 2.07307816e+00
1.28407359e-01 3.26342732e-02 -4.16762292e-01 -2.96123624e-01
4.50786591e-01 5.18760920e-01 -6.05523467e-01 -4.77012813e-01
-3.22907597e-01 4.88497853e-01 8.59388649e-01 8.19834590e-01
-1.09308183e+00 1.26449883e+00 7.08459520e+00 5.23262441e-01
-1.38143921e+00 4.19945046e-02 -3.95823598e-01 -1.13862574e-01
-3.05503041e-01 2.95704901e-01 -8.12739491e-01 -2.84288507e-02
8.98251414e-01 9.23366547e-02 2.99323142e-01 9.22834575e-01
1.02664843e-01 -6.40864968e-01 -1.06984556e+00 9.63703275e-01
9.56481397e-02 -1.44098115e+00 2.52919972e-01 3.45334172e-01
5.02756000e-01 8.65280867e-01 2.26817038e-02 3.69937092e-01
2.17506960e-01 -1.05463648e+00 1.10076261e+00 5.63256323e-01
9.70961928e-01 -8.49761724e-01 5.52804470e-01 2.49791309e-01
-1.45243466e+00 -1.39014363e-01 -4.78630543e-01 -1.74168631e-01
4.29141410e-02 2.27657363e-01 -1.39096510e+00 5.27849138e-01
5.99544704e-01 1.00146782e+00 -1.05397737e+00 1.00255513e+00
-5.86687565e-01 9.64523479e-02 -6.43301368e-01 2.11493075e-01
5.32212317e-01 -1.01814546e-01 6.86095357e-01 1.47787917e+00
2.53322273e-01 -6.31667972e-01 3.61133009e-01 1.21114779e+00
-1.60260014e-02 -1.14127651e-01 -1.51186681e+00 3.02466303e-01
7.16156244e-01 1.25806201e+00 -3.75843883e-01 -1.02353990e-01
-7.02105239e-02 6.20008826e-01 4.18949634e-01 4.78755444e-01
-7.47027695e-01 -5.25720239e-01 8.63496900e-01 -5.37277311e-02
4.27974731e-01 -8.61518443e-01 -3.40361774e-01 -6.05095148e-01
-2.89469659e-01 -1.24079451e-01 4.88481810e-03 -1.17869902e+00
-8.35493982e-01 2.67330319e-01 2.21703187e-01 -1.09187770e+00
-3.34315956e-01 -4.69280481e-01 -3.38724017e-01 1.01564717e+00
-1.83971190e+00 -1.73766410e+00 -7.24994838e-01 5.43206036e-01
1.09466858e-01 2.23279566e-01 7.11786509e-01 5.03361188e-02
2.54473150e-01 3.23647231e-01 1.65065210e-02 -6.34328127e-02
8.44273746e-01 -1.28154230e+00 4.02154177e-01 8.31610084e-01
4.68192697e-01 8.32735121e-01 6.71456575e-01 -6.71596885e-01
-1.18110430e+00 -1.17571676e+00 8.29230249e-01 -6.21267855e-01
4.78575259e-01 -4.89365339e-01 -6.68198943e-01 8.91350448e-01
1.45929843e-01 1.30122036e-01 2.91088045e-01 1.64925233e-02
-4.89995241e-01 -3.23312163e-01 -1.32330513e+00 1.67622671e-01
1.18010151e+00 -1.04701042e+00 -8.04338753e-01 2.27336641e-02
6.23099566e-01 -4.45190609e-01 -4.92455274e-01 7.61464596e-01
5.86918056e-01 -1.18956029e+00 1.08088183e+00 -2.11763799e-01
3.27191921e-03 -5.60467303e-01 -7.29876757e-01 -1.11647129e+00
-1.65226638e-01 3.13992947e-02 1.38116658e-01 7.15486109e-01
-5.19314483e-02 -6.91602409e-01 6.62906885e-01 -2.37646058e-01
-1.72311008e-01 -3.81461978e-01 -1.60441113e+00 -8.40025246e-01
-9.12543014e-02 -5.97414434e-01 6.43682182e-01 6.70838594e-01
-6.58711135e-01 4.08747733e-01 -9.85269174e-02 4.07217175e-01
7.53825545e-01 3.89733046e-01 9.76693571e-01 -1.66349769e+00
4.29305553e-01 -2.49292910e-01 -9.44949448e-01 -9.95468259e-01
4.27816421e-01 -1.19212234e+00 3.37082922e-01 -1.61457253e+00
-2.16966614e-01 -3.47500741e-01 -2.23542541e-01 3.77975821e-01
5.12034655e-01 6.49066508e-01 3.11926633e-01 1.84669539e-01
-5.69564104e-01 7.39280045e-01 7.28061080e-01 -3.33249748e-01
1.23253159e-01 -3.13345730e-01 -2.11221442e-01 5.97904325e-01
5.16648948e-01 -5.45074344e-01 -1.39330462e-01 -4.68687743e-01
2.77790904e-01 -5.05100965e-01 8.87600601e-01 -1.39328885e+00
2.45758399e-01 -9.28616598e-02 5.04662633e-01 -1.24352717e+00
6.72109365e-01 -1.06128263e+00 5.48939034e-02 7.04335570e-01
-1.74192443e-01 1.94195896e-01 3.92866254e-01 8.60355735e-01
-1.98238432e-01 -1.31628625e-02 8.37911248e-01 -5.26291579e-02
-1.26396453e+00 1.81058481e-01 -1.28229216e-01 -2.30042145e-01
7.59313583e-01 -4.60071087e-01 -1.57375664e-01 -4.44339961e-01
-5.41374981e-01 8.01585913e-02 9.31354880e-01 4.54395831e-01
8.36466908e-01 -1.61280882e+00 -3.80842239e-01 6.17789209e-01
5.96911311e-01 8.93029496e-02 -7.23804012e-02 7.93395758e-01
-8.42605174e-01 6.27405941e-01 -4.33618665e-01 -1.35430837e+00
-9.50381041e-01 4.29119080e-01 5.64584017e-01 1.69698760e-01
-6.31193697e-01 5.86746395e-01 3.92725393e-02 -7.23631024e-01
1.30010694e-01 -7.73983777e-01 2.47638792e-01 2.77599663e-01
-7.01801255e-02 2.46560022e-01 7.03108236e-02 -1.23961985e+00
-7.90301621e-01 9.97591019e-01 5.46016335e-01 -2.65797168e-01
1.28696847e+00 -4.54866230e-01 -1.32534564e-01 6.19351804e-01
1.24484587e+00 3.55041921e-01 -1.34344625e+00 -4.04371589e-01
1.79199457e-01 -5.89822650e-01 1.06379144e-01 -7.73798764e-01
-6.26924932e-01 1.18720019e+00 1.03545797e+00 -3.99221927e-01
5.51522613e-01 -1.12189561e-01 4.79794115e-01 7.88701475e-01
8.87742579e-01 -8.50196540e-01 -1.33726627e-01 1.04259980e+00
9.89238620e-01 -1.34577560e+00 3.25779557e-01 3.53705525e-01
-4.32925284e-01 1.04697728e+00 2.44402409e-01 -3.71625096e-01
4.84972417e-01 -5.29176779e-02 2.49168739e-01 -4.39368844e-01
-6.53621852e-02 -5.13565481e-01 2.95148641e-01 8.14909995e-01
-8.84553865e-02 6.43750951e-02 1.63109094e-01 -2.10564628e-01
-1.11186676e-01 -3.12127084e-01 2.58447915e-01 1.11038435e+00
-6.79178357e-01 -7.68328726e-01 -2.83916593e-01 -7.30216503e-02
3.50538015e-01 -1.98788140e-02 -5.68696380e-01 1.20649946e+00
7.09484398e-01 5.05598903e-01 2.12012365e-01 -3.65093708e-01
3.57022524e-01 1.80302188e-01 5.54001331e-01 -6.66251481e-01
-3.02053541e-01 -3.44548285e-01 -2.17125669e-01 -9.42216635e-01
-5.38539112e-01 -6.53439701e-01 -1.33394980e+00 -1.57748207e-01
-6.26593903e-02 -5.86731732e-02 1.20746219e+00 7.86366165e-01
5.33322871e-01 2.08936304e-01 4.49231207e-01 -1.48139918e+00
-2.79226124e-01 -9.67483163e-01 -6.20605230e-01 3.12653661e-01
8.58950853e-01 -1.16746366e+00 -3.16448092e-01 -4.89689469e-01] | [7.58914041519165, -2.142848491668701] |
9bde79e2-28af-48ab-bae0-1c34564257e1 | a-prompt-independent-and-interpretable | null | null | https://aclanthology.org/2021.ccl-1.107 | https://aclanthology.org/2021.ccl-1.107.pdf | A Prompt-independent and Interpretable Automated Essay Scoring Method for Chinese Second Language Writing | “With the increasing popularity of learning Chinese as a second language (L2) the development of an automatic essay scoring (AES) method specially for Chinese L2 essays has become animportant task. To build a robust model that could easily adapt to prompt changes we propose 90linguistic features with consideration of both language complexity and correctness and introducethe Ordinal Logistic Regression model that explicitly combines these linguistic features and low-level textual representations. Our model obtains a high QWK of 0.714 a low RMSE of 1.516 anda considerable Pearson correlation of 0.734. With a simple linear model we further analyze the contribution of the linguistic features to score prediction revealing the model’s interpretability and its potential to give writing feedback to users. This work provides insights and establishes asolid baseline for Chinese L2 AES studies.” | ['Hu Renfen', 'Wang Yupei'] | null | null | null | null | ccl-2021-8 | ['automated-essay-scoring'] | ['natural-language-processing'] | [-4.44835782e-01 1.59828231e-01 -5.08352160e-01 -4.23866451e-01
-1.14095628e+00 -6.89329445e-01 5.65371513e-01 4.52062041e-01
-5.88621736e-01 8.90144944e-01 4.91758823e-01 -5.23187280e-01
-2.25973979e-01 -3.58293563e-01 -3.16998839e-01 8.82182196e-02
3.40409935e-01 4.91119772e-02 -1.23448491e-01 -3.82729739e-01
6.06654465e-01 2.53029138e-01 -1.23261881e+00 3.29483598e-01
1.40399480e+00 6.35420084e-01 2.97834426e-01 6.96012735e-01
-2.37190649e-01 1.13622820e+00 -8.40895474e-01 -7.42324829e-01
-2.37425976e-02 -4.58768755e-01 -7.88830042e-01 -7.98077211e-02
4.45427150e-01 -3.68771344e-01 -3.11663896e-02 4.73687589e-01
2.63731390e-01 -6.65198592e-03 6.86518371e-01 -7.28498578e-01
-1.09265363e+00 8.40262294e-01 -3.43857825e-01 5.52283786e-02
7.30044603e-01 -1.10376023e-01 1.40415990e+00 -1.09103358e+00
5.76606810e-01 6.78202450e-01 9.70687151e-01 3.81429702e-01
-1.19567788e+00 -5.95197439e-01 6.45595938e-02 1.95354700e-01
-8.34151268e-01 -4.27729547e-01 5.83562791e-01 -5.09452760e-01
7.71234393e-01 3.13940704e-01 6.23116851e-01 7.93915153e-01
2.00304657e-01 1.02336192e+00 1.57772601e+00 -7.82642305e-01
-2.27630615e-01 7.11900711e-01 4.17524040e-01 6.80089653e-01
1.90570876e-01 -2.29885191e-01 -9.37301993e-01 1.81556329e-01
2.47380659e-01 -3.46455157e-01 2.06283964e-02 3.93790752e-01
-9.27813351e-01 9.01834965e-01 1.28732607e-01 4.17045772e-01
6.70184046e-02 -8.69127661e-02 3.25214773e-01 7.58401930e-01
3.42225105e-01 1.08953214e+00 -7.06203043e-01 -5.28981745e-01
-1.12905109e+00 3.13752979e-01 8.42171788e-01 6.85694337e-01
2.69119054e-01 1.22694924e-01 -4.05273080e-01 1.15530765e+00
8.71248543e-02 4.07507330e-01 8.07423711e-01 -7.15192735e-01
3.44146967e-01 7.39962459e-01 1.91705592e-03 -9.71115172e-01
-5.40552318e-01 -4.80940253e-01 -2.70522147e-01 -1.72417343e-01
6.33688509e-01 -2.46395111e-01 -1.48530856e-01 1.48934186e+00
-4.13744628e-01 -7.82473564e-01 -2.58246571e-01 5.02299726e-01
9.00646269e-01 5.36457241e-01 4.62467186e-02 -1.75227851e-01
9.50593114e-01 -8.09642375e-01 -5.78716338e-01 -3.14041935e-02
1.13962543e+00 -9.83962834e-01 1.62128007e+00 5.76791465e-01
-1.34289002e+00 -6.99182093e-01 -9.67146695e-01 -4.06374425e-01
-1.38923138e-01 7.99643159e-01 7.02617884e-01 9.79134560e-01
-1.27446902e+00 6.05431676e-01 -4.14895892e-01 -2.58136153e-01
-2.97449827e-02 2.57242471e-01 -2.40742579e-01 2.25051880e-01
-1.10264373e+00 1.21262777e+00 8.87071118e-02 -2.16254085e-01
1.06521873e-02 -6.29707098e-01 -5.97117126e-01 -1.17908277e-01
4.76674363e-02 -7.26377815e-02 1.30697036e+00 -1.14139354e+00
-1.83474982e+00 6.79701447e-01 -7.96504915e-02 7.63343349e-02
5.35341680e-01 -2.86381841e-01 -2.43344992e-01 -4.54004928e-02
1.87587559e-01 2.98639119e-01 2.45869309e-01 -7.27713466e-01
-6.23629630e-01 -1.05089523e-01 2.82068029e-02 2.57307023e-01
-7.83566773e-01 1.94037423e-01 -1.70411348e-01 -9.14180994e-01
4.67342837e-03 -8.05527806e-01 5.30592687e-02 -1.19669318e-01
-1.17612496e-01 -5.24741709e-01 1.79713219e-02 -1.08645296e+00
1.83260155e+00 -1.65699482e+00 -1.75421253e-01 2.92910248e-01
1.25115961e-01 1.31602585e-01 -1.95129454e-01 5.10129750e-01
2.43634447e-01 3.68116319e-01 2.23373055e-01 -1.84570223e-01
1.07466653e-01 -5.76655626e-01 1.85633358e-02 4.75526899e-02
4.46845740e-01 1.22268188e+00 -9.36030924e-01 -3.08693111e-01
-1.40730068e-01 2.06289012e-02 -7.97076106e-01 2.81828642e-01
1.31799847e-01 9.79854763e-02 -3.29116255e-01 5.45347691e-01
1.70063019e-01 -2.61754245e-01 2.09018618e-01 4.47369784e-01
-5.66584289e-01 7.95117855e-01 -6.44229710e-01 1.41066921e+00
-6.64334714e-01 9.73830104e-01 -4.03998971e-01 -2.88433880e-01
1.32136655e+00 1.62553549e-01 3.36293250e-01 -9.28659856e-01
-9.61204544e-02 4.29033995e-01 1.90480053e-01 -4.49237168e-01
6.39031470e-01 -4.32663746e-02 -2.79292136e-01 8.26972604e-01
-1.42227262e-01 -3.72097194e-01 1.33026138e-01 2.44767591e-01
9.07428563e-01 2.65691400e-01 5.06572723e-01 -8.06286097e-01
8.35684478e-01 5.94752319e-02 1.13075748e-02 7.57399976e-01
-1.95547998e-01 2.18359441e-01 5.89792848e-01 -1.43227428e-01
-9.42214429e-01 -6.12226069e-01 -2.25361243e-01 1.25300622e+00
-6.02869689e-01 -7.72790432e-01 -6.73244238e-01 -4.90368068e-01
2.83514336e-02 9.83472347e-01 -4.98766422e-01 -3.16174924e-01
-3.28935385e-01 -4.65605527e-01 5.76538980e-01 4.55207109e-01
2.70783901e-01 -8.28318834e-01 -4.51996297e-01 2.62444973e-01
-2.62813359e-01 -6.65509343e-01 -5.27933359e-01 1.31289631e-01
-8.05541277e-01 -7.99890935e-01 -5.57678342e-01 -8.67046416e-01
2.01636255e-01 2.54714210e-02 1.13472557e+00 4.45660293e-01
2.30826437e-01 4.05730128e-01 -4.49457735e-01 -5.19339442e-01
-4.79923427e-01 4.91447240e-01 4.86629382e-02 -5.55204332e-01
7.46395051e-01 -7.30811730e-02 -1.49218932e-01 5.27334679e-03
-4.27068710e-01 8.66464600e-02 5.18547177e-01 8.82951498e-01
-1.07626893e-01 -4.95768607e-01 8.62563372e-01 -8.70693266e-01
1.21912181e+00 -2.83787847e-01 -3.13324422e-01 5.40113151e-01
-1.08691573e+00 1.07239254e-01 8.41296077e-01 -2.95146435e-01
-1.03239489e+00 -2.40058452e-01 -1.92236140e-01 5.64509034e-01
1.42399549e-01 9.73458230e-01 2.26349652e-01 -1.36701509e-01
6.93701804e-01 5.35766929e-02 -7.14971870e-02 -4.01681155e-01
1.78243145e-01 8.31667185e-01 1.91803992e-01 -9.10525858e-01
7.92792201e-01 -4.02129680e-01 -2.67181158e-01 -1.16061878e+00
-1.10351741e+00 -2.31359243e-01 -1.05745232e+00 -2.78601229e-01
3.67196470e-01 -1.00656807e+00 -9.29931462e-01 3.86337489e-01
-1.03249586e+00 -5.01648426e-01 3.03931697e-03 6.17305577e-01
-4.06427860e-01 4.73305285e-01 -7.98116803e-01 -7.48373687e-01
-2.55347162e-01 -8.90139222e-01 4.04651999e-01 1.54716492e-01
-8.63403141e-01 -1.06364524e+00 3.58644694e-01 5.34080088e-01
4.14512187e-01 -1.20141439e-01 1.14684284e+00 -7.41013467e-01
-2.86998302e-01 -2.88459212e-01 -2.23200247e-01 3.07807982e-01
8.09518993e-03 2.89092034e-01 -8.84725630e-01 -3.70390192e-02
3.49794030e-02 -7.91739881e-01 6.25259399e-01 2.69858956e-01
1.09032559e+00 -4.61949408e-01 4.63985443e-01 4.14994806e-01
1.14779663e+00 -4.53489758e-02 1.95613861e-01 5.00682235e-01
6.28220499e-01 7.24293649e-01 3.88930112e-01 4.36042607e-01
8.28204155e-01 5.42459726e-01 -6.00776911e-01 2.28277922e-01
-3.45801711e-02 -4.26789045e-01 8.36088777e-01 1.66896307e+00
-5.89395612e-02 -1.96642745e-02 -1.11174440e+00 3.18813801e-01
-1.31565571e+00 -6.77407265e-01 -4.49165642e-01 2.20245790e+00
1.15357685e+00 3.19161594e-01 2.34165907e-01 1.79387495e-01
7.48772174e-02 3.03545385e-03 -2.13699229e-02 -9.87031281e-01
-2.26006135e-01 3.32572192e-01 3.06887209e-01 7.20656037e-01
-6.36207044e-01 9.66135979e-01 6.91233349e+00 5.37960649e-01
-1.04131639e+00 -2.54751384e-01 7.25074053e-01 1.75311327e-01
-6.77193403e-01 -2.99155921e-01 -7.82859981e-01 3.46182495e-01
1.19161737e+00 -5.74278116e-01 3.21874887e-01 5.64562082e-01
4.09986466e-01 -8.48449096e-02 -8.62513185e-01 6.15240872e-01
3.68710667e-01 -9.23354745e-01 -3.52494568e-02 -2.58686692e-01
9.63834465e-01 -2.51253515e-01 2.19245106e-01 6.17953300e-01
3.12437534e-01 -1.18279278e+00 1.03543687e+00 5.48460782e-01
9.70872462e-01 -6.52503014e-01 5.26037872e-01 3.75979304e-01
-8.72503877e-01 -2.16323495e-01 -3.04492146e-01 -6.92190588e-01
-3.05216014e-01 7.78499991e-02 -9.84353900e-01 4.32588123e-02
2.86327481e-01 8.93415213e-01 -1.15119827e+00 7.22998977e-01
-5.62891424e-01 1.14069724e+00 6.71884343e-02 -6.29371047e-01
2.93994457e-01 -1.02061674e-01 1.85735784e-02 1.35918939e+00
4.63940918e-01 2.47831479e-01 1.63567215e-01 7.30450094e-01
-4.44235802e-02 9.30505276e-01 -2.63737798e-01 -3.42994064e-01
3.58605623e-01 9.49523032e-01 -5.58192074e-01 -2.98549104e-02
-7.16715276e-01 7.92672276e-01 7.01863110e-01 8.75271261e-02
-4.53721255e-01 -5.09100914e-01 2.02589482e-01 6.78700060e-02
-1.84770852e-01 -4.73482698e-01 -1.12639260e+00 -1.26508152e+00
3.84153835e-02 -9.77789938e-01 2.08252192e-01 -5.49187839e-01
-1.26072502e+00 4.50331807e-01 -4.93269444e-01 -8.93798292e-01
-1.58978313e-01 -9.03878868e-01 -5.90107381e-01 1.13417208e+00
-1.57175219e+00 -9.76185739e-01 -1.75354064e-01 2.88494527e-01
6.09151542e-01 -3.58597696e-01 8.59807730e-01 3.60148624e-02
-4.98903930e-01 1.23392630e+00 2.88562715e-01 4.74435948e-02
1.08573830e+00 -1.52551234e+00 1.66995287e-01 6.26473010e-01
9.55163836e-02 8.17346513e-01 3.58535260e-01 -6.90856755e-01
-8.78848970e-01 -4.69006568e-01 1.97371233e+00 -1.02806783e+00
9.70037162e-01 -3.30484360e-01 -7.99758554e-01 3.91073197e-01
-2.18725987e-02 -9.39762175e-01 1.08275867e+00 3.98382306e-01
-2.27689520e-01 7.44276792e-02 -5.46001196e-01 6.96790934e-01
6.00879908e-01 -9.80919838e-01 -5.52257776e-01 9.89128277e-02
4.27507699e-01 -6.85528815e-02 -9.99012411e-01 4.98292781e-02
9.45736706e-01 -9.71699893e-01 5.54188371e-01 -4.98084813e-01
8.86303902e-01 2.48912990e-01 1.95124760e-01 -1.14016390e+00
-6.76083565e-01 -5.84011018e-01 1.97369352e-01 1.02597857e+00
8.83050323e-01 -1.71502531e-01 7.81428397e-01 1.08061326e+00
-1.82927504e-01 -8.32706451e-01 -3.61698866e-01 -5.91748238e-01
6.41258240e-01 -5.35038054e-01 3.63188326e-01 1.10673296e+00
7.03022063e-01 4.58873749e-01 -3.70952219e-01 -6.37783110e-01
3.77058350e-02 -1.90146908e-01 6.84959352e-01 -1.58492863e+00
-1.55182213e-01 -8.31981421e-01 -2.95367354e-04 -9.40825284e-01
3.29561830e-01 -1.29656589e+00 -1.02303617e-01 -1.18671596e+00
3.89177114e-01 -4.29002196e-01 -4.44032192e-01 4.17858392e-01
-5.17672718e-01 1.69936076e-01 4.82574075e-01 1.88091293e-01
-4.34096843e-01 4.35953259e-01 1.09382153e+00 1.40126333e-01
-4.97063965e-01 2.57045358e-01 -9.88545597e-01 4.82374221e-01
9.02906835e-01 -3.52812111e-01 -3.12678188e-01 -4.81925249e-01
6.82886362e-01 9.36868340e-02 -2.45766133e-01 -6.35409534e-01
1.53571546e-01 -3.23011011e-01 6.22934103e-01 -2.91293234e-01
-1.94046736e-01 -5.29679179e-01 -4.11400586e-01 3.43575567e-01
-9.26692784e-01 5.08416355e-01 2.97849089e-01 -2.16679513e-01
-1.27035424e-01 -5.31996608e-01 3.67110223e-01 -6.96732849e-02
-4.00274009e-01 -8.98246691e-02 -6.16501868e-01 1.58686936e-01
5.83262742e-01 -3.44238073e-01 7.54446257e-03 -6.04088902e-01
-2.27655768e-01 1.51403487e-01 5.76011300e-01 5.70943177e-01
3.64294201e-01 -1.20521712e+00 -1.03742230e+00 1.24291174e-01
1.11350372e-01 -7.92413235e-01 -3.65977347e-01 8.41191828e-01
-7.20015109e-01 9.60458815e-01 -3.14448953e-01 -3.90991010e-02
-1.08485889e+00 -1.11135669e-01 4.91418280e-02 -4.45984185e-01
-1.16357602e-01 8.75392854e-01 -3.16822618e-01 -6.40398800e-01
1.24673456e-01 -2.84915566e-01 -5.84738493e-01 2.81233102e-01
4.93821204e-01 7.00803399e-01 -1.27238594e-03 -5.85218549e-01
-6.38593957e-02 2.52687186e-01 -4.99951020e-02 -2.90365964e-01
1.31065738e+00 -1.10388376e-01 -2.32436806e-01 9.87094700e-01
8.34370971e-01 6.48203552e-01 -9.45826888e-01 4.34515737e-02
8.48205030e-01 -4.96523082e-01 -3.73750776e-02 -1.19408810e+00
-3.55116278e-01 7.83611894e-01 -3.65908332e-02 -2.32085466e-01
7.09526241e-01 -4.81230795e-01 4.14757580e-01 7.98362911e-01
9.46020149e-03 -1.52133906e+00 2.54504800e-01 1.00514734e+00
8.62988830e-01 -1.37448847e+00 3.25419188e-01 5.57199530e-02
-8.71311724e-01 1.59997022e+00 6.72252715e-01 1.08627342e-01
5.29206932e-01 2.19026022e-02 3.24956983e-01 2.16243520e-01
-8.35841894e-01 3.11426640e-01 8.49909067e-01 1.87879056e-01
1.39104986e+00 1.33339405e-01 -1.04624271e+00 1.18181682e+00
-8.31846893e-01 -1.54024418e-02 8.16023648e-01 4.82814163e-01
-4.61952001e-01 -1.17852068e+00 6.62342086e-02 6.87885165e-01
-5.25171041e-01 -3.54664415e-01 -6.93532169e-01 7.43236005e-01
-2.67451137e-01 1.05187511e+00 -2.24425867e-01 -4.13783580e-01
1.85282961e-01 3.82216960e-01 2.35271141e-01 -7.30964065e-01
-1.10880756e+00 -1.30904734e-01 1.07181482e-01 -2.79894263e-01
-2.95607686e-01 -9.24498558e-01 -8.77867579e-01 -4.10687208e-01
-5.09886682e-01 2.38606825e-01 6.56646490e-01 9.63792741e-01
-1.41707852e-01 1.09051563e-01 5.76030731e-01 -1.78702563e-01
-7.83787370e-01 -1.04612696e+00 -6.06255352e-01 1.67463318e-01
2.08914250e-01 3.36457975e-02 -9.81087983e-02 -1.07507959e-01] | [11.251654624938965, 9.443991661071777] |
689c5ccc-5a92-4460-ad7a-244cf81d9ae6 | guiding-interaction-behaviors-for-multi-modal | null | null | https://aclanthology.org/W17-2803 | https://aclanthology.org/W17-2803.pdf | Guiding Interaction Behaviors for Multi-modal Grounded Language Learning | Multi-modal grounded language learning connects language predicates to physical properties of objects in the world. Sensing with multiple modalities, such as audio, haptics, and visual colors and shapes while performing interaction behaviors like lifting, dropping, and looking on objects enables a robot to ground non-visual predicates like {``}empty{''} as well as visual predicates like {``}red{''}. Previous work has established that grounding in multi-modal space improves performance on object retrieval from human descriptions. In this work, we gather behavior annotations from humans and demonstrate that these improve language grounding performance by allowing a system to focus on relevant behaviors for words like {``}white{''} or {``}half-full{''} that can be understood by looking or lifting, respectively. We also explore adding modality annotations (whether to focus on audio or haptics when performing a behavior), which improves performance, and sharing information between linguistically related predicates (if {``}green{''} is a color, {``}white{''} is a color), which improves grounding recall but at the cost of precision. | ['Raymond Mooney', 'Jivko Sinapov', 'Jesse Thomason'] | 2017-08-01 | null | null | null | ws-2017-8 | ['grounded-language-learning'] | ['natural-language-processing'] | [ 3.75716358e-01 2.66912133e-01 1.10827163e-01 -3.01305085e-01
-6.54563487e-01 -7.72644937e-01 3.58247429e-01 7.79862583e-01
-4.57165629e-01 5.76910853e-01 1.58392996e-01 -2.57196337e-01
-3.20595026e-01 -7.67812729e-01 -8.54732990e-01 -3.76879156e-01
-3.47478062e-01 3.67048889e-01 5.56578159e-01 -4.21935797e-01
3.09420116e-02 4.68401074e-01 -2.23418498e+00 7.69119263e-01
1.19972385e-01 1.05110121e+00 4.40151840e-01 5.32446146e-01
-1.04338475e-01 9.64371443e-01 -7.36653805e-01 -2.34514445e-01
-6.37018904e-02 -1.70765698e-01 -9.82451499e-01 -1.99079260e-01
5.43847740e-01 -6.04855955e-01 9.98729616e-02 1.01227140e+00
1.38586804e-01 5.11504829e-01 5.23983419e-01 -1.72096086e+00
-1.03904545e+00 4.29717898e-01 -3.68528783e-01 -1.54683948e-01
1.13511479e+00 2.03467250e-01 1.28375185e+00 -3.86203408e-01
6.17814183e-01 1.51686430e+00 4.55704808e-01 5.99330425e-01
-1.09813023e+00 -7.57346094e-01 4.40500945e-01 5.76834567e-02
-1.37442005e+00 -2.27157563e-01 6.78379536e-01 -5.12049496e-01
1.06052279e+00 4.72185910e-01 5.63593507e-01 9.53842282e-01
-2.32159019e-01 9.42205787e-01 7.84635782e-01 -4.95663434e-01
3.14194709e-01 1.32214606e-01 -1.27291784e-01 8.62310469e-01
2.69763470e-01 6.58612996e-02 -7.85447598e-01 -1.92906857e-01
6.14972651e-01 -5.54061979e-02 -1.92716606e-02 -4.19635445e-01
-1.49071240e+00 5.78023911e-01 4.86346275e-01 2.13833600e-01
-5.40951431e-01 4.82949346e-01 2.65982419e-01 2.51357630e-02
9.00274664e-02 7.52581716e-01 -3.00241232e-01 -2.76919395e-01
-3.25450093e-01 4.30851847e-01 5.53743601e-01 1.41920209e+00
1.07427704e+00 -3.44233215e-01 -2.83101916e-01 6.88464701e-01
3.16888094e-01 8.78267288e-01 -1.03421092e-01 -1.46353745e+00
3.18612814e-01 7.81270742e-01 7.90622056e-01 -8.56790066e-01
-9.26083744e-01 4.20995176e-01 -1.37763590e-01 1.93450972e-01
4.64553505e-01 -9.27198753e-02 -7.83398449e-01 1.96812582e+00
2.96340615e-01 -1.42156973e-01 1.94919080e-01 1.13635302e+00
1.26894009e+00 5.08000314e-01 6.40621603e-01 -2.25354731e-02
1.97492588e+00 -4.14106637e-01 -5.84333360e-01 -2.54412174e-01
8.72204959e-01 -7.89031684e-01 1.64896071e+00 1.90333307e-01
-6.78525090e-01 -5.42796433e-01 -7.09923744e-01 -3.56153816e-01
-7.26655304e-01 4.46869507e-02 8.58153760e-01 3.27082127e-01
-1.00560319e+00 4.46533680e-01 -8.30407917e-01 -7.28453219e-01
1.34591684e-01 2.84071326e-01 -5.63220024e-01 -1.22820273e-01
-1.23664236e+00 8.26571643e-01 4.19079334e-01 -4.27218974e-01
-6.83791041e-01 -4.37690288e-01 -1.26891005e+00 -1.91661432e-01
4.31331456e-01 -5.34216881e-01 1.32401752e+00 -7.17681110e-01
-1.16509867e+00 1.14444196e+00 -2.20785037e-01 -1.13680504e-01
1.10234097e-01 -4.83258396e-01 -6.74099624e-01 4.16047782e-01
4.59328741e-01 1.16559458e+00 4.08682138e-01 -1.60619283e+00
-8.35829377e-01 -4.92414683e-01 9.73489761e-01 3.51471573e-01
-1.96534112e-01 1.14551097e-01 -3.29843104e-01 -2.61057019e-01
4.50327516e-01 -1.14864850e+00 2.53288209e-01 4.51775938e-01
-6.73950255e-01 -4.65202540e-01 8.42170238e-01 -4.97886091e-01
9.16450799e-01 -2.22561002e+00 -7.38077387e-02 1.61932379e-01
-2.19364576e-02 -1.19357191e-01 1.31070957e-01 6.95478380e-01
1.64624602e-01 1.06115572e-01 1.94185197e-01 -1.97515875e-01
3.54013145e-01 5.89400530e-01 2.75764558e-02 2.29731053e-01
8.81208554e-02 6.06335759e-01 -1.29470360e+00 -5.43645322e-01
3.66730362e-01 5.90705931e-01 -5.63307762e-01 1.08631052e-01
-5.58280408e-01 4.26346660e-01 -4.82458532e-01 6.96213841e-01
2.38660112e-01 -2.62694001e-01 8.77466500e-02 -3.58736068e-01
-1.76801547e-01 2.76032537e-01 -1.31325626e+00 1.86565864e+00
-3.55374247e-01 4.07178611e-01 1.14306413e-01 -5.08277655e-01
7.71183252e-01 3.55850905e-01 4.44233030e-01 -8.12921703e-01
-1.15594521e-01 -4.82074432e-02 -1.83226675e-01 -8.76045108e-01
6.09189928e-01 -2.15341657e-01 -4.81759816e-01 3.72797787e-01
-3.74248207e-01 -5.65334201e-01 2.59518236e-01 1.15592629e-01
1.06914568e+00 7.18660295e-01 3.88010293e-02 -1.86438128e-01
2.56706417e-01 5.12441210e-02 1.00955583e-01 6.70073807e-01
-1.19424418e-01 2.63557881e-01 2.36546814e-01 -2.18114629e-01
-7.92514741e-01 -1.27743828e+00 1.77940652e-01 2.04422402e+00
7.61487842e-01 -4.59244281e-01 -2.75307029e-01 2.51768455e-02
1.07110739e-01 9.79706109e-01 -5.99808037e-01 -2.32002616e-01
-4.16116476e-01 2.51570418e-02 4.98171329e-01 9.18077409e-01
4.90426391e-01 -1.37440860e+00 -1.38901317e+00 -1.45685241e-01
-5.58371961e-01 -1.19509399e+00 -2.20385790e-02 3.82719189e-01
-3.61354738e-01 -7.98557818e-01 -5.59338987e-01 -6.89260304e-01
5.68686783e-01 8.45374987e-02 1.09159923e+00 1.21568419e-01
-1.79316357e-01 9.13212597e-01 -5.99640667e-01 -6.82865202e-01
-2.09533736e-01 -4.83186960e-01 8.18132758e-02 -1.96364388e-01
4.00975049e-01 -2.02871695e-01 -5.85570991e-01 2.08606049e-01
-8.43148172e-01 2.14616537e-01 3.65035772e-01 1.64494812e-01
7.55387902e-01 -1.67520940e-01 -8.40473250e-02 -7.45521247e-01
4.46643203e-01 -2.87006617e-01 -2.49745682e-01 2.44704336e-01
1.87183395e-01 1.88007087e-01 2.05885291e-01 -6.24789357e-01
-7.87084520e-01 3.16418186e-02 2.07971931e-02 -6.20231271e-01
-7.29582131e-01 3.70436311e-01 -1.65988326e-01 4.45137262e-01
8.44483972e-01 -1.42770648e-01 -4.88132864e-01 -3.00918788e-01
6.56820178e-01 3.73317569e-01 7.66250551e-01 -1.02177203e+00
5.36705256e-01 7.91621089e-01 1.74716592e-01 -6.34800732e-01
-6.46849096e-01 -7.18464971e-01 -4.02791739e-01 -1.03482723e-01
1.23212910e+00 -8.80037069e-01 -1.48083770e+00 -8.05619955e-02
-1.00841069e+00 -3.34343225e-01 -3.49076688e-01 6.68619692e-01
-8.47260416e-01 7.79462606e-02 -5.29838383e-01 -1.09532857e+00
8.49978104e-02 -8.79930913e-01 1.68866563e+00 1.95892647e-01
-8.19698274e-01 -6.33628666e-01 -5.58900297e-01 6.78753927e-02
4.74431887e-02 4.06077325e-01 1.09653306e+00 -6.35897160e-01
-2.75090963e-01 -4.52446640e-02 -3.64115149e-01 -1.84435323e-01
3.82533252e-01 -1.40212670e-01 -7.85295784e-01 -2.26382956e-01
-7.12720811e-01 -4.55632389e-01 2.37922639e-01 -1.20575152e-01
8.83273423e-01 -1.98387012e-01 -5.84399104e-01 -8.37655552e-03
1.09547520e+00 6.61587179e-01 3.74665111e-01 2.42406994e-01
6.97348773e-01 8.04718256e-01 7.73555696e-01 5.88694155e-01
6.78046107e-01 8.28933358e-01 6.04785204e-01 -6.91803992e-02
-8.11797082e-02 -3.80760759e-01 2.06387505e-01 -2.70284265e-01
-2.03950793e-01 -2.13963985e-01 -9.28738177e-01 5.41678011e-01
-1.60801315e+00 -9.44189847e-01 1.86302245e-03 2.06908894e+00
7.62598634e-01 5.53809963e-02 1.68776795e-01 1.47095978e-01
7.03206837e-01 -1.38983160e-01 -5.51098228e-01 -4.22334135e-01
3.51991057e-02 4.68676649e-02 1.83058500e-01 3.23837608e-01
-1.13445902e+00 9.73832130e-01 4.70260286e+00 3.28055561e-01
-1.07150042e+00 -1.33505911e-01 -9.44461003e-02 -1.03834800e-01
-5.81007361e-01 6.29757866e-02 -5.87884009e-01 2.19931781e-01
3.84033769e-01 3.33525270e-01 4.01149720e-01 8.29865932e-01
4.63766418e-02 -5.11286616e-01 -1.66813576e+00 1.09790719e+00
8.99434835e-02 -8.02538216e-01 1.10887311e-01 -3.59450728e-01
2.44831756e-01 -4.73238528e-01 6.58479705e-02 4.12635475e-01
6.58347666e-01 -8.18606973e-01 1.34240997e+00 4.27094370e-01
8.93354237e-01 -2.28801593e-01 1.80865496e-01 2.70628810e-01
-1.46888041e+00 1.06848413e-02 1.24133587e-01 -2.02142730e-01
8.52707103e-02 3.35028172e-02 -6.84057117e-01 5.61854720e-01
1.12208009e+00 2.12417319e-01 -3.01499099e-01 6.85017288e-01
-4.64911342e-01 1.34309784e-01 -6.59925461e-01 -3.91876340e-01
2.78649703e-02 2.38555625e-01 4.97108221e-01 1.12462354e+00
2.46332243e-01 3.38940144e-01 5.98774076e-01 8.04115713e-01
2.44911358e-01 1.06563650e-01 -6.26137197e-01 -7.80060068e-02
6.08284831e-01 7.51482844e-01 -1.07758093e+00 -2.73807824e-01
-4.71166342e-01 9.77845371e-01 -2.22105696e-03 6.03206158e-01
-8.94965053e-01 -3.55190217e-01 5.97423017e-01 3.62294704e-01
1.48518234e-01 -1.34807676e-01 -1.82182491e-02 -5.20945668e-01
5.72783202e-02 -3.12447876e-01 5.61379492e-01 -1.54157519e+00
-9.21121657e-01 2.29597300e-01 4.64891911e-01 -1.23872459e+00
-1.40143475e-02 -6.34709001e-01 1.24842770e-01 6.62379503e-01
-1.03501546e+00 -1.55346918e+00 -4.36232388e-01 8.79362166e-01
6.99955821e-02 5.11028290e-01 1.19293213e+00 1.95259973e-01
3.01044166e-01 1.99666321e-01 -6.93637669e-01 1.44302025e-01
5.88054419e-01 -1.00293601e+00 -2.48872980e-01 3.87876600e-01
1.52087957e-01 9.06480014e-01 1.07642734e+00 -6.36699557e-01
-1.24364090e+00 -7.52221167e-01 7.69494593e-01 -4.35690731e-01
4.99075681e-01 -2.65965372e-01 -7.95688272e-01 9.04581726e-01
-1.39220774e-01 -6.80009797e-02 7.29974747e-01 4.13654536e-01
-7.28496909e-01 1.66864157e-01 -1.26085567e+00 7.63091564e-01
1.21515799e+00 -9.47627485e-01 -7.38928616e-01 5.21559834e-01
8.23661625e-01 -5.84164083e-01 -7.12181628e-01 6.11550510e-01
6.67389095e-01 -8.32807481e-01 1.00915301e+00 -7.32281506e-01
2.68747985e-01 -5.41700840e-01 -6.16132736e-01 -9.65388298e-01
-2.82849729e-01 -3.33435446e-01 1.80684105e-01 1.03111291e+00
3.20217252e-01 -4.27554041e-01 4.20800865e-01 7.17886090e-01
-4.23294932e-01 -2.88160443e-01 -8.78555238e-01 -8.26742113e-01
-2.18308598e-01 -5.59356451e-01 6.95145607e-01 8.76298606e-01
3.64226609e-01 -2.67128646e-02 -3.09312679e-02 4.52031255e-01
8.50986242e-02 3.12002689e-01 5.16442001e-01 -1.20136261e+00
-3.20477039e-01 -1.98574260e-01 -5.22063136e-01 -9.21991527e-01
1.44348398e-01 -5.83945155e-01 2.66889036e-01 -1.95560908e+00
-5.55851962e-03 -6.64763331e-01 -2.28821307e-01 1.12696767e+00
6.47463351e-02 4.33066696e-01 4.15716439e-01 1.61300395e-02
-9.09764051e-01 -8.88983011e-02 1.28405082e+00 -2.02326760e-01
-1.17326029e-01 -3.02394122e-01 -6.63499355e-01 8.82871628e-01
5.18291533e-01 -7.13015813e-03 -5.50974011e-01 -6.55205309e-01
7.24637508e-01 2.78453439e-01 7.51637220e-01 -9.46265101e-01
1.76727638e-01 -2.07797915e-01 2.25119129e-01 -3.38005096e-01
1.01052904e+00 -9.21004355e-01 2.48866394e-01 3.28389615e-01
-4.71340328e-01 1.58657372e-01 6.35207772e-01 2.45214418e-01
1.51458725e-01 2.31071264e-02 4.25067484e-01 -3.95022660e-01
-1.25478208e+00 -3.41598570e-01 -2.58464932e-01 4.10995185e-02
1.18028080e+00 -2.52186090e-01 -4.15901989e-01 -5.75564861e-01
-1.13567078e+00 2.19573364e-01 3.94476265e-01 7.90340006e-01
3.64892006e-01 -1.28157496e+00 -2.09177375e-01 -4.84931655e-02
6.96238697e-01 4.94936705e-02 1.32403120e-01 4.86011475e-01
-3.42460901e-01 3.12073708e-01 -2.92158902e-01 -6.89601183e-01
-1.29994023e+00 7.15648413e-01 1.26479659e-02 5.03565371e-01
-5.07977903e-01 1.40170074e+00 1.97558999e-01 -1.83701769e-01
4.68531311e-01 -7.62919426e-01 -7.23563954e-02 8.29931796e-02
4.68030870e-02 2.84399778e-01 -7.73471892e-02 -1.05150616e+00
-6.89948082e-01 7.42226422e-01 4.42364722e-01 -1.87942624e-01
6.22010171e-01 -1.67816535e-01 -3.07370052e-02 9.66803551e-01
1.06394684e+00 -5.53595461e-02 -1.07741606e+00 -9.81586501e-02
-2.30882496e-01 -1.20998010e-01 -5.16897619e-01 -1.00793254e+00
-3.32128882e-01 7.61598110e-01 7.27982879e-01 2.23380476e-01
9.08910394e-01 6.87595844e-01 5.35794556e-01 5.04207850e-01
1.06550729e+00 -1.09052491e+00 4.17082638e-01 2.32779309e-01
8.41260612e-01 -1.20964241e+00 7.03557879e-02 -6.32864654e-01
-7.38900065e-01 6.77438736e-01 5.26845038e-01 2.73437232e-01
2.69616097e-01 1.80559218e-01 2.18209531e-02 -4.91550863e-01
-2.86217213e-01 -5.57721317e-01 3.20255578e-01 7.36329377e-01
3.59492391e-01 5.23859680e-01 3.05605046e-02 3.10648978e-01
-4.04534876e-01 -3.06654453e-01 -4.86404821e-02 1.25767374e+00
-5.38757324e-01 -5.87031007e-01 -5.04586399e-01 3.40467721e-01
-2.85486937e-01 4.23853286e-02 -4.27368045e-01 8.53012621e-01
7.63707280e-01 1.31335318e+00 3.28320056e-01 -2.53232658e-01
5.44307768e-01 1.11458236e-02 6.72906935e-01 -9.51229811e-01
-4.07277793e-01 -6.53349236e-02 2.84749508e-01 -6.29728198e-01
-7.40665674e-01 -7.19883919e-01 -2.23512554e+00 5.83231123e-03
-1.09950326e-01 -4.75905053e-02 3.33046794e-01 1.10465348e+00
9.27356109e-02 5.01757562e-01 -3.69212151e-01 -6.74540818e-01
-6.96468204e-02 -6.70189440e-01 -4.83239502e-01 1.18126822e+00
1.52939096e-01 -1.14292824e+00 -1.85455307e-01 3.56580675e-01] | [10.43257999420166, 1.5936479568481445] |
29bf9328-3f2e-49bf-a72a-fdd159d4807b | representation-biases-in-sentence | 2301.13039 | null | https://arxiv.org/abs/2301.13039v1 | https://arxiv.org/pdf/2301.13039v1.pdf | Representation biases in sentence transformers | Variants of the BERT architecture specialised for producing full-sentence representations often achieve better performance on downstream tasks than sentence embeddings extracted from vanilla BERT. However, there is still little understanding of what properties of inputs determine the properties of such representations. In this study, we construct several sets of sentences with pre-defined lexical and syntactic structures and show that SOTA sentence transformers have a strong nominal-participant-set bias: cosine similarities between pairs of sentences are more strongly determined by the overlap in the set of their noun participants than by having the same predicates, lengthy nominal modifiers, or adjuncts. At the same time, the precise syntactic-thematic functions of the participants are largely irrelevant. | ['Sebastian Padó', 'Dmitry Nikolaev'] | 2023-01-30 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [ 8.37060809e-02 2.51508236e-01 -2.89023906e-01 -7.81594217e-01
-3.92292649e-01 -8.22343767e-01 8.23670924e-01 7.48844922e-01
-7.85146058e-01 5.78131676e-01 1.02751172e+00 -5.70861816e-01
-1.93581551e-01 -8.96890759e-01 -5.84053159e-01 -3.37333620e-01
-1.34745613e-01 4.69482869e-01 9.49145630e-02 -7.12217748e-01
-4.36938927e-02 1.56545982e-01 -1.08786058e+00 6.86313272e-01
3.10730845e-01 5.48571348e-01 2.30279550e-01 3.75650704e-01
-2.23615736e-01 6.34162724e-01 -6.43347204e-01 -5.22775114e-01
1.35366783e-01 -3.15491766e-01 -7.38938332e-01 -5.54259479e-01
3.54144603e-01 7.23933056e-02 -3.32850248e-01 7.47587621e-01
3.40210497e-01 2.69571573e-01 8.06842148e-01 -6.11086249e-01
-1.05570924e+00 1.27347708e+00 1.66560840e-02 7.43975818e-01
4.98319149e-01 5.42926118e-02 1.96131539e+00 -7.68534541e-01
7.05745101e-01 1.35393846e+00 3.97172362e-01 6.06658638e-01
-1.53649676e+00 -3.71249080e-01 3.45409691e-01 5.12897149e-02
-1.15566325e+00 -5.98661602e-01 5.61888516e-01 -2.27212161e-01
1.17071068e+00 2.74515629e-01 3.56774479e-01 1.29725242e+00
2.78375864e-01 4.99965847e-01 7.53723443e-01 -4.42410618e-01
4.20840718e-02 2.47256473e-01 5.19220293e-01 3.41175765e-01
2.38614693e-01 9.48032364e-02 -5.61120212e-01 -1.47937844e-03
3.24800551e-01 -1.78135633e-01 -4.25622106e-01 -1.99207231e-01
-1.23732996e+00 1.10839498e+00 7.51179755e-01 7.63211548e-01
-1.75137058e-01 1.80818871e-01 6.10795736e-01 4.85555649e-01
7.21936226e-02 7.27849245e-01 -6.86771989e-01 1.65987909e-01
-4.03766274e-01 2.56775349e-01 6.02749884e-01 7.39300430e-01
5.82641304e-01 -8.49109441e-02 -3.24580848e-01 5.02204537e-01
8.69703963e-02 2.56523579e-01 6.55525267e-01 -4.17636514e-01
6.82648599e-01 3.91925454e-01 -6.33785278e-02 -8.91307414e-01
-3.90570343e-01 -3.04722875e-01 -3.28330249e-01 -4.57167387e-01
3.52282733e-01 -1.17268376e-01 -4.14320886e-01 2.23322105e+00
-1.31765231e-01 -4.22971040e-01 2.79902518e-01 6.41008973e-01
9.11833823e-01 6.09150529e-01 6.12217665e-01 -1.71991125e-01
1.44782424e+00 -2.34044343e-01 -4.57200229e-01 -5.90291262e-01
8.20690751e-01 -3.07530224e-01 1.37226677e+00 -4.87635523e-01
-1.03072941e+00 -4.31661904e-01 -8.47793102e-01 -4.27793890e-01
-3.76580149e-01 -3.16847503e-01 6.61294818e-01 5.42327940e-01
-7.55856752e-01 6.61677003e-01 -4.79094446e-01 -7.92816356e-02
1.70640871e-01 2.72303164e-01 -6.45394266e-01 2.05341116e-01
-1.43990934e+00 1.32653701e+00 3.03377151e-01 -1.25646651e-01
-5.31400144e-01 -8.31652939e-01 -1.32244492e+00 4.03524131e-01
-2.54850119e-01 -6.49927020e-01 1.21922266e+00 -1.00688827e+00
-8.50833416e-01 9.56571758e-01 -4.29135740e-01 -6.62109077e-01
-2.11677104e-01 -1.91041708e-01 -3.30888808e-01 -1.65800497e-01
2.45121762e-01 5.14923751e-01 5.15658498e-01 -7.21543014e-01
-3.71185571e-01 -4.68989283e-01 5.35129845e-01 5.14706314e-01
-3.08536947e-01 3.09912026e-01 3.85526627e-01 -4.47025597e-01
-1.17030866e-01 -3.85588020e-01 -8.19828883e-02 -2.35880554e-01
-2.51069963e-01 -6.18385911e-01 4.75162677e-02 -4.14294362e-01
1.16962564e+00 -2.32430148e+00 2.38067359e-01 -7.93960597e-03
5.13514541e-02 1.33213103e-01 -3.21492612e-01 6.53999329e-01
-2.75716633e-01 3.17600399e-01 -4.23834771e-02 -4.44689533e-03
1.94150746e-01 2.99861670e-01 -6.03877485e-01 4.71999258e-01
4.89443094e-01 9.96380448e-01 -9.30337250e-01 -1.09138906e-01
1.79011002e-01 2.32679442e-01 -5.35526276e-01 -6.65999278e-02
-1.36527553e-01 -1.05248488e-01 -1.15296438e-01 -7.40451440e-02
8.10700208e-02 2.23060220e-01 4.54044312e-01 -7.54927248e-02
3.27180661e-02 1.46582854e+00 -7.01942921e-01 1.00172818e+00
-5.25110424e-01 7.61396170e-01 1.38657212e-01 -7.98359394e-01
7.18566656e-01 2.58013576e-01 -2.61609435e-01 -7.21604466e-01
2.18046367e-01 6.81816787e-02 7.62898982e-01 -2.38729492e-01
4.65510219e-01 -7.22935796e-01 -4.17832434e-01 3.42651844e-01
4.30186056e-02 -1.77586406e-01 3.11671406e-01 4.08099979e-01
1.12979317e+00 -2.72827923e-01 4.21639293e-01 -5.02636850e-01
3.16763669e-01 -1.11000523e-01 7.38515496e-01 5.63162088e-01
-1.66813895e-01 4.57340002e-01 7.18169272e-01 -2.00243801e-01
-1.08266032e+00 -1.36697328e+00 -4.64750141e-01 1.21569157e+00
-1.86629593e-02 -7.54650652e-01 -3.13696146e-01 -6.59761190e-01
2.15752363e-01 1.48685884e+00 -8.03094208e-01 -3.45988810e-01
-6.87390625e-01 -2.11112902e-01 6.57280982e-01 7.95421362e-01
-4.74526435e-01 -1.24194622e+00 -4.26137537e-01 1.84426412e-01
2.13184319e-02 -1.02360892e+00 -5.70735931e-01 5.12698650e-01
-6.06536925e-01 -9.26254272e-01 -1.65173691e-03 -8.99936736e-01
4.97912943e-01 -5.55295832e-02 1.23143411e+00 3.40119153e-02
1.71671569e-01 -1.85942709e-01 -2.58534312e-01 -4.56806630e-01
-3.23621720e-01 -2.79068332e-02 3.15144137e-02 -2.57514685e-01
6.14405751e-01 -3.64663035e-01 -2.10356578e-01 -1.98844299e-01
-7.48295605e-01 -3.01012039e-01 2.11700931e-01 7.79248178e-01
1.43397480e-01 -2.20529348e-01 3.94010782e-01 -1.15669692e+00
9.26948249e-01 -5.72194159e-01 -2.27434430e-02 1.04323260e-01
-9.10196081e-02 4.73062724e-01 8.46803367e-01 -3.22456628e-01
-8.55012715e-01 -3.63052219e-01 -1.44755363e-01 1.64938048e-01
-1.17053010e-01 4.64185625e-01 -2.49552548e-01 8.27801168e-01
8.19001615e-01 -2.88377795e-02 -1.04129419e-01 -2.80655861e-01
5.17846584e-01 3.54657918e-01 1.43072546e-01 -6.04175985e-01
7.44847357e-01 2.88955510e-01 -1.78704664e-01 -8.55100036e-01
-9.07288134e-01 -2.97934234e-01 -4.81887281e-01 5.32549202e-01
8.02955091e-01 -8.98811281e-01 -3.85711551e-01 -2.42281049e-01
-1.18624115e+00 -2.25390464e-01 -5.79188406e-01 5.26657581e-01
-1.76469952e-01 1.33075297e-01 -7.48154759e-01 -3.42620075e-01
-3.50266695e-01 -8.20628524e-01 6.35755301e-01 -1.70971528e-01
-8.43040287e-01 -1.00898004e+00 -2.30226725e-01 -8.63820240e-02
4.24814105e-01 -2.35246152e-01 1.59907937e+00 -9.56496954e-01
-8.04008096e-02 -4.52163592e-02 -4.38626064e-03 4.80400890e-01
1.97764993e-01 -1.83708891e-02 -6.66581333e-01 9.77478847e-02
4.34729084e-02 2.93855872e-02 8.61953616e-01 3.54315400e-01
4.13513720e-01 -4.12027299e-01 -2.55514771e-01 4.36650634e-01
1.31466734e+00 -1.90600291e-01 3.41431320e-01 1.20988250e-01
3.13255847e-01 8.40665460e-01 1.77801579e-01 -1.11602232e-01
3.84300113e-01 3.55445802e-01 1.21165365e-01 4.26361412e-01
-6.14766888e-02 -4.96681660e-01 6.12760425e-01 6.23679817e-01
5.94065011e-01 -1.76152706e-01 -7.65220523e-01 9.28888440e-01
-1.18649578e+00 -9.52709854e-01 -3.86401445e-01 1.95016718e+00
1.05344462e+00 4.89940464e-01 -2.42249757e-01 2.00067431e-01
5.43740809e-01 5.10222077e-01 1.26139179e-01 -8.62243533e-01
-3.74480277e-01 5.76850712e-01 2.94757098e-01 6.61689460e-01
-8.19668055e-01 1.06147647e+00 6.37485218e+00 3.48793954e-01
-6.98260605e-01 4.18730676e-02 1.89107522e-01 -2.74725556e-01
-8.61768484e-01 2.08408460e-01 -8.48661900e-01 3.11308175e-01
1.04928505e+00 -2.98765779e-01 1.43878132e-01 5.05076110e-01
2.01874346e-01 5.10845035e-02 -1.70014071e+00 1.87207758e-01
-8.90422463e-02 -1.30289721e+00 1.10777477e-02 -2.92188466e-01
1.12311698e-01 -3.98785211e-02 -6.79388344e-02 4.33982611e-01
2.99402595e-01 -1.09416926e+00 9.65391219e-01 3.14284004e-02
4.66570377e-01 -6.05237782e-01 6.00251436e-01 3.87606293e-01
-9.08311546e-01 -2.63239771e-01 -6.02906823e-01 -6.35412574e-01
3.04165125e-01 3.68903995e-01 -6.80991232e-01 4.70763352e-03
2.36805469e-01 4.42171186e-01 -4.53614771e-01 4.39422607e-01
-8.84766519e-01 8.15741241e-01 -2.18019769e-01 -4.94781494e-01
2.42912024e-01 2.43070535e-03 5.56702077e-01 1.06640565e+00
-3.13780785e-01 1.74329251e-01 -2.96686590e-01 8.90196145e-01
-1.06867604e-01 2.29251072e-01 -8.52547228e-01 -9.37954634e-02
6.23524904e-01 9.69497383e-01 -3.20328265e-01 -1.58309907e-01
-5.61126113e-01 4.53382879e-01 7.30199277e-01 2.15567276e-01
-4.91117269e-01 -2.50477105e-01 1.11994684e+00 3.79751742e-01
3.71967822e-01 -3.58473659e-01 -3.87832403e-01 -9.08287227e-01
1.46989143e-02 -5.19707561e-01 3.12514544e-01 -7.00987756e-01
-1.46470201e+00 4.24699098e-01 3.00199352e-02 -3.11009228e-01
-1.86537698e-01 -7.62138665e-01 -9.12289858e-01 1.20224762e+00
-1.16343856e+00 -8.04550052e-01 5.82748532e-01 3.78397405e-01
4.37349439e-01 2.05777884e-01 1.21183074e+00 -7.70342723e-02
-3.11311603e-01 3.76624256e-01 -8.56306180e-02 7.15515494e-01
1.93209395e-01 -1.17222416e+00 6.71755552e-01 6.27652407e-01
5.19679606e-01 1.18872178e+00 7.52725422e-01 -4.60413665e-01
-1.11282909e+00 -8.57958555e-01 1.67656922e+00 -8.60877275e-01
7.86486387e-01 -6.86683357e-01 -7.95785129e-01 1.17638254e+00
4.80401069e-01 -1.78756922e-01 8.19476962e-01 9.11611795e-01
-6.89990759e-01 -1.40010908e-01 -8.80061150e-01 7.39012063e-01
1.02136576e+00 -9.17389572e-01 -1.49765491e+00 3.09223145e-01
1.10141122e+00 5.07998578e-02 -3.59346896e-01 2.71811754e-01
9.81528237e-02 -7.27012694e-01 7.55281508e-01 -1.25227904e+00
4.53489751e-01 5.18815070e-02 -4.56156313e-01 -1.43166935e+00
-6.63017571e-01 4.55196537e-02 3.94455850e-01 1.20767236e+00
8.28433752e-01 -4.30472642e-01 4.50040340e-01 6.20092869e-01
-2.08806187e-01 -6.87394679e-01 -9.80682075e-01 -5.91602743e-01
4.30300117e-01 -4.07341599e-01 5.49493313e-01 5.81987083e-01
3.19526196e-01 1.23364723e+00 4.59561080e-01 -1.98092058e-01
1.43275365e-01 1.25966653e-01 2.72309463e-02 -1.11793852e+00
-1.68889388e-01 -4.30735499e-01 -5.20846128e-01 -6.22184038e-01
5.40215552e-01 -1.30340314e+00 2.76467539e-02 -1.51747084e+00
2.20143106e-02 -5.09145737e-01 -2.45561421e-01 4.01968658e-01
-1.37767732e-01 -2.72943199e-01 1.52281672e-01 -1.63018465e-01
-6.56810105e-02 6.68721378e-01 7.79215455e-01 2.74488647e-02
2.74342261e-02 -6.71953261e-02 -1.01100767e+00 7.32873976e-01
8.77555013e-01 -5.23650050e-01 -4.94386882e-01 -9.00735617e-01
4.27943945e-01 -2.13236436e-01 9.80655774e-02 -3.87510628e-01
-1.30821958e-01 -1.91368014e-01 2.04243183e-01 -7.05591664e-02
4.36474025e-01 -6.17472410e-01 -2.83113360e-01 3.80770475e-01
-9.52855110e-01 4.87011850e-01 1.88299567e-01 4.06005979e-01
-8.82703066e-02 -4.77992922e-01 5.52458942e-01 -4.02132928e-01
-4.24480289e-01 -1.87697992e-01 -6.09654665e-01 5.32155156e-01
7.60609746e-01 3.86270769e-02 -2.40369186e-01 -2.55785644e-01
-4.53130603e-01 -1.56848907e-01 2.68241465e-01 6.01575851e-01
5.96216917e-01 -1.00499237e+00 -7.65015841e-01 3.35271955e-01
-6.36368319e-02 -1.60454646e-01 -3.14760834e-01 4.59636986e-01
-7.09557533e-02 4.80814517e-01 -9.20921117e-02 -5.04315272e-02
-1.15186322e+00 5.01722455e-01 2.38695055e-01 8.32735971e-02
-5.17711401e-01 1.25123000e+00 3.76734227e-01 -5.79787612e-01
-1.49677526e-02 -5.57019293e-01 -1.69575483e-01 3.30347359e-01
2.78622717e-01 -1.81731373e-01 1.48403704e-01 -9.36258197e-01
-5.68597913e-01 -1.64579466e-01 -2.24712655e-01 -1.69644028e-01
1.36876988e+00 2.45426372e-01 -1.30567536e-01 6.37375653e-01
1.10205638e+00 1.96757004e-01 -5.36775172e-01 -2.07130790e-01
2.06918493e-01 -3.28150392e-01 -2.35805541e-01 -3.69978875e-01
-5.49069464e-01 9.50826883e-01 -2.35644504e-01 1.53359830e-01
3.36482555e-01 4.50747162e-01 5.60299575e-01 1.36737883e-01
1.88976750e-02 -8.75550508e-01 -3.21707606e-01 8.86490226e-01
8.61653984e-01 -6.57093465e-01 -2.99305141e-01 -4.10433054e-01
-6.99793518e-01 6.91766858e-01 5.23463368e-01 -3.10674697e-01
4.10838306e-01 1.22395061e-01 -7.66413510e-02 -1.94653288e-01
-1.16370785e+00 -3.56053889e-01 -1.42758014e-02 7.26020038e-01
8.40351999e-01 4.38625395e-01 -7.14877248e-01 8.50328028e-01
-7.70826697e-01 -8.95140409e-01 5.28718293e-01 6.33871734e-01
-3.35816681e-01 -8.69975746e-01 7.38024786e-02 6.48055613e-01
-5.89386106e-01 -7.18495011e-01 -4.83246595e-01 7.62432992e-01
-5.94772026e-02 9.78932619e-01 5.13889730e-01 -5.38238660e-02
3.15510273e-01 2.83147186e-01 6.08091533e-01 -1.39349735e+00
-9.10070777e-01 -6.44072831e-01 4.87843513e-01 -1.55978978e-01
5.07927611e-02 -9.33708787e-01 -1.50690925e+00 -1.62579879e-01
-4.04692292e-01 3.22789758e-01 5.55202603e-01 9.78846669e-01
1.86694875e-01 2.83825815e-01 2.71901995e-01 -2.42506251e-01
-1.13417697e+00 -1.15311360e+00 -5.01510382e-01 8.68087232e-01
3.74692321e-01 -3.25217932e-01 -4.10281092e-01 -4.43830490e-01] | [10.487841606140137, 8.974483489990234] |
1a52c672-c45f-4d62-9213-a9aa9bb6d310 | rusentne-2023-evaluating-entity-oriented | 2305.17679 | null | https://arxiv.org/abs/2305.17679v1 | https://arxiv.org/pdf/2305.17679v1.pdf | RuSentNE-2023: Evaluating Entity-Oriented Sentiment Analysis on Russian News Texts | The paper describes the RuSentNE-2023 evaluation devoted to targeted sentiment analysis in Russian news texts. The task is to predict sentiment towards a named entity in a single sentence. The dataset for RuSentNE-2023 evaluation is based on the Russian news corpus RuSentNE having rich sentiment-related annotation. The corpus is annotated with named entities and sentiments towards these entities, along with related effects and emotional states. The evaluation was organized using the CodaLab competition framework. The main evaluation measure was macro-averaged measure of positive and negative classes. The best results achieved were of 66% Macro F-measure (Positive+Negative classes). We also tested ChatGPT on the test set from our evaluation and found that the zero-shot answers provided by ChatGPT reached 60% of the F-measure, which corresponds to 4th place in the evaluation. ChatGPT also provided detailed explanations of its conclusion. This can be considered as quite high for zero-shot application. | ['Natalia Loukachevitch', 'Nicolay Rusnachenko', 'Anton Golubev'] | 2023-05-28 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-4.89377677e-01 5.43494344e-01 -1.83034688e-01 -6.35255396e-01
-9.36356664e-01 -7.13831663e-01 7.52559245e-01 4.98492748e-01
-7.23014593e-01 1.01106191e+00 6.19349241e-01 1.47196829e-01
2.62373656e-01 -4.93372917e-01 -2.42964864e-01 -4.97135639e-01
2.30860695e-01 5.86852849e-01 -1.94475539e-02 -1.00152326e+00
6.68373764e-01 -6.74012629e-03 -1.30513549e+00 8.35804403e-01
5.80402195e-01 7.97085166e-01 -1.86209857e-01 7.98333108e-01
-1.81892395e-01 1.26030576e+00 -1.04511631e+00 -1.01291108e+00
-2.34703630e-01 -3.08066070e-01 -1.23699391e+00 -2.17824727e-01
6.95031062e-02 6.77746177e-01 4.54869419e-01 1.01976705e+00
6.77458763e-01 4.52328980e-01 6.15299165e-01 -7.72485733e-01
-2.54521400e-01 7.80154228e-01 -1.92089871e-01 4.28995758e-01
4.06704754e-01 -4.02321309e-01 1.47450662e+00 -1.24930644e+00
1.28671229e+00 9.39561605e-01 3.78718644e-01 8.15441787e-01
-5.72573304e-01 -1.98741928e-01 -2.40857676e-02 5.47799133e-02
-7.03561783e-01 -2.56680429e-01 1.91569284e-01 -4.70277339e-01
1.47863019e+00 4.47645664e-01 5.66477835e-01 1.03451896e+00
3.65952402e-01 8.56416166e-01 1.31128812e+00 -3.13679665e-01
4.16665494e-01 8.06811213e-01 6.46327317e-01 1.44611239e-01
-2.94437528e-01 -3.39863241e-01 -6.22618854e-01 -2.36671977e-02
-4.46006775e-01 -8.38089168e-01 -5.06408745e-04 3.54104102e-01
-9.11524117e-01 8.66978407e-01 1.77442685e-01 5.75062633e-01
-4.44043428e-01 -4.63979423e-01 1.11994576e+00 6.31908476e-01
1.14837205e+00 8.45438004e-01 -1.26530516e+00 -6.66508734e-01
-5.01927614e-01 2.86057681e-01 1.17915237e+00 5.79981208e-01
3.13152134e-01 -1.76960915e-01 -3.60015959e-01 1.09898877e+00
-4.82924283e-02 5.23049235e-01 6.60361648e-01 -3.92957598e-01
5.57731986e-01 5.48079371e-01 3.42085123e-01 -9.22183514e-01
-7.49075711e-01 -6.33104444e-01 -3.90324205e-01 5.06675169e-02
-1.69402827e-02 -9.47763443e-01 -7.12158799e-01 1.45894206e+00
1.99575335e-01 -5.74153364e-01 7.97973275e-01 7.57681489e-01
1.60343933e+00 1.06626844e+00 2.33620450e-01 -4.27872151e-01
1.43628454e+00 -1.04510343e+00 -9.56179023e-01 -3.36131781e-01
1.21202350e+00 -1.06161439e+00 9.28271234e-01 4.59246010e-01
-9.86740887e-01 -1.79383546e-01 -7.67856240e-01 1.42053828e-01
-7.19962835e-01 1.40864477e-01 3.60339522e-01 6.95014894e-01
-8.57796371e-01 5.02853811e-01 -2.86824703e-01 -3.34947675e-01
-3.53310779e-02 2.38038316e-01 -4.70239699e-01 6.07426524e-01
-1.58462000e+00 1.49156654e+00 3.48519266e-01 -6.97087497e-02
-4.00907308e-01 -4.20737833e-01 -8.03844213e-01 -5.45495898e-02
1.10018343e-01 9.16201994e-02 1.57039106e+00 -1.05471790e+00
-1.62434685e+00 1.50744402e+00 -7.09297881e-02 -6.23222649e-01
2.75100976e-01 -2.29606211e-01 -8.13351393e-01 -2.93751583e-02
4.94639337e-01 3.66636604e-01 1.80313915e-01 -7.03478992e-01
-6.88403428e-01 -9.12462473e-02 -4.56602834e-02 4.36027378e-01
-2.06687167e-01 8.15154254e-01 -5.25027327e-02 -3.38295817e-01
-2.98237443e-01 -9.19894814e-01 -3.25316250e-01 -1.35473120e+00
-3.41165215e-01 -5.42064011e-01 4.20685858e-01 -5.38279295e-01
9.83029127e-01 -2.03353286e+00 2.19711452e-03 -2.18645483e-02
-2.08167389e-01 2.43086576e-01 3.99501212e-02 7.32281148e-01
-5.12662768e-01 1.73456803e-01 2.66712725e-01 -2.76146680e-01
6.33218791e-03 -1.61625922e-01 -5.29226184e-01 2.54454225e-01
2.78350830e-01 6.04885340e-01 -9.82719898e-01 -3.25343251e-01
1.43806580e-02 2.16696069e-01 -2.48844475e-01 -1.71890333e-01
-1.31816179e-01 2.35619202e-01 -5.41380942e-01 3.94549221e-01
3.68808895e-01 1.37153968e-01 1.35323906e-03 -2.16828287e-02
-3.44941467e-01 7.23325729e-01 -4.01232660e-01 1.10137248e+00
-5.73244750e-01 7.40719020e-01 -4.05371338e-01 -6.84547186e-01
1.37327623e+00 8.30826759e-01 1.77150607e-01 -7.82267809e-01
5.80954075e-01 4.17669237e-01 -2.34125704e-01 -5.33285558e-01
1.01165259e+00 -6.30787611e-01 -8.23642552e-01 3.09357457e-02
3.16569507e-01 -5.20482898e-01 6.19253218e-01 3.45048279e-01
9.65294421e-01 -4.60056737e-02 5.26647866e-01 -5.53037465e-01
7.05947578e-01 3.59744459e-01 4.05931979e-01 3.41220468e-01
-3.30303848e-01 3.86127293e-01 1.09509683e+00 -4.19135481e-01
-7.32351542e-01 -3.56975317e-01 -1.18430130e-01 1.18035042e+00
-1.28171310e-01 -7.96185553e-01 -7.26756036e-01 -9.60927427e-01
-9.55990851e-01 1.17444980e+00 -1.09894240e+00 1.43009871e-01
-1.09429441e-01 -1.04005301e+00 3.80528867e-01 7.37535208e-02
1.07402019e-01 -1.53648508e+00 -4.92130816e-01 1.24441087e-01
-6.41169906e-01 -1.24556518e+00 1.26935691e-01 5.67606151e-01
-4.80308205e-01 -7.38739610e-01 -6.78650975e-01 -9.28988934e-01
2.81855941e-01 -5.40910304e-01 1.43967819e+00 -5.07732153e-01
2.46109158e-01 -2.64892466e-02 -9.82239425e-01 -9.49355781e-01
-5.87835789e-01 2.69588351e-01 -2.80447811e-01 -3.40301245e-01
7.01648355e-01 8.23732018e-02 -1.87830299e-01 1.45933807e-01
-1.92560598e-01 -1.33996874e-01 2.17621580e-01 9.30877686e-01
8.90202299e-02 -3.00194174e-01 9.84244704e-01 -1.32242334e+00
6.05347276e-01 -4.86794889e-01 -2.35749498e-01 -5.90133816e-02
-2.55582929e-01 -3.71639282e-01 5.41089714e-01 1.35907739e-01
-1.33675015e+00 -3.62535387e-01 -6.95596516e-01 4.50673938e-01
-1.05285365e-02 7.60779798e-01 1.43817008e-01 4.18412924e-01
1.07529151e+00 -4.13575798e-01 -5.89620411e-01 -1.07832126e-01
1.69754073e-01 8.89081657e-01 1.66519254e-01 -2.71110684e-01
2.11430658e-02 1.54637292e-01 -4.47566450e-01 -9.53337550e-01
-1.70203936e+00 -7.54412532e-01 -1.85994044e-01 -4.57510382e-01
9.40140963e-01 -1.02152777e+00 -7.41718531e-01 3.22883725e-01
-1.10141993e+00 -1.27289265e-01 -5.26027977e-01 5.22342086e-01
-4.87014502e-01 -1.63786784e-01 -9.21524286e-01 -1.05896771e+00
-8.20460677e-01 -6.60608470e-01 8.18745553e-01 2.83358365e-01
-7.05277085e-01 -9.71528292e-01 5.36361635e-01 5.79728246e-01
6.87228218e-02 1.76765457e-01 4.22882050e-01 -1.22247434e+00
8.03216398e-01 -5.98934829e-01 1.98227450e-01 5.44294536e-01
-4.19365019e-01 -2.18702164e-02 -1.22788358e+00 6.98524117e-02
2.44822100e-01 -8.73114586e-01 6.79476202e-01 2.92560250e-01
3.02804500e-01 1.97071731e-01 1.48347244e-01 -4.01003569e-01
1.29192901e+00 2.02590108e-01 8.02176118e-01 7.36293375e-01
8.44205543e-02 1.00561869e+00 1.46823335e+00 6.32988214e-01
6.08770438e-02 3.92305136e-01 3.82306755e-01 -3.60011198e-02
4.39563036e-01 1.38748854e-01 8.10011923e-01 1.10360956e+00
-9.42561179e-02 -5.65368176e-01 -8.76320720e-01 8.00917208e-01
-1.69575405e+00 -1.08304751e+00 -5.78515410e-01 1.65837157e+00
7.83352673e-01 5.21544039e-01 -5.58422096e-02 2.74666883e-02
8.20806801e-01 2.43247300e-01 1.42560095e-01 -1.24267161e+00
-6.40757322e-01 3.48347515e-01 1.52388573e-01 5.04813254e-01
-1.26733971e+00 1.31339228e+00 6.16811991e+00 9.81016994e-01
-8.95903051e-01 2.82134980e-01 8.89884710e-01 -4.06405479e-02
-7.20284730e-02 -1.25571519e-01 -1.02525222e+00 1.74791053e-01
1.58477485e+00 -3.07271093e-01 -5.33784449e-01 8.74718785e-01
2.80231625e-01 -2.26131603e-01 -3.19787890e-01 4.87517774e-01
2.21753672e-01 -1.14176977e+00 -4.68714088e-01 -2.46028468e-01
1.15079427e+00 5.65388739e-01 -4.33044098e-02 8.93432438e-01
2.07720280e-01 -7.10826397e-01 7.96150863e-01 1.86626241e-01
4.52696532e-01 -1.20909464e+00 1.50544071e+00 3.69235396e-01
-5.72045267e-01 2.66683936e-01 -3.68833065e-01 -3.29324633e-01
3.56586605e-01 7.61653304e-01 -9.83630657e-01 5.73153079e-01
7.27258265e-01 1.00844693e+00 -2.93273717e-01 5.60814202e-01
-4.82102841e-01 7.84188509e-01 3.29527669e-02 -8.34948242e-01
6.37257040e-01 -9.72669274e-02 6.48676634e-01 1.56691813e+00
2.07655162e-01 3.40903588e-02 -6.26674369e-02 -2.61423755e-02
-2.55088121e-01 9.71351564e-01 -3.08830231e-01 -8.39427710e-02
-1.81484446e-01 1.64130819e+00 -1.00651824e+00 -6.72276855e-01
-2.83835053e-01 7.88941205e-01 1.93562478e-01 -2.07204983e-01
-6.53343499e-01 -6.77747071e-01 2.39844903e-01 -4.23606336e-01
3.02169621e-01 7.78230250e-01 -4.04889941e-01 -1.18543291e+00
-3.45684767e-01 -7.63696730e-01 2.46145129e-01 -1.09137845e+00
-1.24571884e+00 1.32762015e+00 -3.40240479e-01 -1.23904192e+00
-5.33801541e-02 -7.90284932e-01 -5.21473885e-01 7.04443574e-01
-1.03070927e+00 -7.84158409e-01 3.32895160e-01 8.48174393e-02
6.99431598e-01 -3.77647817e-01 1.25777411e+00 2.39359617e-01
-3.64272326e-01 1.23559706e-01 -9.67288986e-02 1.31804481e-01
9.46432412e-01 -1.56956637e+00 3.38633478e-01 6.03635728e-01
-1.79798186e-01 2.11046025e-01 1.42368817e+00 -7.78922439e-01
-3.53296816e-01 -8.44867051e-01 1.96366382e+00 -6.43603027e-01
1.04267359e+00 -4.88771405e-03 -3.77056986e-01 5.13566971e-01
7.08553195e-01 -4.00978684e-01 1.04375291e+00 5.00582755e-01
-4.76190187e-02 3.71051073e-01 -1.09376526e+00 4.50967312e-01
1.78374141e-01 -4.10864353e-01 -8.65030408e-01 7.37543225e-01
5.41977584e-01 -6.85705483e-01 -8.58977675e-01 2.72890896e-01
9.69578773e-02 -1.09409225e+00 1.96062222e-01 -8.18402886e-01
1.08216572e+00 -7.16756508e-02 -1.38507053e-01 -1.66396344e+00
3.00334960e-01 -4.26887661e-01 6.17183447e-01 1.25300872e+00
1.20379066e+00 -3.51530939e-01 7.18613267e-01 3.40603203e-01
-4.24875081e-01 -8.09383154e-01 -6.26304984e-01 -2.76735783e-01
1.23385720e-01 -6.61659479e-01 -6.59999400e-02 1.15411639e+00
5.83707750e-01 1.10799539e+00 -5.56842446e-01 -2.35041097e-01
-1.22191213e-01 1.36489775e-02 5.19167364e-01 -1.10592890e+00
-2.79738456e-02 -2.95899600e-01 -4.99245346e-01 -2.77051717e-01
2.11196378e-01 -8.82097900e-01 2.02071011e-01 -1.49485373e+00
2.56418914e-01 1.65972590e-01 -2.71214306e-01 4.04456347e-01
-1.62583478e-02 5.36038756e-01 5.32257631e-02 -3.33634317e-01
-8.08628976e-01 3.40848476e-01 1.11444676e+00 1.20865799e-01
-1.78177860e-02 3.07682067e-01 -8.42983603e-01 8.95567119e-01
9.91352201e-01 -6.53257191e-01 -8.61575231e-02 3.04798961e-01
1.10061431e+00 5.74114025e-02 -2.74885029e-01 -5.09309173e-01
-2.19206631e-01 8.22742358e-02 -1.41945924e-03 -9.91547763e-01
5.53771734e-01 -2.80282259e-01 -2.54295945e-01 4.68114793e-01
-5.78917086e-01 -1.55015802e-02 1.54403031e-01 5.02475500e-02
-6.69777691e-01 -7.09849656e-01 7.50719130e-01 -2.10690945e-01
-8.06251287e-01 -3.05243969e-01 -7.47149169e-01 4.61566120e-01
9.75465953e-01 1.82287768e-01 -4.51006979e-01 -5.10545611e-01
-1.21563828e+00 1.93004653e-01 -6.45417967e-05 4.81927127e-01
2.32347086e-01 -7.89062321e-01 -1.04286659e+00 -5.14598906e-01
4.77542311e-01 -6.32989764e-01 5.38000166e-01 8.70834947e-01
-3.89136851e-01 4.11579996e-01 -1.36158302e-01 -7.74785802e-02
-1.64607954e+00 1.52833924e-01 2.35670656e-01 -6.93642020e-01
-2.11100966e-01 1.12822485e+00 -1.01428591e-01 -8.56880426e-01
-2.09466428e-01 4.24440876e-02 -1.11804640e+00 7.58402824e-01
6.61306500e-01 2.10670695e-01 5.51110506e-01 -1.03704488e+00
-3.93347561e-01 3.76582831e-01 -2.27298230e-01 -4.71291721e-01
1.52728879e+00 1.80332530e-02 -3.64810228e-01 9.36299562e-01
1.15102482e+00 4.91635621e-01 -1.02412745e-01 1.79967597e-01
3.24378282e-01 1.43284529e-01 1.06760701e-02 -1.41207039e+00
-7.27183402e-01 7.41536200e-01 1.68560520e-01 5.57313383e-01
6.87502861e-01 -3.83915292e-04 5.13617992e-01 6.14907384e-01
1.48230255e-01 -1.63719738e+00 -2.64655054e-01 1.25319183e+00
8.92072558e-01 -1.40165997e+00 -1.17668301e-01 -3.88136446e-01
-1.59671426e+00 1.02959752e+00 5.82003057e-01 -2.30481058e-01
5.61912298e-01 -9.73295700e-03 6.05643630e-01 -6.47849202e-01
-1.16211104e+00 -1.09406464e-01 2.69354105e-01 1.12701133e-01
9.35820699e-01 2.20856518e-01 -9.75287080e-01 8.20530057e-01
-6.72028661e-01 -3.25717032e-01 9.65960383e-01 7.09745228e-01
-4.10093814e-01 -9.16492105e-01 8.82154852e-02 2.43537456e-01
-1.23806500e+00 -1.99440941e-01 -7.60115504e-01 5.67108393e-01
-2.57817924e-01 1.32307303e+00 -1.82128996e-01 -4.08583492e-01
6.74082696e-01 8.84334818e-02 -1.67236030e-01 -1.00414073e+00
-1.48306406e+00 1.49409622e-01 1.03821468e+00 -1.97815105e-01
-8.17319810e-01 -6.02405071e-01 -1.33432698e+00 -4.77694571e-02
-4.99805927e-01 1.07924938e+00 1.06059778e+00 1.16176701e+00
-1.50985103e-02 5.50668776e-01 7.62155056e-01 -5.84123969e-01
-2.36729428e-01 -1.38069773e+00 -7.55542994e-01 3.54316056e-01
-3.58579420e-02 -1.47162676e-01 -5.87560773e-01 -2.09508196e-01] | [11.234968185424805, 6.933368682861328] |
6f6a0b89-e518-47b3-9827-7c52e959b4c2 | single-image-cloud-detection-via-multi-image | 2007.15144 | null | https://arxiv.org/abs/2007.15144v1 | https://arxiv.org/pdf/2007.15144v1.pdf | Single Image Cloud Detection via Multi-Image Fusion | Artifacts in imagery captured by remote sensing, such as clouds, snow, and shadows, present challenges for various tasks, including semantic segmentation and object detection. A primary challenge in developing algorithms for identifying such artifacts is the cost of collecting annotated training data. In this work, we explore how recent advances in multi-image fusion can be leveraged to bootstrap single image cloud detection. We demonstrate that a network optimized to estimate image quality also implicitly learns to detect clouds. To support the training and evaluation of our approach, we collect a large dataset of Sentinel-2 images along with a per-pixel semantic labelling for land cover. Through various experiments, we demonstrate that our method reduces the need for annotated training data and improves cloud detection performance. | ['M. Usman Rafique', 'Hunter Blanton', 'Connor Greenwell', 'Scott Workman', 'Nathan Jacobs'] | 2020-07-29 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 7.83585250e-01 -2.83335775e-01 1.05436236e-01 -6.45141661e-01
-1.20034015e+00 -9.26899254e-01 8.09562802e-02 3.20056081e-01
-2.89437771e-01 4.79604781e-01 -4.64407742e-01 -4.70378160e-01
1.42994910e-01 -7.30530620e-01 -8.68067443e-01 -4.76240814e-01
-1.71610370e-01 1.30305335e-01 2.82132357e-01 2.14464083e-01
2.67637260e-02 6.42209291e-01 -1.61677742e+00 2.15603322e-01
1.12873721e+00 1.03924596e+00 5.65355361e-01 7.18714416e-01
1.19115233e-01 5.66330373e-01 -5.04919708e-01 6.68213889e-02
4.86074150e-01 -5.55495135e-02 -5.83104312e-01 4.34554487e-01
9.88428593e-01 -5.51364481e-01 3.23656470e-01 1.28812099e+00
2.54045337e-01 -1.70248881e-01 3.62100542e-01 -1.33161640e+00
-6.66532665e-02 3.55992675e-01 -7.03180254e-01 3.21638256e-01
-3.17102253e-01 2.18940318e-01 8.52725148e-01 -8.97726953e-01
3.05647284e-01 7.34199107e-01 9.64575410e-01 1.89985693e-01
-8.71192932e-01 -7.62588501e-01 2.82493949e-01 -3.09319109e-01
-1.53331161e+00 -4.99713749e-01 3.68148714e-01 -5.53373814e-01
7.22266197e-01 4.35538381e-01 8.97887588e-01 3.23953986e-01
-3.92194659e-01 9.97114241e-01 1.40872443e+00 -3.71091247e-01
2.71892995e-01 1.32843792e-01 1.92613229e-01 5.29399693e-01
7.44734406e-01 6.36901567e-03 -2.55804539e-01 -2.04529479e-01
6.17954850e-01 1.22993208e-01 -2.04501048e-01 -3.36940959e-02
-5.25040686e-01 6.66631281e-01 4.41460431e-01 1.50604472e-02
-3.51070642e-01 5.53050637e-01 -8.71952549e-02 -1.79802850e-01
8.09632540e-01 4.60575879e-01 -5.76122701e-01 3.75055403e-01
-1.38332105e+00 3.72181714e-01 4.75547314e-01 9.73508656e-01
1.16122520e+00 2.05422640e-02 2.20081791e-01 5.14947295e-01
4.07880753e-01 1.08064127e+00 -4.06944364e-01 -1.22206688e+00
2.64437914e-01 5.06763577e-01 5.97082913e-01 -6.48670375e-01
-3.82418841e-01 -4.69898522e-01 -3.01946908e-01 4.27655935e-01
1.86765730e-01 -2.07016900e-01 -1.30614448e+00 1.11633670e+00
3.84518504e-01 3.62622231e-01 9.09563899e-02 9.26600873e-01
5.79722643e-01 3.57303321e-01 2.76791334e-01 5.51251844e-02
1.35345590e+00 -7.34391212e-01 -2.71903276e-01 -9.16296899e-01
5.10520756e-01 -5.78189194e-01 8.04683447e-01 1.77329972e-01
-6.35138869e-01 -1.82714656e-01 -9.25924480e-01 3.19510788e-01
-4.40895438e-01 2.97934443e-01 9.15371656e-01 6.86856568e-01
-8.99866760e-01 5.69934666e-01 -9.09945846e-01 -3.64360958e-01
7.60471880e-01 9.35344025e-02 1.41516909e-01 -1.98579013e-01
-7.75830567e-01 7.47848749e-01 3.20077062e-01 4.17850792e-01
-1.12793255e+00 -6.14186287e-01 -7.86202788e-01 -1.51473507e-02
3.60141724e-01 -3.44190449e-01 1.21892965e+00 -1.30740237e+00
-5.22470534e-01 1.05778825e+00 -1.02528885e-01 -5.30673265e-01
3.35437357e-01 -4.56277192e-01 -2.05998942e-01 3.34607452e-01
3.57133716e-01 7.22851396e-01 9.59985256e-01 -1.84584212e+00
-1.10523283e+00 -4.63610888e-01 6.36873245e-02 2.43865147e-01
1.42628886e-02 3.26108575e-01 -3.15477669e-01 -2.54346013e-01
2.57085800e-01 -1.01721823e+00 -6.26809597e-01 8.76000226e-02
-1.52121723e-01 2.78467536e-01 9.49301839e-01 -6.23383105e-01
5.73704064e-01 -2.02433538e+00 -6.67971075e-01 3.69981825e-01
2.15015560e-01 2.55912930e-01 -1.75364628e-01 -9.02383402e-02
3.83635819e-01 4.79052782e-01 -6.16440356e-01 -3.78366679e-01
-3.96081358e-01 3.80237937e-01 -4.30874705e-01 5.08845389e-01
5.77493072e-01 8.62780690e-01 -9.70810115e-01 -5.89095592e-01
2.07730860e-01 2.74427444e-01 -6.93117827e-02 2.27475703e-01
-4.88426059e-01 3.87574345e-01 -3.91238391e-01 1.12986863e+00
1.17944753e+00 -4.24587965e-01 1.29947275e-01 9.39405710e-02
-2.77100176e-01 1.32734358e-01 -1.13599873e+00 1.30239248e+00
-3.13281149e-01 7.45198965e-01 4.54953671e-01 -5.16045928e-01
4.30453271e-01 1.57855581e-02 3.59847605e-01 -5.50394654e-01
5.80855319e-03 3.33308697e-01 -4.76690710e-01 -7.10819542e-01
6.88662291e-01 -9.36921090e-02 1.64125726e-01 2.74497837e-01
-4.30459887e-01 -6.56919718e-01 -1.29085898e-01 9.36006755e-02
9.07288015e-01 2.40919396e-01 -1.56916752e-01 -9.51013789e-02
-1.19074501e-01 6.58507705e-01 5.69532573e-01 1.09104204e+00
-2.08657116e-01 6.88714623e-01 -1.27799332e-01 -3.84963065e-01
-9.45918739e-01 -8.06704104e-01 -1.94783181e-01 1.02082622e+00
4.15180236e-01 6.17968217e-02 -7.58999407e-01 -6.45758033e-01
2.31379643e-01 3.61383587e-01 -3.08823884e-01 5.52889526e-01
-4.12241757e-01 -1.14308953e+00 7.04686522e-01 6.56433940e-01
4.97213274e-01 -8.04938555e-01 -1.09016383e+00 3.05582341e-02
-2.92401016e-01 -1.71453655e+00 9.42088813e-02 3.77801836e-01
-1.00063765e+00 -1.21826589e+00 -2.06287354e-01 -4.38852251e-01
8.57297480e-01 1.01381576e+00 1.28128791e+00 6.72373772e-01
-5.17933607e-01 4.52826589e-01 -4.46753323e-01 -7.56966054e-01
-1.59450859e-01 -1.46837860e-01 -4.27201480e-01 -2.79436707e-01
2.65740633e-01 -3.01091343e-01 -6.77256227e-01 1.73009351e-01
-1.02914679e+00 4.40176055e-02 6.64523900e-01 3.20704609e-01
5.82665622e-01 1.72196746e-01 5.95893785e-02 -9.59965765e-01
1.59270279e-02 -2.95261830e-01 -1.14886212e+00 2.28329837e-01
-6.84993625e-01 -3.87683213e-01 -1.71014056e-01 1.54446647e-01
-9.08984125e-01 6.60475373e-01 2.67400265e-01 -4.42685455e-01
-4.79209334e-01 6.90707564e-01 4.03686166e-02 -6.83346987e-01
5.91134071e-01 8.00148398e-02 -4.43687379e-01 -2.97425240e-01
3.23683619e-01 9.37459350e-01 5.52886963e-01 -3.00590247e-01
1.01558948e+00 9.67379570e-01 -1.78959221e-01 -9.99977708e-01
-1.28658831e+00 -9.14925218e-01 -6.07454479e-01 -4.04115260e-01
8.06349695e-01 -1.51573622e+00 -2.06748888e-01 5.55787265e-01
-9.84348416e-01 -5.72707057e-01 -3.21108173e-03 1.46276370e-01
-1.33961111e-01 3.49841028e-01 -9.62205306e-02 -1.42045557e+00
-6.12883806e-01 -7.94764876e-01 1.52095759e+00 3.09884399e-01
3.35165501e-01 -5.48720896e-01 -3.66689682e-01 7.71369040e-01
5.32219887e-01 4.64207619e-01 2.15223357e-01 -5.45923077e-02
-1.24279165e+00 -2.97144521e-02 -7.26809144e-01 4.80628759e-01
-5.44324555e-02 1.24874994e-01 -1.43905270e+00 -2.49769643e-01
-1.83181792e-01 -4.63040113e-01 1.14370179e+00 4.22056973e-01
1.14198208e+00 -2.49614902e-02 -3.76489341e-01 7.38703072e-01
1.90113914e+00 -1.37706652e-01 3.23733091e-01 4.61345822e-01
8.41370881e-01 5.41066587e-01 9.17669356e-01 3.06813687e-01
4.07544225e-01 1.38328582e-01 8.31310451e-01 -4.96975034e-01
7.42723942e-02 1.67895809e-01 -1.23911500e-01 -2.40661036e-02
-2.04902560e-01 -2.27497920e-01 -1.27846050e+00 8.55200648e-01
-1.88614202e+00 -7.52253890e-01 -3.88951480e-01 1.89950275e+00
5.35753012e-01 8.97135213e-02 -3.03504258e-01 -1.78215563e-01
6.08947635e-01 1.20572284e-01 -5.93213856e-01 4.55117673e-01
-3.40061754e-01 2.41695762e-01 1.39324164e+00 4.70026582e-01
-1.52320111e+00 1.28432071e+00 7.15832853e+00 3.13944697e-01
-1.12184262e+00 3.15482110e-01 4.89888281e-01 8.20818990e-02
-3.24174315e-01 2.45289132e-01 -6.92858815e-01 1.12443231e-01
8.12586486e-01 5.08822024e-01 3.48541349e-01 8.55330229e-01
5.89413047e-01 -7.36664534e-01 -4.45074350e-01 6.96347177e-01
-7.94296246e-03 -1.24140930e+00 -1.83761314e-01 -1.39738947e-01
9.17996466e-01 7.79249012e-01 -1.37377486e-01 -2.34601796e-01
7.37804592e-01 -9.96005595e-01 8.54303002e-01 1.44787252e-01
8.09359372e-01 -2.80722827e-01 7.24237144e-01 1.32643625e-01
-1.22404301e+00 -8.81216004e-02 -1.77870259e-01 -2.15797961e-01
-1.02806248e-01 7.85693407e-01 -9.12562907e-01 4.34433341e-01
8.24160397e-01 5.70741832e-01 -7.06999540e-01 1.34878707e+00
-4.63965863e-01 8.17136407e-01 -6.26781642e-01 3.45386803e-01
3.40805054e-01 -1.48422420e-01 1.80010408e-01 1.18733859e+00
2.58042008e-01 2.69372314e-01 4.74995285e-01 9.02816176e-01
-1.28714904e-01 -2.72651166e-01 -4.80642110e-01 -1.38163343e-01
5.83601177e-01 1.42952299e+00 -1.02792871e+00 -3.80970061e-01
-2.69795120e-01 8.85236263e-01 8.95545911e-03 3.87275696e-01
-7.99903572e-01 -3.47507112e-02 7.64066935e-01 8.82366151e-02
3.51212054e-01 -4.34308738e-01 -5.93163669e-01 -1.08824658e+00
1.46266297e-01 -7.00739026e-01 2.41809562e-02 -1.09179771e+00
-9.40685868e-01 2.31699452e-01 -1.54484332e-01 -1.03380394e+00
1.84664130e-01 -5.31420350e-01 -4.76016730e-01 8.81403208e-01
-2.14242363e+00 -1.65146089e+00 -1.23265529e+00 1.59250051e-01
3.12599093e-01 4.82318372e-01 6.39444530e-01 2.41525859e-01
-2.74792701e-01 -1.45278081e-01 1.18175365e-01 1.39951020e-01
2.81490415e-01 -1.22265995e+00 5.83661079e-01 1.30460119e+00
1.18747786e-01 4.16378006e-02 8.00359130e-01 -7.32542515e-01
-1.28231621e+00 -1.67588842e+00 3.30504596e-01 -5.54091394e-01
5.42645991e-01 -2.58159548e-01 -7.89535522e-01 6.74950957e-01
-2.59512931e-01 3.24199766e-01 4.64161962e-01 -1.12003766e-01
-2.44637594e-01 -1.18941233e-01 -1.16496694e+00 -7.39823580e-02
9.45419312e-01 -6.56420052e-01 -1.06806941e-01 7.09185839e-01
6.13657176e-01 -6.55275822e-01 -5.07289708e-01 6.51003838e-01
5.20276129e-01 -9.16016221e-01 7.30264425e-01 -2.94992000e-01
2.31207281e-01 -5.90520263e-01 -3.19128901e-01 -9.74171221e-01
-2.63945125e-02 -6.53103888e-02 4.11085039e-01 9.39502418e-01
3.90627682e-01 -1.98631704e-01 9.53958154e-01 6.30939603e-01
-1.41375065e-01 -1.24229982e-01 -6.98193848e-01 -8.11145365e-01
-1.31055936e-01 -5.88490844e-01 5.87286711e-01 9.10457492e-01
-1.01383495e+00 -2.04118133e-01 -1.22740798e-01 1.19206655e+00
1.02191997e+00 6.45572662e-01 8.19510102e-01 -1.40106392e+00
1.03721106e-02 -1.86308250e-02 -1.72827438e-01 -6.58415794e-01
-3.71629111e-02 -5.14390409e-01 5.44824958e-01 -1.64645827e+00
4.43660378e-01 -1.03483999e+00 -2.07226407e-02 7.19268262e-01
-4.60985482e-01 5.98395884e-01 2.15750590e-01 4.26308393e-01
-7.32888937e-01 4.11867239e-02 6.79661214e-01 -2.96897590e-01
-6.13543279e-02 -8.96728635e-02 -5.34488857e-01 8.14929366e-01
1.01397288e+00 -8.01955760e-01 4.35369462e-02 -9.90617096e-01
3.68735999e-01 -1.62183404e-01 8.32980156e-01 -1.17817724e+00
1.30119761e-02 -4.97541785e-01 3.85016531e-01 -8.84463966e-01
2.55286902e-01 -9.09848213e-01 8.05354342e-02 2.57820994e-01
1.43438473e-01 -3.09958905e-01 3.87654603e-01 5.92693508e-01
-3.75601426e-02 -4.40973222e-01 9.16558027e-01 -5.46959996e-01
-9.33833063e-01 2.61239916e-01 -2.35124350e-01 -1.04973719e-01
9.25350904e-01 -4.29183505e-02 -3.47620159e-01 -7.84112960e-02
-4.88041759e-01 5.57471514e-01 7.28125572e-01 1.60840139e-01
5.08535028e-01 -5.92384934e-01 -6.75493598e-01 8.09304416e-02
4.47843611e-01 3.83237928e-01 -2.10275650e-02 5.15118837e-01
-6.91435039e-01 8.35659280e-02 1.19818985e-01 -9.25229549e-01
-1.48140824e+00 -2.65008658e-02 7.28412032e-01 6.28858730e-02
-3.50110322e-01 9.06329572e-01 -1.63627952e-01 -3.57237160e-01
-2.11162474e-02 -5.85742831e-01 2.91673958e-01 -6.06702454e-02
3.93728703e-01 -1.00979563e-02 3.36957544e-01 -2.97393024e-01
-4.00239378e-01 3.24674189e-01 2.56927371e-01 -9.17478278e-02
1.52159846e+00 -2.22372055e-01 -2.38125131e-01 2.32304901e-01
4.92661238e-01 -1.49014071e-01 -1.33874238e+00 -1.54729038e-01
9.27841067e-02 -8.11764419e-01 5.49741685e-01 -9.18946445e-01
-1.17936850e+00 5.91715038e-01 9.96473968e-01 2.17755452e-01
1.04905534e+00 7.70178884e-02 5.34272492e-01 4.27829951e-01
6.33950591e-01 -1.03553128e+00 -3.31877172e-01 3.93287838e-01
3.56198519e-01 -1.96731150e+00 1.59374058e-01 -6.33673608e-01
-3.84371638e-01 8.45810890e-01 5.58556616e-01 8.84642825e-02
5.35847306e-01 5.78860939e-01 5.16467094e-01 -6.30162835e-01
-2.46012717e-01 -8.68285298e-01 -1.56182021e-01 4.89616305e-01
1.32776231e-01 3.92137825e-01 2.54610628e-01 7.99063072e-02
1.85219258e-01 1.88311100e-01 6.88876271e-01 1.34222066e+00
-9.27337825e-01 -6.20335877e-01 -7.04806328e-01 7.41732001e-01
-3.75960171e-01 -4.53828037e-01 -4.22422051e-01 4.33107316e-01
4.95285898e-01 1.10861647e+00 9.55884904e-02 -2.79379301e-02
-1.36958566e-02 -2.51637489e-01 3.82815152e-01 -8.58163595e-01
-5.32931447e-01 2.87249573e-02 1.87601134e-01 -4.67984289e-01
-9.99897301e-01 -8.52668762e-01 -1.24577332e+00 1.51844770e-01
-8.85007620e-01 1.05856456e-01 1.10409892e+00 9.88399804e-01
2.25785851e-01 3.30680430e-01 4.63824242e-01 -9.20873821e-01
-3.23999882e-01 -8.01044524e-01 -8.15067053e-01 1.93525463e-01
6.90484822e-01 -5.27712941e-01 -5.05534947e-01 1.86480850e-01] | [9.655633926391602, -1.6225709915161133] |
f740c675-cba2-418a-9e00-25b8229a2087 | self-contrastive-learning-for-session-based | 2306.01266 | null | https://arxiv.org/abs/2306.01266v1 | https://arxiv.org/pdf/2306.01266v1.pdf | Self Contrastive Learning for Session-based Recommendation | Session-based recommendation, which aims to predict the next item of users' interest as per an existing sequence interaction of items, has attracted growing applications of Contrastive Learning (CL) with improved user and item representations. However, these contrastive objectives: (1) serve a similar role as the cross-entropy loss while ignoring the item representation space optimisation; and (2) commonly require complicated modelling, including complex positive/negative sample constructions and extra data augmentation. In this work, we introduce Self-Contrastive Learning (SCL), which simplifies the application of CL and enhances the performance of state-of-the-art CL-based recommendation techniques. Specifically, SCL is formulated as an objective function that directly promotes a uniform distribution among item representations and efficiently replaces all the existing contrastive objective components of state-of-the-art models. Unlike previous works, SCL eliminates the need for any positive/negative sample construction or data augmentation, leading to enhanced interpretability of the item representation space and facilitating its extensibility to existing recommender systems. Through experiments on three benchmark datasets, we demonstrate that SCL consistently improves the performance of state-of-the-art models with statistical significance. Notably, our experiments show that SCL improves the performance of two best-performing models by 8.2% and 9.5% in P@10 (Precision) and 9.9% and 11.2% in MRR@10 (Mean Reciprocal Rank) on average across different benchmarks. Additionally, our analysis elucidates the improvement in terms of alignment and uniformity of representations, as well as the effectiveness of SCL with a low computational cost. | ['Aldo Lipani', 'Xi Wang', 'Zhengxiang Shi'] | 2023-06-02 | null | null | null | null | ['session-based-recommendations'] | ['miscellaneous'] | [ 2.49765486e-01 -2.95224518e-01 -4.00174260e-01 -3.08175862e-01
-6.11111104e-01 -6.24135673e-01 6.99102998e-01 3.44282269e-01
-3.85695815e-01 5.94131529e-01 2.06527025e-01 -2.69273251e-01
-4.78502333e-01 -6.82696640e-01 -8.11886787e-01 -6.22671306e-01
-2.50683010e-01 2.28960142e-01 6.01943433e-02 -3.77477199e-01
3.76622021e-01 1.96747974e-01 -1.66905308e+00 3.01054180e-01
1.08649468e+00 1.25436211e+00 1.31726205e-01 1.95586517e-01
-2.87890192e-02 4.97962296e-01 -3.28629106e-01 -5.70059538e-01
2.55788773e-01 -2.60335565e-01 -5.87526619e-01 -6.93018436e-02
2.51372725e-01 -1.00744531e-01 4.88909567e-03 6.98592842e-01
3.95067841e-01 5.31321645e-01 7.63533533e-01 -1.07591879e+00
-1.03678882e+00 5.93900084e-01 -6.55176580e-01 3.66931595e-02
2.33722866e-01 -1.43086091e-01 1.45061398e+00 -1.07866371e+00
1.44750983e-01 9.50233340e-01 6.83336437e-01 3.36848348e-01
-1.41487336e+00 -5.25099039e-01 5.62784374e-01 1.55267179e-01
-1.30367279e+00 -1.46997720e-01 6.02367043e-01 -4.26632404e-01
8.70133162e-01 5.17392218e-01 4.29579765e-01 9.84378576e-01
-1.34149149e-01 1.02354419e+00 7.65143931e-01 -4.06819165e-01
2.44728431e-01 4.02949065e-01 4.13418412e-01 3.68530452e-01
3.93312097e-01 3.99602950e-02 -3.10880929e-01 -1.69729978e-01
6.80613518e-01 3.33136886e-01 -2.22990826e-01 -6.01831257e-01
-8.15159321e-01 9.14337993e-01 4.44514185e-01 1.73301533e-01
-4.35198218e-01 -1.74835235e-01 3.05719584e-01 1.82024434e-01
7.35208988e-01 7.74943233e-01 -5.55767953e-01 5.53979278e-02
-7.31138587e-01 2.08967015e-01 6.19052470e-01 7.38375306e-01
3.32054287e-01 -8.02801773e-02 -4.00533140e-01 1.15142369e+00
3.40757549e-01 3.02574873e-01 5.86270809e-01 -4.01060134e-01
3.82668763e-01 7.87019908e-01 1.57915309e-01 -8.98916960e-01
-3.22693288e-01 -1.16413414e+00 -8.05936456e-01 -2.14211628e-01
2.23854437e-01 4.71018068e-02 -5.15043974e-01 1.79364967e+00
2.83370782e-02 2.45041564e-01 -8.78686309e-02 7.67167211e-01
7.44112909e-01 5.80135882e-01 1.64815500e-01 -3.91850710e-01
1.20023441e+00 -1.09984159e+00 -4.26951140e-01 -1.90455139e-01
5.44229627e-01 -5.58539867e-01 1.28915381e+00 4.85508412e-01
-1.11554098e+00 -6.14539921e-01 -9.51412320e-01 9.03202370e-02
-2.98425913e-01 2.06129014e-01 8.43483746e-01 6.83959663e-01
-7.63961375e-01 6.17892206e-01 -4.43247378e-01 3.16972984e-03
3.81066114e-01 6.03443325e-01 -1.94034249e-01 3.70292030e-02
-1.12279356e+00 6.15884721e-01 2.18680412e-01 -9.41935852e-02
-1.54114291e-01 -1.23044217e+00 -5.99074066e-01 4.57487881e-01
5.59148550e-01 -6.62886679e-01 1.01623130e+00 -9.70536888e-01
-1.62217593e+00 2.83182740e-01 4.23626751e-02 -4.66664314e-01
3.51824850e-01 -6.53183818e-01 -5.34684837e-01 -3.69280189e-01
-2.41093591e-01 3.79158050e-01 5.02927005e-01 -1.22668755e+00
-6.89287186e-01 -2.73707241e-01 1.38223007e-01 2.07081333e-01
-7.11551547e-01 -2.50736058e-01 -7.69590080e-01 -9.17118549e-01
-1.71338484e-01 -8.96834910e-01 -2.65300035e-01 -1.39858663e-01
4.44196798e-02 -3.59656543e-01 2.78977782e-01 -4.16884363e-01
1.70493925e+00 -2.14239144e+00 1.86941057e-01 3.06472778e-01
5.24835326e-02 5.03147960e-01 -3.56015146e-01 4.08211708e-01
-4.51851077e-02 7.29371756e-02 -3.31317671e-02 -3.53552818e-01
1.88430578e-01 -3.01575996e-02 -2.94900984e-01 3.06141704e-01
1.18015818e-01 9.38066185e-01 -7.97269225e-01 5.77199534e-02
2.35691473e-01 6.46862328e-01 -1.04036593e+00 2.40211576e-01
-1.35023102e-01 1.37489140e-01 -1.69070423e-01 2.64380664e-01
5.21163404e-01 -6.38295472e-01 4.90895629e-01 -1.64612830e-01
4.43586819e-02 6.42633140e-01 -1.28372097e+00 1.46464181e+00
-6.94519818e-01 -6.54223338e-02 -4.37971056e-01 -1.01479483e+00
9.30075705e-01 -6.34203432e-03 5.36991596e-01 -9.73075449e-01
-8.06548744e-02 5.61318621e-02 -5.03130741e-02 -1.06293850e-01
6.62736773e-01 1.02686323e-01 -3.17801982e-02 4.50393021e-01
-4.69316132e-02 5.58580756e-01 1.13536924e-01 1.10176086e-01
6.36485815e-01 6.44998476e-02 4.92385000e-01 -3.79613698e-01
6.77668095e-01 -7.17181206e-01 5.13701737e-01 8.26662719e-01
4.01410550e-01 2.94460207e-01 2.17277229e-01 -1.63895577e-01
-6.60125375e-01 -1.11350667e+00 -1.26967669e-01 1.60899377e+00
1.77249312e-01 -5.56373775e-01 -2.25905731e-01 -9.33712542e-01
1.95359513e-01 9.05972242e-01 -7.26588070e-01 -2.72257298e-01
-4.21854526e-01 -8.88534606e-01 -7.41106458e-03 7.86934316e-01
1.18714580e-02 -7.75753617e-01 -3.84464711e-02 1.67207956e-01
-1.54754305e-02 -6.67794108e-01 -5.65285265e-01 1.35817811e-01
-8.95104408e-01 -7.73061931e-01 -5.46766043e-01 -5.97423673e-01
6.44485533e-01 5.61998785e-01 1.21956766e+00 3.38328704e-02
1.11217164e-01 1.64507344e-01 -6.74009204e-01 -2.70004660e-01
8.02851841e-02 2.44298801e-01 1.74224406e-01 2.29107603e-01
2.36099109e-01 -6.70592546e-01 -8.63714933e-01 4.96800333e-01
-8.68348241e-01 -1.42327413e-01 6.76703811e-01 1.01353586e+00
6.50909305e-01 -1.79666147e-01 9.82576549e-01 -1.21388471e+00
6.05966806e-01 -8.77947211e-01 -5.53187549e-01 2.48281032e-01
-1.23718834e+00 1.72052085e-01 8.79233718e-01 -5.98558486e-01
-9.38503683e-01 -3.41780484e-01 -1.07855201e-01 -2.02448830e-01
2.70723492e-01 5.64122260e-01 -5.11185341e-02 2.07572550e-01
6.04728341e-01 2.47763276e-01 -4.84029427e-02 -8.07884991e-01
5.46710849e-01 6.62072420e-01 2.91960448e-01 -4.99869823e-01
5.93464077e-01 2.07252711e-01 -2.17656955e-01 -4.46227163e-01
-1.02585101e+00 -7.57152915e-01 -3.79533142e-01 1.86146468e-01
7.76944235e-02 -7.42010534e-01 -8.51356924e-01 3.04577220e-02
-4.21838611e-01 3.74843902e-03 -4.44967359e-01 6.46090090e-01
-2.91771740e-01 4.80010837e-01 -4.38975483e-01 -8.57208371e-01
-7.09944487e-01 -9.10851300e-01 7.02536166e-01 1.05445430e-01
-2.56157577e-01 -1.01079631e+00 -2.66995654e-02 4.17266160e-01
5.69640756e-01 -2.80940473e-01 1.01188338e+00 -9.25151646e-01
-1.23111233e-01 -1.70130596e-01 -1.61600754e-01 5.16372085e-01
2.54502982e-01 -4.35144007e-01 -8.38024080e-01 -5.45056701e-01
-3.55031490e-01 -2.41933540e-02 8.79644513e-01 2.96830267e-01
1.40146625e+00 -5.07082939e-01 -8.71639997e-02 4.66730982e-01
1.29072130e+00 1.91299900e-01 5.65662444e-01 2.70856798e-01
4.54750419e-01 5.52719295e-01 7.11731613e-01 6.90046132e-01
2.81322330e-01 1.01700294e+00 3.77617121e-01 -1.00917853e-01
1.32565081e-01 -3.37558836e-01 3.45334500e-01 7.82618165e-01
-1.16409473e-01 -3.93616140e-01 -2.78016418e-01 4.10791337e-01
-2.12459803e+00 -8.65657210e-01 -2.83631906e-02 2.59906983e+00
7.46070445e-01 -5.24557661e-03 4.70709503e-01 2.04312921e-01
4.68817443e-01 -3.86439078e-02 -6.85122490e-01 -2.74393618e-01
1.27129227e-01 3.93443376e-01 3.14550906e-01 1.70843571e-01
-1.16560507e+00 6.15812302e-01 5.69792318e+00 9.24881756e-01
-1.04723728e+00 -9.96338576e-02 4.94435668e-01 -5.03095090e-01
-4.45005864e-01 -3.87613982e-01 -8.28917265e-01 6.03670299e-01
8.91267598e-01 -2.23443657e-01 5.50397575e-01 9.15733933e-01
9.17575583e-02 3.61719310e-01 -1.25059271e+00 9.08442676e-01
1.70782000e-01 -1.29529381e+00 1.60061553e-01 3.14506769e-01
8.68388236e-01 -1.49635434e-01 4.43062305e-01 7.60017991e-01
1.01624727e-01 -8.72994900e-01 5.90810776e-01 4.24952060e-01
5.63693166e-01 -9.19505358e-01 9.17211413e-01 2.82761306e-01
-1.05646741e+00 -3.54912490e-01 -3.92628253e-01 -1.19070701e-01
1.22461449e-02 6.01311326e-01 -6.00575387e-01 8.31533432e-01
6.14213765e-01 8.51015568e-01 -5.69517851e-01 1.05756235e+00
2.90867165e-02 8.11619461e-01 -6.08608611e-02 -1.57492027e-01
1.88960806e-01 -3.53495657e-01 2.57125854e-01 1.27803755e+00
3.12044799e-01 -2.39513610e-02 1.84158757e-01 5.34835875e-01
-2.99059242e-01 6.31723225e-01 -8.08912069e-02 1.66026101e-01
5.30780196e-01 1.17816234e+00 -2.46596560e-01 -1.24058947e-01
-5.94040573e-01 7.03573406e-01 3.96594405e-01 5.48268296e-02
-8.15736055e-01 -7.19919726e-02 7.83329606e-01 3.57444912e-01
7.93609083e-01 2.39290684e-01 -2.74228841e-01 -1.00864065e+00
1.43997610e-01 -9.12272155e-01 4.85325962e-01 -1.82964310e-01
-1.77513182e+00 3.94600570e-01 -2.02388451e-01 -1.60094190e+00
-3.07236761e-01 -4.32458043e-01 -2.82499075e-01 7.18866408e-01
-1.58318174e+00 -1.09292185e+00 8.00897740e-03 3.76076609e-01
5.09184539e-01 -1.52415663e-01 8.59305799e-01 5.41131556e-01
-6.11667931e-01 1.24609447e+00 5.24183393e-01 -2.27913246e-01
6.99134707e-01 -1.19360518e+00 3.27711582e-01 5.95243752e-01
1.72377229e-01 1.03079283e+00 4.26055849e-01 -2.51701891e-01
-1.41653144e+00 -1.22661221e+00 6.70649827e-01 -3.94524008e-01
6.16159737e-01 -2.46823922e-01 -1.01708329e+00 4.44468558e-01
-3.11071664e-01 5.48654087e-02 1.13913214e+00 7.62963116e-01
-7.27519810e-01 -2.92211324e-01 -9.81584072e-01 6.38499677e-01
1.07568860e+00 -2.08149448e-01 -4.28110391e-01 1.82303041e-01
5.34774542e-01 -4.10133740e-03 -1.07616293e+00 5.02728641e-01
9.97515082e-01 -8.62465799e-01 1.24527895e+00 -8.33996117e-01
2.99052000e-01 -1.44381434e-01 -3.12662035e-01 -1.19509590e+00
-8.86382699e-01 -3.43783349e-01 -4.58640665e-01 1.28274250e+00
5.69491982e-01 -6.62021577e-01 5.96073091e-01 4.72488552e-01
-6.91340119e-02 -1.23219264e+00 -4.74772960e-01 -9.37616289e-01
4.46153171e-02 -3.87483895e-01 5.57285428e-01 9.31609571e-01
8.71448144e-02 6.31735325e-01 -6.11701548e-01 1.47142895e-02
3.20030093e-01 2.98456520e-01 6.21628940e-01 -1.59411442e+00
-8.14934373e-01 -6.72798932e-01 -9.01596472e-02 -1.39945436e+00
-4.03405540e-02 -1.08997703e+00 -2.32829332e-01 -1.44851100e+00
3.75760734e-01 -5.74794650e-01 -9.45238829e-01 4.46508408e-01
-4.24279302e-01 2.37908468e-01 2.47453988e-01 2.42665991e-01
-8.31066310e-01 7.03533292e-01 8.97368371e-01 1.09894782e-01
-5.36453664e-01 4.32907790e-01 -1.32096231e+00 5.66546381e-01
6.95540190e-01 -2.88626581e-01 -7.77198195e-01 -1.39115646e-01
3.77902895e-01 -2.88469017e-01 -2.67543159e-02 -4.99534070e-01
-9.61467996e-02 6.76849671e-03 3.74568641e-01 -3.34912926e-01
2.94918269e-01 -6.85589969e-01 4.82414290e-02 2.58662879e-01
-5.85007310e-01 -1.44774407e-01 1.23302795e-01 8.36182177e-01
1.01729900e-01 -1.15354694e-01 6.84522629e-01 2.54006624e-01
-4.80261803e-01 1.30261004e-01 8.37055743e-02 -5.97395748e-02
8.26893568e-01 -1.22658655e-01 -2.42510498e-01 -2.13452816e-01
-3.44433784e-01 1.21706955e-01 1.76283523e-01 7.32533038e-01
3.37968558e-01 -1.23404074e+00 -6.60686612e-01 2.51819164e-01
3.27218115e-01 -4.06896055e-01 3.98056269e-01 7.65325963e-01
1.82081938e-01 5.47800243e-01 -4.09223251e-02 -3.71738195e-01
-1.21702719e+00 6.38164461e-01 -6.85650110e-02 -7.18644142e-01
-4.42289531e-01 7.91000426e-01 4.59393740e-01 -5.18415749e-01
3.00315917e-01 -1.53785884e-01 -5.97935319e-01 1.50736108e-01
6.56009316e-01 5.55288672e-01 2.90602952e-01 -2.49866754e-01
-3.67940307e-01 2.18989745e-01 -6.19790852e-01 3.27562660e-01
1.61878431e+00 -7.93507174e-02 2.45799571e-01 3.10162872e-01
1.07257855e+00 2.05328226e-01 -1.20134878e+00 -5.39440513e-01
-1.26371592e-01 -4.15172458e-01 1.83576092e-01 -1.11150610e+00
-1.01684630e+00 5.40711641e-01 5.78784227e-01 2.88948268e-01
1.19040620e+00 -1.37109190e-01 6.92672133e-01 3.18094343e-01
3.83488744e-01 -8.17742050e-01 8.64645764e-02 3.65308344e-01
8.76942754e-01 -1.12568271e+00 2.54753381e-01 -4.37948406e-01
-7.59619296e-01 5.99822104e-01 4.92812186e-01 -9.30925459e-02
7.27763236e-01 -5.24505638e-02 -4.14980590e-01 1.02199852e-01
-9.20128882e-01 -7.89002180e-02 8.41880977e-01 2.23739117e-01
7.53302336e-01 1.65371284e-01 -6.65184021e-01 1.29291248e+00
-1.20643051e-02 -4.39546816e-02 2.68462952e-02 6.74315035e-01
-2.32361555e-01 -1.29695916e+00 1.52476087e-01 6.96063876e-01
-5.70655107e-01 -3.67623121e-01 -1.24329507e-01 6.81599319e-01
-1.12142079e-01 1.10414505e+00 5.99652156e-02 -5.03985047e-01
4.69156206e-01 -1.31092593e-01 2.51119822e-01 -6.44299090e-01
-8.92483532e-01 1.68483123e-01 -3.97034734e-02 -3.54849726e-01
-2.00090006e-01 -5.06978810e-01 -9.51355875e-01 -1.58908918e-01
-6.63453162e-01 3.49038213e-01 5.25101006e-01 8.61281037e-01
8.13425541e-01 5.65632463e-01 9.10226107e-01 -6.25477314e-01
-8.53031039e-01 -9.93143916e-01 -6.24461234e-01 5.28565705e-01
-1.19589074e-02 -8.40128481e-01 -2.40061074e-01 -2.61912167e-01] | [10.031098365783691, 5.555354118347168] |
75afb196-9f01-4b37-8557-81b2afa48507 | improving-non-native-word-level-pronunciation | 2203.01826 | null | https://arxiv.org/abs/2203.01826v1 | https://arxiv.org/pdf/2203.01826v1.pdf | Improving Non-native Word-level Pronunciation Scoring with Phone-level Mixup Data Augmentation and Multi-source Information | Deep learning-based pronunciation scoring models highly rely on the availability of the annotated non-native data, which is costly and has scalability issues. To deal with the data scarcity problem, data augmentation is commonly used for model pretraining. In this paper, we propose a phone-level mixup, a simple yet effective data augmentation method, to improve the performance of word-level pronunciation scoring. Specifically, given a phoneme sequence from lexicon, the artificial augmented word sample can be generated by randomly sampling from the corresponding phone-level features in training data, while the word score is the average of their GOP scores. Benefit from the arbitrary phone-level combination, the mixup is able to generate any word with various pronunciation scores. Moreover, we utilize multi-source information (e.g., MFCC and deep features) to further improve the scoring system performance. The experiments conducted on the Speechocean762 show that the proposed system outperforms the baseline by adding the mixup data for pretraining, with Pearson correlation coefficients (PCC) increasing from 0.567 to 0.61. The results also indicate that proposed method achieves similar performance by using 1/10 unlabeled data of baseline. In addition, the experimental results also demonstrate the efficiency of our proposed multi-source approach. | ['Zejun Ma', 'Xiaohai Tian', 'Wei Li', 'Kai Wang', 'Shaojun Gao', 'Kaiqi Fu'] | 2022-03-01 | null | null | null | null | ['word-level-pronunciation-scoring'] | ['speech'] | [ 6.60412312e-02 -2.22921386e-01 -3.98724601e-02 -3.00593913e-01
-1.07164431e+00 -3.81094635e-01 2.60130793e-01 -5.89917041e-02
-5.86725593e-01 7.72657812e-01 4.21415597e-01 -2.02415273e-01
3.41268212e-01 -4.51220572e-01 -3.74923766e-01 -8.06613088e-01
2.64231503e-01 1.94545150e-01 -1.12464003e-01 -2.46565625e-01
9.55735818e-02 -6.56033633e-03 -1.52648485e+00 2.31760904e-01
1.24929571e+00 1.01484990e+00 6.33626103e-01 6.51491404e-01
-2.99850494e-01 4.87778515e-01 -9.94003415e-01 -4.42543954e-01
1.33445501e-01 -5.38085341e-01 -1.73393339e-01 -1.11229546e-01
7.15266988e-02 -3.41268599e-01 -9.27434489e-02 1.21570039e+00
7.19334126e-01 3.06484044e-01 4.89340305e-01 -1.00006652e+00
-6.39565349e-01 7.18679309e-01 -2.51280606e-01 -5.43502916e-04
9.87183824e-02 6.75774217e-02 9.36837137e-01 -1.28215587e+00
-5.99588528e-02 1.14509320e+00 4.29991305e-01 6.30904913e-01
-7.08220541e-01 -9.85425353e-01 1.20805442e-01 3.39514881e-01
-1.39663196e+00 -5.49086273e-01 8.41535270e-01 -1.28406435e-01
8.32216799e-01 1.68367684e-01 4.09703225e-01 9.68396425e-01
-2.91190416e-01 8.26447546e-01 1.21272016e+00 -6.19746089e-01
5.64141460e-02 3.09747130e-01 8.10192898e-02 4.60757554e-01
9.52755380e-03 1.14098899e-01 -5.92970192e-01 2.61958293e-03
7.01426923e-01 -2.40331769e-01 -3.43356490e-01 2.19529152e-01
-1.16322911e+00 6.30851448e-01 2.79534429e-01 2.43021116e-01
-3.48789126e-01 -2.43527204e-01 4.03636813e-01 1.99610680e-01
3.36998314e-01 3.06081563e-01 -8.07429135e-01 -3.91638905e-01
-7.43852675e-01 -6.66625798e-02 3.44662160e-01 7.47514188e-01
6.58110380e-01 5.96469879e-01 -1.91590950e-01 1.40526628e+00
2.99290419e-01 6.63443446e-01 1.26682210e+00 -5.07682741e-01
8.09879482e-01 3.59680474e-01 -2.27033142e-02 -7.34861016e-01
-7.38701522e-02 -6.64809585e-01 -9.45010245e-01 -2.29822211e-02
1.91622689e-01 -3.77400666e-01 -1.05417693e+00 2.08369064e+00
9.11329016e-02 4.38143522e-01 4.66096997e-01 8.44421625e-01
8.51036489e-01 1.07136619e+00 2.65371323e-01 -5.04732132e-01
1.14914358e+00 -1.02748764e+00 -1.02190018e+00 -2.27937341e-01
6.89759851e-01 -9.06532645e-01 1.62115133e+00 5.61228633e-01
-8.15591931e-01 -1.01580095e+00 -1.11327362e+00 2.59834290e-01
-2.39450157e-01 6.79152608e-01 4.08880770e-01 7.79656887e-01
-7.18715489e-01 3.35000813e-01 -5.48163056e-01 1.17829740e-01
8.44021142e-03 2.96201855e-01 -3.74113739e-01 -3.33318636e-02
-1.45933378e+00 6.79758191e-01 8.06785643e-01 -6.52441084e-02
-7.15234160e-01 -5.77203751e-01 -6.94564879e-01 1.83156922e-01
2.77830940e-02 5.92768230e-02 1.32834792e+00 -9.27234471e-01
-1.64354599e+00 2.81661898e-01 -1.18432626e-01 -3.18145126e-01
8.44469741e-02 -2.42375597e-01 -7.76266992e-01 -1.33054167e-01
-1.51182100e-01 6.45708442e-01 4.60045040e-01 -9.99735892e-01
-7.29814708e-01 -8.02578479e-02 -3.09586525e-01 6.47493660e-01
-9.46654916e-01 -5.09740487e-02 -3.67195934e-01 -8.95956397e-01
7.12132454e-02 -7.74119198e-01 -3.91360149e-02 -6.63115144e-01
-4.24950063e-01 -3.10151577e-01 4.93461788e-01 -1.03302324e+00
1.49726033e+00 -2.39620614e+00 -1.72629625e-01 8.03731009e-02
-2.74856955e-01 6.55697107e-01 -2.41100341e-01 1.05403900e-01
6.64511845e-02 9.79714766e-02 -4.09930557e-01 -4.58801091e-01
-1.91340372e-01 1.28846675e-01 4.67297202e-03 -1.85560510e-01
2.63363719e-01 5.41230321e-01 -7.78015554e-01 -4.50882345e-01
2.89851993e-01 4.20711547e-01 -5.37673533e-01 4.69487637e-01
8.41207281e-02 4.07479525e-01 -2.11441275e-02 3.41434270e-01
6.71854079e-01 2.53162593e-01 4.02588807e-02 -3.99157144e-02
5.45322895e-02 7.23904967e-01 -1.37121594e+00 1.47629404e+00
-8.14270139e-01 3.61282706e-01 -1.77084178e-01 -8.36015522e-01
1.31862712e+00 6.23749316e-01 5.60219213e-02 -4.97739643e-01
1.82353139e-01 4.52845156e-01 5.87596357e-01 -2.78709888e-01
5.04677355e-01 -1.64526254e-01 -1.23014808e-01 2.21043587e-01
2.79099524e-01 1.38051743e-02 2.75464226e-02 -9.25519466e-02
7.41291225e-01 -1.35090619e-01 2.90945292e-01 -3.69584076e-02
7.01056600e-01 -3.12790304e-01 7.40219831e-01 3.72267276e-01
-2.05893993e-01 6.87040687e-01 -1.93151813e-02 1.22133970e-01
-9.85215366e-01 -1.04535687e+00 -1.42455310e-01 1.13676107e+00
-2.73947358e-01 -2.30900273e-01 -9.08989787e-01 -5.05722821e-01
-3.56658578e-01 6.93051457e-01 -2.06718802e-01 -3.92368793e-01
-6.36699796e-01 -8.41687977e-01 6.17272079e-01 6.33877873e-01
7.50628710e-01 -1.30186224e+00 1.63394064e-01 5.25185704e-01
-5.50505579e-01 -1.07834041e+00 -5.25862753e-01 1.02275185e-01
-6.72280669e-01 -5.38303435e-01 -8.26762676e-01 -1.03274286e+00
4.19746161e-01 1.19641267e-01 7.36954570e-01 8.59827772e-02
3.09422433e-01 -3.96550298e-01 -5.41944027e-01 -2.99324811e-01
-6.42439783e-01 2.87852705e-01 4.61940199e-01 1.56588525e-01
4.76712525e-01 -4.72159594e-01 -2.81636655e-01 3.91397700e-02
-5.71545720e-01 1.91843554e-01 8.84200633e-01 9.26651776e-01
5.77001512e-01 3.57060209e-02 1.25217485e+00 -6.49349689e-01
8.99787664e-01 -5.07810235e-01 -2.03881055e-01 3.29507962e-02
-5.72278798e-01 -9.14856642e-02 1.08909798e+00 -8.95961881e-01
-1.11254597e+00 4.00359221e-02 -5.88681102e-01 -3.08672339e-01
-2.71795183e-01 6.26817107e-01 -7.23983407e-01 4.07404393e-01
3.76430333e-01 5.30796230e-01 -1.93117291e-01 -7.90143132e-01
3.64709944e-01 1.19239342e+00 8.17523599e-01 -3.49004596e-01
5.62200785e-01 -4.87059742e-01 -6.54597521e-01 -7.78937101e-01
-6.17397308e-01 -2.94224858e-01 -2.79696703e-01 4.59118858e-02
5.61106622e-01 -9.44125235e-01 -3.25430274e-01 6.87415004e-01
-1.20497894e+00 1.67906452e-02 -5.20742461e-02 1.02998650e+00
-2.63934821e-01 1.71076804e-01 -6.32143378e-01 -1.08993924e+00
-4.30655509e-01 -1.13210249e+00 5.54531336e-01 3.52447033e-01
-5.88108003e-02 -6.81421757e-01 2.36454252e-02 2.35430986e-01
3.73433679e-01 -1.99106857e-01 9.27779675e-01 -1.08688295e+00
-1.02418527e-01 -2.48723656e-01 8.62405822e-02 1.07093751e+00
4.89600152e-01 -2.34342173e-01 -1.29198897e+00 -5.79553507e-02
3.57198939e-02 -2.65586108e-01 4.62697625e-01 1.42760798e-01
1.32509434e+00 -4.70294356e-01 2.43544385e-01 3.61555427e-01
1.07477129e+00 5.25296807e-01 4.44964528e-01 -9.61177200e-02
7.85217702e-01 4.52526689e-01 7.48155057e-01 2.82355875e-01
2.80269653e-01 5.74626207e-01 1.44418761e-01 -1.92629710e-01
-3.86146426e-01 -4.98211116e-01 5.56254685e-01 1.81869614e+00
1.14419267e-01 -2.68194437e-01 -7.17461765e-01 6.48105621e-01
-1.40746009e+00 -6.29834533e-01 -2.40607843e-01 2.47704148e+00
1.30756152e+00 3.09508175e-01 -5.33594899e-02 7.85442710e-01
9.32366312e-01 -8.78996924e-02 -4.81915593e-01 -4.95565653e-01
-2.44706750e-01 4.90323871e-01 1.25789076e-01 6.51019216e-01
-7.76200175e-01 1.13327384e+00 5.51707983e+00 1.25995600e+00
-1.05356991e+00 3.27238679e-01 5.33515632e-01 -3.89519297e-02
-2.04822600e-01 -3.70739758e-01 -9.09067690e-01 9.20411050e-01
1.19745433e+00 -1.56433046e-01 4.70404059e-01 7.88080871e-01
2.70738602e-01 2.47221753e-01 -6.56079888e-01 1.01524389e+00
7.47599602e-02 -7.10771918e-01 1.80373937e-01 -9.57480669e-02
7.30980515e-01 -1.25116542e-01 1.65657565e-01 7.00556159e-01
2.89198965e-01 -8.60334456e-01 4.87058699e-01 1.27668709e-01
9.37262833e-01 -1.05571115e+00 1.01920795e+00 4.98008370e-01
-1.30505812e+00 -3.79591696e-02 -4.06025380e-01 -2.25463152e-01
1.83344617e-01 6.52842700e-01 -1.12384903e+00 3.53456497e-01
3.28348666e-01 2.44199947e-01 -4.14709002e-01 9.82634008e-01
-4.79027182e-01 1.10813022e+00 -2.50713050e-01 -2.19597042e-01
5.05864136e-02 -7.25820512e-02 9.27186012e-02 1.31952178e+00
5.85635841e-01 2.38260459e-02 7.79803246e-02 2.64810324e-01
-3.04674178e-01 6.79402828e-01 -3.17908525e-01 1.10138562e-02
1.05419266e+00 1.13212693e+00 -6.31085932e-02 -6.20639265e-01
-3.75862271e-01 8.70544791e-01 3.25418919e-01 2.93307036e-01
-8.00732017e-01 -6.89004660e-01 7.91981876e-01 -2.66558856e-01
4.79821749e-02 -2.33293742e-01 -3.70213896e-01 -9.56897914e-01
1.39495492e-01 -9.49366450e-01 -2.93513294e-03 -7.40615249e-01
-1.10550761e+00 8.87802303e-01 -2.98062533e-01 -1.33270490e+00
-4.12111878e-01 -5.09261429e-01 -6.06574893e-01 1.23021889e+00
-1.53647947e+00 -7.57783234e-01 -2.82790393e-01 5.00685990e-01
8.07827473e-01 -4.20167387e-01 9.50979233e-01 5.59779704e-01
-6.12701535e-01 1.07759845e+00 2.90656164e-02 2.72237301e-01
6.81978941e-01 -1.18475592e+00 4.21208918e-01 1.00799334e+00
3.66193891e-01 5.71504533e-01 3.24650884e-01 -4.93893504e-01
-6.49602890e-01 -1.16971302e+00 1.08817124e+00 -6.82930499e-02
3.58108342e-01 -3.49203765e-01 -1.18962550e+00 2.95317471e-01
1.13204896e-01 -2.80251890e-01 9.13112223e-01 -3.95436808e-02
-6.60002753e-02 -2.69354641e-01 -1.05012369e+00 6.02496028e-01
7.31914401e-01 -5.33809543e-01 -7.12286115e-01 1.45850390e-01
1.07799935e+00 -2.11484894e-01 -7.44054139e-01 5.39309740e-01
3.58414620e-01 -7.06002057e-01 7.06420243e-01 -4.01206762e-01
2.34279931e-01 -2.75491983e-01 -4.00625080e-01 -1.72097838e+00
-1.98870093e-01 -2.28824049e-01 -1.47749454e-01 1.71148634e+00
6.33506000e-01 -6.33447289e-01 5.76190293e-01 2.10412428e-01
-4.76110458e-01 -6.56494915e-01 -9.07810986e-01 -8.59260678e-01
1.02291323e-01 -5.95227659e-01 9.36446011e-01 8.77034783e-01
-8.80938116e-03 3.70967299e-01 -6.22241974e-01 1.59588918e-01
2.18692258e-01 -5.24033070e-01 5.86443126e-01 -1.01073408e+00
-4.86013770e-01 -2.41190404e-01 -1.72940791e-01 -1.20429051e+00
7.30341896e-02 -8.65866303e-01 3.67865473e-01 -1.23986220e+00
-1.17953636e-01 -7.20388651e-01 -8.54341865e-01 5.10652900e-01
-7.15875685e-01 2.66989112e-01 2.71724105e-01 1.74349025e-01
8.21680482e-03 7.67914593e-01 1.25298929e+00 1.93994015e-01
-3.85481060e-01 1.40887871e-01 -6.35046721e-01 6.86829388e-01
1.33738840e+00 -4.80216593e-01 -5.42922497e-01 -4.57142413e-01
-3.55435550e-01 5.36922328e-02 -2.69370884e-01 -1.24068224e+00
-7.13728145e-02 -4.98760939e-02 2.54929543e-01 -5.29457867e-01
6.49628460e-01 -5.23082972e-01 -2.52932906e-01 4.60985452e-01
-3.62771362e-01 1.72022000e-01 2.02411428e-01 1.96170017e-01
-4.57926244e-01 -4.73364562e-01 7.68467605e-01 -2.92394701e-02
-5.77927768e-01 1.40727580e-01 -2.83681631e-01 9.58119780e-02
5.28554738e-01 -5.35781309e-02 -1.06453404e-01 -2.72946209e-01
-5.67538142e-01 -1.53880507e-01 8.15314278e-02 5.62417030e-01
5.92833340e-01 -1.84574425e+00 -9.83399808e-01 5.06183743e-01
1.40505493e-01 -1.08757339e-01 1.26942486e-01 4.02643085e-01
-1.30953506e-01 2.18156949e-01 -8.62483159e-02 -3.34928185e-01
-1.06356013e+00 1.69151291e-01 8.82657021e-02 -2.19759867e-01
-3.37292999e-02 9.73692954e-01 2.69857943e-01 -7.58193374e-01
3.10840994e-01 -4.71960962e-01 -4.21508670e-01 -7.23684430e-02
6.30913198e-01 2.88913339e-01 3.55779767e-01 -6.14576280e-01
-2.99533248e-01 3.64293128e-01 -8.19742307e-02 -5.22664189e-01
8.20969522e-01 -9.65296570e-03 3.33605886e-01 5.70545197e-01
1.16520298e+00 3.08588326e-01 -9.91004646e-01 -2.47973680e-01
-2.21878827e-01 -3.21281821e-01 6.22966029e-02 -1.06353951e+00
-1.05396223e+00 1.11850429e+00 6.92920983e-01 -5.41170537e-02
1.23556101e+00 -4.03597504e-01 1.10413349e+00 1.97346821e-01
1.64985076e-01 -1.07982910e+00 1.00661255e-01 5.66579044e-01
6.96605146e-01 -1.25110698e+00 -6.69946432e-01 -2.64452547e-01
-9.13127244e-01 6.79972351e-01 9.85292494e-01 1.58715248e-03
4.19988036e-01 1.96103513e-01 4.26918685e-01 7.01571465e-01
-5.83025157e-01 -2.11384282e-01 1.59882218e-01 7.06595242e-01
6.09111786e-01 3.78224313e-01 -5.15744209e-01 1.06717157e+00
-7.05630660e-01 -3.74359578e-01 4.92231429e-01 4.28269446e-01
-7.62149453e-01 -1.36306190e+00 -4.15542990e-01 4.96229649e-01
-3.82369608e-01 -6.34820223e-01 -1.26026779e-01 3.97955477e-01
3.24250698e-01 1.30381906e+00 1.37280989e-02 -8.23414266e-01
2.94736683e-01 4.07516092e-01 4.72290069e-02 -8.00283492e-01
-5.01710236e-01 6.59170076e-02 -5.45902224e-03 3.31820101e-02
3.82147520e-03 -5.30019462e-01 -1.38027656e+00 2.11617611e-02
-5.38864493e-01 4.25148576e-01 8.54756236e-01 8.47408772e-01
1.83645904e-01 6.18329406e-01 9.15475965e-01 -4.49567765e-01
-6.34314656e-01 -1.65317845e+00 -4.51218367e-01 3.41989398e-01
2.58016884e-01 -6.19592607e-01 -5.20625174e-01 4.62530740e-02] | [14.53555679321289, 6.732218265533447] |
1ce2130d-952d-495f-94bd-657c105a482d | a-maximum-entropy-feature-descriptor-for-age | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Gong_A_Maximum_Entropy_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Gong_A_Maximum_Entropy_2015_CVPR_paper.pdf | A Maximum Entropy Feature Descriptor for Age Invariant Face Recognition | In this paper, we propose a new approach to overcome the representation and matching problems in age invariant face recognition. First, a new maximum entropy feature descriptor (MEFD) is developed that encodes the microstructure of facial images into a set of discrete codes in terms of maximum entropy. By densely sampling the encoded face image, sufficient discriminatory and expressive information can be extracted for further analysis. A new matching method is also developed, called identity factor analysis (IFA), to estimate the probability that two faces have the same underlying identity. The effectiveness of the framework is confirmed by extensive experimentation on two face aging datasets, MORPH (the largest public-domain face aging dataset) and FGNET. We also conduct experiments on the famous LFW dataset to demonstrate the excellent generalizability of our new approach. | ['Xuelong. Li', 'DaCheng Tao', 'Zhifeng Li', 'Jianzhuang Liu', 'Dihong Gong'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['age-invariant-face-recognition'] | ['computer-vision'] | [ 2.73508728e-01 8.27143192e-02 -3.63398254e-01 -7.24678278e-01
-4.03885186e-01 -2.05213502e-01 6.55913234e-01 -4.49567974e-01
-1.88897923e-01 7.09046364e-01 1.97682142e-01 2.86663920e-01
-8.24091807e-02 -7.19114661e-01 -2.99706936e-01 -6.18138015e-01
-4.08570617e-01 4.42613587e-02 -4.11494732e-01 1.22590736e-01
1.20147444e-01 7.31168151e-01 -1.97470021e+00 5.44587597e-02
7.69821763e-01 1.46274149e+00 -1.76020786e-01 5.31251654e-02
1.79794356e-01 4.14943874e-01 -2.05619648e-01 -6.33854747e-01
4.65354443e-01 -4.09919083e-01 -8.10553730e-01 1.94264591e-01
8.21858048e-01 -5.58854580e-01 -6.20932281e-01 1.23481965e+00
6.57859981e-01 4.69366834e-02 8.86212528e-01 -1.39343667e+00
-8.48084807e-01 1.44605860e-01 -5.15321314e-01 9.07272696e-02
5.60218871e-01 -3.01397860e-01 9.91553187e-01 -1.18940759e+00
6.59725070e-01 1.62600863e+00 7.47424722e-01 9.37401593e-01
-1.09089887e+00 -9.41613078e-01 -2.13914201e-01 3.29194754e-01
-1.58729494e+00 -8.98088455e-01 8.55577469e-01 -2.57289618e-01
4.26975071e-01 1.01965211e-01 5.91725290e-01 1.02702737e+00
1.13462046e-01 6.08918309e-01 1.30858922e+00 -3.50885153e-01
2.49877006e-01 -1.35927677e-01 -3.67200859e-02 1.11637807e+00
2.37187177e-01 2.59535015e-01 -8.97572517e-01 -4.49482143e-01
6.69786453e-01 -1.28364325e-01 -1.78733498e-01 -2.20405579e-01
-6.61810458e-01 8.09188664e-01 1.22613862e-01 3.40030462e-01
-3.59072775e-01 -9.76530015e-02 2.03004554e-01 3.78973186e-01
6.28669322e-01 -6.97614765e-03 4.71153930e-02 1.54187664e-01
-1.04426205e+00 1.12299368e-01 6.66175663e-01 6.01594150e-01
8.48837852e-01 -1.23729393e-01 -1.38717562e-01 9.43170667e-01
4.69653547e-01 6.17642939e-01 5.89244068e-01 -1.21261942e+00
-9.79102701e-02 4.84766245e-01 -3.48568082e-01 -1.12778342e+00
-1.97843239e-01 2.88344137e-02 -9.44337845e-01 2.75723994e-01
2.58391798e-01 3.29006165e-01 -8.21788788e-01 2.20054603e+00
2.16293380e-01 3.79741997e-01 -1.58452496e-01 4.24567878e-01
6.93173707e-01 8.36952105e-02 1.84094459e-01 -2.78561205e-01
1.61008906e+00 -2.52070814e-01 -8.37470651e-01 7.22025055e-03
1.86498255e-01 -4.17026281e-01 4.38197136e-01 9.89867300e-02
-1.03135264e+00 -6.22221112e-01 -1.13304162e+00 1.06667839e-01
-3.46728086e-01 2.73922235e-01 9.77411866e-01 9.93487179e-01
-1.16806495e+00 7.07836509e-01 -5.23837566e-01 -4.31728274e-01
9.76254761e-01 7.10876107e-01 -1.04527879e+00 -2.31045678e-01
-1.24831235e+00 6.57272696e-01 1.75402686e-01 -1.11139119e-01
-6.95181847e-01 -6.44265890e-01 -1.03071439e+00 -1.19859077e-01
-3.07130039e-01 -5.61247647e-01 6.33760571e-01 -1.15216947e+00
-1.33479106e+00 1.39530897e+00 -4.63090986e-01 -1.91464975e-01
1.30217984e-01 6.85615391e-02 -6.30851746e-01 5.13414264e-01
8.64408836e-02 8.63042474e-01 1.34683836e+00 -9.82055545e-01
-2.04761654e-01 -9.51498210e-01 -1.14854120e-01 -2.42361382e-01
-8.91346157e-01 2.61422604e-01 -1.38512373e-01 -7.66354680e-01
1.42118543e-01 -9.65627313e-01 2.75613070e-01 4.95558500e-01
2.41217557e-02 -3.96720737e-01 6.99277997e-01 -9.41857696e-01
1.23811853e+00 -2.40317392e+00 2.12755486e-01 4.98086154e-01
5.04707098e-01 2.16099232e-01 -3.46131861e-01 -5.31107597e-02
-4.26420599e-01 2.32924218e-03 -3.14355701e-01 -4.76647794e-01
-5.83930090e-02 2.13413432e-01 -1.69629619e-01 6.46629810e-01
3.93935412e-01 8.26037586e-01 -5.48145056e-01 -6.86824679e-01
-2.33628660e-01 7.23314881e-01 -5.71695685e-01 7.80508965e-02
4.50892538e-01 3.31707180e-01 -4.12354946e-01 9.93750095e-01
1.09164751e+00 2.30019744e-02 1.86987728e-01 -3.54264021e-01
3.09476405e-01 -3.34511280e-01 -1.02033067e+00 1.60884976e+00
-1.76777214e-01 3.69771153e-01 -5.75641990e-02 -9.31531072e-01
1.02432060e+00 3.45540702e-01 7.57831454e-01 -5.36016464e-01
2.86996901e-01 1.34080976e-01 -2.02401713e-01 -3.34698975e-01
1.86254680e-01 -5.75404167e-02 1.57115698e-01 5.21134198e-01
5.50269723e-01 4.38487053e-01 1.05583943e-01 -5.38712274e-03
1.04634893e+00 -1.32364944e-01 4.16209787e-01 -5.59979498e-01
9.79800224e-01 -8.99563789e-01 7.38849044e-01 3.57661933e-01
-6.69442534e-01 4.26598638e-01 3.78641784e-01 -4.45443243e-01
-9.24906909e-01 -1.26280415e+00 -6.22650623e-01 8.46417546e-01
-1.89030603e-01 -6.25723362e-01 -1.02085173e+00 -9.03104007e-01
3.65580916e-01 2.82382853e-02 -1.06342983e+00 -4.24046576e-01
-3.61873239e-01 -7.54652500e-01 7.16912150e-01 5.11257946e-01
8.40500653e-01 -8.74461651e-01 -1.04754552e-01 -2.94157505e-01
-1.93113729e-01 -1.08207440e+00 -7.07489967e-01 -4.99910653e-01
-7.96181262e-01 -1.06242967e+00 -7.82884598e-01 -8.19563866e-01
9.14655507e-01 -3.16969529e-02 8.85161400e-01 2.17075154e-01
-6.36742771e-01 6.34751916e-01 -1.55723214e-01 -1.09139428e-01
-3.58185381e-01 -1.23446360e-01 6.65971160e-01 5.77250779e-01
5.39416373e-01 -7.11613297e-01 -6.47025585e-01 2.05532789e-01
-6.87194705e-01 -4.67034906e-01 5.21557450e-01 8.90647531e-01
4.06901747e-01 -4.70274948e-02 8.81296158e-01 -4.67146486e-01
4.87320960e-01 -3.48544776e-01 -1.87572658e-01 3.68391573e-01
-6.16208315e-01 1.96743637e-01 2.21592098e-01 -4.04630750e-01
-9.62752700e-01 3.22142214e-01 -1.52977765e-01 -4.16363895e-01
-3.81692834e-02 4.34126593e-02 -4.42306370e-01 -6.05064094e-01
3.63034159e-01 2.55057484e-01 5.93435287e-01 -4.52720165e-01
2.94463813e-01 7.49774158e-01 7.62091041e-01 -6.72765732e-01
9.20359731e-01 6.73678041e-01 2.64042526e-01 -8.21730018e-01
-5.32086551e-01 -1.89623356e-01 -8.38868737e-01 -3.41618240e-01
7.41357207e-01 -9.50359344e-01 -9.89449143e-01 7.94862926e-01
-8.06786716e-01 4.31742340e-01 -9.67387930e-02 4.41682160e-01
-7.20715344e-01 6.22273743e-01 -5.41422784e-01 -7.89368808e-01
-4.45416152e-01 -7.89346099e-01 1.00052798e+00 2.87018210e-01
-2.50873297e-01 -8.13385963e-01 4.83029336e-03 2.52603710e-01
2.53024757e-01 2.29257792e-01 8.99577618e-01 -4.83130753e-01
-1.20228492e-01 -3.62527311e-01 -3.30759734e-01 5.58363140e-01
4.99842077e-01 -7.13025182e-02 -1.14896429e+00 -4.89311844e-01
1.19796423e-02 -4.93380308e-01 1.06190908e+00 1.37089199e-04
1.32767797e+00 -8.89396518e-02 -3.34522247e-01 6.57655120e-01
1.10494995e+00 -3.76515612e-02 9.68147099e-01 -1.51694685e-01
2.89511502e-01 5.90407491e-01 3.37501138e-01 5.86155534e-01
2.81337708e-01 7.19380915e-01 6.74566850e-02 1.66844174e-01
-3.62966865e-01 -2.07262278e-01 5.02741814e-01 6.23379886e-01
-5.54747462e-01 3.23582202e-01 -3.92784536e-01 1.70522898e-01
-1.38901126e+00 -1.31736720e+00 6.65817857e-01 2.14817309e+00
8.97288918e-01 -3.93513381e-01 1.74721941e-01 2.37607449e-01
9.21375632e-01 1.19692110e-01 -4.00285482e-01 -8.04430321e-02
-3.20852697e-01 7.94645011e-01 4.75987382e-02 2.91633993e-01
-1.22585809e+00 6.67352259e-01 7.22347355e+00 9.19702351e-01
-8.45278382e-01 -3.16394456e-02 7.58857548e-01 2.89938599e-01
-7.93712027e-03 -3.40231240e-01 -8.09361458e-01 5.54513752e-01
8.21944833e-01 -4.37724650e-01 6.22619271e-01 9.22895133e-01
-3.11747760e-01 1.35178998e-01 -1.10695624e+00 1.30727577e+00
4.73921955e-01 -1.00424325e+00 1.57483727e-01 3.64725918e-01
4.06363159e-01 -6.36310697e-01 5.52512109e-01 2.04147398e-02
-2.06680298e-01 -1.07012963e+00 4.50770676e-01 7.04197288e-01
1.35338986e+00 -8.26712787e-01 2.97651857e-01 -3.56768310e-01
-1.53030038e+00 -2.45096371e-01 -4.74868715e-01 9.00341123e-02
-3.08328241e-01 4.69099909e-01 -3.84685278e-01 5.78892410e-01
6.28833175e-01 8.99948955e-01 -8.40898216e-01 6.50601208e-01
-6.26722630e-03 2.39325002e-01 -1.32280394e-01 5.30921161e-01
-3.52184296e-01 -1.29964024e-01 2.84184992e-01 7.12868512e-01
5.15680671e-01 1.74470976e-01 -2.10419938e-01 7.99242020e-01
-4.60166126e-01 9.53634232e-02 -7.32355297e-01 -1.54512525e-01
5.24104476e-01 1.51109183e+00 -4.98909473e-01 -8.75793174e-02
-4.93068218e-01 1.18230546e+00 4.06268924e-01 7.45913908e-02
-5.05576313e-01 -2.49168724e-01 1.03422248e+00 -1.23195753e-01
2.23957777e-01 -1.19772531e-01 1.85962111e-01 -1.21114743e+00
1.25529175e-03 -9.22105432e-01 4.25993830e-01 -2.56386846e-01
-1.67746806e+00 5.72909951e-01 -6.95559978e-02 -1.08392429e+00
-1.97569624e-01 -8.26159298e-01 -4.69471306e-01 8.42976451e-01
-1.27289832e+00 -1.21290851e+00 -2.64641345e-01 8.46104920e-01
7.36744097e-03 -6.06541455e-01 1.19550669e+00 6.13422632e-01
-6.40469253e-01 1.10774064e+00 -1.90839112e-01 3.10579807e-01
7.80564904e-01 -9.09687638e-01 2.39976794e-01 4.92238462e-01
3.50764394e-02 7.61961937e-01 2.24071220e-01 -5.59726834e-01
-1.39653456e+00 -1.00737858e+00 9.91058707e-01 -3.47229958e-01
3.22385699e-01 -4.46454525e-01 -8.86592507e-01 4.75426912e-01
-9.93540734e-02 2.29878470e-01 1.06165147e+00 1.01396684e-02
-8.26201916e-01 -2.24463940e-01 -1.62337112e+00 2.23795339e-01
1.47871280e+00 -8.80178452e-01 -5.00715017e-01 9.13926028e-03
1.70855775e-01 2.13928908e-01 -1.30658126e+00 7.47775376e-01
1.01356590e+00 -9.20686126e-01 1.15838480e+00 -6.48041248e-01
1.68530703e-01 3.30358706e-02 -4.16494370e-01 -9.31826591e-01
-5.21749854e-01 -5.16451180e-01 -3.73666823e-01 1.46162391e+00
-1.02035500e-01 -7.38789678e-01 7.01332211e-01 5.31077862e-01
4.93012816e-01 -7.23321557e-01 -1.31699824e+00 -1.06506133e+00
-1.20452546e-01 -2.06427723e-02 8.76651943e-01 1.04386353e+00
-1.27483383e-02 -1.07849337e-01 -4.36964661e-01 -5.30604720e-02
9.78106320e-01 -9.57761556e-02 2.22869441e-01 -1.64325166e+00
1.71663731e-01 -2.80472457e-01 -1.14373660e+00 -3.92588049e-01
1.02316153e+00 -1.18239522e+00 -3.52955908e-01 -6.40479207e-01
5.28824925e-01 -1.82913303e-01 -5.68832397e-01 4.96608555e-01
-1.14321865e-01 6.64791107e-01 3.52702737e-02 7.05744997e-02
-2.94298470e-01 7.64657617e-01 9.29231644e-01 -1.47773817e-01
3.19031358e-01 -8.38584080e-02 -5.61482191e-01 6.17856383e-01
7.20811009e-01 -2.34090000e-01 -3.02661419e-01 2.36455768e-01
-2.92560875e-01 -2.84760177e-01 3.47415596e-01 -1.17831612e+00
-1.09228469e-01 2.35489607e-01 8.30970764e-01 -7.65150413e-02
5.47409475e-01 -5.38816273e-01 5.30720130e-02 5.62913537e-01
-1.85425147e-01 -4.76230308e-02 -2.60692537e-02 3.94323915e-01
-2.41622239e-01 -9.10797492e-02 1.08766150e+00 1.45876363e-01
-7.85514832e-01 9.24728870e-01 -9.14100260e-02 -3.52765433e-02
1.00330710e+00 -7.56782666e-02 -1.23813510e-01 -2.44099617e-01
-8.14397454e-01 -2.04087928e-01 5.30819595e-01 5.28825104e-01
8.40682030e-01 -2.09474993e+00 -7.69856513e-01 8.02577198e-01
2.72935927e-01 -1.24625075e+00 2.60010540e-01 5.93445957e-01
2.56820326e-03 1.51065305e-01 -6.69434309e-01 -1.82930276e-01
-1.47406602e+00 5.72624266e-01 3.31417024e-01 5.05641289e-02
-3.17686230e-01 6.78239524e-01 1.38762400e-01 -5.25713973e-02
-2.79857460e-02 3.73113185e-01 -2.56494462e-01 2.96087176e-01
1.03660023e+00 3.93738955e-01 -1.07675411e-01 -1.27261662e+00
-5.64194560e-01 7.88597167e-01 -9.44509059e-02 -8.67056027e-02
1.14370036e+00 -1.12796225e-01 -5.81167638e-01 -5.25209829e-02
1.46467936e+00 -1.53372884e-01 -1.03674555e+00 -2.98110664e-01
-6.68403413e-03 -8.13487828e-01 -1.07197262e-01 -1.80694103e-01
-1.27375627e+00 6.53822780e-01 1.07741702e+00 1.30745051e-02
1.43691981e+00 1.18739784e-01 6.38579130e-01 1.44443244e-01
4.61031199e-01 -1.09589016e+00 1.92144141e-01 1.38650984e-01
9.55244541e-01 -1.09253812e+00 -4.49912101e-02 -4.96759951e-01
-1.71955407e-01 1.22564864e+00 4.74405617e-01 1.13365026e-02
1.00726581e+00 1.96030326e-02 -3.38637352e-01 -4.47897539e-02
-4.18731689e-01 -3.30043614e-01 5.05206466e-01 8.44988346e-01
3.27843845e-01 -9.30442363e-02 -4.33651507e-01 7.81040132e-01
-1.73956946e-01 1.68333828e-01 -9.58046466e-02 8.06459069e-01
-3.70070279e-01 -1.32252431e+00 -2.31613681e-01 5.68908215e-01
-5.87994277e-01 1.00669287e-01 -4.75550950e-01 4.65560496e-01
2.49112159e-01 6.49781108e-01 1.75058752e-01 -5.91224909e-01
-1.34378597e-01 4.57235724e-01 1.12486088e+00 -2.62301266e-01
8.32327008e-02 -5.98358572e-01 -2.03861281e-01 -8.08700383e-01
-6.91938281e-01 -8.33560944e-01 -1.07675576e+00 -3.82423341e-01
2.12629028e-02 -1.57543465e-01 4.39804256e-01 7.08185077e-01
4.44197983e-01 -9.94242728e-02 1.11709142e+00 -7.61781812e-01
-4.84386921e-01 -7.19757915e-01 -9.54900086e-01 7.04116642e-01
1.97983757e-01 -1.09811819e+00 -3.90246391e-01 1.89931646e-01] | [13.31588077545166, 0.6800175309181213] |
8aef60d9-09e7-4b44-b29b-0eb04c928326 | behancepr-a-punctuation-restoration-dataset | null | null | https://aclanthology.org/2022.findings-naacl.149 | https://aclanthology.org/2022.findings-naacl.149.pdf | BehancePR: A Punctuation Restoration Dataset for Livestreaming Video Transcript | Given the increasing number of livestreaming videos, automatic speech recognition and post-processing for livestreaming video transcripts are crucial for efficient data management as well as knowledge mining. A key step in this process is punctuation restoration which restores fundamental text structures such as phrase and sentence boundaries from the video transcripts. This work presents a new human-annotated corpus, called BehancePR, for punctuation restoration in livestreaming video transcripts. Our experiments on BehancePR demonstrate the challenges of punctuation restoration for this domain. Furthermore, we show that popular natural language processing toolkits like Stanford Stanza, Spacy, and Trankit underperform on detecting sentence boundary on non-punctuated transcripts of livestreaming videos. The dataset is publicly accessible at http://github.com/nlp-uoregon/behancepr. | ['Thien Nguyen', 'Franck Dernoncourt', 'Amir Pouran Ben Veyseh', 'Viet Lai'] | null | null | null | null | findings-naacl-2022-7 | ['punctuation-restoration'] | ['natural-language-processing'] | [ 4.34381336e-01 -2.48917595e-01 -1.62611231e-01 -3.15149784e-01
-1.11054182e+00 -7.07294941e-01 3.05790126e-01 -1.01657212e-01
-3.84326100e-01 7.19620705e-01 9.06048477e-01 -3.11629891e-01
3.66608202e-01 -5.30396178e-02 -6.18018270e-01 -3.74974698e-01
-2.16676772e-01 -2.02850074e-01 2.21931830e-01 -1.89284891e-01
3.61501634e-01 8.43355209e-02 -1.59147489e+00 7.25883067e-01
5.94139040e-01 5.24598837e-01 2.60089666e-01 1.23377478e+00
-2.54172117e-01 8.50628376e-01 -6.61220968e-01 -3.98174912e-01
9.73087326e-02 -5.80614746e-01 -1.00120807e+00 3.98022532e-01
4.99868006e-01 -1.48533940e-01 -7.86277413e-01 9.05076325e-01
2.46823236e-01 2.96669185e-01 2.09471643e-01 -1.16321516e+00
-2.87838072e-01 7.70938575e-01 -2.14755863e-01 8.85622025e-01
9.26840842e-01 9.14022923e-02 9.96938229e-01 -9.94729757e-01
8.27681541e-01 9.75578904e-01 6.69397473e-01 4.75797355e-01
-4.13987160e-01 -5.08022547e-01 -2.34274045e-01 4.58839059e-01
-1.62441349e+00 -1.04058123e+00 2.52590030e-01 -4.39948857e-01
1.26041985e+00 4.75092381e-01 3.00097018e-01 1.15057266e+00
3.06797802e-01 9.78704154e-01 6.67600095e-01 -3.57430786e-01
1.61805842e-02 -5.55010021e-01 1.95096865e-01 6.92130864e-01
-4.91339922e-01 -5.18690109e-01 -1.04524493e+00 -1.42715812e-01
5.57936609e-01 -3.29298079e-02 -5.40603757e-01 4.84107703e-01
-1.19926417e+00 2.29433939e-01 -6.30760789e-01 4.30462986e-01
-3.44171494e-01 -7.10128918e-02 8.39738965e-01 5.18718958e-01
5.26511550e-01 1.04713507e-01 -4.53096926e-01 -1.01324260e+00
-1.10143781e+00 -8.32590684e-02 6.69770718e-01 1.12849677e+00
6.21911645e-01 -3.99132162e-01 -1.64499432e-01 8.36990178e-01
-1.28166214e-01 3.86529863e-01 6.66826367e-01 -1.10502172e+00
7.14208424e-01 1.96519524e-01 8.96495059e-02 -7.92723298e-01
-6.22107349e-02 3.77024591e-01 -5.81106484e-01 -6.42668426e-01
3.37580264e-01 -3.88781935e-01 -9.07274425e-01 1.34210229e+00
3.56759548e-01 6.85590208e-01 2.01617539e-01 7.34226882e-01
9.01019573e-01 1.00747788e+00 1.01095922e-01 -3.75048101e-01
1.47873855e+00 -8.77734661e-01 -1.03342223e+00 -1.27467170e-01
6.78226769e-01 -1.08432639e+00 1.17886996e+00 3.42022121e-01
-1.05817103e+00 -3.26777846e-01 -6.72264516e-01 -3.78471732e-01
-1.57625601e-01 -1.25038058e-01 1.03409819e-01 2.99420863e-01
-1.13071442e+00 5.13024509e-01 -8.28922689e-01 -7.06417024e-01
2.03949079e-01 -2.86119282e-02 -5.99205613e-01 3.97560224e-02
-1.33288062e+00 2.39946559e-01 4.26863402e-01 -2.13057026e-02
-6.10878527e-01 -5.79642773e-01 -1.05249798e+00 -1.92528889e-02
6.91926122e-01 -3.81100439e-02 1.74023020e+00 -1.08744800e+00
-1.43934619e+00 8.76907110e-01 -5.56457102e-01 -6.19811237e-01
1.81249112e-01 -4.03252691e-01 -7.42546439e-01 6.82724357e-01
1.43502504e-01 3.29318911e-01 9.48371112e-01 -4.07125533e-01
-8.70573044e-01 9.97234881e-03 -2.53064930e-01 3.80303264e-01
-2.67021209e-01 7.42756307e-01 -8.06176484e-01 -8.56022477e-01
-1.51034877e-01 -6.53865099e-01 9.82832611e-02 -5.64835846e-01
-3.15755635e-01 -2.27103755e-01 8.86712253e-01 -1.34497786e+00
1.59932208e+00 -2.44715643e+00 -1.55060545e-01 -7.49501735e-02
2.56356108e-03 3.36645544e-01 -3.82981122e-01 8.08702588e-01
-3.65094334e-01 3.90205085e-01 -1.52729973e-01 -4.46112454e-01
-1.57836497e-01 1.50924131e-01 -4.17779386e-01 3.79144430e-01
1.49063617e-01 7.61314809e-01 -9.11045134e-01 -6.11298740e-01
-7.31774559e-03 3.61971945e-01 -4.55303848e-01 2.33834416e-01
-7.33021274e-02 2.02984706e-01 -2.34095082e-01 6.48598135e-01
2.63500929e-01 7.95296356e-02 9.69523489e-02 2.78457433e-01
-3.84479254e-01 5.25349915e-01 -8.11730921e-01 1.83140612e+00
-1.01393819e-01 8.82212996e-01 2.15767592e-01 -5.73116243e-01
5.35256565e-01 7.01978862e-01 4.96714026e-01 -4.61126566e-01
1.28387123e-01 -1.83066949e-01 -7.22694457e-01 -1.02553833e+00
1.10889375e+00 8.32524896e-02 -5.16795516e-01 3.46044332e-01
1.74464390e-01 1.31345376e-01 6.74143076e-01 6.66641295e-01
1.60169721e+00 -4.46660537e-03 4.72211689e-01 1.75635487e-01
5.28146684e-01 1.40803814e-01 8.63533080e-01 4.66724008e-01
-5.39085746e-01 8.81052732e-01 6.46609128e-01 -7.98526108e-02
-1.10311699e+00 -9.53672826e-01 2.95190513e-01 1.15879667e+00
-3.35587412e-01 -9.68759298e-01 -1.31044292e+00 -6.67331457e-01
-4.44006115e-01 5.20560920e-01 -2.34048083e-01 1.54680490e-01
-6.56975210e-01 -2.59780020e-01 9.71443534e-01 2.92080015e-01
5.56987524e-01 -1.19610476e+00 -1.32476002e-01 1.29863083e-01
-9.66465890e-01 -1.66118050e+00 -1.15771759e+00 -2.85672247e-01
-5.59624314e-01 -1.14176834e+00 -4.09057438e-01 -7.73971021e-01
2.50792146e-01 4.91119117e-01 7.85967410e-01 1.33757621e-01
-2.94851094e-01 5.51647246e-01 -8.69608283e-01 -8.65328871e-03
-6.79541826e-01 1.67683344e-02 2.14256287e-01 -8.59067068e-02
7.47143567e-01 -4.64179277e-01 -1.08135112e-01 3.19695920e-01
-1.08474612e+00 -5.47001734e-02 1.68047205e-01 2.98890889e-01
5.16715467e-01 1.46768346e-01 5.22254586e-01 -5.99904358e-01
6.54023230e-01 -5.53450763e-01 -2.24513739e-01 1.18257664e-01
2.44557872e-01 -5.36388457e-01 6.69182062e-01 -6.76336586e-02
-1.20768893e+00 -1.65623706e-03 -4.02681291e-01 -5.16716182e-01
-6.40897930e-01 4.25790340e-01 -3.87609750e-02 5.07503569e-01
4.14556384e-01 4.05786723e-01 -8.19382966e-02 -4.67224658e-01
4.18887660e-02 1.28066099e+00 1.00731874e+00 -4.42655869e-02
5.40797353e-01 3.67122024e-01 -7.91510046e-01 -1.63780761e+00
-8.37131798e-01 -1.15057421e+00 -6.67310596e-01 -3.53411078e-01
1.03526139e+00 -1.04082549e+00 -3.81858468e-01 6.41518593e-01
-1.04974401e+00 -4.55706358e-01 -2.20646083e-01 2.25510195e-01
-6.34604454e-01 1.13873875e+00 -1.02811944e+00 -6.31259680e-01
-5.27068496e-01 -5.28774202e-01 9.43034530e-01 2.58923799e-01
-5.09683251e-01 -6.79499030e-01 1.08929709e-01 1.09057760e+00
-3.18380237e-01 -1.15145341e-01 2.41428420e-01 -5.86094439e-01
-2.79626578e-01 -2.35651746e-01 1.67692751e-01 4.78174597e-01
2.40709320e-01 2.90264517e-01 -6.78187609e-01 -4.67859134e-02
-1.69053137e-01 -2.57378399e-01 7.62224913e-01 1.00981966e-01
1.00081205e+00 -4.84179914e-01 1.37234747e-01 3.54544818e-01
7.18319178e-01 1.93390548e-01 1.09163165e+00 1.78398609e-01
4.55101192e-01 4.54402536e-01 1.07215667e+00 7.83396900e-01
4.31934565e-01 2.17651561e-01 -1.26404405e-01 2.96913594e-01
-1.16135016e-01 -3.86459976e-01 1.00733531e+00 1.22351193e+00
9.22685936e-02 -5.52614689e-01 -1.09565473e+00 8.49692762e-01
-1.87371540e+00 -1.42857444e+00 -2.18826428e-01 1.77122450e+00
9.28013623e-01 -6.88132942e-02 1.60816506e-01 2.11020440e-01
1.12969506e+00 2.27425471e-01 -9.94955897e-02 -4.81052309e-01
-1.57209769e-01 8.28581527e-02 3.60561639e-01 4.09148991e-01
-1.41279411e+00 1.42032337e+00 5.99122906e+00 9.38186824e-01
-8.50542247e-01 1.85866803e-01 4.44201976e-01 -1.95407882e-01
3.44235450e-01 -1.44176334e-01 -7.76879311e-01 5.87918818e-01
1.40834796e+00 -3.00448269e-01 4.91718322e-01 5.44048071e-01
9.47599113e-01 -3.01787376e-01 -7.46508777e-01 1.21481812e+00
4.15673792e-01 -1.36525702e+00 -2.65741050e-01 -2.34105304e-01
5.38473248e-01 3.11886042e-01 -2.97593683e-01 1.53652310e-01
-8.96068849e-03 -5.80273151e-01 7.42653131e-01 4.68220323e-01
8.83228004e-01 -7.71937251e-01 6.83996081e-01 1.69481680e-01
-1.15253103e+00 1.16319753e-01 -1.51239082e-01 -1.32108510e-01
5.11047363e-01 5.61762989e-01 -1.09984863e+00 3.60565573e-01
9.13683474e-01 8.57088566e-01 -3.63276064e-01 9.14129317e-01
-4.51775700e-01 9.07650292e-01 -2.96576887e-01 3.18543434e-01
1.35008007e-01 -1.20432012e-01 6.86704993e-01 1.70464385e+00
3.79496515e-01 6.52135313e-01 1.34629920e-01 3.23912427e-02
-2.92236149e-01 1.34093091e-01 -4.09361392e-01 -6.33674979e-01
4.23081756e-01 1.19775772e+00 -7.54822493e-01 -5.49976349e-01
-3.32841098e-01 1.30972624e+00 -4.56011631e-02 3.98178369e-01
-8.76671553e-01 -5.75702906e-01 1.20754886e+00 1.99958071e-01
4.16610718e-01 -5.66430390e-01 3.65134358e-01 -1.51182258e+00
8.87225568e-02 -1.10636485e+00 7.01539636e-01 -9.49278057e-01
-9.74988222e-01 4.39311981e-01 -1.33221522e-01 -1.07845092e+00
-2.17495695e-01 -9.76621360e-02 -6.80696249e-01 1.01072080e-01
-1.21186090e+00 -6.12605453e-01 -2.44472399e-01 8.08630645e-01
1.14185774e+00 9.56176445e-02 4.71086740e-01 4.73625451e-01
-9.84380066e-01 3.81701678e-01 1.48306703e-02 4.99066621e-01
9.06609178e-01 -9.05055523e-01 7.42002308e-01 1.38760066e+00
1.75016671e-01 3.63759696e-01 8.43636453e-01 -9.64704692e-01
-1.27643073e+00 -1.17804003e+00 1.33225536e+00 -3.01045656e-01
9.94967580e-01 -5.75123847e-01 -9.30738509e-01 8.10870051e-01
5.06127954e-01 -5.29252529e-01 1.03612804e+00 -3.39218885e-01
-4.46729735e-02 3.23495090e-01 -8.96531701e-01 8.68189335e-01
1.31462681e+00 -8.61191154e-01 -1.04103971e+00 4.32209760e-01
7.01731086e-01 -4.25294846e-01 -8.58125448e-01 -1.59298722e-02
1.49593592e-01 -8.19449484e-01 5.11582315e-01 -3.85392189e-01
5.55519521e-01 -1.87519804e-01 -6.80067688e-02 -1.05765688e+00
1.20911129e-01 -1.47273517e+00 7.40898103e-02 1.57521462e+00
1.83082148e-01 -2.24314764e-01 7.68981576e-01 4.86931294e-01
-4.99418855e-01 -7.87790567e-02 -9.99195695e-01 -7.03904927e-01
-4.53877449e-01 -8.00378382e-01 3.18813682e-01 1.01191521e+00
6.37342632e-01 3.02205622e-01 -4.83568639e-01 1.95007503e-01
3.70170206e-01 -4.09832299e-01 8.29222679e-01 -4.46932912e-01
1.91898406e-01 2.03435019e-01 -4.99711305e-01 -1.11900413e+00
4.21207458e-01 -5.84260106e-01 3.60822618e-01 -1.37771046e+00
1.29580200e-01 2.43288353e-01 6.78851828e-02 7.01086700e-01
2.28654891e-02 4.28997606e-01 1.48234442e-01 1.30222246e-01
-1.16642463e+00 4.63331640e-01 7.83714712e-01 2.26604283e-01
-3.48384678e-01 -2.80509412e-01 -2.83158869e-01 8.81356835e-01
1.26745129e+00 -4.07194167e-01 -1.76677331e-01 -2.36851960e-01
-3.83063033e-02 2.21365020e-01 1.87734365e-02 -9.01420593e-01
2.53166914e-01 -2.25245729e-01 -1.59287617e-01 -7.56631494e-01
2.73564041e-01 -2.59682238e-01 -5.58846146e-02 -6.76728562e-02
-1.14121757e-01 2.76123703e-01 2.75442153e-01 4.31507677e-01
-3.87139767e-01 -1.69232607e-01 4.44785982e-01 -1.38671279e-01
-1.18637538e+00 1.71788245e-01 -1.34483826e+00 4.95141834e-01
1.00894320e+00 -5.06349683e-01 -3.71087223e-01 -5.89636803e-01
-9.28890526e-01 4.11626518e-01 6.24159396e-01 7.52534688e-01
8.04351091e-01 -8.05291116e-01 -6.80631459e-01 5.28864749e-02
-6.70751035e-02 -2.91692555e-01 6.36032283e-01 9.26875174e-01
-7.44646609e-01 2.85128087e-01 1.37730278e-02 -2.65780628e-01
-2.09147739e+00 3.60020846e-01 4.30494547e-05 7.83202276e-02
-7.15954721e-01 7.21688747e-01 -8.45505595e-02 -5.13713770e-02
2.00720072e-01 -5.68481982e-01 -1.51706442e-01 4.46682423e-02
1.06765747e+00 4.07389075e-01 6.08051606e-02 -1.03080082e+00
-3.52016300e-01 2.29659811e-01 -1.93866163e-01 -1.04610913e-01
1.18459737e+00 -7.39298761e-01 -2.46619120e-01 2.67532617e-01
1.12000883e+00 1.08683355e-01 -1.16139746e+00 -8.11329391e-03
3.74748498e-01 -4.61318970e-01 -9.89207551e-02 -3.65381569e-01
-5.89280307e-01 3.63894135e-01 -6.41097426e-02 9.48262587e-02
1.20794594e+00 6.36960417e-02 1.34386885e+00 5.98763943e-01
1.77729547e-01 -1.40541494e+00 9.25347582e-02 9.64108109e-01
8.29573691e-01 -8.44094098e-01 -3.07523936e-01 -7.12802410e-01
-8.67325366e-01 9.29666996e-01 2.54027367e-01 1.70140520e-01
3.80677760e-01 4.73627061e-01 3.08396250e-01 1.05581954e-01
-8.94110560e-01 -1.72643885e-01 -2.61836469e-01 3.79282027e-01
3.31509948e-01 -2.12099448e-01 -3.30891520e-01 4.85583723e-01
-4.41427916e-01 2.73838252e-01 1.14828646e+00 1.26354063e+00
-4.65796798e-01 -9.69259620e-01 -6.66763365e-01 1.55304849e-01
-1.05721104e+00 -3.75918716e-01 -6.72868907e-01 3.95184219e-01
-1.52724743e-01 1.35678411e+00 3.47386748e-01 -3.53251457e-01
1.94384918e-01 3.59779239e-01 4.25884090e-02 -8.35843682e-01
-7.64747381e-01 3.16838145e-01 6.20208383e-01 -8.24165702e-01
-4.04073030e-01 -1.06803620e+00 -1.72009623e+00 -5.37360430e-01
1.38228104e-01 5.36836922e-01 3.67575198e-01 8.23425829e-01
6.79565191e-01 1.52041271e-01 4.06486064e-01 -5.38287163e-01
-9.20586567e-03 -7.75785506e-01 -5.82870066e-01 4.46055353e-01
3.00956309e-01 1.32627025e-01 -5.44935405e-01 6.73812330e-01] | [10.457326889038086, 0.7553386688232422] |
426cec60-0795-4be0-9d46-2d72f8969c2d | improving-human-sperm-head-morphology | 2202.07191 | null | https://arxiv.org/abs/2202.07191v3 | https://arxiv.org/pdf/2202.07191v3.pdf | Improving Human Sperm Head Morphology Classification with Unsupervised Anatomical Feature Distillation | With rising male infertility, sperm head morphology classification becomes critical for accurate and timely clinical diagnosis. Recent deep learning (DL) morphology analysis methods achieve promising benchmark results, but leave performance and robustness on the table by relying on limited and possibly noisy class labels. To address this, we introduce a new DL training framework that leverages anatomical and image priors from human sperm microscopy crops to extract useful features without additional labeling cost. Our core idea is to distill sperm head information with reliably-generated pseudo-masks and unsupervised spatial prediction tasks. The predicted foreground masks from this distillation step are then leveraged to regularize and reduce image and label noise in the tuning stage. We evaluate our new approach on two public sperm datasets and achieve state-of-the-art performances (e.g. 65.9% SCIAN accuracy and 96.5% HuSHeM accuracy). | ['Danny Z. Chen', 'Yunxia Cao', 'Yiru Zhou', 'Xiaomin Zha', 'Jingjing Zhang', 'Yejia Zhang'] | 2022-02-15 | null | null | null | null | ['sperm-morphology-classification', 'morphology-classification'] | ['computer-vision', 'computer-vision'] | [ 2.67870992e-01 4.09502029e-01 8.97000358e-02 -6.53762102e-01
-9.19445395e-01 -8.64269733e-01 5.33049762e-01 5.49364269e-01
-5.21210313e-01 8.72144103e-01 -9.79957879e-02 -1.73923865e-01
9.13214833e-02 -7.32069910e-01 -6.07615650e-01 -9.21093285e-01
6.00509420e-02 7.87299395e-01 1.81342512e-01 4.76638705e-01
4.38620001e-01 7.49037683e-01 -1.22110963e+00 2.11075827e-01
1.40819359e+00 8.02467227e-01 1.76427081e-01 8.67848039e-01
-3.04012626e-01 2.09347770e-01 -5.57702661e-01 -8.72793078e-01
1.09698936e-01 -3.43419999e-01 -4.65145737e-01 -6.75976351e-02
5.41164458e-01 -9.62781999e-03 2.61215329e-01 9.61786449e-01
8.59326184e-01 -1.95769534e-01 8.25538576e-01 -5.69467902e-01
-4.90329504e-01 4.63546157e-01 -8.30000997e-01 3.26136500e-01
-2.67992526e-01 4.22946721e-01 7.03429461e-01 -6.98206007e-01
8.23610485e-01 8.88624787e-01 7.89920688e-01 6.12994254e-01
-1.54829192e+00 -5.90591371e-01 -3.40714276e-01 -1.39988095e-01
-1.13231540e+00 -4.31952566e-01 3.56312603e-01 -4.02467102e-01
4.53744799e-01 2.82384247e-01 5.28258264e-01 7.17647254e-01
4.35826659e-01 8.82652998e-01 1.05515134e+00 -9.93672162e-02
2.43372947e-01 2.53501651e-03 -1.97542891e-01 6.35989606e-01
6.18658483e-01 -2.58117467e-01 -7.05993116e-01 2.47870326e-01
8.80011022e-01 -1.65637195e-01 4.93062921e-02 -3.23569566e-01
-7.72459447e-01 6.42431796e-01 6.10705078e-01 1.56604424e-01
-2.11778626e-01 -6.20015748e-02 2.47064888e-01 -5.45271695e-01
5.33287466e-01 7.99026310e-01 -2.96595037e-01 -2.12517664e-01
-1.34609914e+00 2.58846451e-02 3.68865222e-01 5.77701747e-01
7.20751941e-01 -1.89199179e-01 -4.08438295e-01 7.40170598e-01
1.07968561e-02 7.76650965e-01 8.42001662e-02 -1.04298544e+00
1.74556985e-01 6.61344290e-01 -1.33396044e-01 -8.60228777e-01
-8.22895229e-01 -1.03998935e+00 -6.41438186e-01 -9.73391905e-03
6.15084529e-01 -1.86431184e-02 -1.40948880e+00 1.16620255e+00
5.14357924e-01 3.79574448e-01 -4.33924109e-01 7.88693190e-01
7.84743845e-01 4.83248055e-01 5.74570820e-02 1.77440390e-01
1.12100708e+00 -7.11550117e-01 -6.36004865e-01 -3.26452732e-01
7.59805083e-01 -3.96418244e-01 7.79548347e-01 4.56839442e-01
-1.22810447e+00 -3.56632978e-01 -8.69545817e-01 -4.64275479e-01
-3.65719169e-01 5.56677759e-01 5.65416038e-01 6.34739578e-01
-1.07661939e+00 7.24820614e-01 -1.50315106e+00 -3.52834105e-01
1.09218693e+00 9.33743834e-01 -4.27309155e-01 1.79063156e-02
-1.77099928e-01 8.22179914e-01 4.12407294e-02 4.54710752e-01
-7.52508044e-01 -1.08937907e+00 -1.06013954e+00 2.57884655e-02
-2.92550903e-02 -6.81494236e-01 8.81852984e-01 -3.50743473e-01
-1.47257006e+00 1.23318160e+00 -3.79231513e-01 -5.36658168e-01
4.73995835e-01 -2.34869227e-01 1.58941388e-01 2.73656517e-01
2.23753721e-01 1.02696955e+00 3.00453812e-01 -1.41905379e+00
-7.63652921e-01 -6.68553889e-01 -5.71773469e-01 1.26940385e-02
-2.27376446e-01 -3.48830193e-01 -1.04022384e+00 -4.06182855e-01
7.07063300e-04 -8.53783906e-01 -2.77722329e-01 1.59173697e-01
-5.31091809e-01 6.80506229e-02 5.50941885e-01 -8.63226354e-01
9.01180208e-01 -1.98655999e+00 3.08480561e-01 1.22565888e-01
4.09710467e-01 5.36980510e-01 -1.74014613e-01 -3.14080834e-01
4.61254716e-01 9.23518911e-02 -2.89996356e-01 -8.91675770e-01
-1.54360622e-01 3.31505358e-01 -1.61256418e-02 9.23276246e-01
4.73121613e-01 1.25072110e+00 -9.29441273e-01 -7.45968223e-01
5.52410901e-01 6.32967591e-01 -6.77692413e-01 2.12400019e-01
-5.12333550e-02 9.72247124e-01 -2.29959607e-01 1.02689886e+00
8.23140979e-01 -4.79010016e-01 2.19985530e-01 -3.34812254e-02
-3.87064144e-02 -1.91659350e-02 -6.60780072e-01 1.99292648e+00
-2.84482092e-01 6.33391201e-01 2.57003754e-01 -6.58902824e-01
6.74614191e-01 -4.68778282e-01 2.86350310e-01 -6.50279045e-01
4.23108459e-01 4.32213634e-01 9.63196754e-02 -4.56050217e-01
2.39648297e-01 -1.22764930e-01 5.54782748e-02 -1.65673524e-01
3.32455516e-01 -1.68560222e-01 1.29150048e-01 -4.11705934e-02
1.09200025e+00 4.17885303e-01 -1.91967115e-01 -5.26582778e-01
4.54924256e-01 1.49677362e-04 9.13075089e-01 6.28601670e-01
-1.82156295e-01 1.15226948e+00 7.20930040e-01 -3.74257416e-01
-9.55398977e-01 -1.21764243e+00 -2.77438849e-01 8.04163337e-01
2.41464347e-01 -4.53479029e-02 -8.36114168e-01 -8.55641186e-01
2.83657819e-01 4.68324304e-01 -1.08665907e+00 4.22087833e-02
-7.96806216e-01 -1.01832819e+00 7.15515137e-01 7.82163143e-01
9.54070687e-02 -8.06771040e-01 -4.07726049e-01 1.84536278e-02
3.25210512e-01 -9.71854329e-01 2.13983492e-03 2.15599447e-01
-7.39824176e-01 -6.63856626e-01 -9.81406450e-01 -6.78493202e-01
9.58818495e-01 -3.83813232e-01 1.03111994e+00 2.17540219e-01
-6.66229844e-01 -4.36878532e-01 -1.86623391e-02 -3.53861153e-01
1.04483657e-01 4.23098207e-01 -8.92866552e-02 6.21091984e-02
5.87159619e-02 -3.14490676e-01 -1.09491158e+00 -1.74304977e-01
-8.49316895e-01 6.69622719e-02 9.77898479e-01 8.96008492e-01
8.48708689e-01 -4.17460561e-01 3.71596128e-01 -1.20178866e+00
3.57954763e-02 -4.33478236e-01 -8.30893278e-01 1.13765791e-01
-4.97698814e-01 -3.21838409e-02 5.91726482e-01 -1.49454087e-01
-1.18772721e+00 1.58700541e-01 -3.08576852e-01 -3.60187113e-01
-1.68433130e-01 2.95981944e-01 -3.35000515e-01 -2.38101780e-01
3.28030229e-01 2.60816902e-01 3.94047014e-02 -6.46031201e-01
3.22019011e-01 3.95323068e-01 1.02097988e+00 -5.45017719e-01
7.29229569e-01 7.78001845e-01 3.77895892e-01 -4.63781267e-01
-7.35980392e-01 -4.07189459e-01 -8.84823918e-01 -1.64394602e-02
1.03496432e+00 -6.23373270e-01 -7.66997516e-01 4.41378802e-01
-6.63168848e-01 -5.93027711e-01 -4.88715582e-02 1.96772322e-01
-1.89755708e-01 2.90439367e-01 -9.69260871e-01 -7.40580916e-01
-3.49925518e-01 -1.18327665e+00 1.72953558e+00 8.00170124e-01
-2.12718844e-02 -1.09385109e+00 -1.85563583e-02 5.56648731e-01
4.65622842e-01 3.09352577e-01 7.81523466e-01 -7.65578389e-01
-5.07245541e-01 -1.68502897e-01 -5.14799714e-01 -1.70405939e-01
5.54330200e-02 -2.80208085e-02 -1.18583786e+00 -2.50528127e-01
-5.56007206e-01 -2.23506823e-01 1.08336771e+00 7.87276983e-01
1.39570630e+00 2.34858170e-01 -7.18894184e-01 1.11787784e+00
1.28602314e+00 -6.09797202e-02 4.68421817e-01 2.60623515e-01
6.77181125e-01 8.07650447e-01 7.36851335e-01 3.25178713e-01
3.55687827e-01 5.13842702e-01 3.83054346e-01 -4.32948411e-01
-4.47330743e-01 -5.74464165e-02 -1.95456594e-01 2.51984984e-01
1.13151208e-01 -3.79371077e-01 -1.34449697e+00 8.40539336e-01
-1.64574337e+00 -3.59374285e-01 -1.78692922e-01 1.87110543e+00
9.54910457e-01 1.80505931e-01 -1.13902941e-01 -3.78956348e-01
6.14942372e-01 -3.31039071e-01 -6.37585461e-01 -1.19498678e-01
-9.21496078e-02 6.55339658e-01 6.18245184e-01 5.47780216e-01
-1.10235524e+00 1.28371334e+00 5.99082327e+00 9.61833000e-01
-1.31119287e+00 -1.39877737e-01 1.19180524e+00 -4.00308818e-01
-1.67978257e-01 -2.73787558e-01 -9.90254641e-01 6.46827161e-01
8.07529509e-01 3.65373641e-01 -2.30423938e-02 2.47313201e-01
1.26161784e-01 -4.03634101e-01 -9.77427840e-01 8.86117816e-01
4.31972109e-02 -1.73979819e+00 -2.01247498e-01 3.53380322e-01
6.25046551e-01 -1.60381962e-02 3.42084736e-01 1.47187620e-01
9.62347016e-02 -1.47193384e+00 1.18843518e-01 5.71445465e-01
1.08047378e+00 -7.04737842e-01 1.24153066e+00 -5.43079339e-03
-8.87111843e-01 2.75756419e-01 -1.92162558e-01 2.23506823e-01
1.77568376e-01 6.57598794e-01 -1.16668594e+00 4.27224666e-01
6.50052249e-01 6.92099333e-01 -1.01729870e+00 1.54791152e+00
-1.32017642e-01 6.81329906e-01 -4.73942906e-01 1.65338397e-01
8.20971653e-02 -1.59007877e-01 1.45434618e-01 1.49587405e+00
3.14578801e-01 -1.30927473e-01 -4.29266274e-01 1.12017357e+00
-1.24454014e-01 -9.87778753e-02 -3.79393339e-01 3.60917076e-02
5.47621548e-01 1.56345880e+00 -1.37158000e+00 9.99982283e-02
-1.97104350e-01 9.93568957e-01 5.05847335e-01 1.15986399e-01
-8.41328323e-01 -2.84073919e-01 5.75889945e-01 1.11939967e-01
2.63740122e-01 1.00865856e-01 -9.52983677e-01 -8.58424306e-01
-1.46435246e-01 -1.41602665e-01 3.08249623e-01 -3.30668896e-01
-1.19968045e+00 8.35688561e-02 -5.33362150e-01 -8.21913540e-01
1.32951096e-01 -8.73577237e-01 -6.25514567e-01 7.51273990e-01
-1.73555756e+00 -1.53120601e+00 -1.42600387e-01 -2.30386272e-01
2.77556419e-01 1.06181562e-01 6.75799429e-01 3.25333208e-01
-7.85688996e-01 6.58352971e-01 2.43379071e-01 2.84197152e-01
8.80788445e-01 -1.85369074e+00 2.18740776e-01 7.50263095e-01
-2.86388695e-01 7.11624205e-01 6.14970446e-01 -7.71937549e-01
-1.64894307e+00 -1.30264080e+00 7.83002794e-01 -8.87316585e-01
5.05924225e-01 -5.61283529e-01 -6.98423147e-01 5.77072859e-01
4.17939909e-02 4.89795655e-01 9.42342818e-01 1.67017728e-02
1.96015090e-01 -1.86869770e-01 -1.36594391e+00 4.85548735e-01
7.10240781e-01 6.73585199e-03 -2.65087694e-01 -2.88547035e-02
1.02607027e-01 -8.38809788e-01 -8.59593451e-01 7.62443066e-01
5.98360121e-01 -7.65832961e-01 8.68075430e-01 -2.59731054e-01
5.79856992e-01 -4.51521277e-01 3.35226446e-01 -1.10227406e+00
-1.86429963e-01 -4.58438754e-01 -1.85397081e-02 1.46370149e+00
5.81823468e-01 -2.96170730e-02 1.46429825e+00 6.79799974e-01
-4.40384269e-01 -1.00528371e+00 -1.02040660e+00 -2.95400858e-01
3.85822803e-01 -5.47561683e-02 3.64327222e-01 6.86338723e-01
-2.20930457e-01 2.09919021e-01 -3.31231579e-02 3.09216082e-01
6.87933683e-01 3.14206839e-01 8.28102231e-01 -1.18575943e+00
-4.97353822e-03 -5.06426334e-01 -5.04250109e-01 -8.39260101e-01
1.65266916e-01 -1.02808261e+00 1.79807216e-01 -1.59042788e+00
4.82276350e-01 -5.60139000e-01 -2.16062903e-01 4.73674566e-01
-5.75230122e-01 8.53622377e-01 7.80933797e-02 -6.56751171e-02
-9.86137092e-01 4.08684134e-01 1.25047135e+00 -2.71921214e-02
-1.68883383e-01 -2.09784210e-01 -7.09956229e-01 6.19010925e-01
5.21698892e-01 -2.44910851e-01 -2.65382621e-02 -7.86257446e-01
-3.37760821e-02 -2.10267797e-01 2.00191215e-01 -9.43801522e-01
4.09066916e-01 3.74598317e-02 1.06793630e+00 -7.07861662e-01
5.23045957e-01 -4.77812588e-01 -1.57291114e-01 4.86984760e-01
-1.32917911e-01 -3.83280098e-01 4.13788974e-01 4.92035300e-01
1.03395343e-01 1.54187858e-01 8.78018677e-01 1.87016338e-01
-3.41018647e-01 1.46576792e-01 -7.51343975e-03 -1.27116248e-01
9.23085809e-01 -1.87124953e-01 -5.28277695e-01 1.57993101e-02
-7.31685340e-01 6.85560524e-01 7.78707385e-01 9.40922871e-02
4.65588629e-01 -8.17818165e-01 -7.31798053e-01 3.20931941e-01
-4.55072410e-02 6.42873466e-01 1.93305641e-01 1.15502942e+00
-9.94369745e-01 6.14615977e-01 -9.37464982e-02 -1.05915940e+00
-9.93144214e-01 3.53509076e-02 3.59536797e-01 -3.35997254e-01
-5.52305937e-01 1.34440017e+00 3.64552855e-01 -6.59196615e-01
2.54878581e-01 -2.26487741e-01 -2.68548667e-01 1.62714384e-02
3.06048602e-01 2.25170910e-01 2.81173199e-01 -6.45897150e-01
-5.02774239e-01 6.43778324e-01 -1.21038936e-01 1.09824091e-01
1.60244215e+00 2.57555731e-02 -1.18830018e-01 4.75778095e-02
9.58346605e-01 4.88097727e-01 -1.49964821e+00 1.51881173e-01
2.61700183e-01 -5.60905218e-01 2.98793782e-02 -1.12239909e+00
-1.22769356e+00 1.12782133e+00 6.57970726e-01 -3.13390255e-01
9.15334582e-01 9.32990015e-02 8.75755787e-01 -1.25139952e-01
1.48721740e-01 -7.43260264e-01 -2.37531871e-01 2.40612209e-01
4.95615825e-02 -1.29816687e+00 1.73919629e-02 -4.53561813e-01
-4.72110033e-01 8.99915218e-01 8.52075338e-01 -5.13335839e-02
2.79055268e-01 7.54675627e-01 7.94584677e-02 -2.18895972e-01
-4.43939984e-01 -2.93362677e-01 3.32075953e-01 6.39697134e-01
5.51273763e-01 -7.20262453e-02 -3.66666794e-01 9.60461259e-01
-1.43762410e-01 -2.01530531e-02 2.03756168e-01 7.91988373e-01
-3.48797977e-01 -1.07908058e+00 -9.09689441e-02 6.01215422e-01
-8.87995839e-01 1.14603676e-01 -4.58036810e-01 4.30299968e-01
3.99124026e-01 5.06332636e-01 3.81405205e-01 -1.54419467e-01
-1.52832165e-01 -2.99831718e-01 5.70625424e-01 -9.86664534e-01
-5.88398755e-01 3.49472404e-01 -1.58939809e-01 -6.30665123e-01
-5.21342568e-02 -6.85245335e-01 -1.67579162e+00 -1.20058306e-01
-3.25950176e-01 2.32121229e-01 6.22962534e-01 9.25177455e-01
6.16646290e-01 4.54653680e-01 1.26113817e-01 -1.16601205e+00
1.15017921e-01 -8.57723236e-01 -5.36393464e-01 2.33944044e-01
4.92140859e-01 -4.94097352e-01 -1.03932098e-01 1.26480028e-01] | [14.762602806091309, -3.1106832027435303] |
b56c3141-6eee-4273-bd82-93fdb1514619 | one-ring-to-bring-them-all-towards-open-set | 2206.03600 | null | https://arxiv.org/abs/2206.03600v2 | https://arxiv.org/pdf/2206.03600v2.pdf | OneRing: A Simple Method for Source-free Open-partial Domain Adaptation | In this paper, we investigate Source-free Open-partial Domain Adaptation (SF-OPDA), which addresses the situation where there exist both domain and category shifts between source and target domains. Under the SF-OPDA setting, which aims to address data privacy concerns, the model cannot access source data anymore during target adaptation. We propose a novel training scheme to learn a (n+1)-way classifier to predict the n source classes and the unknown class, where samples of only known source categories are available for training. Furthermore, for target adaptation, we simply adopt a weighted entropy minimization to adapt the source pretrained model to the unlabeled target domain without source data. In experiments, we show our simple method surpasses current OPDA approaches which demand source data during adaptation. When augmented with a closed-set domain adaptation approach during target adaptation, our source-free method further outperforms the current state-of-the-art OPDA method by 2.5%, 7.2% and 13% on Office-31, Office-Home and VisDA respectively. | ['Joost Van de Weijer', 'Shangling Jui', 'Kai Wang', 'Yaxing Wang', 'Shiqi Yang'] | 2022-06-07 | null | null | null | null | ['universal-domain-adaptation', 'partial-domain-adaptation', 'open-set-learning'] | ['computer-vision', 'methodology', 'miscellaneous'] | [ 3.60053420e-01 2.68974036e-01 -5.91798186e-01 -5.94908953e-01
-9.96852100e-01 -6.76927567e-01 4.41805065e-01 1.05410457e-01
-5.95264256e-01 1.24934947e+00 1.42874476e-02 -2.44460940e-01
1.31961927e-01 -7.03265190e-01 -5.70956767e-01 -5.90672016e-01
3.21380138e-01 6.60579205e-01 2.94965565e-01 6.67189956e-02
-6.68773204e-02 2.00709507e-01 -1.18670428e+00 1.07908905e-01
9.91079390e-01 1.20678151e+00 -3.63050431e-01 3.00572097e-01
-1.63019016e-01 2.61061192e-01 -6.43656433e-01 -6.61455393e-01
7.49193609e-01 -3.04342538e-01 -8.43337774e-01 1.32242486e-01
4.27451432e-01 -4.92875338e-01 -1.55118302e-01 9.73887384e-01
6.03879273e-01 3.77866060e-01 8.06316793e-01 -1.40940845e+00
-9.47026670e-01 1.50392607e-01 -4.72591579e-01 3.76904532e-02
3.35397303e-01 -3.82554412e-01 5.10323226e-01 -6.90954328e-01
7.06508756e-01 1.02214277e+00 4.04882967e-01 1.00015390e+00
-1.36960876e+00 -1.02234280e+00 2.26273343e-01 4.67922306e-03
-1.46147275e+00 -6.99329853e-01 5.66887856e-01 -3.37833792e-01
5.64916730e-01 3.38312648e-02 -1.21377409e-01 1.32916093e+00
-1.90455809e-01 5.72433770e-01 1.20656419e+00 -5.22067904e-01
6.21145010e-01 7.65764475e-01 3.61680031e-01 1.14620835e-01
4.22731996e-01 2.41653338e-01 -3.18696529e-01 -8.54356050e-01
5.55654705e-01 8.59104693e-02 -2.53639638e-01 -1.03280687e+00
-1.01291537e+00 7.32087314e-01 1.08713098e-01 -1.39479682e-01
-1.91711962e-01 -7.98294008e-01 3.78729790e-01 6.01026714e-01
5.44254124e-01 1.11671411e-01 -8.37979674e-01 2.16461197e-01
-7.16715097e-01 1.96445167e-01 1.15604496e+00 1.48968732e+00
7.48096406e-01 -2.56319672e-01 -1.82410687e-01 1.05474353e+00
1.69732437e-01 7.49714196e-01 7.23573744e-01 -8.18894506e-01
5.98067999e-01 4.80259657e-01 4.60920721e-01 -4.58203465e-01
8.90666321e-02 -3.65221858e-01 -7.99512565e-01 8.58198553e-02
7.62798369e-01 -2.99279422e-01 -1.06395721e+00 1.72096705e+00
6.87786758e-01 2.26835832e-01 6.27766669e-01 6.27316058e-01
5.10515511e-01 3.61480445e-01 9.75237340e-02 -4.67129081e-01
1.18312442e+00 -8.92169297e-01 -7.75250733e-01 -2.33259603e-01
4.91531938e-01 -4.62650239e-01 9.82580543e-01 3.90214682e-01
-7.46488750e-01 -3.33101094e-01 -1.00380254e+00 -3.35062444e-02
-7.39224494e-01 -5.76776545e-03 2.49649867e-01 9.81439054e-01
-4.53780442e-01 2.23020181e-01 -4.85896498e-01 -6.00234628e-01
9.48258698e-01 5.60287356e-01 -6.66332066e-01 -4.70333457e-01
-1.15838134e+00 6.33684814e-01 4.76609468e-01 -8.00242126e-01
-7.07342982e-01 -7.92477250e-01 -6.91664577e-01 -3.02483365e-02
6.87121749e-01 -7.38986731e-01 1.28399611e+00 -8.95727456e-01
-1.56953549e+00 1.04616046e+00 -5.28235435e-01 -6.93267703e-01
5.00773370e-01 -9.01344195e-02 -9.01866257e-01 -1.96071923e-01
1.49257362e-01 4.29710567e-01 9.71815526e-01 -1.21513486e+00
-8.84007275e-01 -6.65498555e-01 -1.84490368e-01 1.96869537e-01
-4.88359213e-01 -1.75139412e-01 -3.17849010e-01 -5.78605533e-01
-1.61306411e-01 -7.39972591e-01 -1.20827690e-01 2.78519005e-01
-4.20537204e-01 -1.07490398e-01 9.85799670e-01 -4.94986117e-01
1.37897456e+00 -2.32147193e+00 -1.82350859e-01 4.32673097e-01
1.84268653e-01 6.15157008e-01 2.65739411e-02 -2.86596902e-02
2.81255413e-02 -1.34578049e-01 -6.48340106e-01 -4.49520022e-01
4.92301024e-02 3.58394384e-01 -3.73802245e-01 3.22999746e-01
-2.24564567e-01 3.49508882e-01 -6.80656195e-01 -3.35743934e-01
-1.02143280e-01 1.99806839e-01 -5.92972457e-01 2.40480989e-01
-3.67784984e-02 3.37017983e-01 -3.34492385e-01 8.69263649e-01
1.23073101e+00 -5.04612960e-02 1.31886855e-01 4.09313589e-01
4.32097346e-01 1.68012470e-01 -1.23428607e+00 1.66266704e+00
-2.72040308e-01 1.48296654e-01 1.14303119e-01 -8.64117563e-01
1.19728982e+00 3.70940536e-01 1.81478664e-01 -5.16927421e-01
2.20146049e-02 4.01231259e-01 -3.39012951e-01 -5.08636199e-02
2.60192811e-01 -7.57694319e-02 -2.78035164e-01 2.91010588e-01
1.85009301e-01 5.60095370e-01 -2.97395468e-01 -1.64706439e-01
9.85209823e-01 -1.12845249e-01 8.81125391e-01 -1.25465408e-01
8.02087784e-01 -2.29919869e-02 7.78615177e-01 8.43466580e-01
-6.92229807e-01 6.25150919e-01 1.66088432e-01 -3.32197040e-01
-7.87353218e-01 -1.28684318e+00 -2.78301150e-01 1.14827061e+00
-1.10442527e-02 9.15558860e-02 -7.40688860e-01 -1.41336691e+00
2.32004315e-01 8.45186293e-01 -3.62369269e-01 -2.46007219e-01
-1.67582959e-01 -4.70730633e-01 5.09787619e-01 1.79137483e-01
7.38953769e-01 -4.33949053e-01 2.80593969e-02 2.24470615e-01
-2.36950606e-01 -9.94181037e-01 -8.17097366e-01 2.31918424e-01
-9.06142950e-01 -1.01809037e+00 -9.79270577e-01 -7.56204605e-01
7.81650484e-01 1.70494288e-01 8.13908219e-01 -6.93329811e-01
3.77416044e-01 2.91580647e-01 -2.09957704e-01 -5.52469730e-01
-5.31805873e-01 4.24131155e-01 2.17302054e-01 3.08233529e-01
1.21110356e+00 -7.32050061e-01 -3.15824628e-01 5.20108283e-01
-7.10084677e-01 -6.96694732e-01 5.26966572e-01 8.81740749e-01
7.74359763e-01 -1.91836417e-01 9.94099081e-01 -1.55747390e+00
6.30607486e-01 -9.56766844e-01 -6.52916610e-01 4.02093351e-01
-1.18998766e+00 -8.42684358e-02 7.02185631e-01 -8.42321277e-01
-1.19914961e+00 4.03647989e-01 7.59912059e-02 -6.05667830e-01
-7.00784147e-01 -2.76379921e-02 -7.08121896e-01 -4.19136183e-03
1.02922451e+00 2.13484570e-01 9.02907401e-02 -8.29381466e-01
1.52964935e-01 1.36910605e+00 7.94234872e-01 -4.18464124e-01
9.56198215e-01 4.38665241e-01 -3.60532671e-01 -3.70162159e-01
-6.97270513e-01 -7.58151829e-01 -9.13471162e-01 7.39362299e-01
5.16177773e-01 -1.20424676e+00 -1.62958115e-01 4.66164887e-01
-8.83970022e-01 -5.93392961e-02 -5.84842741e-01 3.93543780e-01
-5.17666698e-01 4.32605833e-01 1.08534671e-01 -7.10501611e-01
-3.47181976e-01 -6.25040174e-01 7.56367862e-01 1.56284645e-01
-1.24555856e-01 -9.63713408e-01 -3.37689072e-02 3.23887378e-01
2.79349774e-01 1.34953395e-01 7.31646299e-01 -1.61492610e+00
-2.13825136e-01 -4.43280846e-01 -4.99189645e-02 5.02279043e-01
4.54709262e-01 -8.88377130e-01 -1.09398866e+00 -3.86127412e-01
5.13320565e-02 -1.03675410e-01 4.01332259e-01 1.84073880e-01
1.15707242e+00 -7.34525204e-01 -5.11750102e-01 6.70511603e-01
1.23341477e+00 4.25502658e-01 5.77129185e-01 3.52580279e-01
3.36300492e-01 2.99758285e-01 9.22804952e-01 7.13373065e-01
5.86353600e-01 6.78935587e-01 2.25888733e-02 4.50112559e-02
9.60410982e-02 -4.81249481e-01 3.49428594e-01 5.16425306e-03
4.23085183e-01 -3.48602355e-01 -6.43664896e-01 7.07194507e-01
-1.66188836e+00 -7.35496640e-01 2.82889307e-01 2.98890710e+00
9.97048020e-01 1.39971180e-02 4.50806737e-01 -1.30280107e-01
7.78185904e-01 -1.90236047e-01 -1.02835929e+00 -3.45908016e-01
4.99194898e-02 2.37934604e-01 9.69441295e-01 4.19078588e-01
-1.42535222e+00 7.02485204e-01 6.15233612e+00 7.95578361e-01
-7.33286619e-01 3.77596349e-01 4.82325166e-01 -1.08103089e-01
-4.89818193e-02 -7.02437460e-02 -1.10592735e+00 5.48474133e-01
1.15945709e+00 -4.73483056e-01 4.57556307e-01 1.33965921e+00
-3.06158453e-01 1.26202807e-01 -1.33567107e+00 9.21623349e-01
6.89735860e-02 -9.14249241e-01 -1.04868740e-01 4.36217248e-01
6.31700337e-01 -1.22775011e-01 4.82037216e-02 6.03121281e-01
5.00944078e-01 -5.26421428e-01 9.94412154e-02 1.08516276e-01
1.07622325e+00 -7.20179439e-01 6.81471050e-01 6.26526713e-01
-6.71643376e-01 -3.40471596e-01 -6.35941684e-01 2.56477594e-01
-2.51353890e-01 5.31807840e-01 -9.20285463e-01 5.97187698e-01
6.18515372e-01 6.06813490e-01 -2.49397486e-01 9.70195770e-01
1.83508217e-01 6.12228274e-01 -5.26812434e-01 4.43330944e-01
-2.84568936e-01 9.79988649e-02 6.71214044e-01 7.93303072e-01
4.31752950e-01 2.26021290e-01 2.68389970e-01 4.94772851e-01
-3.22308064e-01 1.30276054e-01 -8.28381002e-01 1.87566295e-01
1.01866102e+00 5.45425534e-01 1.00742541e-01 -4.01145101e-01
-4.71263349e-01 1.38281798e+00 1.96381941e-01 4.98794556e-01
-5.12180090e-01 -5.33759594e-01 9.91579473e-01 2.16877729e-01
4.68898475e-01 1.81496918e-01 -1.66852519e-01 -1.39977252e+00
1.36691704e-01 -9.27322090e-01 1.08159971e+00 -2.04296276e-01
-1.54178786e+00 5.35034120e-01 1.23900577e-01 -1.52784097e+00
-3.88090789e-01 -2.70016909e-01 -2.28126764e-01 9.95610058e-01
-1.79906619e+00 -1.04264176e+00 1.77585304e-01 1.17985439e+00
3.08759481e-01 -6.15969658e-01 1.32907963e+00 4.41748321e-01
-3.22252601e-01 1.41561449e+00 6.91970944e-01 1.57235906e-01
1.22920489e+00 -9.86067832e-01 1.87372446e-01 8.56237710e-01
-2.00087041e-01 5.71723640e-01 2.91885942e-01 -5.57015359e-01
-8.60950947e-01 -1.41701996e+00 1.17496574e+00 -5.86057544e-01
1.13433160e-01 -5.95151603e-01 -1.15502238e+00 9.36247110e-01
-1.26117706e-01 6.21150315e-01 1.09358883e+00 9.89997908e-02
-7.11280704e-01 -4.01145339e-01 -2.15896487e+00 2.11261109e-01
1.08271980e+00 -2.91408211e-01 -6.59639001e-01 1.55097142e-01
8.04366946e-01 -5.54630876e-01 -1.08457255e+00 2.83925980e-02
3.61252636e-01 -6.49064720e-01 1.09038186e+00 -6.88885808e-01
-1.74963996e-01 -1.84726939e-01 -4.31026965e-01 -1.18501234e+00
-2.17422873e-01 -7.12288976e-01 -6.06790185e-01 1.50231409e+00
4.27512169e-01 -1.18669510e+00 8.60202432e-01 9.64691639e-01
3.48235816e-01 -2.10004076e-01 -1.35499966e+00 -1.04631114e+00
3.36976826e-01 7.19632197e-04 1.03460395e+00 1.00560021e+00
-1.99448541e-01 1.22288167e-01 -5.12688160e-01 5.39357126e-01
8.38172257e-01 1.21056885e-02 8.90935004e-01 -1.39451480e+00
-2.24448398e-01 2.15569556e-01 -2.16566741e-01 -1.19139850e+00
2.67732620e-01 -7.67636597e-01 -2.75698036e-01 -8.58581841e-01
-1.26845077e-01 -6.31559789e-01 -7.07373619e-01 6.54112279e-01
-3.80311087e-02 -1.30406782e-01 -1.22596882e-02 2.72808939e-01
-4.73117858e-01 3.77115220e-01 7.89643884e-01 -4.48555173e-03
-4.96337831e-01 6.09198451e-01 -1.21824598e+00 3.13365132e-01
8.58165443e-01 -7.82577038e-01 -7.51078904e-01 -1.44875661e-01
-7.88727224e-01 -1.32038325e-01 1.12913176e-01 -1.00630999e+00
1.60048261e-01 -3.86173338e-01 4.07903522e-01 -3.35986555e-01
2.14448199e-01 -1.37848234e+00 -5.13900183e-02 1.48231193e-01
-5.26837170e-01 -6.70778334e-01 5.90435602e-02 1.07174778e+00
-6.46013394e-02 -3.68289873e-02 9.70806897e-01 -6.51280070e-03
-7.61220813e-01 5.56508839e-01 -1.68401062e-01 1.13137737e-01
1.25501037e+00 -4.27830040e-01 -2.87500679e-01 -3.04647535e-01
-9.39187646e-01 3.51944178e-01 5.20640254e-01 3.61501902e-01
4.79331940e-01 -1.47443712e+00 -4.84078467e-01 6.14305079e-01
4.75438565e-01 1.94683701e-01 3.13379914e-02 3.29997718e-01
4.15350884e-01 4.53529477e-01 -1.29180729e-01 -2.71888494e-01
-1.35177553e+00 9.11829293e-01 2.02533916e-01 -3.83178532e-01
-2.83614308e-01 7.20284164e-01 1.92982420e-01 -9.85888600e-01
5.19372106e-01 -8.13924000e-02 -4.76637967e-02 -2.08329871e-01
5.76516807e-01 4.84781057e-01 4.52033877e-02 -4.63758707e-01
-4.04686749e-01 6.93619028e-02 -4.38171595e-01 9.81872231e-02
9.99644101e-01 -6.73562288e-01 3.49878877e-01 1.79774031e-01
1.31112111e+00 -7.23086298e-03 -1.28986347e+00 -9.77907777e-01
-4.98334691e-02 -8.01057220e-01 -2.12034047e-01 -1.22833312e+00
-7.60012507e-01 6.45347297e-01 1.02590334e+00 -2.29335889e-01
1.47939217e+00 -1.16222553e-01 9.61863458e-01 4.68893588e-01
4.82984960e-01 -1.18598807e+00 -5.54553807e-01 2.76925862e-01
5.17341733e-01 -1.61092055e+00 -1.29115120e-01 -4.58975643e-01
-8.00706863e-01 7.49205649e-01 7.89497435e-01 2.61886775e-01
8.56343985e-01 -1.11485437e-01 2.83089094e-02 4.54659104e-01
-6.27257168e-01 3.86815961e-03 1.66730866e-01 1.35981560e+00
-2.36292899e-01 1.02180026e-01 4.62785587e-02 1.01550651e+00
9.63656604e-02 4.63727593e-01 4.49458361e-01 9.31834877e-01
-2.27676794e-01 -1.64471662e+00 -5.61535418e-01 4.99771535e-01
-6.19430900e-01 7.79728368e-02 -4.38985914e-01 6.22035801e-01
2.17147708e-01 8.97412837e-01 -4.11814526e-02 -1.85838833e-01
4.54036355e-01 6.13496780e-01 -9.11243353e-03 -6.54746413e-01
-2.37884313e-01 -2.01116025e-01 1.21708378e-01 -3.33655864e-01
-2.25816593e-01 -8.99973273e-01 -1.03349876e+00 -3.36746961e-01
-2.22264275e-01 1.44048482e-01 3.90147507e-01 6.98281884e-01
9.57536101e-01 -1.61802024e-01 7.13719904e-01 -4.05374803e-02
-1.05128050e+00 -6.48122847e-01 -6.52083039e-01 2.81625181e-01
7.21201181e-01 -6.28715336e-01 -2.48701960e-01 2.26056933e-01] | [10.376599311828613, 3.190476894378662] |
ffd0e4bd-b22d-493c-a0d9-0e7b60ca0a94 | nested-scale-editing-for-conditional-image | 2006.02038 | null | https://arxiv.org/abs/2006.02038v1 | https://arxiv.org/pdf/2006.02038v1.pdf | Nested Scale Editing for Conditional Image Synthesis | We propose an image synthesis approach that provides stratified navigation in the latent code space. With a tiny amount of partial or very low-resolution image, our approach can consistently out-perform state-of-the-art counterparts in terms of generating the closest sampled image to the ground truth. We achieve this through scale-independent editing while expanding scale-specific diversity. Scale-independence is achieved with a nested scale disentanglement loss. Scale-specific diversity is created by incorporating a progressive diversification constraint. We introduce semantic persistency across the scales by sharing common latent codes. Together they provide better control of the image synthesis process. We evaluate the effectiveness of our proposed approach through various tasks, including image outpainting, image superresolution, and cross-domain image translation. | ['Jie Min', 'Lingzhi Zhang', 'James C. Gee', 'Yinshuang Xu', 'Tarmily Wen', 'Jiancong Wang', 'Jianbo Shi'] | 2020-06-03 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 6.81358755e-01 -3.12553160e-02 -3.09823304e-01 -1.93498850e-01
-1.12017739e+00 -7.45608270e-01 8.29543293e-01 -2.17301250e-01
-9.62252691e-02 6.31070673e-01 4.73014832e-01 2.81368732e-01
-8.35859962e-03 -6.69234991e-01 -8.60567331e-01 -6.01106226e-01
3.11638445e-01 4.46075983e-02 2.57330924e-01 -3.70022267e-01
3.11262965e-01 5.81980288e-01 -1.40531874e+00 5.12799978e-01
1.10900199e+00 8.34357798e-01 3.74220073e-01 6.41324341e-01
1.07625142e-01 7.17704475e-01 -2.69329011e-01 -2.71368235e-01
5.52399635e-01 -5.69512904e-01 -5.27406812e-01 4.51606750e-01
8.39555025e-01 -3.86643827e-01 -1.52585536e-01 1.21686280e+00
3.38231206e-01 -1.50303140e-01 3.08869720e-01 -1.04527795e+00
-9.70156133e-01 2.01430395e-01 -8.49135399e-01 -1.12945437e-01
3.41705799e-01 1.15965807e-03 1.00685108e+00 -9.97119308e-01
9.86343563e-01 1.35186946e+00 6.21306002e-01 3.80354673e-01
-1.96665406e+00 -5.08906901e-01 1.36579275e-01 -2.81406403e-01
-1.49673247e+00 -6.23582363e-01 9.50894892e-01 -3.65848452e-01
4.65500116e-01 3.34365904e-01 2.39221081e-01 9.26233351e-01
1.71130419e-01 3.11030120e-01 1.38485575e+00 -5.81015050e-01
1.71845555e-01 -7.75370598e-02 -6.46804929e-01 6.55908346e-01
8.79050270e-02 2.41049916e-01 -8.77753437e-01 -1.66927591e-01
1.41678524e+00 -9.04304758e-02 -3.72781664e-01 -7.76530266e-01
-1.45059085e+00 6.97909534e-01 5.21016419e-01 9.50209796e-02
-1.36179224e-01 2.27028087e-01 2.33016580e-01 3.96964788e-01
5.96032619e-01 5.96546173e-01 -4.86759633e-01 2.27638856e-01
-1.31458318e+00 3.57493639e-01 2.76272804e-01 1.22996485e+00
7.87286699e-01 2.02871099e-01 -4.06312495e-01 8.57464492e-01
-1.57511488e-01 3.88603121e-01 4.75036830e-01 -1.43305528e+00
4.43340093e-01 1.97687984e-01 3.63929540e-01 -1.09089136e+00
3.90853360e-02 -5.03142118e-01 -8.55572939e-01 5.07118344e-01
4.06426005e-02 2.12107316e-01 -9.85711932e-01 2.13854313e+00
2.43018255e-01 1.83759004e-01 -6.51416406e-02 6.66717112e-01
2.40733758e-01 5.89613616e-01 -2.22290412e-01 -3.44417125e-01
1.40998924e+00 -1.16073477e+00 -8.17295671e-01 -2.13071093e-01
-4.14693095e-02 -9.60189402e-01 1.28821313e+00 2.07592651e-01
-1.45495200e+00 -7.12925076e-01 -1.07597804e+00 -5.17042041e-01
9.78639051e-02 1.36417970e-01 4.55504298e-01 4.71124321e-01
-1.16918671e+00 6.16342485e-01 -6.97895706e-01 -1.35018975e-01
3.72018635e-01 -3.48661989e-02 -4.64908272e-01 -2.04155192e-01
-8.74537349e-01 6.71411574e-01 -1.28430743e-02 -5.31391561e-01
-6.95350945e-01 -1.08005095e+00 -9.34864342e-01 -1.01306878e-01
1.84407279e-01 -1.04477286e+00 9.81921673e-01 -1.13654029e+00
-1.54698789e+00 1.04953468e+00 -3.75291854e-01 -3.08435500e-01
7.96456158e-01 -1.44458443e-01 -2.49104053e-01 2.93524116e-01
4.93611664e-01 9.78011012e-01 1.25877690e+00 -1.57279360e+00
-4.76378649e-01 -1.28091946e-01 -1.14561534e-02 4.30468619e-01
-1.62540331e-01 -9.79517028e-02 -7.19637275e-01 -1.39700985e+00
3.40847760e-01 -1.01946020e+00 -2.16801524e-01 5.74606597e-01
-1.40328035e-01 7.19858646e-01 6.43362582e-01 -7.98850894e-01
1.20631409e+00 -2.33375072e+00 6.52348459e-01 -1.39113203e-01
1.79427296e-01 -3.21313620e-01 -3.04490417e-01 3.33737165e-01
-5.02512045e-02 -1.22676447e-01 -5.37550390e-01 -6.60900414e-01
-2.24278942e-01 1.06356852e-01 -5.48625946e-01 3.45633239e-01
2.65279144e-01 8.29601467e-01 -8.28048110e-01 -4.64962065e-01
2.99179256e-02 5.87048948e-01 -1.05662632e+00 1.82905961e-02
-1.84838384e-01 7.90951431e-01 -1.85492590e-01 5.18355250e-01
8.38929415e-01 -4.32458043e-01 4.86515537e-02 -4.97956008e-01
-2.51297444e-01 9.95253678e-03 -1.08874977e+00 2.34200740e+00
-6.47764325e-01 4.61512148e-01 2.14440688e-01 -4.93233711e-01
7.65422285e-01 2.71129757e-02 2.31720164e-01 -6.37184918e-01
-4.18780982e-01 2.59534359e-01 -4.01440829e-01 7.91866407e-02
6.46703362e-01 -2.71633953e-01 -5.30217364e-02 1.47295311e-01
-9.76514220e-02 -4.87211168e-01 5.25955409e-02 2.39267796e-01
6.77586496e-01 2.96595067e-01 3.06216180e-01 -4.50408340e-01
3.54832828e-01 -2.60519147e-01 3.87973338e-01 7.15820909e-01
8.10618848e-02 1.17997050e+00 3.15197498e-01 -1.59438908e-01
-1.54514611e+00 -1.40457499e+00 -2.78202057e-01 9.82092440e-01
3.13472807e-01 -2.45490298e-01 -7.65260816e-01 -2.58924961e-01
-1.62358627e-01 5.16125321e-01 -7.72670746e-01 2.99973355e-04
-5.75960517e-01 -3.19722980e-01 3.28088522e-01 4.58709806e-01
7.18271971e-01 -6.88672125e-01 -4.93521184e-01 7.76067153e-02
-3.81865412e-01 -1.26814675e+00 -1.08737326e+00 -1.90588921e-01
-1.05030334e+00 -5.65816939e-01 -1.07213080e+00 -8.85580778e-01
7.90210247e-01 4.28423315e-01 1.26559317e+00 -1.63541347e-01
-2.84088463e-01 4.69871536e-02 -1.15926012e-01 3.15759182e-01
-4.71105814e-01 7.93928280e-02 -1.47088587e-01 1.18346401e-01
-6.80878878e-01 -1.00413120e+00 -8.85784030e-01 3.83081675e-01
-1.23370445e+00 5.46073377e-01 4.69581902e-01 8.77544701e-01
8.25657964e-01 -4.41951267e-02 3.91613841e-01 -7.23555267e-01
4.97341454e-01 5.35892174e-02 -5.77470839e-01 2.13795587e-01
-5.10292470e-01 3.66867781e-01 5.96281290e-01 -5.48727691e-01
-1.24776185e+00 -2.61969715e-02 1.26448378e-01 -5.02960324e-01
2.28498489e-01 2.66727600e-02 -1.35054693e-01 -2.49281719e-01
7.54865527e-01 3.62052709e-01 -7.59290606e-02 -4.34797227e-01
7.91358769e-01 1.06292576e-01 8.02924931e-01 -7.34685540e-01
8.83992910e-01 9.94243324e-01 1.06965445e-01 -5.17773449e-01
-8.48657668e-01 2.09486764e-02 -6.37035608e-01 1.85213879e-01
8.30117047e-01 -1.27763271e+00 -4.90759760e-02 4.81770337e-01
-9.89704013e-01 -2.38335267e-01 -5.84743083e-01 6.69702739e-02
-7.92807341e-01 3.76902074e-01 -7.03177333e-01 -1.10989377e-01
-1.07996352e-02 -1.26504135e+00 1.52453637e+00 -2.15266794e-01
-2.04830274e-01 -7.87137330e-01 9.74148288e-02 1.74275160e-01
6.06994331e-01 4.44088757e-01 7.95082510e-01 3.02587032e-01
-8.70740771e-01 2.03229263e-01 -4.43952411e-01 2.71700799e-01
3.09216142e-01 -2.01424927e-01 -7.48519123e-01 -5.04653513e-01
-1.21920586e-01 -1.92327127e-01 9.58430588e-01 3.71504843e-01
1.03892338e+00 -3.13245237e-01 5.47401328e-03 1.08815789e+00
1.54659188e+00 -2.41361618e-01 6.83183610e-01 3.32015246e-01
7.66860962e-01 5.69595516e-01 3.82483542e-01 3.63738537e-01
2.80030131e-01 1.13579524e+00 1.48317784e-01 -3.10428143e-01
-5.64214647e-01 -4.30684060e-01 2.19969451e-01 6.33414567e-01
1.59237668e-01 1.59537509e-01 -5.31319141e-01 4.87336576e-01
-1.67596805e+00 -9.56481338e-01 3.81331325e-01 2.04820156e+00
1.26815248e+00 -5.93344867e-02 -9.05566067e-02 -6.21175021e-02
7.00218856e-01 4.61210191e-01 -5.74427664e-01 -1.14176370e-01
-4.68642503e-01 1.41910568e-01 6.54647171e-01 9.56427455e-01
-8.96674693e-01 1.00975347e+00 6.92340469e+00 1.15240777e+00
-1.16199458e+00 2.29650885e-01 7.55090654e-01 -1.95071116e-01
-6.64910734e-01 9.35320649e-03 -4.95404571e-01 3.73583555e-01
2.51034617e-01 -1.64473221e-01 7.09361494e-01 6.51674449e-01
9.66561362e-02 1.41509712e-01 -9.61049378e-01 9.60195124e-01
2.02727884e-01 -1.72757256e+00 2.60002047e-01 -5.14900982e-02
1.41863704e+00 -2.83924848e-01 4.07843709e-01 -2.75920480e-01
2.79434621e-01 -8.53142381e-01 1.09087622e+00 4.04017329e-01
1.45516014e+00 -6.84396803e-01 -1.17588535e-01 -6.59248754e-02
-1.27282310e+00 5.80210425e-03 -1.72483668e-01 2.24838912e-01
2.81282902e-01 5.63349903e-01 1.02913052e-01 5.44984519e-01
5.34982026e-01 5.14032900e-01 -5.56450009e-01 4.09353793e-01
-2.98150122e-01 3.01310774e-02 -1.37910202e-01 9.03996706e-01
-6.47827983e-02 -2.38612786e-01 5.26582599e-01 9.52934265e-01
5.09622157e-01 1.25880092e-02 1.74736440e-01 1.02415228e+00
-1.99899361e-01 -1.55033812e-01 -3.60722899e-01 3.85276794e-01
6.94155335e-01 8.75093699e-01 -8.06460857e-01 -2.59929597e-01
-2.69362658e-01 1.76039910e+00 1.33121565e-01 4.75196540e-01
-9.34444308e-01 -1.34280309e-01 7.57906020e-01 2.32147858e-01
5.79466403e-01 -2.74290860e-01 -6.28973246e-01 -1.39573681e+00
1.84110567e-01 -1.06090117e+00 1.17877191e-02 -9.92150724e-01
-1.11952531e+00 6.83605850e-01 2.19623055e-02 -1.34270406e+00
-1.65499121e-01 -1.81464359e-01 -2.11683571e-01 8.69604647e-01
-1.56908596e+00 -1.37826622e+00 -2.47357443e-01 5.94033659e-01
6.58891439e-01 -3.13652828e-02 7.16303170e-01 3.84044856e-01
-6.86466694e-02 6.85975075e-01 2.44587049e-01 -2.99545974e-01
9.79508221e-01 -9.15664256e-01 6.17888629e-01 9.10849750e-01
2.44984012e-02 7.14322686e-01 8.47770691e-01 -6.09506369e-01
-1.09595311e+00 -1.17746198e+00 7.76969850e-01 -4.16354895e-01
4.33673114e-01 -4.13072646e-01 -7.04195917e-01 5.44738352e-01
7.63224065e-02 3.61281186e-01 2.42104068e-01 -2.66219020e-01
-8.09575617e-01 -9.48328972e-02 -1.37432492e+00 8.89517546e-01
1.31233728e+00 -8.98449838e-01 -1.61241412e-01 6.85025677e-02
1.21530914e+00 -5.26926756e-01 -7.78189361e-01 4.10986006e-01
6.97585940e-01 -1.15537393e+00 1.33502924e+00 -1.12583771e-01
9.14263606e-01 -5.58313131e-01 -4.04842705e-01 -1.07337403e+00
-6.08441114e-01 -8.09007406e-01 1.54841930e-01 1.19088936e+00
1.51496768e-01 -3.65188599e-01 5.97220361e-01 2.91164309e-01
9.08862501e-02 -5.83909035e-01 -8.16440880e-01 -8.59702587e-01
5.85612319e-02 -1.77699630e-03 6.25507772e-01 9.90473628e-01
-4.07421798e-01 1.23513773e-01 -7.08164394e-01 2.61791348e-01
9.49738026e-01 4.38499451e-01 6.34042621e-01 -6.24095142e-01
-5.92918992e-01 -3.44327152e-01 -2.20090255e-01 -1.33378780e+00
-8.61495957e-02 -6.27451241e-01 -1.48383915e-01 -1.14622998e+00
2.76320130e-01 -1.96447834e-01 -2.33826581e-02 6.39025718e-02
-1.21163808e-01 8.40056956e-01 4.47422713e-01 4.80272442e-01
-3.56260359e-01 5.90046585e-01 1.62960029e+00 2.31355503e-01
-1.16320759e-01 -6.05786026e-01 -8.68727982e-01 6.29468799e-01
6.15078330e-01 -2.87392825e-01 -5.72234511e-01 -6.96727693e-01
2.19958037e-01 3.32434326e-01 3.32923651e-01 -9.19672251e-01
-9.83948410e-02 -3.22794646e-01 3.29142511e-01 -1.78785443e-01
3.93193960e-01 -6.41323507e-01 4.20704275e-01 2.48340577e-01
-7.74425685e-01 8.77429992e-02 1.59613580e-01 5.71738780e-01
-2.53594428e-01 2.09470019e-01 1.20449841e+00 8.27783346e-02
-3.70269388e-01 2.70824581e-01 1.42367676e-01 1.67514145e-01
7.61620045e-01 -3.23006511e-01 -3.55251245e-02 -4.69904870e-01
-6.31138265e-01 -2.11605757e-01 1.08815694e+00 4.37468320e-01
4.51120913e-01 -1.64231467e+00 -7.39734769e-01 5.05195379e-01
1.78195104e-01 -4.20907661e-02 2.81652063e-01 4.90174323e-01
-5.01026750e-01 1.32544294e-01 -2.66737282e-01 -6.37488067e-01
-1.20359528e+00 4.85517114e-01 1.15415744e-01 -4.39846516e-01
-5.80733418e-01 9.12811279e-01 7.07873166e-01 -2.55335838e-01
2.36380901e-02 -3.01710457e-01 3.38804692e-01 -2.02446640e-01
4.77652907e-01 1.03822403e-01 -1.37830615e-01 -5.99433184e-01
-1.31460518e-01 9.70013320e-01 9.81230056e-04 -6.42168880e-01
1.13191068e+00 -5.20025015e-01 -4.95866016e-02 2.50317872e-01
1.20203793e+00 5.30209839e-01 -2.00317264e+00 -2.29010388e-01
-4.39685613e-01 -9.35298026e-01 3.21413353e-02 -7.41114140e-01
-9.89120662e-01 4.69305366e-01 5.42167783e-01 -1.31863698e-01
1.42142284e+00 -5.65769784e-02 7.18953907e-01 -1.63218886e-01
4.45317775e-01 -7.79433727e-01 3.80273640e-01 2.49520943e-01
1.29142296e+00 -1.07840478e+00 2.43620142e-01 -6.39745176e-01
-6.09693468e-01 6.51805460e-01 2.04699785e-01 -4.37755674e-01
4.40759361e-01 4.36619401e-01 -1.71911895e-01 1.57066986e-01
-6.18008733e-01 6.28447980e-02 3.91817093e-01 5.47840118e-01
4.76962656e-01 -6.16837516e-02 -2.85144329e-01 1.15037955e-01
-2.34367177e-01 -5.38073741e-02 2.99347311e-01 7.27489889e-01
-2.57861465e-01 -1.21128070e+00 -4.63176519e-01 -1.85960978e-01
-4.13752049e-01 -4.52006608e-01 2.49046888e-02 5.06003916e-01
-6.81770180e-05 5.36713600e-01 7.56877810e-02 2.13609766e-02
2.37267375e-01 -2.07735017e-01 8.75851929e-01 -4.28513467e-01
-9.24926326e-02 2.60897368e-01 -1.85330749e-01 -8.40261042e-01
-3.87990266e-01 -5.56698859e-01 -8.59439731e-01 -2.46534035e-01
-1.28540188e-01 -2.67820597e-01 4.34503853e-01 4.21882838e-01
6.91715360e-01 5.67099452e-01 7.63550699e-01 -1.05842614e+00
-4.61484075e-01 -4.82172221e-01 -4.08362806e-01 6.76591516e-01
5.95426083e-01 -4.16066378e-01 -3.59032780e-01 5.25757551e-01] | [11.536538124084473, -0.5822892189025879] |
c16cbe6e-3b88-48e4-9bc0-8306fc5004d4 | groundnet-segmentation-aware-monocular-ground | 1811.07222 | null | https://arxiv.org/abs/1811.07222v4 | https://arxiv.org/pdf/1811.07222v4.pdf | GroundNet: Monocular Ground Plane Normal Estimation with Geometric Consistency | We focus on estimating the 3D orientation of the ground plane from a single image. We formulate the problem as an inter-mingled multi-task prediction problem by jointly optimizing for pixel-wise surface normal direction, ground plane segmentation, and depth estimates. Specifically, our proposed model, GroundNet, first estimates the depth and surface normal in two separate streams, from which two ground plane normals are then computed deterministically. To leverage the geometric correlation between depth and normal, we propose to add a consistency loss on top of the computed ground plane normals. In addition, a ground segmentation stream is used to isolate the ground regions so that we can selectively back-propagate parameter updates through only the ground regions in the image. Our method achieves the top-ranked performance on ground plane normal estimation and horizon line detection on the real-world outdoor datasets of ApolloScape and KITTI, improving the performance of previous art by up to 17.7% relatively. | ['Xi Li', 'Kris Kitani', 'Xinshuo Weng', 'Yunze Man'] | 2018-11-17 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 3.18810403e-01 1.34770334e-01 9.78595540e-02 -3.03348631e-01
-1.08272660e+00 -8.31877649e-01 3.35746109e-01 3.35765123e-01
-4.06137794e-01 3.82862777e-01 -1.34941682e-01 -1.47693425e-01
2.33392477e-01 -1.01381481e+00 -8.93543839e-01 -6.83465719e-01
-3.05366904e-01 3.92169595e-01 5.81905365e-01 -1.16851358e-02
5.82078993e-01 3.93410683e-01 -1.36555707e+00 -1.28743097e-01
1.12099802e+00 1.34323287e+00 -1.00462213e-02 1.00889540e+00
1.87420726e-01 4.20433581e-01 -2.24199384e-01 -1.03165194e-01
5.06914318e-01 -4.60390700e-03 -4.39682841e-01 4.03361946e-01
9.01228070e-01 -6.99281275e-01 6.42378554e-02 9.20310438e-01
4.01090264e-01 1.98948428e-01 6.50888085e-01 -7.85599470e-01
5.43507040e-01 -3.00728768e-01 -1.18830419e+00 7.03035221e-02
2.63089836e-01 1.74262494e-01 1.04036868e+00 -6.58205867e-01
2.52635062e-01 8.41884494e-01 8.28034520e-01 -3.12640607e-01
-9.08199549e-01 -4.58044380e-01 7.11772978e-01 -2.70510077e-01
-1.46828556e+00 -1.70111313e-01 5.68164945e-01 -6.66961968e-01
9.02368605e-01 1.36018610e-02 6.89752996e-01 1.79083511e-01
1.33546486e-01 6.05296493e-01 7.99683452e-01 -2.41991714e-01
1.86064661e-01 -4.41595584e-01 -1.05828166e-01 9.29938197e-01
2.40080848e-01 -1.16219446e-01 -6.71949029e-01 1.21973706e-02
9.42488194e-01 -2.16340557e-01 -3.68805468e-01 -2.43127853e-01
-1.10294688e+00 5.90789616e-01 3.09721142e-01 -5.81875384e-01
-4.65575188e-01 2.99063861e-01 -1.57384351e-01 -1.21465966e-01
8.43525290e-01 2.77145028e-01 -3.91048640e-01 4.05432545e-02
-1.03184879e+00 7.30059445e-01 5.87683320e-01 9.91298676e-01
1.23918581e+00 -2.99232572e-01 -4.30603363e-02 4.23688203e-01
7.33937025e-01 9.74322975e-01 -1.99236870e-01 -1.28335142e+00
8.56778920e-01 5.69069862e-01 4.02542949e-01 -1.07238257e+00
-3.91886592e-01 -4.79334086e-01 -3.07837576e-01 2.17057422e-01
5.92476785e-01 -4.60484713e-01 -1.09661472e+00 1.23069870e+00
8.37184548e-01 4.08055305e-01 -1.01742767e-01 9.04760897e-01
3.30495536e-01 7.99547553e-01 -3.71535093e-01 3.07074606e-01
1.09874606e+00 -1.17160451e+00 -6.79376572e-02 -9.09926057e-01
6.55857444e-01 -7.18492627e-01 5.16774118e-01 4.82020795e-01
-9.54421818e-01 -1.55656114e-01 -1.15837491e+00 -7.78773800e-02
1.35464072e-01 2.59678096e-01 5.12743771e-01 1.66982055e-01
-1.07006538e+00 5.34876406e-01 -1.12276673e+00 -7.15555698e-02
8.67151320e-02 1.22265570e-01 2.52689362e-01 -1.37149421e-02
-6.56466663e-01 2.81164736e-01 1.16616286e-01 2.21145317e-01
-7.86333442e-01 -8.37915957e-01 -9.27711844e-01 -9.30861533e-02
5.81177711e-01 -8.00909281e-01 1.31785357e+00 -4.99197394e-01
-1.37086594e+00 9.96469200e-01 -3.69905084e-01 -6.22799158e-01
6.58921480e-01 -6.04882002e-01 1.19068168e-01 3.37950557e-01
3.81143332e-01 6.48070633e-01 8.58523309e-01 -1.30222273e+00
-1.42177403e+00 -4.88220066e-01 5.76951653e-02 6.48758173e-01
2.39574373e-01 -6.29914165e-01 -9.19626236e-01 -3.32415462e-01
7.62455165e-01 -8.34161222e-01 -4.11768287e-01 3.02054703e-01
-7.15216994e-01 3.54926586e-01 3.40337902e-01 -7.64046788e-01
1.10908449e+00 -2.12777257e+00 6.22043246e-03 3.42905045e-01
4.23918396e-01 -3.77874613e-01 2.15766162e-01 1.14271231e-01
6.27003193e-01 -1.70021001e-02 -4.70106781e-01 -5.59195101e-01
-3.46325517e-01 -1.23502977e-01 -3.00396740e-01 6.39882982e-01
1.94131508e-01 3.79663080e-01 -8.64591956e-01 -3.20044488e-01
3.02462071e-01 2.02375337e-01 -7.31202900e-01 1.50227115e-01
-4.72893238e-01 4.12977248e-01 -7.20366478e-01 5.86897552e-01
1.13498771e+00 -4.16985303e-01 -8.18345919e-02 -4.02238145e-02
-5.36048591e-01 3.40840608e-01 -1.37017071e+00 1.62811577e+00
-4.10590380e-01 4.82031405e-01 1.42925590e-01 -3.54268521e-01
8.40177000e-01 -2.42241889e-01 5.46044290e-01 -6.04393125e-01
-2.92874724e-02 2.61470228e-01 -4.14203197e-01 -2.41999999e-01
5.17012775e-01 1.77467510e-01 -6.53797528e-03 4.13068421e-02
-5.54303706e-01 -7.08763242e-01 -7.61309341e-02 -3.04940548e-02
8.40221047e-01 3.90980750e-01 1.61598384e-01 -6.94376305e-02
4.10418928e-01 2.59301782e-01 6.62398815e-01 6.71747267e-01
1.59313083e-01 9.41766024e-01 5.55434704e-01 -2.70316809e-01
-8.88556004e-01 -1.23972249e+00 1.58677831e-01 6.06765807e-01
6.05665803e-01 -3.74057561e-01 -7.53630817e-01 -4.39163357e-01
8.99058804e-02 2.66932040e-01 -4.04123068e-01 3.56384426e-01
-4.99944508e-01 -6.52986288e-01 2.09565505e-01 3.45344067e-01
8.26396823e-01 -1.66602358e-01 -9.29392993e-01 2.47443885e-01
-2.97348917e-01 -1.43161857e+00 -3.59889030e-01 1.85597032e-01
-1.00737858e+00 -9.67080235e-01 -7.19326377e-01 -4.06196445e-01
5.63809991e-01 3.84637773e-01 1.09522676e+00 7.67246932e-02
-4.35007624e-02 7.36303702e-02 -2.30199963e-01 -3.98995280e-01
3.89593691e-01 1.43100157e-01 -5.74016213e-01 1.92852944e-01
-2.77941406e-01 -5.42314887e-01 -8.40584695e-01 3.73371542e-01
-3.74620199e-01 4.80951935e-01 1.33181602e-01 4.30675566e-01
1.17556548e+00 3.65051404e-02 -2.16574922e-01 -6.94429040e-01
-4.11891043e-01 -4.60790247e-01 -1.34436631e+00 -1.75629720e-01
-4.18725520e-01 -8.94857720e-02 1.60262018e-01 1.44054085e-01
-9.99043286e-01 1.91884130e-01 -7.34352022e-02 -2.79148430e-01
1.05863765e-01 3.84929866e-01 2.74263620e-02 -2.81090081e-01
3.73437643e-01 5.79149872e-02 -3.47751975e-01 -3.21049720e-01
9.40721259e-02 3.33923966e-01 6.55124724e-01 -4.97411907e-01
9.50966537e-01 1.02041888e+00 3.38297635e-01 -1.14030182e+00
-1.36561358e+00 -7.65074134e-01 -5.58675110e-01 -1.87706888e-01
1.07763886e+00 -1.34360576e+00 -4.12342489e-01 1.02439201e+00
-1.16631770e+00 -8.34402323e-01 8.28791857e-02 3.21272701e-01
-3.89227480e-01 2.91667849e-01 -3.51242423e-01 -7.99934328e-01
-2.90663093e-01 -8.23548615e-01 1.67142236e+00 3.29455018e-01
2.57725656e-01 -9.27153111e-01 9.81272385e-02 2.61784941e-01
-6.67237490e-02 7.04350233e-01 6.70982182e-01 2.08537415e-01
-1.30375934e+00 -1.74917206e-01 -3.93672854e-01 1.33345410e-01
-1.11658826e-01 1.44374743e-01 -1.20223212e+00 -1.48248911e-01
-1.67059422e-01 2.27960050e-02 9.85111177e-01 8.61362338e-01
8.92117202e-01 -9.84051898e-02 -4.33590919e-01 1.47281432e+00
1.69651318e+00 -8.26608688e-02 3.47672552e-01 6.29291534e-01
9.47066844e-01 5.58536649e-01 1.02437973e+00 8.17645252e-01
8.00609887e-01 5.76586664e-01 6.97508812e-01 -1.47882834e-01
3.03310249e-02 -2.64261246e-01 -9.04848352e-02 2.78237402e-01
-1.06954448e-01 -6.88468874e-01 -1.00441039e+00 4.77617383e-01
-1.75908136e+00 -2.61742473e-01 -3.96268010e-01 2.58202314e+00
1.76133260e-01 3.53870481e-01 -1.33023173e-01 -1.83984101e-01
4.62966442e-01 4.75576162e-01 -8.11867654e-01 1.85456321e-01
-1.49187386e-01 -1.35249197e-01 1.30350339e+00 8.58269870e-01
-1.27160990e+00 1.04291320e+00 6.00659752e+00 3.93631071e-01
-1.16843033e+00 -3.75690192e-01 5.93088567e-01 -2.58108765e-01
-4.53966260e-01 1.64802745e-01 -1.27068007e+00 3.72337371e-01
6.54711068e-01 3.11989859e-02 2.12665692e-01 5.67075431e-01
3.48732620e-01 -7.49556303e-01 -6.47980571e-01 8.40299606e-01
-7.23004043e-02 -1.35766363e+00 -2.37403974e-01 2.61876553e-01
1.04149806e+00 5.24355710e-01 -1.13068588e-01 -1.76629230e-01
2.11972743e-01 -5.21124005e-01 9.58316147e-01 3.86210948e-01
6.94693923e-01 -7.57319033e-01 4.39325035e-01 3.12918067e-01
-1.42037511e+00 1.43860340e-01 8.99178442e-03 -2.25991786e-01
5.21788955e-01 8.57328892e-01 -7.72351086e-01 6.92612350e-01
7.74676621e-01 9.48242724e-01 -4.68367636e-01 1.45735991e+00
-5.90391636e-01 5.92193007e-01 -9.74083126e-01 3.97945076e-01
4.90973920e-01 -3.03525418e-01 5.96631885e-01 7.25876033e-01
4.17636722e-01 1.95660636e-01 4.21220481e-01 6.29647136e-01
8.60600471e-02 -6.75962958e-03 -3.22844356e-01 3.88092309e-01
4.86193329e-01 1.07632887e+00 -8.32943439e-01 -2.09462598e-01
-4.54005510e-01 1.09074926e+00 2.82381065e-02 4.40201223e-01
-9.40931201e-01 -2.71613955e-01 8.91124606e-01 4.48789537e-01
4.60756779e-01 -4.17030543e-01 -7.04871595e-01 -9.88131821e-01
3.10901850e-01 -3.42999279e-01 2.78241843e-01 -6.40036225e-01
-6.47721767e-01 3.46639872e-01 -2.70787403e-02 -1.39894712e+00
-2.35121235e-01 -4.54928815e-01 -6.76766396e-01 1.08793211e+00
-2.04319978e+00 -9.51348424e-01 -8.25398266e-01 6.17003106e-02
4.69428360e-01 4.86939818e-01 2.68869668e-01 2.03030333e-01
-5.01833498e-01 1.59291893e-01 1.34318188e-01 -9.45039764e-02
4.27109092e-01 -1.31823444e+00 1.02352691e+00 1.13648987e+00
-3.41224015e-01 -3.32155935e-02 5.45563400e-01 -8.46301377e-01
-1.11951804e+00 -1.18075037e+00 6.86842203e-01 -1.58629507e-01
5.05469978e-01 -2.35988408e-01 -8.54672611e-01 6.11538291e-01
-4.71441954e-01 1.37298137e-01 5.64952753e-02 -2.16862410e-01
-4.05125655e-02 -6.15349077e-02 -9.36539888e-01 4.23467129e-01
9.07426715e-01 -1.73592225e-01 1.17770121e-01 3.67019653e-01
9.10199761e-01 -1.20483959e+00 -5.18877864e-01 4.29344028e-01
4.10196006e-01 -1.07426965e+00 1.09165406e+00 2.08395734e-01
4.12273854e-01 -6.18133187e-01 -2.64371395e-01 -9.85069990e-01
1.53118586e-02 -7.68264949e-01 -1.05661489e-01 9.19738233e-01
6.26528561e-01 -6.43764615e-01 1.25855577e+00 3.06027144e-01
-3.73088807e-01 -8.90779316e-01 -6.74693346e-01 -4.59315091e-01
-2.96272159e-01 -6.58239603e-01 4.59300905e-01 3.80631328e-01
-8.45231056e-01 3.09227943e-01 -1.28824458e-01 1.00815380e+00
9.35814440e-01 4.71012294e-01 9.57431018e-01 -1.06587982e+00
-4.28518683e-01 -2.36525178e-01 -2.07334861e-01 -1.78633094e+00
-2.46594235e-01 -5.54182112e-01 4.72691566e-01 -1.65572357e+00
-3.49620342e-01 -4.82062191e-01 3.16828668e-01 2.64677703e-01
-2.56765515e-01 5.25227040e-02 -1.46840096e-01 6.95131719e-02
-4.27344531e-01 4.05249178e-01 1.20283592e+00 1.21991418e-01
-5.20034254e-01 2.50555575e-01 -3.31349015e-01 1.12696612e+00
6.05785847e-01 -3.35403979e-01 -3.61840904e-01 -1.06392348e+00
3.96085978e-01 1.98321685e-01 3.00718486e-01 -1.22866917e+00
2.49950483e-01 -3.80322576e-01 3.58150065e-01 -1.09914124e+00
4.98032391e-01 -3.11714321e-01 -3.74211073e-01 1.66238863e-02
1.49981722e-01 -2.00012952e-01 2.77260453e-01 6.45923734e-01
-2.98584867e-02 2.09678896e-02 8.93514216e-01 3.93564142e-02
-8.37352872e-01 6.64053738e-01 -2.40750402e-01 3.40501338e-01
9.91608620e-01 -3.82712096e-01 -2.51830053e-02 -2.33647630e-01
-1.77225515e-01 7.45091140e-01 7.06918597e-01 3.10739167e-02
6.27965510e-01 -6.03728175e-01 -6.29252017e-01 3.44824284e-01
2.42406517e-01 1.02890337e+00 3.56483132e-01 7.18715906e-01
-1.06133342e+00 9.11293253e-02 3.90264302e-01 -9.38968003e-01
-8.71439278e-01 -2.10045546e-01 7.27907777e-01 -2.03449234e-01
-8.90659094e-01 1.31647360e+00 3.30929726e-01 -2.20400393e-01
1.99021816e-01 -7.80931234e-01 1.01440623e-01 -8.29963982e-02
3.85444611e-01 5.51441133e-01 1.00051723e-01 -3.32893044e-01
-3.13244343e-01 1.20774758e+00 1.28591195e-01 -3.67324680e-01
1.03933716e+00 -5.39632857e-01 1.63784713e-01 2.68852204e-01
1.07708228e+00 1.43291265e-01 -2.00037885e+00 -1.41531965e-02
-1.84914887e-01 -7.29060948e-01 5.63111305e-01 -5.45676053e-01
-1.00257897e+00 8.44448924e-01 3.65119725e-01 -2.92905331e-01
1.02579629e+00 -1.28080621e-01 1.01134515e+00 2.59104490e-01
4.39641386e-01 -9.05583084e-01 -1.33013010e-01 7.26925552e-01
3.24663639e-01 -1.21499264e+00 1.94254383e-01 -9.10344839e-01
-3.50551337e-01 1.00903070e+00 6.31653011e-01 -3.29869896e-01
4.88747120e-01 4.02556330e-01 -8.11762922e-03 -3.56319964e-01
-5.04813790e-01 -3.21714342e-01 3.24844837e-01 4.04682815e-01
1.07786909e-01 -7.18372241e-02 3.16990823e-01 1.50446698e-01
-2.08830178e-01 -2.62619942e-01 4.72579807e-01 9.13213551e-01
-8.80094230e-01 -6.47023618e-01 -4.82411593e-01 4.99824882e-01
-2.05219164e-01 -1.84622988e-01 5.73423468e-02 5.17551243e-01
7.51550943e-02 7.01798022e-01 5.09731114e-01 -2.60338336e-01
3.86211216e-01 -6.06410742e-01 2.98103184e-01 -8.64139676e-01
-5.78210242e-02 2.24963263e-01 2.51986772e-01 -7.67081141e-01
1.16807498e-01 -7.08502650e-01 -1.48674655e+00 2.29406748e-02
-4.39796634e-02 -8.54551643e-02 7.05798864e-01 9.18562531e-01
2.34157190e-01 3.11779499e-01 7.48026669e-01 -1.25157976e+00
-6.68296888e-02 -5.60742915e-01 -6.58982158e-01 -2.61777371e-01
7.85757661e-01 -5.33372283e-01 -7.95635581e-01 -1.38403326e-02] | [8.372243881225586, -2.5666520595550537] |
45fd2aca-8a3d-417e-8a3f-f1f79b8ccb40 | evaluating-the-effectiveness-of-pre-trained | 2302.10199 | null | https://arxiv.org/abs/2302.10199v1 | https://arxiv.org/pdf/2302.10199v1.pdf | Evaluating the Effectiveness of Pre-trained Language Models in Predicting the Helpfulness of Online Product Reviews | Businesses and customers can gain valuable information from product reviews. The sheer number of reviews often necessitates ranking them based on their potential helpfulness. However, only a few reviews ever receive any helpfulness votes on online marketplaces. Sorting all reviews based on the few existing votes can cause helpful reviews to go unnoticed because of the limited attention span of readers. The problem of review helpfulness prediction is even more important for higher review volumes, and newly written reviews or launched products. In this work we compare the use of RoBERTa and XLM-R language models to predict the helpfulness of online product reviews. The contributions of our work in relation to literature include extensively investigating the efficacy of state-of-the-art language models -- both monolingual and multilingual -- against a robust baseline, taking ranking metrics into account when assessing these approaches, and assessing multilingual models for the first time. We employ the Amazon review dataset for our experiments. According to our study on several product categories, multilingual and monolingual pre-trained language models outperform the baseline that utilizes random forest with handcrafted features as much as 23% in RMSE. Pre-trained language models reduce the need for complex text feature engineering. However, our results suggest that pre-trained multilingual models may not be used for fine-tuning only one language. We assess the performance of language models with and without additional features. Our results show that including additional features like product rating by the reviewer can further help the predictive methods. | ['Dimitar Shterionov', 'Javad PourMostafa Roshan Sharami', 'Ali Boluki'] | 2023-02-19 | null | null | null | null | ['feature-engineering', 'xlm-r'] | ['methodology', 'natural-language-processing'] | [-4.47470784e-01 4.81878668e-02 -6.78878307e-01 -5.06146073e-01
-9.32700276e-01 -8.05668533e-01 9.63683963e-01 5.90416431e-01
-6.80505276e-01 6.06388867e-01 3.33925873e-01 -6.45687222e-01
4.58144955e-02 -6.51914001e-01 -3.00006956e-01 -1.28122613e-01
3.79657418e-01 4.03040886e-01 -2.50192821e-01 -6.71344817e-01
4.85615641e-01 2.82148033e-01 -1.34482539e+00 3.38792801e-01
1.09054768e+00 7.80904949e-01 2.94640332e-01 3.55785698e-01
-4.86967295e-01 8.73491883e-01 -4.49307501e-01 -8.37538183e-01
4.83762711e-01 -1.28169030e-01 -3.24093997e-01 -1.45523891e-01
3.53595704e-01 -1.89579785e-01 4.13640350e-01 6.35283411e-01
2.50472397e-01 -7.92579651e-02 6.26628876e-01 -8.64655554e-01
-1.11450267e+00 8.68115306e-01 -7.86652625e-01 1.08776003e-01
4.64306593e-01 -8.67307335e-02 1.59010530e+00 -1.20979512e+00
6.30154371e-01 1.05224800e+00 6.98053598e-01 -1.38408933e-02
-8.08869481e-01 -5.41541874e-01 4.78282154e-01 -2.14838982e-03
-9.77610588e-01 -2.78104603e-01 6.21290684e-01 -6.35881186e-01
1.11234450e+00 2.82223932e-02 4.95543927e-01 5.82072139e-01
3.97296041e-01 4.47771102e-01 1.50445700e+00 -6.35487795e-01
-2.69013294e-03 8.57691169e-01 4.21620399e-01 3.55920225e-01
5.11930287e-01 -1.99942738e-01 -5.47654748e-01 -1.05422974e-01
7.79429823e-02 3.06671858e-02 3.61657232e-01 -6.45546243e-02
-8.64280164e-01 1.24302626e+00 1.81232855e-01 4.56862003e-01
-4.78032440e-01 -3.25997442e-01 5.03241837e-01 6.20795071e-01
1.03267813e+00 8.02652240e-01 -1.02240717e+00 -1.93907514e-01
-1.07354414e+00 -7.36212432e-02 9.11705196e-01 6.48994803e-01
1.05441165e+00 -1.52389258e-01 1.92340925e-01 1.16117477e+00
5.22393346e-01 5.76664627e-01 6.64371252e-01 -3.64896923e-01
4.13199276e-01 8.91062200e-01 -8.71711131e-03 -1.09773326e+00
-3.85796368e-01 -4.16458815e-01 -3.26415181e-01 -1.26641810e-01
1.83122277e-01 -1.75901756e-01 -6.42674923e-01 1.09711671e+00
-2.48290077e-02 -6.94586575e-01 -5.82456477e-02 6.69839442e-01
9.21331882e-01 5.94085813e-01 1.11930870e-01 -3.86264563e-01
1.19076955e+00 -1.17901766e+00 -5.38124144e-01 -4.00606751e-01
8.89990032e-01 -1.33359289e+00 1.49929106e+00 6.09211922e-01
-8.93003404e-01 -6.81694210e-01 -9.95133817e-01 -1.41129941e-01
-7.39377618e-01 5.94252884e-01 8.98928404e-01 7.60398507e-01
-1.01606536e+00 3.68252426e-01 -3.51717979e-01 -4.89000797e-01
-1.78841442e-01 3.40795308e-01 -3.60758156e-01 -1.44891828e-01
-1.22203481e+00 1.48119402e+00 -4.34252679e-01 1.35692075e-01
-3.98715436e-01 -5.43658674e-01 -8.23629022e-01 -3.44532073e-01
1.52568102e-01 -2.22716987e-01 1.20517850e+00 -1.31245029e+00
-1.34193850e+00 7.57237852e-01 -3.05391908e-01 -2.14782417e-01
1.65832520e-01 -3.95823568e-01 -6.85589373e-01 -2.44915828e-01
4.63193953e-01 4.74613279e-01 5.43405294e-01 -9.84400511e-01
-7.74384737e-01 -2.77859896e-01 2.78351039e-01 1.26592204e-01
-4.39715922e-01 3.75313491e-01 -3.27837288e-01 -4.78864670e-01
-3.37110519e-01 -9.59863782e-01 -5.19900262e-01 -5.47768116e-01
3.40472199e-02 -2.61814415e-01 2.31727675e-01 -7.85822749e-01
1.41759026e+00 -1.73937666e+00 -4.27380234e-01 2.11656064e-01
-3.34389023e-02 1.90572709e-01 -4.50664431e-01 5.45170605e-01
-4.99184877e-02 5.28356910e-01 3.83948594e-01 -1.46066040e-01
-4.29097004e-02 -5.84943481e-02 -2.08108217e-01 1.60928413e-01
2.15337038e-01 1.05565143e+00 -7.02977896e-01 -2.16438040e-01
7.75519833e-02 5.38801730e-01 -3.66492391e-01 -1.13405742e-01
2.08162263e-01 -8.10096636e-02 -2.72866964e-01 6.24180317e-01
3.64611268e-01 -1.31128222e-01 8.58238712e-02 2.00034752e-02
-3.28689963e-01 9.85689223e-01 -8.14112723e-01 1.08176041e+00
-1.13647699e+00 4.98488396e-01 -1.29863545e-01 -5.23673773e-01
1.24295080e+00 -5.59280999e-02 4.26490933e-01 -1.03842556e+00
-1.92277487e-02 6.13572598e-01 1.87436596e-01 -1.46737412e-01
8.76896799e-01 -2.76509464e-01 -1.32468998e-01 7.30963588e-01
-2.39648357e-01 -2.22827911e-01 5.78670144e-01 3.16821426e-01
7.60187745e-01 -2.07401086e-02 5.94565809e-01 -2.61927277e-01
6.88583076e-01 1.10787138e-01 2.22822547e-01 5.01456380e-01
-3.51723619e-02 4.01978672e-01 4.97541189e-01 -3.34049702e-01
-9.92293000e-01 -4.60068345e-01 -1.64848551e-01 1.37140083e+00
-3.64915937e-01 -7.68150151e-01 -3.07066530e-01 -8.81266594e-01
2.49906451e-01 7.90612221e-01 -5.80366492e-01 6.31977469e-02
-1.49212927e-01 -7.17705131e-01 4.28566262e-02 3.28431636e-01
-1.42230153e-01 -9.77267683e-01 -7.66630545e-02 2.17056930e-01
1.77594706e-01 -8.74683440e-01 -4.49267626e-01 4.03956890e-01
-8.66041422e-01 -8.82346630e-01 -7.31633782e-01 -5.58696985e-01
5.49383938e-01 4.34503198e-01 1.43333161e+00 -6.35817498e-02
6.51502162e-02 4.83365685e-01 -6.26304865e-01 -5.05808830e-01
-4.62132335e-01 5.21551073e-01 6.67003691e-02 -3.80475312e-01
9.79937494e-01 -1.65442318e-01 -2.72148639e-01 3.21312606e-01
-5.57515919e-01 -4.69888836e-01 8.56795192e-01 6.58407807e-01
3.37146491e-01 -2.15329245e-01 1.22065163e+00 -1.26213491e+00
1.18554533e+00 -5.56160390e-01 -3.62007052e-01 2.96738058e-01
-1.34081388e+00 1.41037226e-01 5.72874904e-01 -3.70598823e-01
-8.58740985e-01 -1.97444960e-01 -1.15095265e-01 4.14534777e-01
2.38532871e-01 1.09793305e+00 2.28427008e-01 4.77618314e-02
7.48319924e-01 -3.43545973e-01 3.69053297e-02 -4.31251466e-01
5.86356461e-01 8.39733243e-01 -3.93464863e-01 -3.44142646e-01
5.95312774e-01 -3.36116590e-02 -4.78410363e-01 -7.52071023e-01
-8.79145384e-01 -9.08888876e-01 -8.16189706e-01 -6.45047203e-02
2.93793082e-01 -1.08163655e+00 -2.95301795e-01 1.17458077e-02
-9.73618329e-01 6.16325177e-02 -2.96017259e-01 5.29301584e-01
5.82588688e-02 3.80074054e-01 -7.48091161e-01 -8.13842952e-01
-5.83321810e-01 -1.25496459e+00 9.12880123e-01 -6.40028790e-02
-5.99834442e-01 -1.19293749e+00 2.82344788e-01 6.97061777e-01
6.31100535e-01 -3.32806230e-01 8.37480307e-01 -9.27999675e-01
-1.16024531e-01 -5.41339040e-01 -8.07202905e-02 7.26916671e-01
3.67096364e-01 2.90166378e-01 -7.71483064e-01 3.37321237e-02
-7.88230598e-02 -3.78299147e-01 9.02128518e-01 3.63571852e-01
1.72677457e-01 -1.92102194e-01 1.85449108e-01 -2.79945076e-01
1.21153021e+00 -7.12382328e-03 3.21867585e-01 5.79578102e-01
7.25255728e-01 1.00004506e+00 1.19323015e+00 4.12392527e-01
7.51673281e-01 4.86698717e-01 9.98454019e-02 -1.27934739e-01
1.90275759e-01 -2.10556835e-01 9.07935917e-01 1.43406034e+00
1.23089336e-01 1.87124670e-01 -7.75967002e-01 6.18232489e-01
-1.49469340e+00 -5.46104372e-01 -3.86078924e-01 2.32840538e+00
9.78131473e-01 2.62443036e-01 2.29358852e-01 -8.40932429e-02
3.39361638e-01 1.63898125e-01 -2.33777791e-01 -1.16275895e+00
-1.49109557e-01 2.85389513e-01 6.26037180e-01 8.07269752e-01
-6.94400966e-01 9.38184679e-01 6.25927496e+00 5.82711160e-01
-1.21511388e+00 1.61396906e-01 7.51206160e-01 -1.81772500e-01
-8.02853227e-01 2.33373255e-01 -1.12431788e+00 1.09336354e-01
1.11402440e+00 -1.98005155e-01 2.75967360e-01 1.13919854e+00
5.51271856e-01 -3.82096708e-01 -9.24938321e-01 7.08139360e-01
4.20262605e-01 -8.29211533e-01 -3.97761464e-02 1.37930959e-01
9.14199412e-01 2.83531874e-01 1.66177288e-01 6.25050783e-01
3.89957964e-01 -1.03753126e+00 6.40219331e-01 2.58682698e-01
4.02370363e-01 -8.59520197e-01 1.00225770e+00 2.53799587e-01
-1.09027767e+00 1.02291713e-02 -6.48313522e-01 -3.94730330e-01
1.98215649e-01 1.09614706e+00 -8.53957534e-01 4.24755335e-01
3.80481154e-01 1.27299845e+00 -1.07707584e+00 5.52291811e-01
-3.20391804e-01 6.74375236e-01 -1.63740367e-01 -3.92330080e-01
3.85897338e-01 -4.59404290e-01 -1.06067561e-01 1.14801967e+00
3.88737142e-01 -3.99715245e-01 1.69112086e-01 4.18363780e-01
1.71528593e-01 1.03409076e+00 -7.79876530e-01 -3.77086669e-01
1.01221472e-01 1.65151310e+00 -5.29891551e-01 -6.18241578e-02
-8.63444924e-01 5.09781182e-01 1.94393262e-01 6.62977770e-02
-4.41357344e-01 -1.10517636e-01 2.04615146e-01 4.19671744e-01
4.74349298e-02 -1.28994405e-01 -5.77443361e-01 -1.03168643e+00
9.79252979e-02 -1.18469512e+00 6.91542104e-02 -7.16621935e-01
-1.55537283e+00 8.02293062e-01 -3.78889382e-01 -1.08523154e+00
-4.28244501e-01 -8.64570558e-01 -2.38367617e-01 1.15633392e+00
-1.90353656e+00 -1.34384239e+00 2.14110985e-01 9.99219939e-02
5.92156529e-01 -4.02277857e-01 5.89837015e-01 4.61081117e-01
-2.17901185e-01 4.89707202e-01 -1.07480019e-01 -3.81881505e-01
1.16006744e+00 -1.25698149e+00 2.68131554e-01 5.48816860e-01
3.95543247e-01 1.12453341e+00 4.30056006e-01 -8.56310487e-01
-1.01714969e+00 -6.43137395e-01 1.66186750e+00 -8.68587911e-01
1.05181003e+00 -1.91578001e-01 -5.99504173e-01 3.21868271e-01
5.13327301e-01 -5.49428582e-01 1.01650560e+00 9.40070271e-01
-4.74372625e-01 -3.36823016e-01 -5.70307434e-01 4.42339540e-01
2.81856120e-01 -7.38335431e-01 -4.12502617e-01 2.91952878e-01
4.82650548e-01 3.92122298e-01 -1.00337493e+00 1.76167905e-01
7.15825677e-01 -7.87135541e-01 5.14528036e-01 -3.16680104e-01
7.95015037e-01 -1.43967092e-01 3.10846586e-02 -1.41011333e+00
-2.59899735e-01 -2.93327332e-01 4.89781737e-01 1.42441201e+00
1.19351721e+00 -6.07093036e-01 4.26101089e-01 6.96616650e-01
1.43516630e-01 -9.09753740e-01 -2.89110720e-01 -5.52731037e-01
3.54629666e-01 -6.12822711e-01 3.59709829e-01 8.45965445e-01
2.51895815e-01 8.94033372e-01 -2.89534271e-01 -5.13947010e-01
-2.75373310e-01 -5.59272133e-02 9.57720280e-01 -1.26614940e+00
-1.84730574e-01 -5.96637309e-01 -1.13021780e-03 -7.61837602e-01
3.61935914e-01 -9.14756238e-01 -1.44950613e-01 -1.68537414e+00
3.45033526e-01 -7.02303886e-01 -2.48680711e-01 5.65355301e-01
-2.39989534e-01 1.83805555e-01 2.15497985e-01 2.25219950e-01
-5.24556756e-01 1.96442172e-01 1.13364124e+00 -1.48967445e-01
-3.26193810e-01 2.34688550e-01 -1.37491405e+00 6.54308856e-01
7.37506807e-01 -3.85348290e-01 -5.59834003e-01 -1.47282198e-01
1.19736481e+00 -5.05192518e-01 -2.40510777e-01 -1.46758348e-01
-4.10889722e-02 -2.25906800e-02 6.28334880e-02 -5.13758183e-01
-8.34170729e-02 -8.32186639e-01 -1.88272640e-01 -3.98493670e-02
-6.36092246e-01 6.56845391e-01 3.56399380e-02 2.26573750e-01
-4.60803747e-01 -5.67336023e-01 2.33793914e-01 -1.39913663e-01
-4.56677496e-01 -1.85475841e-01 -4.84140694e-01 -1.20976865e-01
6.38127089e-01 -1.49515986e-01 -2.70267099e-01 -7.15317249e-01
-3.03014755e-01 4.76471223e-02 5.37163794e-01 8.41384590e-01
1.86004668e-01 -1.01054049e+00 -7.54665971e-01 -5.46448343e-02
4.07177985e-01 -6.96274936e-01 -2.54737169e-01 8.49068046e-01
-2.65592515e-01 8.30679119e-01 1.41710147e-01 -1.45071089e-01
-1.23857129e+00 5.88803530e-01 -2.30622459e-02 -8.91046941e-01
2.22942978e-01 4.66553539e-01 -4.26837280e-02 -8.97460043e-01
-1.25756502e-01 -4.52554047e-01 -7.08239913e-01 7.31507719e-01
3.89287919e-01 4.16708767e-01 4.40681010e-01 -8.52456510e-01
-3.02666098e-01 6.55698836e-01 -5.32657325e-01 -3.42574388e-01
1.38914537e+00 -2.33986571e-01 -4.42857236e-01 8.15057695e-01
1.05671036e+00 7.02425361e-01 -5.64601660e-01 -2.80477647e-02
2.87363976e-01 -1.95544943e-01 1.13657057e-01 -1.17721403e+00
-1.06339121e+00 9.16181564e-01 2.03241736e-01 2.57466912e-01
8.82936358e-01 -1.50948599e-01 3.88555646e-01 3.80344629e-01
4.65445161e-01 -1.31190526e+00 -6.07645251e-02 6.18726373e-01
8.41553509e-01 -1.57345974e+00 4.18191433e-01 -2.28628367e-01
-1.22224474e+00 8.77531886e-01 4.98125553e-01 1.62700890e-03
9.58847940e-01 1.31453112e-01 5.78882277e-01 -9.80747044e-02
-8.64942193e-01 -2.12212384e-01 6.67517662e-01 3.19563776e-01
1.23384309e+00 4.19805236e-02 -1.03624773e+00 8.90547693e-01
-3.32412809e-01 -1.85499281e-01 6.58745766e-01 6.05521679e-01
-3.07481498e-01 -1.65666056e+00 -2.66758166e-02 8.07643116e-01
-7.41501629e-01 -6.45332575e-01 -7.71812916e-01 8.26577306e-01
4.53933775e-02 1.33223414e+00 -4.54696149e-01 -2.47277275e-01
3.36102754e-01 -1.15187675e-01 -6.16771355e-02 -1.10031545e+00
-1.29138803e+00 2.23212704e-01 3.57448906e-01 -1.19090848e-01
-6.77682400e-01 -8.03529143e-01 -7.58268058e-01 -2.26121694e-01
-8.27211261e-01 3.24217767e-01 9.95977938e-01 8.80561590e-01
3.57674301e-01 8.31901357e-02 7.48733640e-01 -4.20115203e-01
-6.76333666e-01 -1.13815641e+00 -6.12661064e-01 1.86135963e-01
-6.43621245e-03 -3.41468990e-01 -5.51050067e-01 -8.35281461e-02] | [11.137591361999512, 6.926249027252197] |
5e82cd64-0ad0-4990-856b-70bc6cf61e65 | videocapsulenet-a-simplified-network-for | 1805.08162 | null | http://arxiv.org/abs/1805.08162v1 | http://arxiv.org/pdf/1805.08162v1.pdf | VideoCapsuleNet: A Simplified Network for Action Detection | The recent advances in Deep Convolutional Neural Networks (DCNNs) have shown
extremely good results for video human action classification, however, action
detection is still a challenging problem. The current action detection
approaches follow a complex pipeline which involves multiple tasks such as tube
proposals, optical flow, and tube classification. In this work, we present a
more elegant solution for action detection based on the recently developed
capsule network. We propose a 3D capsule network for videos, called
VideoCapsuleNet: a unified network for action detection which can jointly
perform pixel-wise action segmentation along with action classification. The
proposed network is a generalization of capsule network from 2D to 3D, which
takes a sequence of video frames as input. The 3D generalization drastically
increases the number of capsules in the network, making capsule routing
computationally expensive. We introduce capsule-pooling in the convolutional
capsule layer to address this issue which makes the voting algorithm tractable.
The routing-by-agreement in the network inherently models the action
representations and various action characteristics are captured by the
predicted capsules. This inspired us to utilize the capsules for action
localization and the class-specific capsules predicted by the network are used
to determine a pixel-wise localization of actions. The localization is further
improved by parameterized skip connections with the convolutional capsule
layers and the network is trained end-to-end with a classification as well as
localization loss. The proposed network achieves sate-of-the-art performance on
multiple action detection datasets including UCF-Sports, J-HMDB, and UCF-101
(24 classes) with an impressive ~20% improvement on UCF-101 and ~15%
improvement on J-HMDB in terms of v-mAP scores. | ['Yogesh S Rawat', 'Mubarak Shah', 'Kevin Duarte'] | 2018-05-21 | videocapsulenet-a-simplified-network-for-1 | http://papers.nips.cc/paper/7988-videocapsulenet-a-simplified-network-for-action-detection | http://papers.nips.cc/paper/7988-videocapsulenet-a-simplified-network-for-action-detection.pdf | neurips-2018-12 | ['multiple-action-detection'] | ['computer-vision'] | [ 5.96000366e-02 1.68639511e-01 -4.13894922e-01 -1.49936318e-01
-5.97413898e-01 -6.09224617e-01 3.49201173e-01 -2.82259285e-01
-4.88347977e-01 1.71662793e-01 5.74888647e-01 2.26621002e-01
1.62114769e-01 -4.18099344e-01 -8.42561245e-01 -7.40810335e-01
-5.11806130e-01 -7.94422030e-02 6.69690311e-01 3.30381960e-01
-4.94610593e-02 4.53359574e-01 -1.04337025e+00 8.04754674e-01
5.23382604e-01 1.47640491e+00 2.60606140e-01 7.08353281e-01
2.22673729e-01 1.31653583e+00 -5.15605211e-01 -2.06113368e-01
4.41290468e-01 -3.78600448e-01 -7.50739694e-01 3.94896030e-01
5.54171920e-01 -5.32876849e-01 -7.54292667e-01 1.07013059e+00
4.28001046e-01 -1.04062870e-01 3.86755615e-01 -9.12729681e-01
-3.18632394e-01 5.80082595e-01 -8.29804912e-02 2.49876425e-01
4.27755833e-01 3.63106072e-01 5.70762992e-01 -7.28546858e-01
8.80548179e-01 1.31817734e+00 6.13301039e-01 7.07208693e-01
-6.26424491e-01 -3.19456458e-01 2.51212060e-01 -1.41056046e-01
-1.04479897e+00 -1.41368315e-01 5.54129660e-01 -4.46262985e-01
1.10663223e+00 1.12219676e-01 1.07751727e+00 1.16902769e+00
4.41535503e-01 1.42091262e+00 5.08332551e-01 1.92921236e-01
1.45904437e-01 -2.08463654e-01 -4.21658754e-01 1.06865036e+00
1.51710108e-01 -1.97639003e-01 -2.70660043e-01 3.49835873e-01
1.37128401e+00 2.72428244e-01 -6.32179797e-01 -5.12509346e-01
-1.61265814e+00 6.63261831e-01 8.82462442e-01 2.23879009e-01
-1.91405520e-01 4.37015176e-01 6.29097879e-01 5.17666191e-02
3.09900701e-01 6.73262358e-01 -5.95820963e-01 -7.22226351e-02
-7.97730565e-01 1.37299195e-01 9.97663736e-01 7.58085012e-01
-1.62060007e-01 -2.28122827e-02 -6.88219368e-01 5.93829870e-01
5.07438838e-01 1.47670835e-01 6.39494479e-01 -1.22535467e+00
5.02572536e-01 9.05974567e-01 2.54266430e-02 -7.22740531e-01
-6.70569181e-01 -6.51911974e-01 -8.16299796e-01 4.12334532e-01
5.28950870e-01 -8.65927488e-02 -1.18380225e+00 1.29831123e+00
2.20976472e-01 3.45460981e-01 -1.87215339e-02 1.19140923e+00
1.11916232e+00 9.99698862e-02 5.62063158e-02 -2.47910470e-02
1.20060587e+00 -1.77209520e+00 -4.60544884e-01 -2.09082842e-01
9.20710802e-01 -6.28574908e-01 5.14030218e-01 4.10705447e-01
-1.12243795e+00 -6.56320155e-01 -1.15962410e+00 1.57781262e-02
-2.38393739e-01 5.62265158e-01 9.30737853e-01 5.38924932e-01
-1.03597629e+00 8.23463082e-01 -1.37876284e+00 -8.90101641e-02
1.01648879e+00 7.65909493e-01 -5.67873478e-01 -1.19842798e-01
-7.60482073e-01 4.75192130e-01 3.92094731e-01 1.81885347e-01
-1.41572976e+00 -4.92605865e-01 -1.07331038e+00 4.05778922e-02
5.09482026e-01 -7.00744510e-01 1.29549420e+00 -8.64110351e-01
-1.52446139e+00 7.82137871e-01 2.72887945e-01 -7.21460938e-01
8.65600228e-01 -1.24039873e-01 -1.48927584e-01 6.12156808e-01
4.66910191e-02 9.43991184e-01 5.79391658e-01 -4.73483473e-01
-5.03924787e-01 -1.18841767e-01 4.28169698e-01 1.08115666e-01
-7.29511529e-02 7.17290491e-02 -9.69464660e-01 -1.03716886e+00
3.36024493e-01 -1.01364064e+00 -3.19272369e-01 7.57232666e-01
-4.32987750e-01 -3.94448578e-01 6.28255546e-01 -4.68879133e-01
9.42759335e-01 -2.11049819e+00 2.19201028e-01 -3.37991893e-01
4.32215571e-01 1.88845485e-01 -2.08872944e-01 -3.84238750e-01
-1.35730028e-01 3.04732490e-02 -2.21258849e-02 -5.77888310e-01
-2.98850060e-01 -2.48411484e-02 2.40865260e-01 9.57376003e-01
1.58942223e-01 8.59667599e-01 -9.87324715e-01 -4.65902060e-01
5.84865153e-01 4.72386211e-01 -8.12171578e-01 3.01607341e-01
-3.52300853e-01 6.30465150e-01 -6.48267508e-01 1.10984445e+00
5.20824671e-01 -5.54012477e-01 2.61039659e-02 -3.25886995e-01
2.96912193e-02 1.45466074e-01 -1.14891970e+00 2.16904283e+00
-2.71009784e-02 3.67347509e-01 1.66453406e-01 -9.28475142e-01
4.46032643e-01 4.98652875e-01 1.10215461e+00 -2.41533503e-01
3.45761091e-01 3.24852139e-01 1.64605916e-01 -8.23172390e-01
-1.28232419e-01 4.71963763e-01 4.00792509e-02 -6.91489130e-02
2.87555039e-01 1.47826284e-01 3.20199400e-01 -3.48061472e-02
1.27054715e+00 6.55309498e-01 1.29510731e-01 2.03856044e-02
7.34382212e-01 1.43094771e-02 7.52399802e-01 7.34960794e-01
-5.48774898e-01 8.12633276e-01 7.19204843e-01 -6.27666891e-01
-6.22855246e-01 -8.16036642e-01 -9.80728939e-02 5.27382910e-01
4.20097083e-01 -4.86763000e-01 -7.76807129e-01 -1.13946784e+00
1.44094959e-01 -2.07873926e-01 -7.58180499e-01 -2.40903404e-02
-7.33455598e-01 -6.25184298e-01 3.50547642e-01 8.96158814e-01
9.29485261e-01 -1.08152270e+00 -6.18269205e-01 3.90189946e-01
9.00427103e-02 -1.53133440e+00 -5.45815527e-01 8.00177082e-02
-1.03036070e+00 -1.43686843e+00 -9.57550585e-01 -9.42017496e-01
7.44685590e-01 1.54914007e-01 6.87596023e-01 -8.91972482e-02
-4.73940462e-01 2.58057296e-01 -3.15228730e-01 -4.09764871e-02
-1.69108927e-01 -3.29924107e-01 -9.46983844e-02 3.46139707e-02
7.83319622e-02 -1.84218496e-01 -1.16521156e+00 3.95756245e-01
-6.94411695e-01 -1.44127265e-01 8.32183242e-01 4.38571930e-01
7.16299653e-01 -2.93492436e-01 -2.90007256e-02 -4.31809634e-01
-1.22363575e-01 -2.43463069e-01 -5.74599147e-01 2.83155233e-01
1.26310989e-01 -2.12550331e-02 5.47159433e-01 -6.98223531e-01
-8.10551643e-01 7.07987487e-01 -1.59015208e-02 -8.63546848e-01
1.43794511e-02 1.85756814e-02 7.24571338e-03 -6.62944973e-01
4.15409297e-01 1.66020542e-01 2.35869270e-02 -5.71555912e-01
1.67007342e-01 5.66195190e-01 5.38846195e-01 7.02492520e-02
1.71568468e-01 7.89443612e-01 -8.98801908e-02 -3.69010121e-01
-1.05256474e+00 -6.72863245e-01 -6.76921785e-01 -4.71980423e-01
1.61623681e+00 -1.37999713e+00 -9.13979113e-01 6.56835139e-01
-1.14333653e+00 -3.35167617e-01 3.60813998e-02 1.04892182e+00
-5.75101256e-01 4.33434248e-01 -8.43871176e-01 -3.61863345e-01
-3.16155374e-01 -1.76474512e+00 1.25589275e+00 2.24168956e-01
-3.12446523e-02 -8.06995451e-01 -4.60968912e-01 4.96880502e-01
3.62308890e-01 6.88650548e-01 2.48607904e-01 -6.71892226e-01
-1.18104649e+00 -4.16003644e-01 -7.79810995e-02 4.67873961e-01
-1.29836518e-02 -3.71990204e-01 -8.53475094e-01 -4.13129896e-01
1.81491971e-01 -3.98320675e-01 1.13621926e+00 9.20121133e-01
1.64395654e+00 2.61597075e-02 -6.14055872e-01 1.14525151e+00
1.22768319e+00 2.32818693e-01 5.27025461e-01 3.39915812e-01
7.09185421e-01 9.00245532e-02 2.75621980e-01 4.08768237e-01
-8.24582577e-02 6.70530915e-01 8.93515110e-01 -1.56205729e-01
-6.60918832e-01 3.10227603e-01 6.62294328e-01 3.54039848e-01
-1.40277058e-01 -2.36620270e-02 -4.58762586e-01 3.55709612e-01
-1.72788858e+00 -7.73467302e-01 -9.80858803e-02 2.01849341e+00
5.03084779e-01 3.52208883e-01 1.42417839e-02 -4.19292539e-01
3.88580412e-01 3.43025148e-01 -5.36053777e-01 2.99193770e-01
6.33876771e-02 -3.27841267e-02 6.94840670e-01 4.68664058e-02
-2.20208240e+00 7.59920716e-01 5.82703972e+00 4.90690112e-01
-1.20541954e+00 1.21677309e-01 5.77549160e-01 -4.25729007e-01
5.61825395e-01 -4.56963331e-01 -8.02361786e-01 3.87801319e-01
3.19651753e-01 3.64267915e-01 1.05270028e-01 1.07584584e+00
3.40863407e-01 -1.29696429e-01 -1.17981732e+00 1.04373252e+00
3.55390340e-01 -1.64340746e+00 -1.65940300e-01 -1.74949646e-01
6.79405093e-01 3.95185173e-01 -2.09993094e-01 3.04860413e-01
2.15961496e-04 -9.32623565e-01 6.34314716e-01 4.17507261e-01
7.27771103e-01 -1.69700161e-01 8.01989973e-01 -1.21166699e-01
-1.47510350e+00 -4.83452827e-01 -2.71184772e-01 2.55305260e-01
1.01012386e-01 6.83650374e-02 -6.11577392e-01 3.46565336e-01
8.54394794e-01 1.33030581e+00 -6.25902116e-01 1.77755094e+00
-3.58618319e-01 1.83925256e-01 -1.83652431e-01 -1.31659787e-02
5.71104288e-01 2.72230774e-01 5.48029363e-01 1.07757890e+00
2.76184320e-01 -2.34080732e-01 7.82371759e-01 7.44276822e-01
-3.33563060e-01 -2.58298129e-01 -2.06332535e-01 -9.99468863e-02
-1.79205745e-01 1.36324394e+00 -9.93550420e-01 -4.60742056e-01
-8.29267800e-01 1.16134870e+00 1.96935590e-02 1.23832047e-01
-1.10578156e+00 -1.46705344e-01 7.49743998e-01 -4.11741361e-02
6.91719353e-01 3.38581428e-02 4.63555902e-02 -1.35904968e+00
2.55367160e-01 -6.09515548e-01 2.95129031e-01 -5.77403426e-01
-9.50408936e-01 4.84714031e-01 -3.65723163e-01 -1.87925613e+00
8.45938027e-02 -1.30529928e+00 -3.51307839e-01 3.22706968e-01
-1.27373838e+00 -1.08883512e+00 -5.65613389e-01 6.01523638e-01
8.12637150e-01 -1.34516701e-01 6.26730919e-01 3.72647494e-01
-6.38086021e-01 4.71630186e-01 -1.36137336e-01 9.35949206e-01
6.64311886e-01 -1.28809392e+00 -1.72206741e-02 8.10909033e-01
-6.83761090e-02 3.16899598e-01 1.52618498e-01 -6.07401073e-01
-1.44276786e+00 -1.46163511e+00 3.24152738e-01 -4.09549832e-01
6.51009679e-01 -3.13868910e-01 -4.17718768e-01 6.50174797e-01
2.54875012e-02 7.55311310e-01 3.76412153e-01 -6.07066691e-01
-4.29750308e-02 6.67158365e-02 -1.02224576e+00 3.87617230e-01
1.22981632e+00 -2.64489681e-01 -3.59016031e-01 1.02476180e+00
8.87492359e-01 -9.12659585e-01 -1.15179944e+00 4.36803520e-01
4.01643217e-01 -8.01672816e-01 1.07729959e+00 -5.36561370e-01
6.21687949e-01 -4.45348859e-01 1.59104928e-01 -6.51432991e-01
-2.29341745e-01 -5.29302895e-01 -2.82942384e-01 2.07718745e-01
3.31653029e-01 -1.81047583e-03 1.06626630e+00 3.92671406e-01
-7.36926138e-01 -1.26813471e+00 -1.06219304e+00 -7.51408577e-01
-2.61121452e-01 -3.01759839e-01 -4.59036566e-02 6.69312537e-01
-3.41573246e-02 -3.41685921e-01 -1.64719850e-01 1.15386166e-01
4.75619882e-01 1.83063019e-02 3.18412840e-01 -7.64074802e-01
-6.73301518e-01 -5.08098483e-01 -8.78053427e-01 -1.48342168e+00
-2.93349326e-01 -1.12718666e+00 -8.56900513e-02 -1.69280124e+00
2.23532736e-01 1.99460670e-01 -3.04748088e-01 5.45180678e-01
7.97441751e-02 1.83762476e-01 2.62337536e-01 2.89729804e-01
-1.06740522e+00 2.46028855e-01 1.87227225e+00 -4.58375394e-01
-2.77301133e-01 1.62719294e-01 -3.06257810e-02 9.22771096e-01
2.68891543e-01 -3.36677194e-01 -1.20572835e-01 -5.43158889e-01
-3.12463313e-01 2.73352057e-01 5.73928475e-01 -1.55610156e+00
3.88575435e-01 4.84071523e-01 9.69174206e-01 -4.25645828e-01
4.35337782e-01 -6.46280646e-01 -2.50254601e-01 9.74648654e-01
-3.62577200e-01 -4.68953699e-01 -8.37067440e-02 7.10154414e-01
-3.91184330e-01 3.44346762e-02 8.17090034e-01 -6.11004114e-01
-9.52800989e-01 6.44788504e-01 -3.03130031e-01 -1.65469632e-01
1.28443074e+00 -2.34961554e-01 -2.43643746e-01 -4.51575741e-02
-1.30893731e+00 3.71892780e-01 1.28828064e-01 6.77795470e-01
6.07046366e-01 -1.19618678e+00 -4.58371043e-01 8.22990462e-02
1.66613474e-01 1.64536059e-01 3.22121501e-01 1.48529208e+00
-8.39896560e-01 5.79451203e-01 2.74268314e-02 -7.75039315e-01
-1.17591238e+00 4.01032567e-01 1.10348177e+00 -3.84348035e-01
-8.90713334e-01 1.39643550e+00 2.14583904e-01 -2.18042061e-01
8.23321223e-01 -9.38647270e-01 -4.99654472e-01 -1.57768071e-01
5.61597288e-01 -7.48188049e-02 -2.14716688e-01 -2.62249976e-01
-5.29237390e-01 4.93082702e-01 2.99546495e-02 5.54174006e-01
1.38372231e+00 5.23686886e-01 -1.22317515e-01 -2.87281692e-01
1.48127830e+00 -5.03308833e-01 -1.79612017e+00 4.68694791e-02
-2.05314770e-01 -3.46677959e-01 1.54356644e-01 -9.11679089e-01
-1.59300017e+00 6.23233557e-01 8.18453610e-01 -1.83094308e-01
9.35209394e-01 3.11077237e-01 7.60442674e-01 2.71177649e-01
1.90237448e-01 -9.94990230e-01 4.70248342e-01 3.40523750e-01
9.79338169e-01 -1.40346670e+00 1.23288706e-01 -6.69689894e-01
-4.14583862e-01 1.43051231e+00 9.03585017e-01 -3.49632919e-01
5.94965994e-01 3.43520761e-01 -1.45754054e-01 -3.74090075e-01
-3.16607803e-01 -3.84857692e-02 4.16482389e-01 1.52985916e-01
4.84703600e-01 4.48528789e-02 -1.80381596e-01 5.72047472e-01
4.83655006e-01 2.25415468e-01 4.33181584e-01 8.51187050e-01
-4.93721843e-01 -7.56170869e-01 -1.56634375e-01 5.57156146e-01
-7.58576035e-01 1.37839437e-01 -3.28843415e-01 9.04678702e-01
4.76855487e-01 6.15072489e-01 7.30313063e-02 -3.00673455e-01
2.68652678e-01 -3.32596093e-01 5.05310118e-01 -6.45951211e-01
-7.46050537e-01 5.22447288e-01 8.56398791e-02 -1.31582773e+00
-8.39761674e-01 -6.02222979e-01 -1.42697024e+00 4.45378065e-01
9.94401635e-04 -4.23721582e-01 3.40770245e-01 8.82750928e-01
3.02651107e-01 8.80924165e-01 4.15626168e-01 -9.67479646e-01
-6.16700888e-01 -1.04310942e+00 -3.97134393e-01 3.94715607e-01
3.09547991e-01 -8.33001494e-01 -5.61092615e-01 9.80692580e-02] | [9.158905982971191, 0.017797349020838737] |
22e5a61a-6e05-4314-83e6-3e89f8f19456 | a-semi-autoregressive-graph-generative-model | 2306.12018 | null | https://arxiv.org/abs/2306.12018v1 | https://arxiv.org/pdf/2306.12018v1.pdf | A Semi-Autoregressive Graph Generative Model for Dependency Graph Parsing | Recent years have witnessed the impressive progress in Neural Dependency Parsing. According to the different factorization approaches to the graph joint probabilities, existing parsers can be roughly divided into autoregressive and non-autoregressive patterns. The former means that the graph should be factorized into multiple sequentially dependent components, then it can be built up component by component. And the latter assumes these components to be independent so that they can be outputted in a one-shot manner. However, when treating the directed edge as an explicit dependency relationship, we discover that there is a mixture of independent and interdependent components in the dependency graph, signifying that both aforementioned models fail to precisely capture the explicit dependencies among nodes and edges. Based on this property, we design a Semi-Autoregressive Dependency Parser to generate dependency graphs via adding node groups and edge groups autoregressively while pouring out all group elements in parallel. The model gains a trade-off between non-autoregression and autoregression, which respectively suffer from the lack of target inter-dependencies and the uncertainty of graph generation orders. The experiments show the proposed parser outperforms strong baselines on Enhanced Universal Dependencies of multiple languages, especially achieving $4\%$ average promotion at graph-level accuracy. Also, the performances of model variations show the importance of specific parts. | ['Ping Li', 'Mingming Sun', 'Ye Ma'] | 2023-06-21 | null | null | null | null | ['dependency-parsing'] | ['natural-language-processing'] | [-1.44049689e-01 6.17397726e-01 -1.44423060e-02 -5.03084183e-01
-4.37349826e-01 -6.08226120e-01 4.04407084e-01 -8.40389132e-02
-3.44179221e-03 5.39422989e-01 5.03547370e-01 -5.58370411e-01
7.44478256e-02 -9.26820815e-01 -1.00256145e+00 -7.33519554e-01
-3.23558927e-01 6.71034873e-01 1.40773162e-01 -2.68254369e-01
-4.93166298e-01 2.06367016e-01 -1.01457560e+00 4.76390570e-02
9.69418943e-01 3.58982474e-01 2.99971312e-01 6.34411514e-01
-4.95158434e-01 7.88858891e-01 -6.27082229e-01 -9.17320549e-01
1.41918659e-01 -3.88310701e-01 -5.14497936e-01 -9.00985152e-02
-4.43344340e-02 -2.71526515e-01 -3.95370007e-01 1.08148539e+00
1.66647121e-01 -1.83165729e-01 5.44750273e-01 -1.15567112e+00
-9.13226545e-01 1.49646187e+00 -7.66202152e-01 1.14146195e-01
1.43528460e-02 -2.56648928e-01 1.46964693e+00 -4.99418885e-01
4.15829182e-01 1.50563574e+00 3.84592980e-01 3.09956372e-01
-1.09708142e+00 -6.89219773e-01 8.74284744e-01 -9.71737951e-02
-8.77688825e-01 -2.44698331e-01 8.81741583e-01 -3.00076127e-01
1.24081206e+00 -4.18998413e-02 4.15890545e-01 1.16120672e+00
4.25443828e-01 7.13792205e-01 6.77317858e-01 -4.96907115e-01
3.22048850e-02 -1.88038960e-01 7.04661191e-01 7.06628442e-01
3.93326133e-01 7.87118508e-04 -4.47785884e-01 3.07687349e-03
7.24123418e-01 -3.20488751e-01 -3.22812535e-02 -1.91427171e-01
-6.50164127e-01 7.80477464e-01 4.51294690e-01 4.10163850e-01
-3.08473319e-01 3.50129530e-02 3.38490427e-01 1.50363818e-01
3.62826020e-01 -1.12457089e-01 -6.05093837e-01 1.09833024e-01
-3.85795832e-01 -1.06119387e-01 8.60458374e-01 1.34208179e+00
7.50639200e-01 2.87682980e-01 4.92981113e-02 7.19411433e-01
5.20475864e-01 2.38907650e-01 4.74257886e-01 -2.55830407e-01
8.10220182e-01 7.64420450e-01 -3.43849182e-01 -9.24291730e-01
-4.59675074e-01 -5.41309834e-01 -1.06095707e+00 -3.45079571e-01
2.47665405e-01 -5.35691857e-01 -1.23014975e+00 2.01686978e+00
2.08358765e-01 2.20942482e-01 1.17625222e-01 5.40110648e-01
7.88466096e-01 6.55487835e-01 5.54706275e-01 -3.75615925e-01
1.41417801e+00 -1.11779571e+00 -9.44132745e-01 -6.63103104e-01
6.54935002e-01 -5.50055087e-01 7.23332763e-01 1.69862300e-01
-9.71180737e-01 -6.67376816e-01 -1.03928792e+00 -8.04905519e-02
-1.06540449e-01 1.17759608e-01 1.16597283e+00 8.30129087e-01
-9.60634530e-01 4.12689924e-01 -1.17410457e+00 2.06667840e-01
-4.16586734e-02 3.64162266e-01 -5.35190642e-01 -1.25748768e-01
-1.34660900e+00 5.40694773e-01 5.35070717e-01 4.98878717e-01
-5.20795822e-01 -4.02530611e-01 -1.05254960e+00 3.70377481e-01
2.56342381e-01 -6.70814872e-01 8.99435401e-01 -9.72694099e-01
-1.43844020e+00 2.58634210e-01 -3.81489515e-01 -4.56209689e-01
1.73242599e-01 -4.33414191e-01 -4.57750887e-01 -1.84330851e-01
6.12539761e-02 1.34524375e-01 7.35149384e-01 -1.07137001e+00
-6.75831378e-01 -5.69996297e-01 3.12168509e-01 2.12730139e-01
-2.86742955e-01 1.85739949e-01 -9.37970698e-01 -5.93880713e-01
4.53132659e-01 -1.03273642e+00 -3.31786841e-01 -1.13560867e+00
-6.02260888e-01 -3.06359738e-01 4.61983144e-01 -8.68811965e-01
1.41670346e+00 -2.06173038e+00 3.85750771e-01 -4.21007946e-02
1.66563705e-01 1.82925612e-02 -2.20963851e-01 4.68581736e-01
-5.47186077e-01 3.83432955e-01 -6.34607375e-02 -4.43095803e-01
-1.71085194e-01 6.22787476e-01 -3.00992876e-01 1.18397988e-01
4.89380777e-01 9.84470427e-01 -9.45497572e-01 -4.26357567e-01
-1.43159956e-01 4.91644561e-01 -4.18770999e-01 3.85022908e-01
-1.80068552e-01 3.87823194e-01 -5.23583889e-01 3.41645598e-01
8.06378186e-01 -9.15531069e-02 8.02239537e-01 -1.59706116e-01
2.21622050e-01 4.47702557e-01 -1.25224209e+00 1.50107253e+00
-3.92582893e-01 -4.72483449e-02 1.49209678e-01 -1.06094217e+00
9.47552025e-01 3.96793097e-01 1.15844108e-01 -2.52549946e-01
-7.22009018e-02 -1.03756152e-01 3.98789912e-01 -1.03178784e-01
3.20947975e-01 -1.37153983e-01 -2.72162616e-01 2.32043192e-01
4.26637322e-01 2.26729080e-01 5.23012578e-01 5.82942009e-01
1.20513582e+00 4.20136780e-01 1.05151936e-01 -1.38334110e-01
3.89033109e-01 -2.88813829e-01 1.01068604e+00 6.04173899e-01
2.76340634e-01 4.95103627e-01 1.06349814e+00 -2.72616655e-01
-4.89560276e-01 -1.08238244e+00 3.26465964e-01 1.31306851e+00
-1.23551190e-01 -6.90186024e-01 -6.33141458e-01 -9.77977931e-01
-4.09374237e-01 6.90073073e-01 -6.33619547e-01 -5.03746644e-02
-9.70632255e-01 -1.11666453e+00 2.78443724e-01 7.50870228e-01
2.15388741e-03 -8.89057100e-01 4.21779044e-03 4.63897735e-01
-9.37175825e-02 -1.29028952e+00 -2.58325726e-01 7.41486192e-01
-8.36105824e-01 -9.84286845e-01 -3.02248806e-01 -8.28189671e-01
9.67636466e-01 1.30117491e-01 1.39085186e+00 5.68986870e-02
3.43058586e-01 -1.89443510e-02 -6.16234899e-01 -2.26218522e-01
-5.39663434e-01 1.87842831e-01 -1.30035087e-01 -2.47449093e-02
2.07893729e-01 -9.90806460e-01 -1.44946054e-01 7.53316609e-03
-7.66174018e-01 1.23544827e-01 8.58622611e-01 9.03965175e-01
5.47626436e-01 2.45693386e-01 3.32979560e-01 -1.39339983e+00
5.93550503e-01 -5.78215778e-01 -6.10125601e-01 3.96128237e-01
-4.32217091e-01 4.88450319e-01 9.27039027e-01 -2.94927806e-01
-1.38339448e+00 2.81289428e-01 -2.31940985e-01 -3.23478639e-01
4.34583761e-02 7.82341719e-01 -6.79842770e-01 5.64899981e-01
1.40953958e-01 -1.81576461e-02 -5.55011690e-01 -5.46751618e-01
7.12707639e-01 1.90814450e-01 3.18874180e-01 -8.21579993e-01
8.41926396e-01 -1.05099894e-01 1.48556670e-02 -5.93215883e-01
-5.62003195e-01 -1.78349927e-01 -8.05720210e-01 1.39847636e-01
8.43409598e-01 -1.01410973e+00 -2.78803319e-01 4.32548404e-01
-1.48763037e+00 1.84492506e-02 4.80492972e-02 4.28309232e-01
4.32847328e-02 6.74810588e-01 -8.95198405e-01 -9.36187506e-01
-2.80107468e-01 -1.27394617e+00 8.41377497e-01 2.88573384e-01
1.05819739e-01 -9.12408054e-01 -2.62965797e-03 3.71586829e-02
-4.49246578e-02 -7.02791065e-02 1.40640068e+00 -9.53004718e-01
-4.23089266e-01 -1.98860720e-01 -6.32420331e-02 4.01017845e-01
9.89929959e-02 3.09842944e-01 -6.87059999e-01 -5.83223850e-02
-1.18328802e-01 5.80076575e-02 8.11281621e-01 4.36677486e-01
5.17290354e-01 -1.41925290e-01 -4.25172478e-01 3.92458886e-01
1.11946833e+00 3.03217620e-01 4.85552639e-01 -2.39046574e-01
9.90541875e-01 6.93866313e-01 9.24350396e-02 -3.74127477e-02
6.14240348e-01 3.61252248e-01 4.72670585e-01 2.03243438e-02
1.22162122e-02 -6.44568682e-01 5.42713046e-01 1.48902965e+00
-2.72761196e-01 -5.54506600e-01 -7.61867166e-01 4.25490290e-01
-1.87891519e+00 -4.88399029e-01 -5.79723656e-01 2.07708073e+00
4.51953441e-01 4.49633151e-01 -1.34280175e-01 -1.05915256e-01
8.21634769e-01 2.36439675e-01 -1.90898165e-01 -5.95220983e-01
-1.92965686e-01 3.86850625e-01 4.69228953e-01 5.35514951e-01
-9.02953744e-01 1.25840032e+00 6.19429159e+00 4.52279508e-01
-8.13179493e-01 -2.31468901e-02 6.78051949e-01 4.02544707e-01
-6.15206659e-01 4.17587638e-01 -1.09853590e+00 3.51968616e-01
1.08044600e+00 -5.23893535e-02 1.30820662e-01 9.04640973e-01
-1.99898824e-01 1.70081899e-01 -9.28810120e-01 3.89060259e-01
-2.63024360e-01 -7.36723900e-01 2.18569025e-01 -3.61194974e-03
3.43126059e-01 4.69605736e-02 -3.65579277e-01 7.85917401e-01
9.92421806e-01 -9.18289900e-01 5.31335652e-01 1.51175573e-01
1.65826440e-01 -7.72003889e-01 6.62559032e-01 4.90675777e-01
-1.48032081e+00 -1.88675635e-02 -4.18402046e-01 -1.95907459e-01
4.10215020e-01 8.35277379e-01 -5.17968297e-01 1.19600105e+00
5.47899187e-01 3.37087959e-01 -3.69969279e-01 3.13050002e-01
-8.66044641e-01 7.98963547e-01 -2.68509030e-01 1.85589194e-01
2.10104436e-02 -7.32743621e-01 3.39038670e-01 1.17737603e+00
1.83321983e-01 9.02146846e-02 3.42101566e-02 5.61444104e-01
-1.06713444e-01 2.20773458e-01 -3.15725923e-01 -2.27583036e-01
4.34696555e-01 1.36278403e+00 -8.39734793e-01 -2.60340273e-01
-7.39645123e-01 8.79498899e-01 7.77765155e-01 4.14371789e-01
-9.74601388e-01 -1.52877674e-01 3.51179093e-01 -1.99436620e-01
5.49936593e-01 -4.93350208e-01 1.03811584e-02 -1.29423404e+00
2.63507426e-01 -8.01356018e-01 5.21433592e-01 -4.65228707e-01
-1.28503060e+00 1.02503395e+00 7.73814246e-02 -5.13212442e-01
-4.77101833e-01 -5.55477798e-01 -5.94241798e-01 1.05449843e+00
-1.34098065e+00 -1.47743356e+00 1.40186608e-01 4.08877283e-01
4.61334884e-01 1.84694603e-01 1.01375628e+00 9.31893364e-02
-9.31532025e-01 6.15969300e-01 -3.25664639e-01 4.24654454e-01
4.14211422e-01 -1.35814500e+00 8.93448591e-01 1.45921862e+00
5.54624498e-01 8.43273580e-01 4.85030174e-01 -8.72832060e-01
-1.45651293e+00 -8.46360862e-01 1.09042048e+00 -2.50753224e-01
8.23350191e-01 -6.41494751e-01 -1.03602266e+00 1.16615820e+00
1.32031247e-01 7.50540048e-02 4.17154759e-01 6.90051377e-01
-5.37297904e-01 -2.91414410e-02 -4.80721384e-01 5.62809944e-01
1.20466149e+00 -1.02423668e-01 -6.52540505e-01 2.21211329e-01
9.86674607e-01 -4.27292228e-01 -8.08191717e-01 5.34587085e-01
3.37320775e-01 -1.11436033e+00 6.22440517e-01 -7.93915510e-01
5.70995390e-01 -9.82251540e-02 1.12820685e-01 -1.37724006e+00
-6.68871701e-01 -6.84736252e-01 -2.19797835e-01 1.72763455e+00
7.55066276e-01 -6.47230983e-01 7.36272752e-01 8.27246547e-01
-3.02471519e-01 -5.19940197e-01 -8.19483817e-01 -5.63609540e-01
2.16006204e-01 -4.34036493e-01 6.33201957e-01 6.69733047e-01
-6.94917068e-02 1.00334668e+00 -4.91641700e-01 5.88475585e-01
2.44411483e-01 3.42716351e-02 6.78745568e-01 -1.11447704e+00
-8.25927436e-01 -1.15644515e-01 -1.71502158e-01 -1.42733455e+00
4.15885955e-01 -7.33565569e-01 7.97555447e-02 -1.60330999e+00
2.98126098e-02 -5.76353908e-01 -2.93089002e-01 5.53665400e-01
-6.71713710e-01 -4.88607913e-01 4.69783619e-02 1.14102326e-01
-1.57887474e-01 3.39374840e-01 1.04174137e+00 5.13826162e-02
-3.18144351e-01 1.40057117e-01 -9.74122524e-01 7.12985754e-01
5.49519837e-01 -5.45282304e-01 -6.33920074e-01 -8.32922697e-01
3.27483237e-01 4.31576669e-01 -4.33288068e-01 -6.33689761e-01
1.19341657e-01 5.14385961e-02 1.14723936e-01 -3.21781099e-01
1.88801080e-01 -7.20160604e-01 4.00381356e-01 1.30457312e-01
5.23859449e-02 3.31669122e-01 -5.39522544e-02 8.57932150e-01
-2.64315605e-01 -2.26375818e-01 2.91668355e-01 -2.42006153e-01
-5.24089396e-01 2.87969142e-01 -1.15389429e-01 -4.39319685e-02
6.90265298e-01 2.50404060e-01 -2.25059390e-01 -1.98603764e-01
-7.41912723e-01 1.54326767e-01 -8.35837498e-02 6.43014014e-01
8.49512070e-02 -8.97695899e-01 -8.12801659e-01 2.74447560e-01
-2.83337176e-01 3.91739607e-01 3.59582335e-01 5.75947225e-01
-1.13803916e-01 2.22348958e-01 7.44520053e-02 -3.33540827e-01
-1.12207711e+00 6.97089374e-01 -1.92120839e-02 -9.94850278e-01
-5.10729194e-01 1.04152405e+00 6.00126982e-01 -3.66117030e-01
4.51793335e-02 -3.14392209e-01 -4.62552756e-01 -4.81055379e-02
2.27210566e-01 -1.65704489e-01 1.34562865e-01 -7.34715521e-01
-2.06489548e-01 2.69744784e-01 -3.78341436e-01 2.32394993e-01
1.41787159e+00 3.57473232e-02 -3.87186408e-01 3.22876722e-01
7.60085762e-01 4.57943261e-01 -1.04146564e+00 1.56954620e-02
8.45257714e-02 3.40817757e-02 -1.65822685e-01 -5.19883633e-01
-1.29790413e+00 7.69258082e-01 -6.43210933e-02 4.66506898e-01
1.02240384e+00 1.07036762e-01 5.78782499e-01 -1.50451316e-02
2.53050297e-01 -6.11424506e-01 -3.71715784e-01 7.03629792e-01
6.44316435e-01 -8.16840053e-01 5.21521224e-03 -9.92951632e-01
-7.68723845e-01 1.06432366e+00 5.43563545e-01 -3.24916244e-01
7.21586764e-01 7.25035548e-01 -1.09951422e-02 1.37221590e-01
-8.05303574e-01 -7.59013370e-02 2.23273784e-01 7.16089010e-01
8.01315546e-01 2.54077077e-01 -5.04416883e-01 1.27906775e+00
-3.18061173e-01 -5.36040306e-01 2.95090884e-01 5.68851769e-01
-6.73063099e-02 -1.61785376e+00 -1.27090327e-02 2.15548486e-01
-5.91522217e-01 -4.14696366e-01 -1.92870051e-01 9.37016368e-01
-6.56394064e-02 1.07977581e+00 -7.96915516e-02 -3.94768864e-01
3.86826247e-01 1.76961720e-01 3.34256232e-01 -7.41393268e-01
-5.54763377e-01 3.46173108e-01 3.21405888e-01 -2.29496732e-01
-1.57774344e-01 -2.04971790e-01 -1.35389113e+00 1.53703809e-01
-6.43322527e-01 4.56839770e-01 6.19919121e-01 9.55659449e-01
4.27416772e-01 8.80711377e-01 4.87186253e-01 -6.08791292e-01
-4.84739780e-01 -1.14241028e+00 -5.43375671e-01 2.40912870e-01
-2.29673147e-01 -4.95274991e-01 -4.60366756e-01 1.22658394e-01] | [10.33139705657959, 9.58723258972168] |
ebdcff51-b747-46b8-92f9-5e895542a54b | dp-2-nilm-a-distributed-and-privacy | 2207.00041 | null | https://arxiv.org/abs/2207.00041v1 | https://arxiv.org/pdf/2207.00041v1.pdf | DP$^2$-NILM: A Distributed and Privacy-preserving Framework for Non-intrusive Load Monitoring | Non-intrusive load monitoring (NILM), which usually utilizes machine learning methods and is effective in disaggregating smart meter readings from the household-level into appliance-level consumption, can help analyze electricity consumption behaviours of users and enable practical smart energy and smart grid applications. Recent studies have proposed many novel NILM frameworks based on federated deep learning (FL). However, there lacks comprehensive research exploring the utility optimization schemes and the privacy-preserving schemes in different FL-based NILM application scenarios. In this paper, we make the first attempt to conduct FL-based NILM focusing on both the utility optimization and the privacy-preserving by developing a distributed and privacy-preserving NILM (DP2-NILM) framework and carrying out comparative experiments on practical NILM scenarios based on real-world smart meter datasets. Specifically, two alternative federated learning strategies are examined in the utility optimization schemes, i.e., the FedAvg and the FedProx. Moreover, different levels of privacy guarantees, i.e., the local differential privacy federated learning and the global differential privacy federated learning are provided in the DP2-NILM. Extensive comparison experiments are conducted on three real-world datasets to evaluate the proposed framework. | ['Xizhong Chen', 'Qian Wang', 'Fanlin Meng', 'Shuang Dai'] | 2022-06-30 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [-2.03797951e-01 -2.90912718e-01 -1.21898070e-01 -7.14489162e-01
-8.50584507e-01 -5.36343396e-01 5.11518180e-01 1.19813316e-01
-1.49428397e-01 9.59588349e-01 2.27621153e-01 -3.59162211e-01
-2.59654880e-01 -9.95686710e-01 -2.33617589e-01 -1.25535965e+00
-2.80698150e-01 1.03327490e-01 -4.70156103e-01 2.77089506e-01
-4.88614254e-02 2.50783086e-01 -1.08304894e+00 6.04191050e-02
1.17530215e+00 1.36055553e+00 -4.87396866e-01 1.41019359e-01
1.49029955e-01 8.05460036e-01 -5.47489464e-01 -6.88695073e-01
7.76207745e-01 -2.15180457e-01 -6.87052906e-01 -2.76799649e-01
-1.60462409e-01 -8.39597225e-01 -2.30451554e-01 1.41698432e+00
8.06032121e-01 1.88860193e-01 3.28105181e-01 -2.22171378e+00
-4.87799883e-01 9.44902539e-01 -5.65299809e-01 -2.10625947e-01
2.73234874e-01 4.19498324e-01 9.59693849e-01 -3.44205163e-02
-1.17038667e-01 1.05739391e+00 6.82694077e-01 1.94230840e-01
-1.24843562e+00 -9.52491879e-01 -4.14142646e-02 7.81972945e-01
-1.32620227e+00 -2.63423532e-01 8.29202712e-01 -5.13692051e-02
9.91638422e-01 7.19114423e-01 4.19415325e-01 5.89664102e-01
4.44800586e-01 1.18076253e+00 1.52122152e+00 -4.13015485e-02
6.96771204e-01 2.82665938e-01 2.15798840e-01 -5.83032556e-02
5.55820942e-01 2.11959273e-01 -2.90127009e-01 -7.24194229e-01
8.19007680e-02 3.47141832e-01 -4.76386070e-01 -5.23384154e-01
-6.48600876e-01 9.98604953e-01 2.11567551e-01 1.65294617e-01
-4.70755517e-01 -4.27708596e-01 9.77729142e-01 4.84544903e-01
3.67940396e-01 -1.38198689e-01 -8.61362398e-01 1.54928431e-01
-9.26209986e-01 -1.23261707e-02 1.34026182e+00 8.82902980e-01
9.29254532e-01 2.64632195e-01 -1.13482736e-01 2.27423757e-01
3.32617581e-01 1.66248783e-01 8.68864596e-01 -8.41367662e-01
4.13324714e-01 6.95159912e-01 -2.71613467e-02 -7.65917301e-01
-3.52789283e-01 2.17129081e-01 -1.41871476e+00 3.33054245e-01
3.51424958e-03 -5.87531030e-01 7.09598213e-02 1.59927940e+00
4.34722394e-01 2.33354464e-01 4.81686831e-01 5.91260672e-01
4.60791796e-01 6.37015343e-01 2.38139406e-01 -4.56203789e-01
1.29916477e+00 -6.02898002e-01 -9.90415990e-01 5.24983823e-01
8.22386980e-01 -8.92599225e-02 6.85534596e-01 4.19424117e-01
-7.55569160e-01 1.15488060e-01 -1.06639075e+00 2.01508835e-01
-8.01772416e-01 -3.29379588e-01 6.77281559e-01 9.56445456e-01
-9.49631333e-01 4.80226278e-01 -6.70486569e-01 -1.82182521e-01
8.61795962e-01 6.64782166e-01 -4.77090359e-01 -5.04408497e-03
-1.43121016e+00 6.48371577e-01 7.55200684e-01 -2.15405062e-01
-7.74891317e-01 -7.32645452e-01 -9.12438929e-01 2.89991349e-01
1.21469744e-01 -3.97414297e-01 1.21717703e+00 -4.86953080e-01
-1.28781855e+00 4.05408591e-01 4.35665935e-01 -1.14167356e+00
7.69983947e-01 1.18658200e-01 -5.93864441e-01 -1.36869773e-01
-5.24910912e-02 -9.95185599e-02 4.46916312e-01 -9.62471187e-01
-1.04833186e+00 -8.13952804e-01 -2.26800963e-01 5.51213324e-02
-4.74296451e-01 -3.41712177e-01 5.75160563e-01 -3.58162522e-01
-5.59074163e-01 -2.53762752e-01 -1.51289001e-01 -2.40634769e-01
-4.93899435e-01 -4.24616188e-01 1.58912265e+00 -1.06011271e+00
1.14488423e+00 -2.10360098e+00 -6.77383780e-01 4.03323501e-01
6.59125447e-02 4.83748913e-01 1.86804712e-01 4.18345273e-01
-1.78229939e-02 -6.61354885e-02 -3.94972265e-01 -5.00858605e-01
8.68030429e-01 3.76579583e-01 -7.59786516e-02 8.14128458e-01
-6.92372859e-01 1.19964099e+00 -7.92634070e-01 -3.48849267e-01
8.14864755e-01 4.11566496e-01 -8.31719339e-02 4.64137256e-01
-3.74766737e-02 2.98076451e-01 -4.87452775e-01 7.94098735e-01
1.24198329e+00 2.36257520e-02 3.90022874e-01 -3.08039397e-01
-9.70401913e-02 -3.89309227e-02 -1.26910710e+00 1.23969877e+00
-5.69630206e-01 1.74925581e-01 4.33148146e-01 -1.15379608e+00
8.47694278e-01 6.93769097e-01 8.77710044e-01 -8.33258986e-01
5.76864660e-01 4.38306406e-02 -4.28284943e-01 -1.78033054e-01
1.68052197e-01 1.91802144e-01 -1.89120069e-01 8.86620879e-01
-5.98554574e-02 4.31173623e-01 -2.22474605e-01 -3.22114348e-01
9.54621494e-01 -2.12033093e-01 9.20903444e-01 -5.12823403e-01
9.75761175e-01 -6.13001287e-01 8.89515340e-01 5.16712606e-01
-7.71651149e-01 -2.65755624e-01 1.62960976e-01 -5.60033023e-01
-4.24132884e-01 -7.92296052e-01 -1.72900736e-01 7.73993790e-01
2.22956855e-03 -2.77861536e-01 -7.97242641e-01 -9.52284455e-01
4.33539510e-01 1.35242105e+00 -1.16274312e-01 -3.40786785e-01
-1.58761576e-01 -1.17711937e+00 5.67501545e-01 2.71146566e-01
1.38902676e+00 -1.02991617e+00 -8.28063786e-01 2.65466452e-01
-6.77947178e-02 -6.86977327e-01 -4.50312644e-01 4.31576341e-01
-4.14913982e-01 -1.16161716e+00 -5.18619791e-02 -5.03644407e-01
2.87729532e-01 3.26087959e-02 9.50276554e-01 -5.94014764e-01
-6.12561069e-02 3.85110378e-01 -9.51543227e-02 -4.59688544e-01
-1.81241140e-01 1.91493273e-01 1.42541498e-01 5.58357239e-01
1.12740862e+00 -1.04185796e+00 -7.53850639e-01 1.19451568e-01
-9.53812718e-01 -4.41298276e-01 4.43386257e-01 6.17670953e-01
4.92051065e-01 7.90179491e-01 8.47848713e-01 -9.00942147e-01
8.60087574e-01 -7.20598102e-01 -9.35171902e-01 3.72948170e-01
-1.41394663e+00 -1.96377024e-01 1.10814202e+00 -1.49057448e-01
-1.10079134e+00 1.36089057e-01 -3.17912817e-01 -1.71075299e-01
-4.06344771e-01 -1.25739276e-01 -1.23222852e+00 -2.61338830e-01
-1.66386306e-01 5.60340583e-01 -2.14460284e-01 -7.30806470e-01
4.36441660e-01 1.11063516e+00 6.53670490e-01 -3.34266633e-01
7.11790800e-01 3.95394862e-01 -1.29915491e-01 -4.00009453e-01
-1.07036203e-01 -3.87035102e-01 -3.43433291e-01 3.34267795e-01
7.09360600e-01 -9.26416457e-01 -1.42757165e+00 9.75480437e-01
-7.70493209e-01 3.16926278e-02 -6.57017589e-01 3.03115815e-01
-7.16481149e-01 5.91602862e-01 -5.33016682e-01 -1.04141569e+00
-1.19634402e+00 -1.16094410e+00 5.63503504e-01 3.29144120e-01
-4.84233312e-02 -1.23518205e+00 5.19771576e-02 1.84758127e-01
5.24724424e-01 5.60579181e-01 1.00934279e+00 -1.31051707e+00
-5.02791703e-01 -4.43372071e-01 -1.11843511e-01 5.40054083e-01
6.54106200e-01 -9.46927607e-01 -1.11136985e+00 -9.12980199e-01
6.29929960e-01 -2.90555865e-01 1.25371888e-01 1.87937543e-01
1.32919693e+00 -1.28234875e+00 7.54628107e-02 1.14611423e+00
1.86376500e+00 5.05809724e-01 6.74554169e-01 4.94865716e-01
5.09402275e-01 2.09830314e-01 1.49307549e-01 1.03630102e+00
7.52295434e-01 5.93724214e-02 6.27264678e-01 1.88858304e-02
7.94026077e-01 -4.27462429e-01 3.24004471e-01 6.20920658e-01
7.27789819e-01 -3.11884433e-01 -8.15961957e-02 4.73347038e-01
-2.18102670e+00 -8.08495402e-01 2.97490388e-01 2.22194910e+00
7.42715061e-01 -4.12801683e-01 1.65649772e-01 4.82204050e-01
4.40864652e-01 4.21504289e-01 -1.25113249e+00 -6.38876796e-01
-3.73889536e-01 4.27416004e-02 8.72995198e-01 3.10437948e-01
-1.22146523e+00 1.05196990e-01 5.21857595e+00 7.64109552e-01
-8.04294229e-01 4.09482688e-01 7.54587770e-01 1.45992100e-01
-2.06894130e-01 -1.16397373e-01 -4.78676826e-01 7.73341119e-01
8.67872119e-01 -8.57980549e-01 7.13403702e-01 1.05558360e+00
2.75502533e-01 1.47740200e-01 -1.18674099e+00 1.32407820e+00
-3.44369531e-01 -1.02279615e+00 -1.27228007e-01 3.65114093e-01
7.91226387e-01 7.90816615e-04 -2.29903743e-01 1.95722893e-01
1.03742683e+00 -7.86197722e-01 3.70483212e-02 3.29751641e-01
5.85468352e-01 -1.29440784e+00 9.50991988e-01 4.64212805e-01
-1.40344381e+00 -4.66545045e-01 -8.64605084e-02 1.28859267e-01
-9.82824527e-03 3.60717207e-01 -2.29584903e-01 1.17381048e+00
1.04978538e+00 6.83556259e-01 -3.34486485e-01 7.48446763e-01
-3.30186673e-02 5.93752325e-01 -4.64124173e-01 1.96287856e-01
1.26811713e-01 -5.76047301e-01 2.50294983e-01 1.02980065e+00
5.28038740e-02 -7.24097937e-02 2.44943693e-01 8.47387254e-01
-1.55903310e-01 2.34910265e-01 -5.33233941e-01 3.92464876e-01
6.17369831e-01 1.55196643e+00 3.52205597e-02 -1.35325715e-02
-7.29534626e-01 1.02047539e+00 -1.02005720e-01 2.93464005e-01
-3.76404643e-01 -4.16389048e-01 1.23869252e+00 -3.21640998e-01
2.23985180e-01 2.92309940e-01 -4.44712073e-01 -1.32004058e+00
-1.36901319e-01 -9.95147049e-01 9.83448744e-01 -8.07444751e-02
-2.05867934e+00 1.00250589e-02 1.96846593e-02 -9.22881842e-01
-4.65968281e-01 3.65279764e-02 -1.18685329e+00 9.09627616e-01
-1.77568114e+00 -1.26582134e+00 5.90025447e-02 1.26385045e+00
1.18701439e-02 -3.05883169e-01 1.01118541e+00 3.46908152e-01
-6.73911750e-01 1.06011307e+00 9.89166737e-01 1.72232404e-01
9.85876545e-02 -1.29913056e+00 2.48347893e-01 8.59264195e-01
-2.08898112e-01 8.54798406e-02 2.53592521e-01 -2.46045560e-01
-1.56547058e+00 -1.46161950e+00 6.00394666e-01 2.10954979e-01
3.45419317e-01 -5.45216680e-01 -1.00834298e+00 9.21994567e-01
6.50680065e-01 1.66913471e-03 9.22228932e-01 -5.89341044e-01
-4.45599742e-02 -6.50268853e-01 -2.35304976e+00 1.34889036e-01
3.37510169e-01 -4.85877275e-01 -3.11500371e-01 2.41961673e-01
6.81866109e-01 2.45907560e-01 -1.24156642e+00 1.93554685e-01
2.65073955e-01 -1.06514227e+00 7.16719866e-01 -2.25976810e-01
-7.74804294e-01 -3.31476033e-01 -6.49942100e-01 -1.24121261e+00
-3.08691919e-01 -1.11849499e+00 -8.32568228e-01 1.96846187e+00
-5.08323491e-01 -1.48882246e+00 7.97215760e-01 1.01919782e+00
5.55929780e-01 -6.39756680e-01 -1.39605522e+00 -7.45759666e-01
1.74648196e-01 -2.84347951e-01 1.92455161e+00 1.01770735e+00
-4.54698242e-02 -2.37258732e-01 -4.47348118e-01 3.35993856e-01
1.25392354e+00 2.08308458e-01 5.68937182e-01 -9.00353134e-01
3.73475067e-02 -1.79626584e-01 -5.79068422e-01 -5.15823126e-01
4.14093792e-01 -8.99708688e-01 -4.56393957e-01 -1.22734666e+00
1.18667586e-03 -2.23711249e-03 -9.34433341e-01 8.42051327e-01
3.88503522e-01 -1.62869990e-01 3.31267029e-01 6.12577461e-02
-4.14840013e-01 8.43805611e-01 3.19849759e-01 -4.43133444e-01
-2.37131007e-02 3.32780927e-01 -9.54821885e-01 6.04272664e-01
1.26211190e+00 -1.88416883e-01 -5.82112849e-01 -1.11253232e-01
-6.94632113e-01 -9.82395262e-02 2.69147456e-01 -7.22682714e-01
2.77590275e-01 -1.73177913e-01 3.58528733e-01 -9.61380243e-01
-2.36945510e-01 -1.40045798e+00 4.23361301e-01 6.96313560e-01
-1.43383804e-03 6.97430298e-02 -8.67846683e-02 3.10697049e-01
4.86338884e-03 1.08541645e-01 8.58506322e-01 -2.73516357e-01
-8.53575230e-01 6.73114300e-01 -8.61719698e-02 -1.82554368e-02
1.26354420e+00 -5.37033528e-02 -2.07368746e-01 -5.63984334e-01
-2.88384706e-01 9.28360999e-01 4.79782343e-01 1.80637807e-01
2.55539954e-01 -1.62932193e+00 -7.10102022e-01 5.94082057e-01
-1.92516863e-01 -3.41929980e-02 1.37666672e-01 4.05862421e-01
1.40128285e-01 5.10879397e-01 -1.29502118e-01 -2.19295710e-01
-1.08125961e+00 7.43548751e-01 5.21944046e-01 -6.41367972e-01
-6.63402021e-01 6.47890717e-02 1.13865100e-01 -7.23828375e-01
6.31358922e-01 -1.25298858e-01 1.40181109e-01 -3.46125327e-02
6.56074524e-01 1.05330527e+00 2.58243501e-01 -6.74303651e-01
-4.78854597e-01 3.61762568e-02 1.56062007e-01 4.07180607e-01
1.32134235e+00 -5.43887138e-01 -4.58227336e-01 1.37939885e-01
1.71789026e+00 -6.52063310e-01 -1.25063992e+00 -4.01206255e-01
1.18729271e-01 -5.03590584e-01 1.31313413e-01 -9.59016562e-01
-1.51149368e+00 5.32536089e-01 1.13425171e+00 2.27897778e-01
1.68783212e+00 -6.14823759e-01 1.44317627e+00 2.32185751e-01
8.33116949e-01 -1.03332281e+00 -9.89017904e-01 -1.77359238e-01
1.65578097e-01 -1.09293795e+00 7.62540549e-02 3.61377358e-01
-4.31778848e-01 5.66434622e-01 4.66049939e-01 2.31043503e-01
1.05005872e+00 4.89375144e-01 -4.79556993e-02 2.95701861e-01
-5.66306889e-01 4.15343493e-01 -3.63544673e-01 1.08618617e+00
-3.02369118e-01 4.91214663e-01 -3.64835918e-01 1.08028162e+00
-2.84942687e-01 2.64702916e-01 1.43785506e-01 1.12251270e+00
3.41580093e-01 -1.17360902e+00 -4.45894688e-01 7.28920877e-01
-6.05187953e-01 2.98440307e-01 -1.55529797e-01 4.26692873e-01
4.15393442e-01 1.06851780e+00 -1.57121494e-01 -4.49630201e-01
2.37788856e-01 1.74118772e-01 -4.38903738e-03 3.72385353e-01
-7.80765057e-01 -4.97501045e-01 -5.64528108e-01 -5.18405318e-01
-1.78508133e-01 -8.90060782e-01 -1.12731409e+00 -8.36169004e-01
-2.64757067e-01 3.08320463e-01 3.02066416e-01 8.08892071e-01
3.24565858e-01 -8.29991028e-02 1.31917906e+00 -6.03433609e-01
-1.20106947e+00 -5.23555458e-01 -1.23201871e+00 5.42980850e-01
4.68253255e-01 9.75611135e-02 -5.92073441e-01 -4.36347336e-01] | [5.888927459716797, 2.797316789627075] |
bc8da434-17b1-47f7-b805-b44ec3aafa2c | writing-style-aware-document-level-event | 2201.03188 | null | https://arxiv.org/abs/2201.03188v1 | https://arxiv.org/pdf/2201.03188v1.pdf | Writing Style Aware Document-level Event Extraction | Event extraction, the technology that aims to automatically get the structural information from documents, has attracted more and more attention in many fields. Most existing works discuss this issue with the token-level multi-label classification framework by distinguishing the tokens as different roles while ignoring the writing styles of documents. The writing style is a special way of content organizing for documents and it is relative fixed in documents with a special field (e.g. financial, medical documents, etc.). We argue that the writing style contains important clues for judging the roles for tokens and the ignorance of such patterns might lead to the performance degradation for the existing works. To this end, we model the writing style in documents as a distribution of argument roles, i.e., Role-Rank Distribution, and propose an event extraction model with the Role-Rank Distribution based Supervision Mechanism to capture this pattern through the supervised training process of an event extraction task. We compare our model with state-of-the-art methods on several real-world datasets. The empirical results show that our approach outperforms other alternatives with the captured patterns. This verifies the writing style contains valuable information that could improve the performance of the event extraction task. | ['Lixin Cui', 'Lu Bai', 'Yue Wang', 'Zhuo Xu'] | 2022-01-10 | null | null | null | null | ['document-level-event-extraction'] | ['natural-language-processing'] | [ 1.38007358e-01 2.40701139e-02 -6.14803255e-01 -5.39540529e-01
-3.15215230e-01 -4.91799533e-01 9.48904455e-01 6.00558937e-01
-4.86542881e-01 7.23098040e-01 5.65805793e-01 -8.83086026e-02
-2.70739287e-01 -7.84520447e-01 -4.00100827e-01 -6.91413462e-01
4.85463172e-01 3.99376541e-01 5.97668588e-01 -1.65858120e-01
5.50543308e-01 1.14139780e-01 -1.77027166e+00 7.71882832e-01
6.85245872e-01 8.05331349e-01 1.49964571e-01 5.35552539e-02
-5.51269829e-01 1.23172915e+00 -9.40390348e-01 -3.59519541e-01
-7.83428401e-02 -4.04263228e-01 -9.71948862e-01 7.62593299e-02
-1.65379107e-01 -1.37858629e-01 1.00737356e-01 1.02242267e+00
3.02361578e-01 -1.64602607e-01 9.43935156e-01 -1.28360224e+00
-2.60103971e-01 9.97616231e-01 -7.47921824e-01 2.53020793e-01
3.37737203e-01 -7.32481539e-01 1.28001070e+00 -6.72030270e-01
7.17036486e-01 1.25156677e+00 2.66067624e-01 2.39140257e-01
-5.26127577e-01 -5.61329365e-01 5.43977261e-01 4.53165799e-01
-1.13640130e+00 -9.65475515e-02 8.90159786e-01 -5.12512147e-01
4.64970797e-01 2.87899315e-01 2.21981004e-01 1.20917261e+00
2.43992552e-01 1.06190765e+00 1.22939003e+00 -7.54724920e-01
-3.50804371e-03 4.29460377e-01 6.85687840e-01 4.13995653e-01
5.24363875e-01 -4.51737195e-01 -5.87420285e-01 -2.72766858e-01
2.55775303e-01 3.24771017e-01 -6.91289455e-02 1.85060441e-01
-1.13153148e+00 6.76756084e-01 -2.45506257e-01 6.57496274e-01
-6.19240999e-02 -3.40214521e-01 8.54159176e-01 1.25225484e-01
5.66205084e-01 1.32035553e-01 -7.43516684e-01 1.13034837e-01
-7.43498564e-01 2.17736140e-01 7.45319843e-01 9.46773708e-01
3.97983640e-01 -4.71478283e-01 -5.15596807e-01 7.90784836e-01
5.53243577e-01 -4.57036346e-02 7.04846203e-01 -4.18139338e-01
5.39974034e-01 1.01037788e+00 1.62115484e-01 -9.68885183e-01
-4.22519982e-01 -3.38898987e-01 -6.34970605e-01 -3.38008642e-01
4.93493885e-01 -6.38864487e-02 -4.08736050e-01 1.48244178e+00
5.03588498e-01 -1.13125227e-01 -6.50216490e-02 5.18186092e-01
9.13823009e-01 5.84799290e-01 7.53239170e-02 -4.76091802e-01
1.98524594e+00 -9.00427341e-01 -1.29669714e+00 1.17609492e-02
4.38556701e-01 -9.38844442e-01 8.92240405e-01 7.66372025e-01
-5.31938910e-01 -4.83104736e-01 -1.03128016e+00 1.63518518e-01
-4.81294483e-01 4.85316902e-01 6.15710318e-01 5.05247295e-01
-1.36129767e-01 4.79040802e-01 -6.34048402e-01 -1.94410950e-01
1.58586547e-01 -7.94124976e-02 2.32763365e-02 2.38547891e-01
-1.62093937e+00 7.14000642e-01 7.81405091e-01 -1.02902472e-01
-5.03658593e-01 -2.70077169e-01 -5.75398982e-01 2.27051806e-02
8.54086280e-01 -1.12284176e-01 1.16777492e+00 -7.53258705e-01
-9.83844697e-01 7.57822990e-01 -2.56822079e-01 -8.82656127e-02
3.94899100e-01 -3.00979167e-01 -6.65953517e-01 1.20922230e-01
4.99211133e-01 -2.47471079e-01 8.71194661e-01 -1.25807190e+00
-1.10689974e+00 -3.58699083e-01 9.52464119e-02 1.01371571e-01
-6.05006814e-01 7.27267385e-01 -2.74926066e-01 -9.28690612e-01
1.11177586e-01 -5.73411703e-01 1.01941854e-01 -6.32896721e-01
-7.09815919e-01 -1.00526834e+00 8.83699179e-01 -4.17215765e-01
1.69313622e+00 -1.98434234e+00 -2.13409707e-01 1.07966848e-01
2.16440782e-01 -2.37907827e-01 5.79941928e-01 5.65479398e-01
-5.66014759e-02 2.96908230e-01 2.19071373e-01 -2.83388615e-01
2.10488215e-02 2.71793485e-01 -5.22248209e-01 3.27116758e-01
-6.01060092e-02 2.95041680e-01 -9.54863608e-01 -1.05651438e+00
-1.94590613e-01 2.52479725e-02 -1.05056301e-01 2.35492766e-01
-3.07235777e-01 1.38278991e-01 -8.38723004e-01 7.16070950e-01
4.14952189e-01 -4.45488334e-01 4.54622447e-01 -4.19288963e-01
-1.35599986e-01 5.40353656e-01 -1.51426220e+00 1.25473189e+00
1.06164988e-03 2.26861656e-01 -2.12186262e-01 -1.02715921e+00
7.88934529e-01 5.21894693e-01 6.23754740e-01 -1.85036510e-01
2.48663545e-01 1.72323585e-01 -5.18896542e-02 -5.35445273e-01
6.90069020e-01 -1.27160996e-01 -4.18602735e-01 6.90200329e-01
-3.52725834e-02 5.21496058e-01 5.87332129e-01 3.40974361e-01
1.01839554e+00 6.59493655e-02 5.94310462e-01 -2.84217745e-01
4.92774069e-01 -8.43642652e-02 9.07206535e-01 6.76131606e-01
2.52185538e-02 4.87813145e-01 7.48978555e-01 -2.78751761e-01
-6.78221464e-01 -4.75437939e-01 -4.21352297e-01 1.34296834e+00
2.90181130e-01 -8.80471647e-01 -4.89956051e-01 -1.18798149e+00
-1.80297658e-01 4.52096909e-01 -6.59704208e-01 1.14715256e-01
-5.74061513e-01 -1.07086003e+00 6.90660357e-01 4.20735180e-01
3.62190157e-01 -1.27776361e+00 -4.84295756e-01 3.89659792e-01
-4.68057483e-01 -1.04373813e+00 -2.68483758e-01 5.55229247e-01
-4.68116671e-01 -1.33934820e+00 -2.76623905e-01 -7.48215437e-01
7.06539571e-01 6.55101091e-02 9.64033961e-01 -4.00413983e-02
2.00705439e-01 -1.67775840e-01 -9.77675915e-01 -8.94592524e-01
-3.83393586e-01 2.14329556e-01 -1.63873460e-03 3.45434397e-01
5.45915663e-01 -3.81212413e-01 -3.66541535e-01 5.29988706e-01
-1.04278100e+00 -4.46825512e-02 6.25704288e-01 7.42353082e-01
4.13737744e-01 7.87016869e-01 5.41884303e-01 -1.52870762e+00
7.48191535e-01 -6.46681011e-01 -1.79360986e-01 4.27627742e-01
-8.88351798e-01 3.25997829e-01 7.60598123e-01 -6.08742714e-01
-1.40942562e+00 -2.53205657e-01 1.90232590e-01 8.46501440e-02
-3.66697401e-01 6.56854451e-01 -5.38603246e-01 7.86373794e-01
3.79642457e-01 2.55689979e-01 -5.24306536e-01 -7.33978748e-01
-1.61436111e-01 9.17656004e-01 1.03804469e-01 -9.28314984e-01
4.73652095e-01 5.61152577e-01 -1.80075213e-01 -3.16873044e-01
-1.54135120e+00 -8.03481996e-01 -5.19781291e-01 -1.22472569e-01
5.37861466e-01 -7.70292759e-01 -5.26966989e-01 4.13300574e-01
-1.43052244e+00 2.77503848e-01 2.78935190e-02 4.56378251e-01
7.03706732e-03 4.31391686e-01 -7.92493045e-01 -9.28361714e-01
-1.94690675e-01 -7.53531992e-01 1.20393431e+00 8.44611228e-02
-4.67772752e-01 -8.57076466e-01 2.92991623e-02 2.99566835e-01
-1.06430843e-01 2.97854781e-01 1.11771953e+00 -1.10079432e+00
-1.70169324e-01 -2.69813612e-02 -7.00150849e-03 1.00444019e-01
4.13475037e-01 3.63514945e-03 -9.61115658e-01 1.09557554e-01
4.09281820e-01 -1.20845221e-01 8.99185658e-01 7.73626566e-02
1.33082461e+00 -5.74910045e-01 -6.25168324e-01 -1.01077564e-01
1.22073638e+00 4.46640253e-01 4.34783340e-01 5.71815729e-01
7.54420877e-01 8.97960186e-01 1.03645074e+00 7.97505736e-01
4.85286653e-01 5.34259915e-01 1.60080612e-01 9.97119471e-02
2.09990904e-01 -3.67442906e-01 4.27280992e-01 8.68700206e-01
-6.71682283e-02 -5.10749519e-01 -6.39465451e-01 5.56600928e-01
-2.17300487e+00 -1.19967699e+00 -3.48911822e-01 1.66795409e+00
1.39408803e+00 6.59148753e-01 4.31038812e-02 7.78874755e-01
9.40546393e-01 3.03054422e-01 -1.57012865e-01 -1.18898220e-01
-1.32726625e-01 -2.24028945e-01 2.55769759e-01 -6.28215224e-02
-1.20667696e+00 5.95493972e-01 5.56241179e+00 1.13664758e+00
-6.86982036e-01 1.45477325e-01 5.34946144e-01 2.86771655e-01
-2.21649632e-01 1.41353890e-01 -1.49122131e+00 9.26955819e-01
7.13185251e-01 -1.35522738e-01 -3.00241828e-01 7.57406592e-01
3.24680090e-01 -1.75579980e-01 -1.23826158e+00 6.99114859e-01
3.16492766e-01 -9.70676839e-01 1.84373289e-01 1.43576190e-01
3.42345744e-01 -5.15338302e-01 -4.96770442e-01 2.08630577e-01
2.53738284e-01 -6.02634132e-01 1.04776633e+00 4.44913745e-01
4.07653511e-01 -6.35532498e-01 1.04242074e+00 7.45115936e-01
-1.26759708e+00 -3.30070709e-03 -1.47023320e-01 -1.64050505e-01
1.07640453e-01 1.22125340e+00 -6.92038417e-01 8.16482544e-01
6.22812033e-01 9.49681222e-01 -5.90222418e-01 6.09537184e-01
-4.88059253e-01 9.89037752e-01 -1.80973202e-01 -2.62035280e-01
-1.25061676e-01 1.55166455e-03 2.39220664e-01 1.31026912e+00
1.19495258e-01 3.12789120e-02 4.11288559e-01 6.06729686e-01
2.53857188e-02 3.88348222e-01 -3.44820887e-01 1.13447569e-01
6.29398644e-01 1.26939607e+00 -1.13583505e+00 -5.67572415e-01
-4.74594802e-01 5.17690063e-01 1.00051844e-02 -9.95616391e-02
-7.92240620e-01 -6.18068874e-01 2.76887100e-02 2.71876931e-01
1.14791885e-01 1.30617261e-01 -3.17698538e-01 -1.21522701e+00
3.91760796e-01 -8.28829527e-01 8.15095663e-01 -4.45811808e-01
-1.43318629e+00 5.21171093e-01 2.60661572e-01 -1.50638521e+00
-5.70144169e-02 -6.52552783e-01 -6.13881171e-01 3.43643188e-01
-1.58386815e+00 -1.05607939e+00 -1.51656598e-01 4.65601802e-01
7.65820265e-01 -1.75763726e-01 5.13058841e-01 5.37124932e-01
-6.14570379e-01 4.88855213e-01 -8.44387487e-02 5.57959437e-01
1.06882620e+00 -1.41167915e+00 -2.38692299e-01 7.36163199e-01
1.19554728e-01 7.67024636e-01 7.06489980e-01 -8.49810123e-01
-8.01241875e-01 -9.57703471e-01 1.45159435e+00 -5.82974911e-01
5.82862496e-01 -3.59915197e-01 -9.10489023e-01 5.16543925e-01
3.19219500e-01 -2.48996764e-01 9.74543691e-01 3.55606318e-01
-3.34101200e-01 -9.29097384e-02 -1.00079823e+00 2.50993639e-01
9.18503284e-01 -2.74256885e-01 -1.15183854e+00 3.86398435e-01
5.78371048e-01 -2.04601511e-01 -6.61418080e-01 1.58429325e-01
4.02804852e-01 -6.49933040e-01 6.43302321e-01 -4.50863749e-01
6.97313070e-01 -6.15887344e-01 8.07299986e-02 -1.00679493e+00
-2.14300796e-01 -3.13742697e-01 -3.59685212e-01 1.86642253e+00
3.67338449e-01 -4.66867924e-01 4.27041620e-01 3.10697794e-01
2.50424743e-02 -5.49296021e-01 -5.38545251e-01 -6.12306714e-01
-3.36962998e-01 -3.27723801e-01 7.67667651e-01 1.34124899e+00
2.80392915e-01 7.38970816e-01 -4.45198238e-01 6.10224493e-02
4.41115826e-01 4.11158860e-01 2.35148534e-01 -1.70166087e+00
-2.91247904e-01 -2.24104628e-01 -2.41478644e-02 -7.76652575e-01
3.11839074e-01 -8.28939497e-01 -1.72602296e-01 -1.55248618e+00
5.99833488e-01 -5.01259685e-01 -5.66408813e-01 6.95776582e-01
-3.69627386e-01 -4.74015862e-01 -1.86797157e-01 5.64480722e-01
-1.02529085e+00 3.31130356e-01 1.12593520e+00 -2.41941497e-01
-1.26355201e-01 2.08036348e-01 -1.04022789e+00 1.03072274e+00
7.63467312e-01 -9.42382336e-01 -3.24814975e-01 -1.14253916e-01
5.84258080e-01 -2.37434015e-01 -1.40573934e-01 -5.04317045e-01
2.34171718e-01 -4.89456236e-01 3.23035538e-01 -8.15657318e-01
-2.40418047e-01 -8.93620253e-01 -1.54871091e-01 1.64595544e-01
-5.07375300e-01 3.85293402e-02 -3.41501594e-01 5.29682100e-01
-4.88243729e-01 -6.27655029e-01 2.81191438e-01 -2.55324960e-01
-4.63876069e-01 -1.36865765e-01 -4.12896425e-01 2.03922555e-01
8.00764322e-01 -3.73823494e-02 -3.99909139e-01 5.76303080e-02
-3.47921044e-01 1.38303518e-01 6.46917373e-02 5.52940786e-01
3.02043825e-01 -1.43809497e+00 -7.76710510e-01 -2.73985509e-03
3.73937398e-01 2.18352363e-01 -3.32896262e-01 5.19262135e-01
1.89232398e-02 1.44327909e-01 1.65136755e-01 -2.89037138e-01
-1.47459531e+00 4.24574822e-01 -1.18407920e-01 -8.40686381e-01
-3.20624053e-01 2.88231432e-01 6.52453452e-02 -7.41689205e-02
2.64579803e-01 -5.91862977e-01 -1.08446002e+00 6.72640860e-01
5.93928635e-01 3.35030526e-01 1.62248254e-01 -3.50742042e-01
-4.16619539e-01 3.09344828e-01 -5.13410509e-01 -7.90871680e-02
1.20634317e+00 -5.70406802e-02 -4.78400677e-01 7.64390111e-01
8.45375776e-01 3.76421779e-01 -6.69424891e-01 -4.18199033e-01
5.80537140e-01 -3.26330960e-01 -1.63223237e-01 -6.90125644e-01
-7.43534625e-01 4.68602628e-01 7.19360709e-02 6.46160960e-01
8.09235990e-01 2.87629366e-01 5.35215914e-01 2.42652953e-01
5.20793140e-01 -1.48501790e+00 2.34074965e-01 6.35621846e-01
6.28245592e-01 -1.18945789e+00 2.73135662e-01 -7.12063074e-01
-5.88199735e-01 1.05111182e+00 6.64814472e-01 2.51064539e-01
8.25956106e-01 4.39502358e-01 -9.99235213e-02 -2.83204287e-01
-6.97787583e-01 -4.89660129e-02 3.39053810e-01 8.83992240e-02
7.64474452e-01 -4.33446020e-02 -9.46313798e-01 1.28864181e+00
-7.11659938e-02 -1.87082857e-01 4.94388551e-01 1.20018828e+00
-4.92467672e-01 -1.57805288e+00 -5.10885894e-01 7.15797484e-01
-1.03282082e+00 -2.93991878e-03 -5.03178716e-01 4.48777080e-01
6.15792155e-01 1.11084545e+00 -1.97459474e-01 -1.08678214e-01
4.05669093e-01 2.25850254e-01 1.79366812e-01 -9.53942537e-01
-7.95623779e-01 2.60888368e-01 2.38181189e-01 -5.56131490e-02
-7.62698352e-01 -8.23701143e-01 -1.44895422e+00 2.22360373e-01
-5.90783119e-01 5.00301659e-01 1.88931897e-01 1.20252872e+00
6.03316985e-02 8.15013945e-01 7.93079317e-01 -2.56155670e-01
-4.79277015e-01 -1.08445323e+00 -7.46321619e-01 7.20016718e-01
1.33022860e-01 -8.98886561e-01 -4.82241809e-01 2.36563042e-01] | [9.169719696044922, 9.187223434448242] |
4e7eab4b-49c3-461b-a556-237d628242cf | korean-specific-dataset-for-table-question | 2201.06223 | null | https://arxiv.org/abs/2201.06223v2 | https://arxiv.org/pdf/2201.06223v2.pdf | Korean-Specific Dataset for Table Question Answering | Existing question answering systems mainly focus on dealing with text data. However, much of the data produced daily is stored in the form of tables that can be found in documents and relational databases, or on the web. To solve the task of question answering over tables, there exist many datasets for table question answering written in English, but few Korean datasets. In this paper, we demonstrate how we construct Korean-specific datasets for table question answering: Korean tabular dataset is a collection of 1.4M tables with corresponding descriptions for unsupervised pre-training language models. Korean table question answering corpus consists of 70k pairs of questions and answers created by crowd-sourced workers. Subsequently, we then build a pre-trained language model based on Transformer and fine-tune the model for table question answering with these datasets. We then report the evaluation results of our model. We make our datasets publicly available via our GitHub repository and hope that those datasets will help further studies for question answering over tables, and for the transformation of table formats. | ['Kyungkoo Min', 'Hansol Jang', 'Hyun Kim', 'Myoseop Sim', 'Jooyoung Choi', 'Changwook Jun'] | 2022-01-17 | null | https://aclanthology.org/2022.lrec-1.657 | https://aclanthology.org/2022.lrec-1.657.pdf | lrec-2022-6 | ['unsupervised-pre-training'] | ['methodology'] | [-1.36748657e-01 1.71664089e-01 -1.74956769e-01 -4.87170488e-01
-1.67360008e+00 -1.02659523e+00 1.96649522e-01 5.93110740e-01
-2.00948477e-01 8.35834920e-01 6.06661916e-01 -5.56025565e-01
-3.94355878e-02 -1.26634574e+00 -7.77065277e-01 2.53886133e-01
3.56578976e-01 9.55744386e-01 6.42227888e-01 -7.11805701e-01
1.11426786e-01 -3.09264034e-01 -1.36460805e+00 1.17366207e+00
1.07026649e+00 1.34070945e+00 -9.06871930e-02 6.62984967e-01
-9.66749609e-01 1.57658505e+00 -7.20115066e-01 -1.06366074e+00
1.10554807e-01 -2.53591716e-01 -1.46867311e+00 -2.43191689e-01
4.72247779e-01 -3.11729580e-01 -3.60118568e-01 7.47944593e-01
2.99567908e-01 8.25217459e-03 4.21532482e-01 -1.20824182e+00
-1.21476686e+00 1.00086737e+00 2.43837729e-01 1.35767460e-01
8.95695031e-01 -2.49650151e-01 1.15816104e+00 -8.58148038e-01
7.54741430e-01 1.21794319e+00 2.74101347e-01 4.69889283e-01
-7.03837335e-01 -3.08819175e-01 -4.63598929e-02 3.14507484e-01
-1.06438661e+00 -4.47654784e-01 5.80223382e-01 -1.60367861e-01
1.04171693e+00 7.11918592e-01 7.15725691e-05 8.84306490e-01
-5.66785336e-02 7.09616125e-01 6.38031423e-01 -3.47818524e-01
-9.85528156e-02 5.23601174e-01 2.45296896e-01 5.65662682e-01
3.51679325e-01 -9.84914124e-01 -6.73985541e-01 -1.26075163e-01
2.34264836e-01 -2.71486461e-01 -2.03075469e-01 -2.42723912e-01
-1.47198951e+00 7.69291759e-01 3.12853813e-01 1.17306322e-01
3.42021398e-02 -1.85571268e-01 6.22827649e-01 6.04724824e-01
7.80217648e-02 5.45599520e-01 -9.81064796e-01 -2.01896295e-01
-2.13302001e-02 5.51207185e-01 1.33162689e+00 1.51571155e+00
6.14126086e-01 -7.79748619e-01 -3.82120907e-01 9.73640680e-01
2.31850356e-01 6.20958745e-01 3.89489204e-01 -1.05563378e+00
1.53929222e+00 1.34284449e+00 2.72705078e-01 -1.14361525e+00
-2.04147041e-01 3.27011168e-01 -6.80698276e-01 -1.06133151e+00
7.92860091e-01 7.06387386e-02 -3.37214917e-01 1.23661041e+00
5.06889224e-01 -1.00790823e+00 5.12401998e-01 4.61497784e-01
1.55791497e+00 8.24406207e-01 -1.14220180e-01 1.59448728e-01
1.86882162e+00 -1.11676776e+00 -1.17005062e+00 -1.31848991e-01
1.07827270e+00 -8.76866102e-01 1.65092731e+00 1.21914461e-01
-8.71428549e-01 -5.63670695e-01 -5.24080217e-01 -9.53315854e-01
-1.11132622e+00 1.48215979e-01 3.16926301e-01 6.02342844e-01
-6.78465188e-01 -2.02901617e-01 -4.50681984e-01 -6.53738976e-01
2.14836612e-01 -7.53831342e-02 -2.98086792e-01 -3.28220338e-01
-1.35966671e+00 5.93673766e-01 4.34067935e-01 8.61914232e-02
-2.13880867e-01 -7.59221017e-01 -1.07716537e+00 -4.63320464e-02
8.82010996e-01 -7.79941261e-01 1.38108242e+00 -6.41648918e-02
-9.62707460e-01 1.00283206e+00 -2.79943824e-01 -2.49177128e-01
9.10536200e-02 -8.81540254e-02 -6.82617784e-01 2.67460067e-02
4.93726313e-01 4.69356537e-01 -4.77447137e-02 -1.14425528e+00
-4.85856712e-01 -5.03567159e-01 6.85717940e-01 4.10377719e-02
-3.18998456e-01 1.25873953e-01 -8.61125529e-01 -4.99276996e-01
2.08138093e-01 -4.61150974e-01 8.28162804e-02 -1.55136034e-01
-6.00543857e-01 -4.55069989e-01 4.36629027e-01 -1.18897080e+00
1.67728591e+00 -1.87006557e+00 -4.04792964e-01 -1.13596790e-01
4.49253879e-02 -2.24328309e-01 -1.52065396e-01 9.25919950e-01
2.94658303e-01 4.38756466e-01 -3.24059010e-01 1.77198678e-01
3.19772065e-01 2.13639885e-01 -8.13441932e-01 -6.34978533e-01
2.33525306e-01 1.27227151e+00 -7.02832758e-01 -8.06032777e-01
-4.33311462e-01 -1.82757646e-01 -6.63931608e-01 4.60784763e-01
-7.46127963e-01 -3.43134301e-03 -7.40771830e-01 9.97817397e-01
5.10225534e-01 -4.48376864e-01 2.33806640e-01 -6.21632673e-02
4.61145610e-01 9.37942803e-01 -1.14040613e+00 1.69065833e+00
-3.52810413e-01 1.01402648e-01 -1.93141311e-01 -4.42867339e-01
1.05242360e+00 1.73295513e-01 1.88842624e-01 -1.04982269e+00
-1.06039375e-01 4.10655618e-01 -3.22072625e-01 -8.96265984e-01
8.65617156e-01 2.78705209e-01 -6.16288424e-01 3.88566375e-01
-1.52144700e-01 -5.85316062e-01 8.30115378e-01 4.54879463e-01
1.07024431e+00 -1.16928905e-01 3.79793830e-02 -1.23259105e-01
1.17811942e+00 5.81639647e-01 -1.05009489e-01 6.25110865e-01
5.31141311e-02 4.95256245e-01 7.49619246e-01 -5.99571466e-01
-6.61867142e-01 -1.01377821e+00 -1.32650703e-01 1.06224298e+00
-2.53400635e-02 -9.81817842e-01 -7.59246826e-01 -8.82237792e-01
2.00574949e-01 7.00421870e-01 -6.60710454e-01 -1.82145312e-02
-6.79815829e-01 -2.85962254e-01 4.68433261e-01 6.62289739e-01
6.36021733e-01 -1.30812025e+00 2.03496814e-02 2.70210892e-01
-9.07703519e-01 -1.32099545e+00 -6.33991301e-01 1.17482357e-01
-7.22737849e-01 -1.40225494e+00 -1.06005669e-02 -1.05456400e+00
4.86652970e-01 6.22986741e-02 1.86250305e+00 2.86139190e-01
2.15236604e-01 5.15580535e-01 -4.78675365e-01 -5.58531761e-01
-4.25580025e-01 5.97131729e-01 -3.95816952e-01 -2.76856720e-01
7.34390378e-01 -2.62409952e-02 -2.89063931e-01 4.26801682e-01
-1.19654691e+00 -1.60923481e-01 -2.71421839e-02 3.62817228e-01
7.31139898e-01 -9.91903245e-02 7.05974579e-01 -1.43863988e+00
9.38124537e-01 -5.45825839e-01 -6.83573067e-01 8.17931533e-01
-2.46250376e-01 3.92507613e-01 8.96304846e-01 -2.96507441e-02
-9.50430214e-01 -4.45184350e-01 -4.18765366e-01 3.04180145e-01
2.91045625e-02 7.22106218e-01 -6.73478842e-01 4.78953570e-01
5.45908928e-01 8.24617594e-02 -3.15302730e-01 -7.10992277e-01
5.52531779e-01 8.49995613e-01 4.41576630e-01 -9.21835363e-01
9.55276608e-01 2.34759778e-01 -6.27305925e-01 -1.44186810e-01
-1.24602389e+00 -4.29003716e-01 -5.45426130e-01 -2.37653442e-02
1.01916552e+00 -8.39167058e-01 -9.34588194e-01 1.51511841e-02
-8.66502643e-01 -2.04398140e-01 -4.26717907e-01 -3.05081040e-01
-4.34527248e-01 3.63232545e-03 -6.34491801e-01 -5.41669726e-01
-3.13525826e-01 -8.86375189e-01 1.16912770e+00 -7.10478872e-02
-3.06505144e-01 -9.59623039e-01 9.64927226e-02 1.39965117e+00
3.90514880e-01 -6.18084446e-02 1.65021718e+00 -8.41772676e-01
-9.77619469e-01 -3.05212468e-01 -1.78657368e-01 8.43537040e-03
2.46118918e-01 -3.42889577e-01 -6.65240228e-01 2.04858080e-01
-8.46167728e-02 -1.03518391e+00 6.02609992e-01 -3.86487454e-01
1.71826553e+00 -5.72586954e-01 1.22899890e-01 1.93388239e-01
1.39504707e+00 3.29854071e-01 7.16622114e-01 7.08351672e-01
7.66183615e-01 1.01977587e+00 6.78254068e-01 1.03856474e-01
1.48757541e+00 3.33447754e-01 1.59971312e-01 3.42541635e-01
1.35957897e-01 -9.87628400e-01 -1.44294128e-01 1.47561240e+00
5.76090395e-01 -5.02076805e-01 -1.08077061e+00 5.31916082e-01
-1.66588378e+00 -6.46120071e-01 -2.32618511e-01 1.96547806e+00
1.22910798e+00 1.74441591e-01 -6.00818731e-02 2.08757773e-01
1.35662973e-01 -7.42249758e-05 -4.80851293e-01 -3.45212787e-01
-2.33239874e-01 -1.99581422e-02 1.03406906e-01 4.32097644e-01
-9.97484326e-01 7.99027979e-01 5.90395021e+00 6.03139281e-01
-3.28530371e-01 -1.81006998e-01 8.74261439e-01 2.33471707e-01
-8.22015762e-01 -6.77458048e-02 -8.69941592e-01 3.10549945e-01
1.09760439e+00 -3.79598647e-01 4.59826618e-01 6.67054176e-01
-3.29841018e-01 -1.21285282e-01 -1.18154359e+00 1.04056859e+00
1.44147739e-01 -1.45565975e+00 6.75882638e-01 -2.99764037e-01
4.01270926e-01 -5.89907527e-01 -1.30017817e-01 9.56988990e-01
1.70504063e-01 -9.99187768e-01 6.37749135e-01 6.08274996e-01
6.00243866e-01 -5.59803724e-01 9.23120379e-01 4.58495885e-01
-1.32592034e+00 -5.71534187e-02 -5.16994715e-01 3.32714319e-02
-2.63698786e-01 1.18476339e-01 -4.57373470e-01 7.39483237e-01
1.09429336e+00 3.46513927e-01 -1.22667134e+00 4.03689384e-01
1.50549393e-02 4.52425510e-01 -1.71854600e-01 -4.49876726e-01
-1.38068810e-01 -2.56013870e-01 -3.37528735e-01 7.51387060e-01
9.75735933e-02 2.89530009e-01 2.95547619e-02 7.01619208e-01
-7.38212466e-01 5.27639270e-01 -5.56374967e-01 -2.47407943e-01
4.47380126e-01 1.13164198e+00 -6.39503121e-01 -5.87637663e-01
-7.65872300e-01 4.18133706e-01 6.28871799e-01 2.60846645e-01
-5.45624077e-01 -7.38987625e-01 3.98428977e-01 2.32327998e-01
2.35833243e-01 3.46890055e-02 -4.32370812e-01 -1.64523864e+00
7.90290415e-01 -1.63349712e+00 1.10299110e+00 -1.13357937e+00
-1.42918909e+00 8.31369042e-01 -9.22421291e-02 -9.26029563e-01
-3.88294399e-01 -6.24219477e-01 1.17116891e-01 7.36169159e-01
-1.26227677e+00 -8.54960740e-01 -4.27880108e-01 6.28619611e-01
5.59772730e-01 -5.58729097e-02 1.02569962e+00 5.86470246e-01
-4.09761101e-01 6.13651991e-01 2.55025923e-03 7.22359061e-01
8.90350282e-01 -1.45313323e+00 6.51844621e-01 5.70782423e-01
3.46160620e-01 8.65628362e-01 1.91374674e-01 -5.21941066e-01
-1.96453559e+00 -1.19906843e+00 1.58256733e+00 -1.57314563e+00
9.29666162e-01 -1.02632260e+00 -1.30036139e+00 8.78360987e-01
3.36708069e-01 -7.76777267e-02 9.06838298e-01 8.34580362e-02
-5.68434298e-01 -5.10514081e-01 -1.08243513e+00 4.37533021e-01
9.02274191e-01 -9.01839614e-01 -8.26289833e-01 5.34513712e-01
1.28778219e+00 -6.99331701e-01 -1.17009330e+00 2.55646199e-01
1.82549477e-01 -5.10314822e-01 8.93095911e-01 -9.79247749e-01
7.83080518e-01 -4.78943169e-01 -6.19259298e-01 -6.89456642e-01
1.34302482e-01 -2.62670159e-01 -3.34335625e-01 1.58520210e+00
1.04131484e+00 -4.32264268e-01 7.69435644e-01 9.28593993e-01
2.20441550e-01 -9.84392285e-01 -7.19765902e-01 -4.46946949e-01
3.25094700e-01 -4.29927886e-01 1.10566533e+00 7.46287286e-01
4.62352894e-02 6.73659384e-01 3.61940235e-01 3.13947275e-02
1.84825584e-01 5.12645960e-01 9.45934892e-01 -7.45662093e-01
4.75260206e-02 4.28681970e-02 2.11880133e-01 -1.15184295e+00
-1.44313991e-01 -1.07453179e+00 -1.48977503e-01 -1.98319137e+00
1.88182846e-01 -3.97817731e-01 1.72852069e-01 4.58624631e-01
-5.11499822e-01 4.37578298e-02 1.48773089e-01 1.29834995e-01
-9.62257802e-01 4.92404580e-01 1.37742817e+00 -3.24421287e-01
1.61386743e-01 -2.36942261e-01 -1.25097454e+00 2.09835380e-01
6.58233762e-01 -5.45994699e-01 -5.54118872e-01 -8.27944338e-01
7.36331046e-01 4.11112577e-01 -1.89295352e-01 -8.24707329e-01
3.18506300e-01 -1.72743008e-01 3.28053057e-01 -1.02443004e+00
6.23198673e-02 -9.23381925e-01 -3.84406149e-01 2.56693289e-02
-7.07869828e-01 6.65085852e-01 2.84948409e-01 2.01379031e-01
-6.86338544e-01 -4.59192619e-02 1.15252696e-01 -2.35100642e-01
-2.32916832e-01 3.58182311e-01 -2.72542894e-01 1.08324671e+00
3.83896887e-01 2.30121583e-01 -6.19603097e-01 -6.52480960e-01
-3.52730751e-01 8.69820654e-01 1.13518640e-01 8.47591996e-01
4.13178653e-01 -1.52070093e+00 -5.51441193e-01 5.43484129e-02
7.47755349e-01 2.23411217e-01 -6.99415011e-03 8.50886181e-02
-7.97043800e-01 9.25371230e-01 1.89339191e-01 -3.84744853e-02
-6.58486485e-01 9.09171999e-01 2.35707849e-01 -6.10238314e-01
-5.58630228e-02 6.67362332e-01 -2.08138868e-01 -1.35418963e+00
3.34578246e-01 -9.79741752e-01 -4.50358450e-01 1.59332812e-01
7.10595787e-01 -5.26479557e-02 4.51170921e-01 -4.16285060e-02
-4.47477549e-01 2.44308263e-01 2.66517680e-02 8.18060711e-02
1.04059422e+00 -1.55901358e-01 -5.97909570e-01 7.65694499e-01
1.23134029e+00 3.60603392e-01 -2.98975408e-01 -3.39541882e-01
4.59494412e-01 -2.43931949e-01 -8.40930283e-01 -1.08158195e+00
-7.74738908e-01 5.60463309e-01 -8.15151259e-02 4.44555759e-01
1.18881667e+00 2.84747094e-01 1.22628081e+00 1.30965185e+00
5.02468824e-01 -8.90178561e-01 1.33834571e-01 7.86791146e-01
1.14193892e+00 -1.41604507e+00 -4.27553445e-01 -6.13374770e-01
-5.01062930e-01 1.04301584e+00 9.63301063e-01 4.39262986e-01
5.91045856e-01 1.56618372e-01 7.03165174e-01 -2.30724409e-01
-1.18453789e+00 -1.57967508e-01 3.17645967e-01 6.32293284e-01
6.72275960e-01 -2.17513785e-01 -5.28168231e-02 1.11592162e+00
-9.45112824e-01 -9.58344191e-02 4.54062968e-01 1.11641359e+00
-2.53512383e-01 -1.33124995e+00 -4.30334270e-01 8.69245648e-01
-3.85990560e-01 -3.79495025e-01 -7.10654497e-01 7.77586043e-01
-2.37803921e-01 1.31786668e+00 3.42636593e-02 -2.36834675e-01
9.30729270e-01 3.72590512e-01 1.68532059e-01 -7.17183352e-01
-7.86918879e-01 -8.54420424e-01 4.75310206e-01 -4.74175811e-01
9.21787843e-02 -3.01520944e-01 -1.14491355e+00 -3.28671664e-01
1.60280049e-01 7.06894219e-01 1.85068026e-01 5.81997812e-01
3.38369220e-01 3.27300280e-01 2.45924726e-01 7.41555154e-01
-4.57147896e-01 -1.00462162e+00 -2.45010450e-01 5.27336180e-01
9.30212662e-02 -4.35619801e-02 -1.32566124e-01 3.05974036e-01] | [10.076653480529785, 7.889129161834717] |
d8f3122e-7813-446f-88ca-be9cb7c1b6b9 | acrofod-an-adaptive-method-for-cross-domain | 2209.10904 | null | https://arxiv.org/abs/2209.10904v1 | https://arxiv.org/pdf/2209.10904v1.pdf | AcroFOD: An Adaptive Method for Cross-domain Few-shot Object Detection | Under the domain shift, cross-domain few-shot object detection aims to adapt object detectors in the target domain with a few annotated target data. There exists two significant challenges: (1) Highly insufficient target domain data; (2) Potential over-adaptation and misleading caused by inappropriately amplified target samples without any restriction. To address these challenges, we propose an adaptive method consisting of two parts. First, we propose an adaptive optimization strategy to select augmented data similar to target samples rather than blindly increasing the amount. Specifically, we filter the augmented candidates which significantly deviate from the target feature distribution in the very beginning. Second, to further relieve the data limitation, we propose the multi-level domain-aware data augmentation to increase the diversity and rationality of augmented data, which exploits the cross-image foreground-background mixture. Experiments show that the proposed method achieves state-of-the-art performance on multiple benchmarks. | ['Wei-Shi Zheng', 'Shiyong Li', 'Song Xie', 'Yunmu Huang', 'Lingxiao Yang', 'Yipeng Gao'] | 2022-09-22 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 4.97134864e-01 -3.82988483e-01 -9.62238088e-02 -1.80350974e-01
-8.41566503e-01 -1.94537073e-01 4.95668471e-01 -2.29252890e-01
-3.92493308e-01 7.85547078e-01 2.96899471e-02 2.88844526e-01
6.82736337e-02 -5.31211138e-01 -4.48618293e-01 -8.60699415e-01
5.17492831e-01 3.73906821e-01 1.13932288e+00 -1.65213943e-02
2.25871205e-01 3.00175875e-01 -1.68013084e+00 2.78868139e-01
9.92251337e-01 1.08396912e+00 4.17938471e-01 2.44226232e-01
-5.29448807e-01 5.92239261e-01 -7.00311303e-01 2.41085943e-02
4.31732506e-01 -6.16843343e-01 -3.03166270e-01 7.07367361e-01
1.47335589e-01 -5.19373596e-01 -5.27832061e-02 1.47772479e+00
6.02708936e-01 1.84230223e-01 4.50187445e-01 -1.19669032e+00
-4.17834014e-01 2.10172668e-01 -1.02011466e+00 6.36570573e-01
-1.25730306e-01 5.25520980e-01 3.07841718e-01 -1.19705379e+00
4.61688906e-01 1.32501316e+00 2.56359607e-01 8.39228809e-01
-1.03743756e+00 -8.72123301e-01 5.77092350e-01 2.77432650e-01
-1.54871464e+00 -6.55029058e-01 1.02189338e+00 -4.66083229e-01
1.11271173e-01 1.21813931e-01 5.50320446e-01 1.11250687e+00
-4.68991518e-01 8.78037333e-01 1.09779847e+00 -3.28529149e-01
4.55525994e-01 4.34975564e-01 1.42739087e-01 4.24412221e-01
3.41812670e-01 9.99535769e-02 -3.50835502e-01 -2.06750274e-01
5.48723698e-01 8.17573816e-03 -2.27279648e-01 -6.84926808e-01
-1.19063079e+00 5.90913534e-01 -1.11044543e-02 2.32624754e-01
-3.02209616e-01 -3.78211766e-01 5.92934430e-01 -1.77490845e-01
4.89647597e-01 3.18504013e-02 -5.17729223e-01 6.77385926e-02
-8.05542231e-01 1.02027319e-01 2.30410174e-01 1.04644680e+00
7.45132267e-01 3.44258130e-01 -3.96079391e-01 9.92127955e-01
2.07947686e-01 4.63965744e-01 6.23096645e-01 -4.74742055e-01
6.65597141e-01 8.24978054e-01 4.31645334e-01 -7.27599442e-01
-2.13904912e-03 -6.42605960e-01 -5.90347648e-01 2.61160523e-01
4.49131042e-01 -1.24979310e-01 -1.15162790e+00 1.62193394e+00
7.97499120e-01 4.47625309e-01 1.03249460e-01 1.12357366e+00
7.99393177e-01 6.28572762e-01 1.15880266e-01 -7.20210910e-01
1.30026281e+00 -9.16354299e-01 -7.64761150e-01 -5.39735794e-01
2.86163747e-01 -7.66555727e-01 1.15459085e+00 1.74214378e-01
-7.59247482e-01 -8.00270140e-01 -1.11691618e+00 6.34629190e-01
-2.11196795e-01 3.00202549e-01 1.40123710e-01 5.79208851e-01
-1.45360619e-01 4.11023274e-02 -4.97651696e-01 -3.93843621e-01
7.15121210e-01 1.19464464e-01 -5.65934591e-02 -2.77601689e-01
-9.93553877e-01 5.28741181e-01 9.37278032e-01 -5.58053739e-02
-1.01284325e+00 -7.88823843e-01 -5.40471792e-01 -1.73412994e-01
9.31361973e-01 -2.61683941e-01 1.02642596e+00 -1.29618824e+00
-1.30566812e+00 7.25061715e-01 -1.57021359e-02 -2.26030305e-01
6.90995932e-01 -2.41855562e-01 -7.51935244e-01 -1.64978147e-01
1.39575377e-01 3.36921632e-01 8.89643013e-01 -1.53258049e+00
-1.20486605e+00 -4.94638592e-01 -4.51518893e-01 3.80342543e-01
-6.23033524e-01 1.14468075e-01 -8.21695209e-01 -6.71829283e-01
3.88346538e-02 -5.68600893e-01 -3.00746500e-01 9.56494361e-03
-3.28712612e-01 2.25378051e-02 1.16372728e+00 -3.26437145e-01
1.26547992e+00 -2.36480546e+00 -2.41932780e-01 -1.58715680e-01
7.94446394e-02 8.14308584e-01 -1.29367352e-01 -2.25584865e-01
1.66370824e-01 -1.71208695e-01 -3.09147418e-01 4.33229543e-02
-3.40572923e-01 8.16904530e-02 -3.05516392e-01 4.07985568e-01
4.09346133e-01 4.06855762e-01 -1.00606644e+00 -8.50605369e-01
2.99799591e-01 -1.06619753e-01 -2.26501122e-01 3.52643400e-01
-4.38624769e-01 5.17870009e-01 -7.41143107e-01 9.07881320e-01
1.27603555e+00 -7.89965764e-02 -2.50127375e-01 -2.95594007e-01
-1.61518767e-01 -2.24192038e-01 -1.65678000e+00 1.31943464e+00
3.03307235e-01 5.30788600e-02 1.41966194e-01 -9.49951530e-01
1.09878051e+00 -6.07977472e-02 5.70050359e-01 -5.83674669e-01
1.96297348e-01 2.71156132e-01 8.55211541e-02 -4.50939476e-01
2.98029572e-01 -2.03878298e-01 1.00572363e-01 2.41305027e-02
-1.13900155e-01 4.12433505e-01 8.13717246e-02 -3.42688449e-02
1.01498771e+00 1.61224738e-01 5.00990570e-01 -1.36107489e-01
7.92031229e-01 2.76924342e-01 1.19001532e+00 7.14375317e-01
-7.59589076e-01 4.99918282e-01 2.27361396e-01 -1.95664868e-01
-9.11674917e-01 -9.86275971e-01 1.35692254e-01 1.08333254e+00
6.11979008e-01 2.12703004e-01 -5.38365901e-01 -1.02889395e+00
-1.66432545e-01 5.86842954e-01 -5.22835433e-01 -3.74892116e-01
-5.31995535e-01 -1.09095514e+00 3.83485943e-01 5.07668912e-01
8.03443968e-01 -1.00261152e+00 -6.36318266e-01 2.20608801e-01
-1.26302168e-01 -1.15222800e+00 -6.17352962e-01 1.41169176e-01
-7.74596572e-01 -1.04029596e+00 -8.72093797e-01 -7.15649664e-01
7.10839868e-01 6.51511550e-01 6.96579993e-01 -1.71378776e-01
-1.54440567e-01 -1.76433884e-02 -4.60679293e-01 -3.99586409e-01
-3.94519985e-01 -2.53592581e-01 7.05666319e-02 4.37450826e-01
7.17529774e-01 -1.54689059e-01 -4.68807340e-01 5.98590970e-01
-8.67950797e-01 -3.99201885e-02 7.96238661e-01 9.31044400e-01
7.96341181e-01 1.53138295e-01 7.42365658e-01 -7.97718525e-01
2.66712308e-01 -5.24019539e-01 -7.67897427e-01 3.00148368e-01
-4.02077496e-01 -3.00945848e-01 5.78012347e-01 -1.04585803e+00
-1.44000387e+00 4.27508354e-01 2.65882730e-01 -7.45004535e-01
-4.15163875e-01 -8.15501064e-02 -6.93041086e-01 1.80975180e-02
8.07173967e-01 5.20471334e-01 -1.29748300e-01 -5.56208432e-01
1.59521371e-01 7.18263686e-01 6.50391042e-01 -4.21273261e-01
9.59914505e-01 5.69156647e-01 -2.76131749e-01 -5.54143131e-01
-8.26640069e-01 -7.46596158e-01 -4.80207831e-01 -2.98891664e-01
6.93183482e-01 -9.03929830e-01 -8.94628763e-02 5.58882117e-01
-9.88261878e-01 -4.54035029e-02 -3.81033361e-01 4.66019213e-01
-2.43051142e-01 5.25163054e-01 -4.54958938e-02 -1.13134313e+00
-2.34947413e-01 -9.88239348e-01 8.07877183e-01 5.42625606e-01
3.26612413e-01 -3.17763656e-01 -1.78259164e-01 1.17519438e-01
6.88985214e-02 2.52784431e-01 5.01850247e-01 -1.10771883e+00
-6.14554465e-01 -7.60814026e-02 -4.82550591e-01 3.46419066e-01
4.32485163e-01 -1.29649535e-01 -8.35447431e-01 -1.77602932e-01
2.16462091e-01 -2.01118320e-01 7.57761776e-01 1.42675295e-01
9.51694131e-01 -1.07259624e-01 -5.93639553e-01 4.21874017e-01
1.40063214e+00 5.06274641e-01 3.44282329e-01 2.78949559e-01
6.04310155e-01 4.95487839e-01 1.34424877e+00 8.28325093e-01
-1.04738571e-01 6.46321774e-01 4.01176006e-01 -3.62566747e-02
-2.18718886e-01 -1.45299077e-01 2.90274978e-01 3.43268096e-01
4.36926723e-01 -1.19805738e-01 -7.46018767e-01 8.62997234e-01
-1.94401896e+00 -9.14061844e-01 -1.96156979e-01 2.44656730e+00
7.48858690e-01 5.78521073e-01 4.81186092e-01 -4.46788408e-02
1.33389187e+00 -1.24794831e-02 -9.93567169e-01 4.32082713e-01
-1.99323043e-01 -2.93913156e-01 5.07366776e-01 2.63772085e-02
-1.20352864e+00 9.96900499e-01 5.43215656e+00 1.17819905e+00
-9.71780002e-01 3.14051509e-01 4.12404269e-01 -1.29099071e-01
1.23436950e-01 -3.41661200e-02 -1.14705813e+00 9.25620794e-01
2.90123940e-01 -2.02073142e-01 2.44649500e-01 9.70610857e-01
1.83994174e-02 -1.05494589e-01 -6.90868497e-01 8.05263400e-01
1.40412837e-01 -1.04451752e+00 -2.10572369e-02 -6.49456307e-02
7.18279421e-01 -2.69754887e-01 -1.12058744e-01 4.97787833e-01
2.19478995e-01 -1.03429988e-01 7.18232930e-01 3.41369480e-01
5.92034936e-01 -6.25950038e-01 5.24840593e-01 6.39665306e-01
-1.27856028e+00 -3.07446808e-01 -6.38800859e-01 3.85825664e-01
1.75894666e-02 5.68839073e-01 -7.53733635e-01 4.26046461e-01
5.68090439e-01 4.41627949e-01 -6.50995672e-01 1.25791502e+00
2.55885661e-01 5.35158277e-01 -3.06038588e-01 7.99637139e-02
-2.00378597e-02 1.27847180e-01 8.64226580e-01 1.10871673e+00
2.62118727e-01 2.29962215e-01 5.13004124e-01 8.71394277e-01
3.67140859e-01 1.37105614e-01 -3.91093671e-01 1.42226130e-01
7.82181919e-01 1.12365437e+00 -8.08828950e-01 -5.84809363e-01
-5.92824936e-01 9.54939008e-01 3.58300097e-02 3.41476530e-01
-1.07816708e+00 -1.55820340e-01 4.56242472e-01 1.43416986e-01
5.18145323e-01 2.48330146e-01 -1.92022696e-01 -1.05506301e+00
1.11524716e-01 -9.60539103e-01 6.75568342e-01 -5.31933963e-01
-1.47918212e+00 3.93176705e-01 -8.06758739e-03 -1.71821678e+00
2.87364632e-01 -3.51537019e-01 -5.34904242e-01 7.36845791e-01
-1.43144083e+00 -1.04856050e+00 -6.32336736e-01 7.20730364e-01
9.00973439e-01 -3.92645717e-01 3.05982620e-01 6.21015012e-01
-6.65954411e-01 5.65807939e-01 9.86338407e-02 -4.71373647e-02
8.68940234e-01 -7.86643267e-01 5.49953915e-02 1.16277277e+00
-4.25072223e-01 2.99656600e-01 7.41024554e-01 -9.06287909e-01
-9.75293458e-01 -1.39507830e+00 1.28946111e-01 -9.84917283e-02
4.58390266e-01 -2.04066619e-01 -1.16902995e+00 3.60873640e-01
-3.50560635e-01 4.66274291e-01 3.62997085e-01 -5.25937974e-01
-1.06134370e-01 -4.73234683e-01 -1.32983947e+00 5.80972254e-01
9.73947644e-01 9.51935723e-02 -6.46285832e-01 2.03330651e-01
7.63995111e-01 -3.74924272e-01 -3.48012567e-01 6.96867406e-01
2.21786216e-01 -9.20112252e-01 9.51041281e-01 -5.46501577e-01
1.35888476e-02 -8.57025802e-01 -3.17447156e-01 -9.95325804e-01
-4.49712515e-01 -4.00276870e-01 -3.98845136e-01 1.74019241e+00
-6.40260130e-02 -4.84352022e-01 9.81032491e-01 4.08712476e-01
-7.94353709e-02 -4.86338049e-01 -9.33963716e-01 -1.08112836e+00
-3.08009475e-01 -7.54287913e-02 6.06086135e-01 7.48482406e-01
-4.24173206e-01 2.06383526e-01 -6.06492221e-01 3.79550695e-01
8.02878976e-01 1.38391703e-01 1.01060295e+00 -1.17712879e+00
-2.67545015e-01 -3.00162166e-01 -2.52498865e-01 -1.16948378e+00
-2.54925996e-01 -2.48257592e-01 5.70826113e-01 -1.10186779e+00
5.69339275e-01 -4.71554101e-01 -6.53790057e-01 9.99822170e-02
-6.26545787e-01 6.01873510e-02 1.69388160e-01 3.20316404e-01
-9.71961915e-01 6.93629742e-01 1.23951209e+00 -1.07530385e-01
-4.18310761e-01 1.88475043e-01 -7.45960355e-01 6.45331085e-01
6.05889380e-01 -6.08289480e-01 -3.85874420e-01 -8.42212811e-02
-6.08102560e-01 -6.91505224e-02 3.41696054e-01 -1.30387974e+00
3.97346504e-02 -6.97075605e-01 5.52611947e-01 -8.88326108e-01
3.20971996e-01 -9.76420760e-01 -8.61253664e-02 5.34254730e-01
-1.62722473e-03 -4.68641132e-01 2.33516693e-01 9.92000103e-01
-8.66618007e-02 -3.11136842e-01 1.31345928e+00 -2.18726903e-01
-1.21890116e+00 2.67978787e-01 -2.52022475e-01 2.41548330e-01
1.40996194e+00 -3.66374612e-01 -4.45522100e-01 1.29333600e-01
-3.84928107e-01 3.90046239e-01 5.66535473e-01 4.74735469e-01
4.28560317e-01 -1.42156577e+00 -6.10195220e-01 2.17175886e-01
5.15296876e-01 -3.28059383e-02 3.91895980e-01 7.29439020e-01
9.43989530e-02 5.32110408e-03 -1.93317458e-01 -7.21517265e-01
-1.30054557e+00 1.08831036e+00 3.21519792e-01 -1.67285711e-01
-5.02239347e-01 8.01653326e-01 5.08337796e-01 -1.57805845e-01
1.56498238e-01 9.22354832e-02 -1.97794393e-01 -2.89064255e-02
7.44838655e-01 5.04907906e-01 -3.08609217e-01 -5.86043298e-01
-6.27980709e-01 4.01477605e-01 -2.02165261e-01 4.60513011e-02
9.61328566e-01 -3.13184649e-01 3.49850655e-01 3.47817481e-01
6.97182775e-01 1.59780178e-02 -1.74125659e+00 -6.15922034e-01
6.69317544e-02 -8.80568445e-01 -2.00832710e-01 -8.60978246e-01
-8.58641446e-01 7.36617148e-01 1.04365110e+00 4.92050834e-02
1.32062018e+00 -1.18121773e-01 7.66344368e-01 -3.22416574e-02
5.83550744e-02 -1.37869906e+00 4.72602844e-01 2.88429618e-01
4.02289689e-01 -1.36711466e+00 9.04840007e-02 -5.12974918e-01
-9.20023859e-01 8.73773575e-01 1.34692121e+00 9.94614437e-02
3.92881155e-01 2.16183618e-01 -1.10645900e-02 6.66051656e-02
-5.45564234e-01 -6.66641951e-01 1.73266709e-01 8.91352654e-01
-2.49590188e-01 -3.19967449e-01 -3.53869706e-01 9.11807299e-01
7.59277761e-01 1.70935780e-01 3.75311911e-01 9.34287310e-01
-9.52309251e-01 -7.45706797e-01 -8.73085499e-01 3.91182065e-01
-3.64645720e-01 1.86205298e-01 -3.17287564e-01 6.51724696e-01
5.98313212e-01 9.68279779e-01 -1.34693906e-01 -3.36226135e-01
4.60603118e-01 1.12763621e-01 2.09257111e-01 -6.79198563e-01
-8.63202214e-02 3.92090738e-01 -1.86925709e-01 -2.04913422e-01
-2.70962119e-01 -7.25980937e-01 -1.15993690e+00 2.67635345e-01
-7.19272971e-01 -1.24561042e-01 1.96640596e-01 9.25183535e-01
3.60990703e-01 6.37730300e-01 5.36584258e-01 -6.10736966e-01
-7.91702628e-01 -1.04139221e+00 -4.48958814e-01 5.04699230e-01
3.14265221e-01 -9.13710833e-01 -3.20381075e-01 1.40950426e-01] | [9.38917064666748, 1.4963964223861694] |
e5449212-db5f-4bb0-84d1-44d5e6cb91b6 | segmentation-is-all-you-need | 1904.13300 | null | https://arxiv.org/abs/1904.13300v3 | https://arxiv.org/pdf/1904.13300v3.pdf | Segmentation is All You Need | Region proposal mechanisms are essential for existing deep learning approaches to object detection in images. Although they can generally achieve a good detection performance under normal circumstances, their recall in a scene with extreme cases is unacceptably low. This is mainly because bounding box annotations contain much environment noise information, and non-maximum suppression (NMS) is required to select target boxes. Therefore, in this paper, we propose the first anchor-free and NMS-free object detection model called weakly supervised multimodal annotation segmentation (WSMA-Seg), which utilizes segmentation models to achieve an accurate and robust object detection without NMS. In WSMA-Seg, multimodal annotations are proposed to achieve an instance-aware segmentation using weakly supervised bounding boxes; we also develop a run-data-based following algorithm to trace contours of objects. In addition, we propose a multi-scale pooling segmentation (MSP-Seg) as the underlying segmentation model of WSMA-Seg to achieve a more accurate segmentation and to enhance the detection accuracy of WSMA-Seg. Experimental results on multiple datasets show that the proposed WSMA-Seg approach outperforms the state-of-the-art detectors. | ['Thomas Lukasiewicz', 'Zhenghua Xu', 'Weiyang Wang', 'Zehua Cheng', 'Yuxiang Wu'] | 2019-04-30 | null | null | null | null | ['head-detection', 'robust-object-detection'] | ['computer-vision', 'computer-vision'] | [ 2.82719493e-01 -6.65540621e-02 -3.01692192e-03 -5.43172359e-01
-1.02271569e+00 -3.71248513e-01 4.12166476e-01 3.27222735e-01
-6.42711520e-01 2.02692226e-01 -3.83212328e-01 2.37691216e-02
3.68744940e-01 -7.79112875e-01 -7.07320273e-01 -8.08618188e-01
3.27468663e-01 2.05189481e-01 1.40096962e+00 -9.23550036e-03
2.93728650e-01 4.21347201e-01 -1.27061677e+00 2.36826673e-01
1.03673649e+00 1.08423924e+00 6.08526707e-01 4.71894979e-01
-4.14512992e-01 2.89612144e-01 -5.79112887e-01 -2.69964039e-01
3.91414911e-01 -3.47186923e-01 -6.17293000e-01 3.53894889e-01
4.05054450e-01 -4.57570046e-01 1.26500219e-01 1.24580562e+00
2.25778297e-01 1.04328386e-01 7.55766094e-01 -1.00086129e+00
-1.82384223e-01 4.75201249e-01 -1.20128930e+00 9.01007131e-02
-2.03455314e-01 2.14328002e-02 6.94741189e-01 -1.08913505e+00
2.97482669e-01 1.30098462e+00 6.17435932e-01 4.04362559e-01
-1.05137491e+00 -6.85830891e-01 3.78246099e-01 -2.62173414e-01
-1.52379310e+00 -2.21262276e-01 6.31461561e-01 -3.13331157e-01
3.17968547e-01 2.63303965e-01 5.18646240e-01 3.54849994e-01
1.35683000e-01 1.20295465e+00 9.49256718e-01 -5.10356724e-01
2.51242429e-01 2.14809716e-01 3.61969262e-01 9.31215405e-01
4.67435569e-01 -4.09386277e-01 -2.23945752e-01 -1.17780864e-01
9.72574711e-01 1.24891894e-02 -7.19342083e-02 -3.97953123e-01
-1.07355535e+00 6.21564686e-01 6.91370845e-01 3.40436548e-01
-3.29813033e-01 1.69099297e-03 3.46095771e-01 -4.26814497e-01
3.77383471e-01 7.47602992e-03 -1.01694889e-01 5.35477996e-01
-1.32769465e+00 2.83855110e-01 3.51442009e-01 9.95589137e-01
8.24694097e-01 -1.42275169e-01 -3.55835646e-01 9.14920032e-01
4.65205908e-01 2.88059503e-01 2.92477041e-01 -6.51679575e-01
3.71141374e-01 9.61090207e-01 2.44709820e-01 -8.84053349e-01
-5.36817491e-01 -5.68272948e-01 -6.81982040e-01 2.97095865e-01
5.60513914e-01 1.64260194e-02 -1.34469688e+00 1.30142355e+00
5.26836991e-01 1.92899689e-01 -8.34526345e-02 1.17748511e+00
8.57893169e-01 7.25178123e-01 3.74931574e-01 -9.82463956e-02
1.52263796e+00 -9.83145177e-01 -6.65703535e-01 -5.55460036e-01
5.53408086e-01 -7.73670733e-01 9.30292964e-01 1.60147995e-01
-1.01143634e+00 -7.53420353e-01 -1.10539317e+00 -3.14102918e-02
-2.05471426e-01 6.03637040e-01 3.41731578e-01 6.20395243e-01
-6.86546326e-01 6.42580464e-02 -1.06747472e+00 -4.72963244e-01
8.07736099e-01 2.56238878e-01 -1.01198345e-01 9.64122713e-02
-6.61743879e-01 7.37545907e-01 8.70347500e-01 3.26389045e-01
-7.59320796e-01 -2.31001094e-01 -8.06578815e-01 -4.43403386e-02
6.92448437e-01 -2.47288659e-01 1.17470598e+00 -8.22986722e-01
-9.82348323e-01 9.38327670e-01 -2.28452161e-01 -3.68882835e-01
5.78707874e-01 -1.62612423e-01 -4.46222201e-02 2.14229390e-01
1.79305539e-01 1.00275469e+00 8.02957296e-01 -1.71614385e+00
-1.05801022e+00 -3.35809410e-01 -1.61291510e-01 1.07894041e-01
-2.64794558e-01 2.05780417e-01 -9.56582129e-01 -6.94808185e-01
6.56540990e-01 -6.85021758e-01 -4.91171092e-01 1.00588180e-01
-7.74587989e-01 -3.72276187e-01 1.19308650e+00 -5.09934843e-01
1.30089617e+00 -2.06611586e+00 -3.58732134e-01 2.17787951e-01
1.36918202e-01 5.73180497e-01 -1.17143847e-01 -7.37286583e-02
5.54532528e-01 -7.97552541e-02 -7.37109601e-01 -6.39679551e-01
-2.66050339e-01 1.21974960e-01 -8.41787606e-02 4.79340762e-01
4.72935289e-01 6.60583317e-01 -5.63309252e-01 -1.02804637e+00
3.74964625e-01 1.80558145e-01 -3.21372718e-01 2.05642074e-01
-2.39951909e-01 1.48334369e-01 -6.94165826e-01 1.10359418e+00
1.02326465e+00 -1.56451881e-01 -1.05723992e-01 -3.86933029e-01
-3.37753832e-01 -4.40239608e-01 -1.57174611e+00 1.46094632e+00
2.62086913e-02 2.83129632e-01 3.60683978e-01 -8.69321585e-01
1.13059235e+00 -1.79688469e-01 1.14087634e-01 -3.98901701e-01
3.37419689e-01 4.19831067e-01 6.01254171e-04 -4.19169813e-01
7.53035843e-01 8.64747316e-02 -2.95786094e-02 1.53855840e-02
-1.55350029e-01 -1.23212911e-01 3.37816834e-01 2.42886350e-01
5.37791014e-01 1.98163196e-01 3.45435500e-01 -4.25802886e-01
7.52042413e-01 1.20868102e-01 1.04389167e+00 9.50017691e-01
-5.65781057e-01 9.47460592e-01 1.78578332e-01 -1.88379139e-01
-7.41226137e-01 -9.67851222e-01 -3.65478069e-01 1.15714300e+00
7.62970507e-01 -4.94039953e-02 -1.27369642e+00 -7.63648152e-01
-3.12820703e-01 4.33341533e-01 -2.93264300e-01 1.55682102e-01
-6.56222641e-01 -1.09366989e+00 4.86824036e-01 7.90821970e-01
1.02753937e+00 -1.00605559e+00 -6.92634344e-01 3.91713530e-01
-1.19830236e-01 -1.17834353e+00 -5.98065555e-01 8.89655501e-02
-7.14405954e-01 -9.92318749e-01 -9.70945179e-01 -1.07633901e+00
8.50505829e-01 3.82075906e-01 4.56717163e-01 2.15889886e-01
-3.59532326e-01 -2.27056473e-01 -2.75568396e-01 -4.32102531e-01
-2.77828604e-01 -2.16096044e-02 -1.72653094e-01 2.74952263e-01
3.90204847e-01 1.53212190e-01 -7.40525782e-01 4.59203690e-01
-1.21341562e+00 3.17958742e-02 9.47940111e-01 6.47395611e-01
8.13565493e-01 -7.89298937e-02 6.92039967e-01 -8.44922841e-01
2.51445860e-01 -3.63246724e-02 -8.93676519e-01 2.97775656e-01
-3.13772202e-01 -3.24661881e-01 2.52026856e-01 -4.65011060e-01
-1.22361505e+00 5.46824932e-01 -2.00255901e-01 -1.87582776e-01
-4.35998201e-01 1.76597103e-01 -3.02304804e-01 -3.08913201e-01
4.29866344e-01 3.76701474e-01 -1.70512006e-01 -5.30831695e-01
1.88775867e-01 7.70472407e-01 7.85149276e-01 -2.59564072e-01
6.70518458e-01 6.64404571e-01 -1.55351475e-01 -8.74922633e-01
-9.38866496e-01 -8.69829059e-01 -7.69583583e-01 -2.67529279e-01
1.07479846e+00 -8.23100507e-01 -3.02169472e-01 7.55982935e-01
-1.21650386e+00 -2.12143466e-01 1.66090205e-01 3.34067613e-01
-2.38704771e-01 4.11476791e-01 -6.63495302e-01 -1.15129089e+00
-3.43002379e-01 -1.41730571e+00 1.36044490e+00 7.21474767e-01
5.96505478e-02 -6.13505900e-01 -4.91188109e-01 4.74485517e-01
2.06750855e-01 3.16812247e-01 4.21114296e-01 -8.40871990e-01
-9.20192599e-01 -4.11338925e-01 -7.27226794e-01 3.49384755e-01
-7.60491565e-02 8.17926452e-02 -1.02027404e+00 -2.04475343e-01
-3.46367717e-01 -1.44695342e-01 1.22334623e+00 6.95650697e-01
1.21499288e+00 2.03298122e-01 -7.79834330e-01 3.53769749e-01
1.40198064e+00 2.98826873e-01 4.70180929e-01 4.24430788e-01
7.36294329e-01 5.45224130e-01 1.15457308e+00 2.28791550e-01
2.38273904e-01 4.86085176e-01 5.05515337e-01 -6.22089326e-01
-7.57825226e-02 -1.35676995e-01 2.61017770e-01 3.79016668e-01
1.31907418e-01 -3.50714684e-01 -8.20223570e-01 6.73755109e-01
-1.88039005e+00 -4.41279829e-01 -5.79938352e-01 2.00160265e+00
6.06068730e-01 5.09704053e-01 3.29912931e-01 7.34918416e-02
1.03333998e+00 1.54518515e-01 -4.50695693e-01 -1.56136174e-02
-1.03414491e-01 -1.31797478e-01 6.09461367e-01 3.32971364e-01
-1.60326457e+00 1.18357563e+00 5.44739819e+00 1.21275878e+00
-9.13237989e-01 2.22716764e-01 6.86129391e-01 3.98773074e-01
2.81836420e-01 -1.10120460e-01 -1.15743995e+00 4.17729139e-01
8.49003792e-02 3.90691072e-01 -5.43011308e-01 1.00576949e+00
3.61540765e-01 -5.40754557e-01 -6.36455297e-01 7.76229143e-01
2.98408326e-02 -1.08067489e+00 2.63585802e-02 -2.67315179e-01
7.56291270e-01 -2.28728890e-01 -2.93226689e-01 2.17358887e-01
-1.64488062e-01 -5.11511922e-01 9.08474088e-01 3.25134724e-01
4.00879025e-01 -7.68737257e-01 9.52085972e-01 6.39195681e-01
-1.52304602e+00 -9.35809128e-03 -5.29548764e-01 4.46737200e-01
3.13178450e-01 6.72368884e-01 -6.89431310e-01 3.98097545e-01
6.92818105e-01 1.47303939e-01 -7.05086529e-01 1.64132929e+00
-5.52448556e-02 6.84797347e-01 -4.23952132e-01 -1.31038919e-01
5.27125537e-01 -1.64749578e-01 5.62129796e-01 1.62808371e+00
2.02906609e-01 2.45684206e-01 7.38467038e-01 1.07474351e+00
-5.73809296e-02 2.86986530e-01 5.47944345e-02 2.93199986e-01
5.98766685e-01 1.52215862e+00 -1.57313800e+00 -4.18689787e-01
-2.79947251e-01 1.01144421e+00 3.31289545e-02 3.32073033e-01
-7.24246860e-01 -5.27528584e-01 2.37698063e-01 2.50601500e-01
2.81707495e-01 -1.80392876e-01 -3.54694039e-01 -7.46714890e-01
-1.55743454e-02 -4.38606530e-01 4.14570063e-01 -4.79871482e-01
-1.04073620e+00 4.22246218e-01 -4.64964937e-03 -1.11933887e+00
3.73789430e-01 -5.15811265e-01 -8.61443639e-01 7.00628340e-01
-1.49225879e+00 -1.54306638e+00 -3.90225708e-01 2.45393321e-01
8.73588085e-01 1.87869102e-01 1.92142844e-01 2.61818349e-01
-8.16675246e-01 5.23368299e-01 -2.32661247e-01 4.13979918e-01
6.28730118e-01 -1.09292257e+00 2.90629983e-01 1.26596427e+00
3.45252752e-02 4.16967988e-01 5.01283765e-01 -8.06542277e-01
-8.98626447e-01 -1.39562130e+00 3.48485172e-01 -2.21701786e-02
9.56305116e-02 -4.69919860e-01 -1.30862296e+00 4.10174400e-01
-9.28328112e-02 2.67365128e-01 1.96927130e-01 -4.04988050e-01
4.18185890e-02 -9.61267874e-02 -1.15424383e+00 4.46914017e-01
6.07699692e-01 -2.70236451e-02 -5.67065120e-01 2.41361037e-01
6.52891755e-01 -5.50359130e-01 -3.60979944e-01 6.69586658e-01
3.15076441e-01 -9.64996874e-01 8.53390813e-01 1.21009879e-01
-3.57588939e-02 -9.11640108e-01 7.76174143e-02 -7.47353315e-01
-1.06552646e-01 -1.65296406e-01 1.54944479e-01 1.52876329e+00
3.06561440e-01 -3.93456429e-01 9.52886820e-01 4.38413352e-01
-3.18197936e-01 -7.18495071e-01 -8.29672337e-01 -7.36298144e-01
-1.93228155e-01 -4.49268788e-01 3.74348074e-01 7.90764630e-01
-4.41815913e-01 -1.30312696e-01 -6.43918365e-02 5.59915900e-01
8.98868620e-01 5.41152395e-02 6.75071180e-01 -1.07648301e+00
7.82174543e-02 -5.65469682e-01 -4.23062950e-01 -1.17266762e+00
-2.74278760e-01 -5.55588961e-01 6.42688513e-01 -1.79413450e+00
4.51641738e-01 -4.95089769e-01 -3.24480176e-01 4.04148340e-01
-6.24270797e-01 6.55265927e-01 1.76706642e-01 2.04217732e-01
-8.20105076e-01 4.13290024e-01 1.24193072e+00 5.03795594e-02
-4.21983331e-01 1.97141334e-01 -3.37543219e-01 1.13895535e+00
6.87050939e-01 -4.56737518e-01 -7.23935515e-02 -5.11817671e-02
-5.08501470e-01 -2.58070976e-01 4.33426678e-01 -1.03508830e+00
4.58079576e-01 -2.43031979e-02 4.09157634e-01 -1.23520529e+00
3.26089859e-01 -5.35749853e-01 -4.92669731e-01 4.33256596e-01
-2.11169764e-01 -5.41444600e-01 2.81345755e-01 6.49305284e-01
-2.44005471e-01 -4.47670102e-01 1.14420772e+00 8.47007986e-03
-1.10679364e+00 1.81143299e-01 -4.74178523e-01 -2.68919021e-01
1.23634708e+00 -4.01370078e-01 -2.53783643e-01 2.01814920e-01
-4.82794970e-01 4.80547011e-01 3.23135167e-01 3.28151613e-01
7.91677654e-01 -9.86261249e-01 -6.17302358e-01 1.94525301e-01
2.77148038e-01 4.75461930e-01 2.17405781e-01 7.91544735e-01
-6.17101669e-01 1.24355741e-01 1.97373986e-01 -8.54285181e-01
-1.37633979e+00 1.44122332e-01 3.23851019e-01 1.19028419e-01
-5.55561900e-01 1.05927074e+00 5.63270569e-01 -3.04257959e-01
2.88504958e-01 -6.03574038e-01 -1.80368841e-01 -2.95446161e-02
5.69604814e-01 3.22127312e-01 -2.15543397e-02 -6.27807021e-01
-4.26204562e-01 5.78018546e-01 -3.10113460e-01 7.62837827e-02
8.10797155e-01 -1.63951278e-01 -4.85149100e-02 1.68379366e-01
6.32298291e-01 -7.57610723e-02 -1.42213464e+00 -2.72667229e-01
1.77128792e-01 -3.87423247e-01 3.40926200e-01 -6.14573359e-01
-1.02191329e+00 7.88608074e-01 7.62583792e-01 5.13142273e-02
1.13990068e+00 1.69262022e-01 9.76947129e-01 1.49963573e-01
2.34442025e-01 -1.15290880e+00 3.41124311e-02 9.70172584e-02
5.84822953e-01 -1.47514999e+00 -2.03164574e-02 -8.31934571e-01
-5.88455379e-01 1.10811853e+00 1.07805896e+00 -4.13625687e-02
3.33040029e-01 4.33901727e-01 1.98864207e-01 -2.70798244e-02
3.51802073e-02 -5.95476508e-01 4.40681905e-01 3.63175780e-01
1.81283668e-01 -7.94618055e-02 -4.59463447e-01 7.76285648e-01
5.82059205e-01 -2.88273692e-01 4.20057029e-01 1.00395799e+00
-1.18676984e+00 -8.73700917e-01 -8.06567550e-01 4.91476238e-01
-5.04874647e-01 2.34408095e-01 -2.97288120e-01 9.83759940e-01
3.62116396e-01 9.52497900e-01 1.18355006e-01 -9.07402337e-02
1.72235236e-01 -9.00133476e-02 1.93158433e-01 -7.06654251e-01
-4.30291116e-01 5.85180342e-01 -2.34813660e-01 -4.04070556e-01
-5.50410271e-01 -6.01918280e-01 -1.82176244e+00 2.99721658e-01
-9.06458974e-01 -2.06769839e-01 6.18302405e-01 1.02197695e+00
-9.85941812e-02 5.11940598e-01 1.85445055e-01 -9.91088331e-01
-2.30213329e-01 -9.28542435e-01 -6.66032732e-01 1.25074565e-01
2.75764912e-01 -6.02488697e-01 -8.26053470e-02 1.22005433e-01] | [9.324237823486328, 0.7451703548431396] |
8776e8b4-6464-4eaf-8013-ab0746863f09 | failure-detection-for-motion-prediction-of | 2301.04421 | null | https://arxiv.org/abs/2301.04421v2 | https://arxiv.org/pdf/2301.04421v2.pdf | Failure Detection for Motion Prediction of Autonomous Driving: An Uncertainty Perspective | Motion prediction is essential for safe and efficient autonomous driving. However, the inexplicability and uncertainty of complex artificial intelligence models may lead to unpredictable failures of the motion prediction module, which may mislead the system to make unsafe decisions. Therefore, it is necessary to develop methods to guarantee reliable autonomous driving, where failure detection is a potential direction. Uncertainty estimates can be used to quantify the degree of confidence a model has in its predictions and may be valuable for failure detection. We propose a framework of failure detection for motion prediction from the uncertainty perspective, considering both motion uncertainty and model uncertainty, and formulate various uncertainty scores according to different prediction stages. The proposed approach is evaluated based on different motion prediction algorithms, uncertainty estimation methods, uncertainty scores, etc., and the results show that uncertainty is promising for failure detection for motion prediction but should be used with caution. | ['Hong Wang', 'Jun Li', 'Liang Peng', 'Yanchao Xu', 'Wenbo Shao'] | 2023-01-11 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [-1.44128531e-01 1.33691728e-01 -2.56818384e-01 -4.23930109e-01
-3.01267087e-01 -2.71729797e-01 7.72260666e-01 2.22066641e-01
-2.96462059e-01 9.55614448e-01 -1.19188003e-01 -5.41433811e-01
-4.74066079e-01 -8.51821899e-01 -4.31484967e-01 -7.39352882e-01
-9.83394589e-03 3.11566442e-01 6.20646954e-01 -2.67386828e-02
4.52069938e-01 6.81102753e-01 -2.07767320e+00 -2.07517207e-01
1.06845415e+00 1.13990235e+00 2.45238155e-01 4.77289140e-01
-6.20704070e-02 7.90345371e-01 -4.87533778e-01 -3.32927145e-02
-3.04652583e-02 3.21915261e-02 -4.86385912e-01 -3.81524235e-01
-4.54444766e-01 -2.19986886e-01 -2.63175722e-02 1.07375896e+00
-1.46996111e-01 2.86224753e-01 8.49314868e-01 -1.78160906e+00
4.85894859e-01 4.06144321e-01 2.60292441e-01 1.50642395e-01
1.13362320e-01 3.23699117e-01 4.80667830e-01 -6.75120354e-01
3.48078221e-01 9.19308782e-01 4.99251544e-01 4.19649929e-01
-7.48968542e-01 -4.10251826e-01 -3.93878706e-02 6.53488576e-01
-1.47055185e+00 -4.03038651e-01 6.45785391e-01 -7.19142258e-01
7.68416405e-01 3.13753188e-01 5.54854453e-01 6.27710164e-01
1.15178716e+00 5.30996501e-01 8.13997090e-01 -6.41812533e-02
6.99852467e-01 1.00414693e-01 3.82500142e-02 2.48300225e-01
6.99954689e-01 6.14902854e-01 -4.42570776e-01 -8.35877750e-03
-4.34833467e-02 -2.57395834e-01 -3.53871062e-02 -3.68786126e-01
-8.52499425e-01 6.60533607e-01 1.22705489e-01 1.39576823e-01
-3.32551181e-01 2.60012358e-01 2.07635701e-01 3.81504260e-02
7.86327943e-02 1.47431031e-01 -2.27002740e-01 -4.50789005e-01
-8.10723662e-01 5.94195902e-01 5.64279377e-01 8.00022542e-01
7.22442687e-01 2.23300919e-01 -2.46581942e-01 1.71231970e-01
8.02682400e-01 5.72813570e-01 2.67413765e-01 -1.12820446e+00
1.18095815e-01 4.98954147e-01 6.78832114e-01 -1.08841276e+00
-6.69507205e-01 3.64528522e-02 -6.25891209e-01 7.40126491e-01
1.19909093e-01 -6.85555041e-02 -7.69084692e-01 1.32608283e+00
1.88697025e-01 3.02182287e-01 2.96288848e-01 7.82436848e-01
4.09833193e-01 6.65299416e-01 1.99070916e-01 -2.92469680e-01
8.48920703e-01 -4.20308769e-01 -9.66554165e-01 -3.44684333e-01
8.55689228e-01 -6.07491791e-01 3.75434101e-01 5.68779230e-01
-6.96057141e-01 -6.59634948e-01 -1.38567746e+00 5.62398374e-01
-1.97471991e-01 -1.97144732e-01 2.75927365e-01 7.64259696e-01
-6.14839852e-01 8.25390637e-01 -1.07160807e+00 2.01210063e-02
-9.59915966e-02 3.33394378e-01 -1.08943895e-01 -1.68370545e-01
-1.35405529e+00 1.53402960e+00 6.87672973e-01 5.32460213e-01
-7.55931258e-01 -9.20684189e-02 -9.76833224e-01 -2.40472361e-01
1.29250705e-01 -4.20152128e-01 1.16373062e+00 2.65497286e-02
-1.26557076e+00 -1.30283520e-01 -2.92947561e-01 -7.97977686e-01
7.15752780e-01 -2.11628392e-01 -7.13114262e-01 -1.92649513e-01
5.67347221e-02 5.44303179e-01 9.19345379e-01 -1.17359018e+00
-9.85013127e-01 3.50268409e-02 -2.71046549e-01 1.03479652e-02
4.70257193e-01 -2.73392469e-01 -1.34271970e-02 -8.20887834e-02
3.99175256e-01 -1.10708582e+00 -6.07892990e-01 -3.12615186e-01
-1.31459847e-01 -2.14455441e-01 8.93050790e-01 -3.36306870e-01
1.45709145e+00 -2.02393198e+00 -4.58344594e-02 3.45284373e-01
-2.00912714e-01 -4.74128723e-02 4.40589130e-01 1.98421165e-01
5.50877213e-01 1.10598907e-01 -4.94687527e-01 -1.07210480e-01
1.11818820e-01 5.64174235e-01 -2.44605362e-01 3.11731011e-01
5.03176093e-01 5.39169669e-01 -6.80516958e-01 -3.50544095e-01
8.64394426e-01 2.30295867e-01 -5.72865643e-02 2.38505349e-01
-2.95296431e-01 4.70984221e-01 -8.06452513e-01 5.10011971e-01
6.63588583e-01 3.80900145e-01 -2.38750607e-01 7.56859481e-02
-4.34753448e-01 1.17782101e-01 -1.23099279e+00 8.63426983e-01
-3.41591001e-01 7.13320792e-01 -2.82192618e-01 -6.82624042e-01
1.15886402e+00 2.18918905e-01 3.67864400e-01 -3.54547709e-01
2.47736573e-01 4.47593778e-01 1.63797528e-01 -6.36498570e-01
1.00815189e+00 -1.74823403e-01 -4.15448695e-01 -1.39291985e-02
-4.79807794e-01 -6.52287841e-01 -4.42680018e-03 -2.41307572e-01
9.01527941e-01 2.54668504e-01 1.69781342e-01 -2.70534545e-01
7.53212154e-01 3.25484037e-01 7.90952861e-01 4.89796877e-01
-6.41474426e-01 3.27945143e-01 3.50492805e-01 -4.77760762e-01
-7.89090574e-01 -7.52666235e-01 -4.28995520e-01 2.13775150e-02
6.18190587e-01 -2.06846967e-01 -4.07549798e-01 -4.29492414e-01
1.65652260e-01 1.35289752e+00 -3.58513325e-01 -6.23242557e-01
-2.28622794e-01 -7.68151402e-01 1.25913054e-01 5.68900347e-01
3.29303592e-01 -7.17083752e-01 -9.38277125e-01 5.13724029e-01
-1.12637840e-01 -1.10599172e+00 4.25601393e-01 2.19466016e-01
-8.94542456e-01 -8.29458117e-01 -7.04690143e-02 1.01353884e-01
4.01408881e-01 2.53468543e-01 7.57525384e-01 2.23871991e-01
2.60562122e-01 2.19661891e-01 -4.59778845e-01 -5.83095133e-01
-8.20742190e-01 -4.55330074e-01 3.61148387e-01 -2.39825264e-01
4.15040821e-01 -2.17100717e-02 -4.61995453e-01 8.36864173e-01
-8.12150598e-01 -3.00865740e-01 2.87208855e-01 6.30881011e-01
5.12899578e-01 8.06023419e-01 6.26664698e-01 -2.81607181e-01
6.21622264e-01 -6.55289888e-01 -9.65177715e-01 -3.14491987e-02
-1.05873525e+00 2.97924161e-01 2.28615537e-01 1.88470796e-01
-1.12461627e+00 1.98621809e-01 -1.82698518e-01 -2.80197263e-01
-3.66809964e-01 6.21422887e-01 -1.48206472e-01 -1.19788401e-01
4.25144851e-01 -3.52913700e-02 4.33391109e-02 2.79863216e-02
7.26804789e-03 4.95458663e-01 3.01319450e-01 -2.01706856e-01
5.48041761e-01 3.25403720e-01 5.03627479e-01 -7.50979006e-01
-7.73412436e-02 -3.02881926e-01 -6.15323424e-01 -8.19842458e-01
8.29995096e-01 -8.32603514e-01 -4.42227125e-01 3.87165725e-01
-1.10095406e+00 -3.36403586e-02 6.08578622e-02 7.62924731e-01
-7.09661126e-01 4.55467403e-01 8.61321464e-02 -1.36256218e+00
2.60845065e-01 -1.61760736e+00 6.64943099e-01 1.78723022e-01
-5.85071623e-01 -8.85816038e-01 -2.43318886e-01 2.13438913e-01
2.49464631e-01 1.58255592e-01 5.53157568e-01 -2.53385752e-01
-8.14677298e-01 -6.37169540e-01 1.76657334e-01 2.96903491e-01
-1.74195021e-01 4.36041772e-01 -8.49796772e-01 1.01759858e-01
1.96999401e-01 1.19040281e-01 7.99524069e-01 4.99635726e-01
7.10671723e-01 2.36064475e-02 -5.58914423e-01 -9.84768048e-02
1.26069951e+00 3.90633583e-01 8.20431650e-01 5.06402373e-01
1.82416320e-01 9.75307584e-01 1.48927236e+00 6.75501525e-01
3.67002159e-01 6.21899366e-01 8.05129230e-01 7.88127244e-01
4.14004356e-01 -8.83682724e-03 5.20199716e-01 4.99146909e-01
-2.29906552e-02 -3.78045887e-01 -1.16136849e+00 3.84181499e-01
-2.11227727e+00 -8.15217614e-01 -4.90895063e-01 2.52051997e+00
-1.05901239e-02 7.60295749e-01 -1.07783042e-01 5.24594605e-01
4.84046578e-01 -5.06209247e-02 -3.71529996e-01 -6.00408018e-01
2.43831024e-01 -6.29591048e-01 6.43951654e-01 7.11670637e-01
-1.03553438e+00 7.48135984e-01 7.05403662e+00 7.14533687e-01
-9.82061982e-01 -1.47552371e-01 4.26772803e-01 2.54499882e-01
-6.05061173e-01 2.42210120e-01 -1.00927901e+00 8.05377662e-01
1.27470684e+00 -3.60159904e-01 -2.98813134e-01 8.80797446e-01
5.73002934e-01 -9.24060524e-01 -7.72790134e-01 5.49332261e-01
-4.14776236e-01 -1.10250556e+00 -3.65286767e-01 -1.00239888e-01
5.05235374e-01 -7.94816837e-02 -9.29072946e-02 1.64953202e-01
2.64003366e-01 -1.04397953e+00 8.35327566e-01 1.09559882e+00
9.21360962e-03 -1.06628072e+00 1.28677595e+00 7.18837261e-01
-1.14631987e+00 -1.44908860e-01 -5.33568263e-01 -3.61668676e-01
6.01902783e-01 7.92635798e-01 -1.01055253e+00 7.48400807e-01
5.26426971e-01 4.52029794e-01 -2.91759759e-01 1.15836108e+00
-3.67105901e-01 3.75482768e-01 -5.67595840e-01 -5.28917134e-01
1.72323793e-01 -1.40211150e-01 6.96859002e-01 5.51721811e-01
7.78803885e-01 2.88875075e-03 -9.87246856e-02 8.22264671e-01
9.68100369e-01 -2.93041945e-01 -8.26074719e-01 1.58471197e-01
7.31027365e-01 7.69504607e-01 -7.17783868e-01 -9.45944712e-02
-2.62650013e-01 4.35547173e-01 -3.35637718e-01 1.04480959e-01
-8.48538816e-01 -1.13060325e-02 7.36008108e-01 3.30053903e-02
-1.35091975e-01 -5.22863150e-01 -6.69569969e-01 -6.13558829e-01
7.84004033e-02 -6.99408501e-02 6.34122938e-02 -8.20985138e-01
-6.32621050e-01 6.85658514e-01 3.20305973e-01 -1.78645813e+00
-9.16827857e-01 -6.49751425e-01 -7.20259130e-01 7.58301675e-01
-1.44913602e+00 -5.49563766e-01 -9.15369615e-02 -6.42465577e-02
3.21924955e-01 -1.11218706e-01 5.38197994e-01 -1.14140347e-01
-6.40178204e-01 2.62316726e-02 -3.49867418e-02 -8.51109326e-01
1.89244196e-01 -7.37196743e-01 1.80051640e-01 1.06343687e+00
-5.70147753e-01 -8.17627162e-02 1.42759836e+00 -9.91512597e-01
-1.39131153e+00 -1.21003115e+00 9.09459233e-01 -7.37331510e-01
6.24265671e-01 3.95344794e-01 -9.62911487e-01 2.09965855e-01
-3.57446611e-01 -1.52287617e-01 3.75414968e-01 -2.61988014e-01
4.29687947e-01 1.32437930e-01 -1.29002249e+00 6.13756299e-01
4.83057082e-01 -1.36194870e-01 -6.10956430e-01 -1.96552336e-01
5.62208593e-01 -1.00512058e-01 -9.15817082e-01 9.80236888e-01
5.32042027e-01 -1.08098650e+00 5.52666247e-01 -4.31903526e-02
-5.27824312e-02 -7.26084411e-01 -2.94558614e-01 -1.06503177e+00
-3.18330467e-01 -8.51499587e-02 -5.54119535e-02 8.89961839e-01
7.46104360e-01 -5.47653794e-01 7.24305332e-01 1.36014843e+00
-3.62051964e-01 -3.76214176e-01 -1.54637480e+00 -1.01866901e+00
8.74627084e-02 -1.36720574e+00 7.40737915e-01 2.97668159e-01
5.46917655e-02 -2.64413983e-01 -4.39478815e-01 6.62736237e-01
5.70609391e-01 -1.43349722e-01 6.61905944e-01 -1.33572805e+00
2.40028232e-01 -5.27148426e-01 -9.43578780e-01 -5.18267214e-01
1.60500675e-01 -1.95959374e-01 5.93792617e-01 -1.50738788e+00
-5.06361842e-01 -4.81771439e-01 -2.63484567e-01 5.17610200e-02
-9.70415547e-02 -1.19337715e-01 5.92224188e-02 3.55354846e-01
-2.75391996e-01 8.47174644e-01 7.42778897e-01 -1.72431365e-01
-8.17963034e-02 5.96192300e-01 7.13375211e-02 9.38181996e-01
1.04735482e+00 -5.16116202e-01 -4.73119527e-01 1.97750419e-01
4.45251435e-01 3.28362256e-01 2.55442679e-01 -1.57329440e+00
3.76041889e-01 -5.36841750e-01 5.23627698e-02 -1.04118764e+00
3.44471693e-01 -1.16189349e+00 5.87089717e-01 7.51057565e-01
1.02480218e-01 -1.62362218e-01 4.94434327e-01 7.47814775e-01
-4.54397082e-01 -7.47290075e-01 5.17345846e-01 3.54277700e-01
-1.31010997e+00 4.54272255e-02 -1.03749001e+00 -6.74399137e-01
1.34478581e+00 -3.61483634e-01 1.51652530e-01 -3.62305254e-01
-8.69281530e-01 5.79595685e-01 4.79490787e-01 5.45242310e-01
8.01726937e-01 -1.37964582e+00 -5.23027182e-01 1.44416541e-01
4.05604839e-01 -9.59481448e-02 4.63333011e-01 6.03058577e-01
-5.35371959e-01 6.23568237e-01 -1.75147071e-01 -8.77731860e-01
-8.49754393e-01 5.55664062e-01 1.66464493e-01 1.28661677e-01
-2.18865991e-01 4.76339579e-01 -2.93763012e-01 -8.19766745e-02
-2.55765058e-02 -5.77620566e-01 -4.97189403e-01 -1.59202293e-01
4.55056280e-01 6.85808122e-01 1.46001220e-01 -8.86148572e-01
-4.45320576e-01 4.99561101e-01 4.99359965e-01 -3.43737274e-01
5.29555023e-01 -6.05729222e-01 1.33900449e-01 6.75567865e-01
3.56336802e-01 -3.33906144e-01 -1.34016931e+00 3.66077960e-01
4.44618404e-01 -3.84884000e-01 4.19000834e-01 -5.30700684e-01
-4.60617691e-01 8.52092683e-01 5.89407623e-01 2.52847075e-01
7.68839002e-01 -3.09057027e-01 3.75653505e-01 5.12302458e-01
9.05616999e-01 -1.28040743e+00 -6.66519046e-01 5.00267267e-01
9.65830684e-01 -1.58131146e+00 -6.67563528e-02 -5.10304809e-01
-7.13331223e-01 1.10861456e+00 6.56962872e-01 3.20072263e-01
1.05742252e+00 2.78632104e-01 -8.16926956e-02 3.01006943e-01
-9.49233592e-01 -1.10848829e-01 4.28641528e-01 5.69125056e-01
6.74474761e-02 4.49503332e-01 -6.54391110e-01 6.85633779e-01
-7.26067182e-03 2.07114056e-01 6.42223656e-01 9.77937996e-01
-1.00515568e+00 -1.02712703e+00 -6.54973328e-01 4.24224079e-01
1.27406627e-01 4.41624969e-01 2.43441463e-02 5.68546891e-01
2.58673698e-01 1.43445790e+00 7.49226362e-02 -9.44086015e-01
2.63868123e-01 1.63680300e-01 -9.56667289e-02 -2.07088947e-01
2.22822130e-01 -3.20829839e-01 4.70874816e-01 -5.76550364e-01
-2.60561675e-01 -8.21062922e-01 -1.52079785e+00 -2.66330719e-01
-3.62493396e-01 2.95629054e-01 8.79139781e-01 1.20137501e+00
3.81172955e-01 3.75840068e-01 5.88521302e-01 -6.91245973e-01
-4.88181949e-01 -8.46223354e-01 -4.88833994e-01 -2.18120262e-01
1.76334769e-01 -1.18545461e+00 -5.48081160e-01 -5.04103482e-01] | [5.54959774017334, 1.409559965133667] |
eb2d940d-9718-4fb8-adea-8c8037875e18 | stagnet-an-attentive-semantic-rnn-for-group | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Mengshi_Qi_stagNet_An_Attentive_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Mengshi_Qi_stagNet_An_Attentive_ECCV_2018_paper.pdf | stagNet: An Attentive Semantic RNN for Group Activity Recognition | Group activity recognition plays a fundamental role in a variety of applications, e.g. sports video analysis and intelligent surveillance. How to model the spatio-temporal contextual information in a scene still remains a crucial yet challenging issue. We propose a novel attentive semantic recurrent neural network (RNN), namely stagNet, for understanding group activities in videos, based on the spatio-temporal attention and semantic graph. A semantic graph is explicitly modeled to describe the spatial context of the whole scene, which is further integrated with the temporal factor via structural-RNN. Benefiting from the 'factor sharing' and 'message passing' mechanisms, our model is able to extract discriminative spatio-temporal features and to capture inter-group relationships. Moreover, we adopt a spatio-temporal attention model to attend to key persons/frames for improved performance. Two widely-used datasets are employed for performance evaluation, and the extensive results demonstrate the superiority of our method. | ['Luc van Gool', 'Jiebo Luo', 'Jie Qin', 'Yunhong Wang', 'Mengshi Qi', 'Annan Li'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['group-activity-recognition'] | ['computer-vision'] | [ 1.99451163e-01 -4.99483824e-01 -2.13392481e-01 -3.63724768e-01
-1.58128470e-01 -7.40443170e-02 5.09523749e-01 5.74139096e-02
-4.09366131e-01 2.18597084e-01 7.85448194e-01 2.14141305e-03
-5.11187613e-01 -6.01903915e-01 -6.78642809e-01 -7.14151561e-01
-2.95985907e-01 -2.35374168e-01 4.44764704e-01 -2.02864796e-01
1.40665382e-01 2.95912683e-01 -1.48597932e+00 2.66851753e-01
5.93277156e-01 9.22629416e-01 4.64365751e-01 5.29809177e-01
6.48219585e-02 1.36942422e+00 -4.12377596e-01 -3.93041372e-02
-2.31280476e-01 -7.11694837e-01 -5.54657280e-01 4.46323425e-01
-1.12036772e-01 -1.58076733e-01 -8.53684843e-01 8.58818233e-01
1.80490449e-01 6.65086985e-01 8.82470757e-02 -1.07205403e+00
-3.04031014e-01 2.39332020e-01 -4.59963977e-01 6.24702215e-01
4.17970210e-01 2.38013625e-01 1.00621712e+00 -5.80544591e-01
5.28673768e-01 1.26611316e+00 2.88457096e-01 2.18842193e-01
-6.92222416e-01 -5.62727273e-01 7.44440675e-01 8.03598642e-01
-1.28042746e+00 -2.92512387e-01 9.74016607e-01 -2.58839518e-01
7.14175403e-01 1.18754975e-01 1.01060724e+00 1.17185771e+00
2.61221409e-01 1.07570517e+00 6.18880808e-01 7.14177713e-02
1.88075863e-02 -4.72441047e-01 1.23974621e-01 7.97756612e-01
-2.31304675e-01 -1.71620741e-01 -8.49859774e-01 2.45685816e-01
1.08696198e+00 6.69380248e-01 -2.90216118e-01 -3.45964015e-01
-1.31502199e+00 6.90416694e-01 7.82693386e-01 6.28572702e-01
-6.39378369e-01 5.04974902e-01 4.77906883e-01 -3.44395228e-02
5.86323917e-01 1.42171532e-02 -7.88707212e-02 -2.00619549e-01
-6.98616743e-01 8.96762460e-02 2.86134660e-01 6.92575574e-01
6.28459632e-01 6.88784346e-02 -5.07340431e-01 7.24608839e-01
3.25839877e-01 2.70287395e-01 3.48178655e-01 -9.20063078e-01
6.00486815e-01 9.01696920e-01 -9.72468555e-02 -1.54800630e+00
-4.37518030e-01 -6.71974480e-01 -1.10538077e+00 -3.96848530e-01
8.81344825e-02 3.52205560e-02 -7.56295383e-01 1.69552469e+00
2.83428788e-01 9.51438606e-01 -2.09205583e-01 1.14757264e+00
7.04130471e-01 7.70544350e-01 3.85818958e-01 -3.21408778e-01
1.55252874e+00 -1.12060177e+00 -7.64699280e-01 -2.40707070e-01
4.85440314e-01 -2.31210604e-01 6.92284584e-01 1.66975290e-01
-7.58865297e-01 -8.38469982e-01 -7.86277235e-01 -7.04796314e-02
-1.60192966e-01 4.94611189e-02 6.61033690e-01 2.53331102e-02
-8.14587057e-01 4.17152077e-01 -1.11215293e+00 -3.52399111e-01
6.16356730e-01 2.31035799e-01 -3.87131393e-01 -3.17770928e-01
-1.18328857e+00 2.32048824e-01 4.26099360e-01 4.20376658e-01
-9.54357803e-01 -2.23545432e-01 -1.02766097e+00 2.02978238e-01
8.65490317e-01 -7.42938638e-01 6.92397654e-01 -9.72639382e-01
-1.21761191e+00 3.00827801e-01 -5.17919362e-01 -5.61852217e-01
2.89342582e-01 -5.18378317e-01 -4.75029707e-01 4.88039851e-01
1.67436600e-01 2.10100859e-01 7.23469317e-01 -8.02306950e-01
-6.80992544e-01 -4.90769207e-01 1.60346106e-01 5.60911953e-01
-3.26328009e-01 4.03036445e-01 -9.66084540e-01 -1.03089321e+00
1.22222409e-01 -6.81786776e-01 -5.67272246e-01 -2.26263031e-01
-1.58480689e-01 -3.32792252e-01 8.95790994e-01 -9.71797884e-01
1.53825295e+00 -2.30242276e+00 5.24429381e-01 1.82949528e-01
3.13666701e-01 3.21332753e-01 -8.96287784e-02 4.44961399e-01
7.81739801e-02 -1.99339330e-01 -1.87135693e-02 -3.73253584e-01
-2.23019749e-01 3.36512059e-01 -1.42103121e-01 4.50753987e-01
1.44163415e-01 1.15822148e+00 -1.07565475e+00 -3.69981468e-01
4.58285600e-01 7.12707520e-01 -4.68257695e-01 2.90718466e-01
-2.37524658e-01 7.27780163e-01 -9.11857307e-01 2.79886544e-01
8.76449421e-02 -5.61702669e-01 8.64226371e-02 -1.71169877e-01
-1.08536929e-01 9.48602557e-02 -1.06954718e+00 2.13078570e+00
-1.09370284e-01 5.04173100e-01 -1.14332609e-01 -1.23799503e+00
6.34668827e-01 2.01546937e-01 5.55119097e-01 -1.07007802e+00
1.47051334e-01 -2.94675082e-01 -2.61852175e-01 -7.12885976e-01
5.33536315e-01 2.79967755e-01 -1.24687687e-01 3.12423408e-01
1.10168159e-01 8.32181096e-01 1.64640352e-01 4.32487249e-01
1.03151560e+00 1.91670105e-01 1.09455630e-01 -1.40972450e-01
7.15495229e-01 -4.14413095e-01 8.63764405e-01 5.02674162e-01
-1.48856401e-01 3.72895867e-01 5.06595790e-01 -7.57301748e-01
-5.71630955e-01 -5.02816319e-01 7.04973698e-01 1.37151492e+00
5.64766228e-01 -6.51925802e-01 -6.14589274e-01 -5.73882639e-01
-4.92587119e-01 2.40325078e-01 -7.62068093e-01 -3.92711103e-01
-8.28966975e-01 -6.05001271e-01 2.89161727e-02 7.33215034e-01
7.93559492e-01 -1.15046406e+00 -6.12409532e-01 3.70237738e-01
-4.57049042e-01 -1.24149752e+00 -6.81037664e-01 -4.29401517e-01
-6.16100132e-01 -1.18887758e+00 -6.42951965e-01 -6.26631498e-01
4.46857244e-01 7.65685081e-01 9.18290496e-01 3.46680164e-01
-1.65052831e-01 4.55828547e-01 -6.32996142e-01 4.45715943e-03
8.90727490e-02 6.40756935e-02 -1.56930968e-01 6.84241533e-01
2.65311539e-01 -6.87773287e-01 -9.72163796e-01 6.26781881e-01
-9.25045371e-01 3.71855646e-01 5.34243703e-01 6.20384514e-01
5.33684373e-01 1.42528355e-01 3.96235287e-01 -6.80973887e-01
2.72928566e-01 -5.43883145e-01 -3.06554854e-01 2.69617975e-01
2.24066928e-01 -1.76950812e-01 4.21382546e-01 -2.10710347e-01
-9.86729920e-01 -1.73019454e-01 3.14956382e-02 -7.33014047e-01
-1.03760809e-01 5.30799091e-01 -5.13887107e-01 1.93283901e-01
1.48189247e-01 4.75133389e-01 -1.42330259e-01 -4.46952909e-01
5.51425926e-02 7.34513849e-02 5.86794972e-01 -2.99333036e-01
5.18453777e-01 6.52961075e-01 6.71852455e-02 -8.51324439e-01
-9.06848609e-01 -8.05903554e-01 -5.16673744e-01 -4.36793029e-01
1.21838439e+00 -1.15585411e+00 -9.16355193e-01 5.50730228e-01
-1.01494908e+00 -2.83080488e-01 7.87693076e-03 7.25226939e-01
-4.91179466e-01 5.44453740e-01 -6.03567660e-01 -7.20248163e-01
-4.20608521e-02 -9.21245873e-01 1.06568527e+00 2.96494722e-01
1.55632228e-01 -1.08596420e+00 -1.58387199e-01 5.38217187e-01
1.02571629e-01 3.70380789e-01 4.94105458e-01 -4.29584056e-01
-1.04683709e+00 9.51912925e-02 -4.19854492e-01 1.56651765e-01
1.27191812e-01 -3.92208785e-01 -5.75488746e-01 -1.79507226e-01
-8.29615369e-02 1.94524199e-01 1.02755857e+00 4.82982546e-01
1.39783978e+00 -3.05232108e-01 -4.09212291e-01 6.46302044e-01
9.84153330e-01 1.38753787e-01 8.01030219e-01 -3.65977995e-02
1.27521574e+00 6.32553160e-01 7.20711529e-01 4.42404628e-01
5.82264841e-01 7.44780362e-01 3.83465499e-01 -1.06322050e-01
8.28657895e-02 -4.89095777e-01 3.05049181e-01 9.44785535e-01
-5.60459256e-01 -2.74289906e-01 -7.36462414e-01 5.49954772e-01
-2.42783976e+00 -1.26335585e+00 3.49530764e-02 1.93111813e+00
-2.59424988e-02 2.70329807e-02 2.84698159e-01 8.96697566e-02
7.51006782e-01 6.42114758e-01 -4.10424352e-01 3.44911397e-01
-2.04748362e-02 -2.51338691e-01 1.39565721e-01 2.42194340e-01
-1.20465910e+00 1.02774310e+00 4.97452402e+00 1.02040458e+00
-8.27108026e-01 2.27093939e-02 6.79728031e-01 -9.56375301e-02
-7.10814595e-02 -1.65824428e-01 -4.57495838e-01 6.22431636e-01
6.91224694e-01 6.95559308e-02 3.72081310e-01 4.36026633e-01
6.15009248e-01 -1.89313886e-03 -5.90189934e-01 1.07107651e+00
3.12326044e-01 -1.28329062e+00 1.74073547e-01 6.02025986e-02
4.40282971e-01 -1.42591149e-01 -1.16260991e-01 3.01703811e-01
1.28183559e-01 -9.73034322e-01 4.92293775e-01 9.21491265e-01
4.04150516e-01 -1.03567731e+00 6.42258584e-01 4.31037009e-01
-1.71107054e+00 -1.99215695e-01 -1.28336117e-01 -1.17104076e-01
3.88740748e-01 3.49913239e-01 -2.05856562e-01 9.05951679e-01
7.99984097e-01 1.65849543e+00 -4.81641233e-01 8.89665782e-01
-2.94501126e-01 5.69387972e-01 3.25549841e-02 -1.42698688e-02
3.83672923e-01 -4.26177979e-01 6.16834164e-01 1.11978972e+00
1.85887724e-01 4.27764803e-01 4.06334817e-01 4.03999805e-01
-2.43820599e-03 3.10944114e-02 -4.77836460e-01 -6.63859695e-02
5.26911989e-02 1.01735520e+00 -9.19619739e-01 -3.85178536e-01
-4.44277644e-01 1.02648354e+00 2.03104064e-01 5.60582995e-01
-9.68436062e-01 -1.31034464e-01 7.89515555e-01 -3.35292593e-02
4.88152504e-01 -5.32348096e-01 4.08822954e-01 -1.34312916e+00
1.93329364e-01 -7.44984984e-01 6.20670199e-01 -8.24497759e-01
-8.31985474e-01 4.80120778e-01 -8.06381181e-02 -1.31123734e+00
-2.15753257e-01 -2.04489738e-01 -6.23758435e-01 5.21880627e-01
-1.43108332e+00 -1.35551405e+00 -5.60271740e-01 9.98650312e-01
7.61488676e-01 -1.10758930e-01 3.33997339e-01 4.62567151e-01
-8.09433281e-01 1.24487303e-01 -3.68971378e-01 4.58885282e-01
1.72595695e-01 -7.43651211e-01 3.20787519e-01 1.09836960e+00
4.45521533e-01 5.44336259e-01 4.33615595e-01 -5.76550424e-01
-1.35019565e+00 -1.41004932e+00 7.67909646e-01 -3.17309827e-01
6.35391951e-01 -4.37597334e-01 -1.09877694e+00 7.71332920e-01
3.72853242e-02 1.49678096e-01 6.39521241e-01 8.11495930e-02
-2.22837746e-01 -1.56697065e-01 -3.23479325e-01 6.01790667e-01
1.67364180e+00 -7.03287721e-01 -4.64544624e-01 2.16334656e-01
1.01069570e+00 -2.36462235e-01 -5.73153615e-01 4.41049248e-01
3.81429344e-01 -1.11240542e+00 1.07730460e+00 -6.29174829e-01
3.85947257e-01 -4.62682635e-01 5.30067608e-02 -1.13418651e+00
-5.74510694e-01 -6.70102715e-01 -2.42899567e-01 1.04277647e+00
-3.19037557e-01 -4.37856495e-01 8.31808865e-01 2.22029120e-01
-2.77401835e-01 -6.35572076e-01 -8.68143260e-01 -3.74987960e-01
-8.88782382e-01 -6.31280601e-01 5.43098271e-01 8.42549682e-01
-3.45076323e-01 5.11230886e-01 -8.14655185e-01 3.47642511e-01
3.79121155e-01 1.17663473e-01 7.49096811e-01 -1.02088428e+00
-4.27353829e-01 -2.97605187e-01 -8.71576250e-01 -1.45829546e+00
1.61754653e-01 -4.27379042e-01 -2.02203274e-01 -1.49458456e+00
3.48890394e-01 -7.77419237e-03 -7.23799527e-01 2.67554313e-01
-4.54679698e-01 1.67605683e-01 3.22365880e-01 8.35231170e-02
-1.48168492e+00 8.89163077e-01 1.18037641e+00 -2.01567654e-02
-1.72745198e-01 5.82085773e-02 -5.84955692e-01 6.52373970e-01
4.96875614e-01 -2.17532635e-01 -6.44966602e-01 -4.81839865e-01
2.39233866e-01 3.89739752e-01 7.98206508e-01 -1.22162366e+00
4.12131369e-01 -2.65426368e-01 2.59231538e-01 -4.93046880e-01
4.75087613e-01 -8.50122571e-01 1.16606705e-01 3.48204285e-01
-1.71306416e-01 1.32020593e-01 -9.99278948e-02 1.16165566e+00
-5.96534491e-01 4.20707375e-01 2.84985423e-01 -7.51233175e-02
-1.04504466e+00 7.85690963e-01 -2.84199387e-01 -7.62153342e-02
1.04568100e+00 -2.25895733e-01 -7.45647252e-02 -6.07569098e-01
-8.37593555e-01 5.21289170e-01 2.30372578e-01 6.46860361e-01
7.69664168e-01 -1.41021442e+00 -5.64788043e-01 2.66574472e-01
1.99336484e-01 6.90714270e-02 1.02270925e+00 9.04082775e-01
-2.68090546e-01 3.16862643e-01 -7.37769008e-02 -7.47109532e-01
-1.38356173e+00 5.75914860e-01 1.65201604e-01 -3.80170107e-01
-7.95047939e-01 7.03301132e-01 6.45074308e-01 2.87727833e-01
1.69541284e-01 -4.19720858e-01 -6.57733619e-01 -8.69256854e-02
8.63588333e-01 3.60279918e-01 -4.11523610e-01 -1.01038146e+00
-2.24892139e-01 7.44806051e-01 1.46070525e-01 2.03072906e-01
1.50040126e+00 -3.02197009e-01 -4.31127064e-02 4.92891729e-01
1.05495036e+00 -3.28058690e-01 -1.54751992e+00 -6.19651020e-01
3.62574426e-03 -5.99936724e-01 -1.70313101e-02 -2.16320157e-01
-1.22999036e+00 9.93957698e-01 2.34153584e-01 6.35791570e-02
1.33664906e+00 8.41490775e-02 7.79227674e-01 1.76198334e-01
3.45432639e-01 -8.68851364e-01 4.99586642e-01 4.73896682e-01
7.76588976e-01 -1.01720119e+00 -6.68907091e-02 -5.62073529e-01
-7.41913795e-01 6.93373859e-01 3.60341698e-01 -2.17094600e-01
5.24857700e-01 -4.30998623e-01 -3.03896338e-01 -4.16523337e-01
-6.67960823e-01 -3.06360781e-01 5.66870213e-01 3.00593227e-01
3.79044823e-02 -8.80388990e-02 1.17900245e-01 6.86716735e-01
3.68221939e-01 3.43608260e-02 -5.93687445e-02 7.87884057e-01
-3.09302092e-01 -7.46473312e-01 -6.96403533e-02 2.71617174e-01
-3.32194597e-01 2.98345000e-01 1.25377653e-02 5.31298161e-01
1.91293672e-01 9.19416726e-01 2.27862656e-01 -5.25062561e-01
2.49604017e-01 -1.68129325e-01 9.26322117e-02 -5.17498016e-01
-5.48482537e-01 3.15301210e-01 -8.58512670e-02 -9.80828226e-01
-1.02610779e+00 -6.89613104e-01 -1.06074977e+00 -1.11854129e-01
1.79851264e-01 2.02108219e-01 2.14245126e-01 1.12824476e+00
6.14125729e-01 9.23652172e-01 6.70934975e-01 -7.28616655e-01
3.40428978e-01 -8.05702746e-01 -5.64277112e-01 7.02071369e-01
3.44325006e-01 -7.64764547e-01 5.42915352e-02 1.16943657e-01] | [8.385503768920898, 0.6594128012657166] |
c7bd3e15-acb5-4a49-b092-f79f67666f66 | visual-object-tracking-in-first-person-vision | 2209.13502 | null | https://arxiv.org/abs/2209.13502v1 | https://arxiv.org/pdf/2209.13502v1.pdf | Visual Object Tracking in First Person Vision | The understanding of human-object interactions is fundamental in First Person Vision (FPV). Visual tracking algorithms which follow the objects manipulated by the camera wearer can provide useful information to effectively model such interactions. In the last years, the computer vision community has significantly improved the performance of tracking algorithms for a large variety of target objects and scenarios. Despite a few previous attempts to exploit trackers in the FPV domain, a methodical analysis of the performance of state-of-the-art trackers is still missing. This research gap raises the question of whether current solutions can be used ``off-the-shelf'' or more domain-specific investigations should be carried out. This paper aims to provide answers to such questions. We present the first systematic investigation of single object tracking in FPV. Our study extensively analyses the performance of 42 algorithms including generic object trackers and baseline FPV-specific trackers. The analysis is carried out by focusing on different aspects of the FPV setting, introducing new performance measures, and in relation to FPV-specific tasks. The study is made possible through the introduction of TREK-150, a novel benchmark dataset composed of 150 densely annotated video sequences. Our results show that object tracking in FPV poses new challenges to current visual trackers. We highlight the factors causing such behavior and point out possible research directions. Despite their difficulties, we prove that trackers bring benefits to FPV downstream tasks requiring short-term object tracking. We expect that generic object tracking will gain popularity in FPV as new and FPV-specific methodologies are investigated. | ['Christian Micheloni', 'Giovanni Maria Farinella', 'Antonino Furnari', 'Matteo Dunnhofer'] | 2022-09-27 | null | null | null | null | ['visual-tracking', 'human-object-interaction-detection', 'visual-object-tracking'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.51712699e-02 -3.47166151e-01 -4.11993116e-01 5.45201898e-02
-1.76328689e-01 -7.62127638e-01 6.71563745e-01 -2.25114107e-01
-6.05495393e-01 4.89432842e-01 -6.16072603e-02 -1.09933019e-01
-7.93801174e-02 -7.76266083e-02 -7.31629372e-01 -5.60033858e-01
-1.94406211e-01 4.10175264e-01 7.55048454e-01 3.62936668e-02
-1.66678149e-02 6.98681116e-01 -2.00150752e+00 -3.13488543e-02
1.86846033e-01 8.04065347e-01 3.70324314e-01 9.11307573e-01
1.73779204e-01 3.85348082e-01 -7.44091690e-01 -5.62395155e-01
4.68162954e-01 -7.69403502e-02 -4.61549282e-01 1.88372359e-01
1.07622612e+00 -2.44097605e-01 -2.95552015e-01 1.00017989e+00
6.16301060e-01 -1.01861674e-02 2.85736293e-01 -1.81412053e+00
-4.74494040e-01 1.40576810e-01 -4.52608198e-01 6.22114420e-01
4.08867657e-01 6.37034714e-01 6.61137819e-01 -6.25848055e-01
7.53541112e-01 1.35870123e+00 9.40035701e-01 8.88580024e-01
-1.21533275e+00 -7.14154780e-01 2.39147812e-01 1.18636012e-01
-1.16331375e+00 -5.29763877e-01 4.76881266e-01 -8.61420810e-01
8.89688492e-01 4.18284863e-01 7.94726431e-01 1.47535145e+00
1.31323665e-01 8.30446422e-01 1.17780268e+00 -3.94899368e-01
-1.01091012e-01 4.24242347e-01 5.16786397e-01 5.05736828e-01
7.78586328e-01 7.46887565e-01 -6.37214899e-01 -9.56588238e-02
6.11455619e-01 -2.76058227e-01 -4.21955615e-01 -7.53990650e-01
-1.41994298e+00 5.96772492e-01 2.24359393e-01 4.43203360e-01
-1.30921364e-01 4.42210346e-01 5.48396111e-01 3.98053415e-03
2.05710098e-01 2.87199259e-01 -3.84002715e-01 -3.72391850e-01
-9.24782932e-01 6.03927433e-01 6.05490446e-01 1.12267911e+00
6.47299364e-02 -1.16167732e-01 -7.39434659e-01 4.45998371e-01
5.78573108e-01 8.22306097e-01 6.01134188e-02 -8.74831557e-01
1.26957059e-01 2.40956411e-01 4.55214232e-01 -8.80723119e-01
-4.90905613e-01 -4.12438512e-01 -1.47066876e-01 5.66564143e-01
8.70664954e-01 3.69048677e-02 -7.12274671e-01 1.60036802e+00
5.26032925e-01 4.10219938e-01 -4.51452225e-01 1.05233037e+00
9.06371474e-01 1.20819367e-01 2.73716718e-01 -2.15313628e-01
1.84796846e+00 -9.13866580e-01 -9.37302470e-01 -8.17539096e-02
4.94389534e-01 -8.85404110e-01 7.99950719e-01 3.33983928e-01
-8.85205448e-01 -8.76914978e-01 -8.30905259e-01 2.28190958e-01
-3.46612185e-01 2.43957043e-01 5.62174737e-01 1.27813601e+00
-1.05412257e+00 2.83710510e-01 -1.00123179e+00 -8.42810810e-01
5.48971772e-01 5.87311149e-01 -2.17965052e-01 2.32781246e-01
-7.77979314e-01 1.13013935e+00 3.30266982e-01 1.19400620e-01
-9.90822494e-01 -8.33036125e-01 -5.17787874e-01 -4.25507367e-01
6.80764914e-01 -8.05974782e-01 1.45175695e+00 -6.09512269e-01
-1.02646887e+00 1.04580665e+00 -2.84071267e-01 -6.91910148e-01
7.87121952e-01 -4.49740767e-01 -4.50016975e-01 -6.00521155e-02
3.08943652e-02 6.24329746e-01 9.32377040e-01 -1.18531609e+00
-8.23791027e-01 -2.53911316e-01 1.22123845e-02 -2.03143209e-01
-9.69473273e-02 4.65668589e-01 -7.97023952e-01 -5.56106687e-01
-7.21708477e-01 -1.28783464e+00 9.11261067e-02 4.36807215e-01
-2.36048952e-01 -4.01317000e-01 1.15204871e+00 -2.50180453e-01
1.14691472e+00 -1.97669935e+00 -7.64335766e-02 -3.06117982e-01
3.62490565e-01 8.89217198e-01 -1.75959675e-03 2.60540158e-01
1.40645981e-01 -3.06678236e-01 3.31552565e-01 -4.36938196e-01
1.31049395e-01 1.16230071e-01 -2.42904499e-01 7.16442227e-01
-1.63875431e-01 1.21830523e+00 -9.61847305e-01 -5.29061675e-01
7.58185685e-01 5.60994923e-01 -2.26435244e-01 3.83874662e-02
-3.24198544e-01 5.31453252e-01 -3.65099400e-01 6.82311118e-01
5.77517927e-01 -1.93444774e-01 -3.18167180e-01 -3.81572813e-01
-3.23868960e-01 -3.41805428e-01 -1.15082002e+00 1.34784925e+00
9.14141908e-02 1.08556509e+00 6.29755557e-02 -6.97481215e-01
5.00619411e-01 3.08450460e-01 5.19807756e-01 -4.27597225e-01
3.41707081e-01 -1.80210304e-02 1.88345000e-01 -5.56122124e-01
5.53361058e-01 1.72647923e-01 3.57672155e-01 4.53746468e-02
-6.23805821e-03 6.48321211e-01 4.03700411e-01 4.45157178e-02
9.85236347e-01 4.13308501e-01 3.04959416e-01 -2.28290007e-01
5.64918399e-01 2.53301084e-01 1.54025689e-01 1.16317427e+00
-8.59286368e-01 3.86847913e-01 -9.25817788e-02 -4.61961746e-01
-7.87740946e-01 -1.01265967e+00 -3.79753292e-01 1.04466653e+00
3.44734848e-01 -5.27202487e-01 -7.42026448e-01 -6.53286397e-01
2.77772933e-01 3.89575988e-01 -8.10767353e-01 1.63452793e-03
-5.38604975e-01 -6.08148575e-01 6.08617961e-01 7.55971313e-01
4.21450734e-01 -1.11480355e+00 -1.20122898e+00 1.92931652e-01
1.16121002e-01 -1.46958351e+00 -4.73904043e-01 -3.14642400e-01
-7.79867887e-01 -1.41574013e+00 -8.62744093e-01 -3.04920137e-01
2.53394842e-01 7.08737850e-01 1.15570164e+00 1.99416161e-01
-3.82001370e-01 1.14119852e+00 -2.71502733e-01 -8.27852786e-01
-2.27492645e-01 -1.31295741e-01 4.76106972e-01 3.38242687e-02
7.65213668e-01 1.05677638e-02 -5.77938974e-01 6.92278028e-01
-3.61285567e-01 -2.40510598e-01 3.73011470e-01 3.73829484e-01
3.15085143e-01 -3.35808307e-01 8.25249404e-02 -6.27642930e-01
3.79806697e-01 -2.37978548e-02 -1.08459842e+00 3.91025662e-01
-4.50164497e-01 -8.40332881e-02 -6.19549975e-02 -8.28485966e-01
-8.37844431e-01 2.12157488e-01 1.82797700e-01 -8.20108116e-01
-1.85110301e-01 -2.27833644e-01 -6.76715299e-02 -5.39061427e-01
7.49912620e-01 1.19049691e-01 -2.37835608e-02 -5.17911851e-01
3.50709051e-01 4.00223911e-01 7.48733580e-01 -6.01091683e-01
1.07143474e+00 7.27028847e-01 -3.13568451e-02 -1.05200863e+00
-7.12896407e-01 -8.60568702e-01 -5.64130902e-01 -7.32064486e-01
1.13871086e+00 -7.65020490e-01 -1.23005617e+00 4.40782100e-01
-1.26570332e+00 -1.76834136e-01 -2.41455913e-01 6.14977002e-01
-5.22464335e-01 4.30811614e-01 -8.53125900e-02 -1.14652634e+00
-7.77498037e-02 -1.27989340e+00 1.28187668e+00 3.83359402e-01
-3.51924002e-01 -9.80146766e-01 2.60147184e-01 5.31883359e-01
3.04235995e-01 1.95226461e-01 -1.39849767e-01 -4.55056071e-01
-9.60039437e-01 -2.10000128e-01 -1.58613190e-01 1.49140984e-01
-6.85887411e-02 -1.30824465e-03 -1.25310552e+00 -6.01349652e-01
-2.07990795e-01 1.31590202e-01 7.64398396e-01 8.35749090e-01
7.33372688e-01 2.19274819e-01 -1.03423023e+00 5.05818605e-01
1.18173611e+00 2.83142984e-01 3.76023114e-01 5.32999933e-01
8.04486930e-01 4.16544974e-01 8.03006291e-01 9.70674306e-02
1.13886133e-01 1.41903853e+00 5.14571428e-01 1.43272474e-01
-6.36037946e-01 4.79231924e-02 4.41300333e-01 -6.28836676e-02
-5.30754328e-01 -2.55666137e-01 -8.25944781e-01 5.64747632e-01
-1.85011530e+00 -1.09546578e+00 -5.33432722e-01 2.29626656e+00
1.85073331e-01 2.86849707e-01 6.35612726e-01 -5.49392290e-02
8.32389295e-01 3.27996463e-02 -2.72837639e-01 2.07734957e-01
1.36876330e-01 -9.54389423e-02 7.60266185e-01 2.80325979e-01
-1.35962176e+00 8.44655812e-01 6.52711391e+00 5.54378331e-01
-9.52307403e-01 3.13699991e-01 -3.04037660e-01 -3.47348452e-01
5.66684902e-01 -1.93643615e-01 -1.57713401e+00 6.46815777e-01
9.45810676e-01 -1.75377149e-02 1.11066766e-01 9.26235497e-01
3.34028780e-01 -1.39309287e-01 -1.28642237e+00 1.34686399e+00
9.89422873e-02 -1.22860467e+00 -3.72006565e-01 3.96508187e-01
3.55771691e-01 6.40746951e-02 1.21243037e-01 3.10758740e-01
-6.55599544e-03 -8.04430425e-01 8.03617656e-01 5.23329556e-01
4.10617143e-01 -2.20801696e-01 5.60087323e-01 1.44048646e-01
-1.52954686e+00 -4.06318307e-02 -3.55300933e-01 4.75283936e-02
3.69059712e-01 1.68999523e-01 -6.88068271e-01 4.76478547e-01
9.98310328e-01 7.25942731e-01 -9.81717408e-01 1.59932303e+00
1.83454618e-01 5.12813389e-01 -2.36648962e-01 4.75207046e-02
7.02332109e-02 1.34153366e-01 9.46438491e-01 1.26480913e+00
6.79856315e-02 -1.42765701e-01 2.38435954e-01 6.31400108e-01
2.52990782e-01 -2.53675759e-01 -8.47121298e-01 2.54319925e-02
3.63347024e-01 1.15141976e+00 -9.79244411e-01 -2.71433383e-01
-7.33824313e-01 6.65550113e-01 7.77750015e-02 1.98757157e-01
-1.09326756e+00 3.96655619e-01 9.23179924e-01 4.17787403e-01
5.84279120e-01 -2.19958201e-01 4.74141613e-02 -1.02553916e+00
9.64311883e-02 -8.46884370e-01 3.45270455e-01 -7.07832813e-01
-1.14513850e+00 4.51901346e-01 5.32366216e-01 -1.38305175e+00
-2.90184580e-02 -9.36722875e-01 -3.05090070e-01 5.60074687e-01
-1.17884636e+00 -1.31047952e+00 -5.11030912e-01 5.33380926e-01
5.76729298e-01 -2.56670654e-01 3.67410481e-01 3.43641132e-01
-4.53039855e-01 5.10362506e-01 -7.70826340e-02 1.39590696e-01
6.82793081e-01 -9.54051852e-01 3.98351222e-01 1.00427616e+00
4.85666662e-01 6.88348353e-01 1.15286303e+00 -7.34245062e-01
-1.48079848e+00 -9.31533933e-01 6.36995733e-01 -1.50923932e+00
4.41372454e-01 -5.00761151e-01 -8.04132283e-01 8.70489955e-01
1.33334413e-01 4.16732401e-01 4.64898795e-01 1.19663574e-01
-1.00900002e-01 9.48522836e-02 -7.65427530e-01 4.33117628e-01
1.43097126e+00 -2.74551153e-01 -7.44460046e-01 3.61798763e-01
4.81276751e-01 -5.38714826e-01 -6.35285735e-01 3.53387713e-01
8.57904255e-01 -9.96327639e-01 1.40811229e+00 -5.99052846e-01
-4.60927993e-01 -6.52020395e-01 4.84810732e-02 -7.97174752e-01
-3.04380625e-01 -4.70504999e-01 -5.57207406e-01 1.23328292e+00
-2.55057067e-01 -4.38915074e-01 9.07438576e-01 5.36119580e-01
-1.73784830e-02 -3.82012755e-01 -9.16812718e-01 -1.41685414e+00
-3.75925690e-01 -7.37481296e-01 5.30105531e-02 4.81446177e-01
-6.50457263e-01 -3.46965678e-02 -7.30509520e-01 1.93111211e-01
1.05743873e+00 -1.81014225e-01 1.23128533e+00 -1.56700444e+00
-3.05627376e-01 -4.37338293e-01 -8.25141132e-01 -8.83609593e-01
3.42408195e-02 -6.58599675e-01 -1.89625695e-01 -1.14450598e+00
2.50293374e-01 -2.34417692e-02 -7.35601690e-03 1.78705290e-01
-2.42695272e-01 5.26157916e-01 7.05314219e-01 2.35369533e-01
-7.54944026e-01 2.02050582e-01 1.18788552e+00 -2.09859207e-01
-5.20846397e-02 3.17889214e-01 -3.62324357e-01 5.67212462e-01
3.84432137e-01 -5.91175139e-01 -4.32820261e-01 -1.06782138e-01
-2.33274207e-01 -4.63387370e-01 1.00645983e+00 -1.36307657e+00
4.78771120e-01 1.21535778e-01 3.90245140e-01 -7.93989062e-01
4.63594705e-01 -1.08815849e+00 4.99962926e-01 7.64479995e-01
3.50260288e-02 -3.57813016e-02 5.38573086e-01 7.00617254e-01
2.08803356e-01 -1.49627596e-01 8.26110363e-01 -4.51378040e-02
-1.04883206e+00 2.77777970e-01 -1.92593887e-01 1.09969154e-01
1.48569381e+00 -6.53785467e-01 -3.13142687e-01 1.99996401e-02
-6.80023968e-01 1.89243659e-01 5.68577290e-01 8.55914235e-01
2.59642988e-01 -1.33089852e+00 -5.79597712e-01 1.53994635e-02
2.97999501e-01 -6.06759310e-01 9.63537991e-02 1.07553124e+00
-1.54521644e-01 8.87050331e-01 -3.34568053e-01 -1.10862148e+00
-2.03948236e+00 1.05312002e+00 4.19089258e-01 -9.71280411e-02
-9.13938165e-01 6.20398521e-01 3.33531469e-01 7.21143931e-02
4.65659499e-01 -2.80162722e-01 -2.27480710e-01 4.56842333e-02
7.92378485e-01 5.76486468e-01 -1.22129373e-01 -1.00455201e+00
-6.69921219e-01 8.99787009e-01 -2.03424301e-02 9.77385640e-02
8.07962656e-01 -1.33185029e-01 5.01852393e-01 3.02200437e-01
6.93975270e-01 -1.99457780e-01 -1.40146160e+00 -1.48456395e-01
1.40019044e-01 -7.78009057e-01 -1.13107845e-01 -6.88348413e-01
-1.14016902e+00 6.67471170e-01 1.18237484e+00 4.70343977e-02
7.48514831e-01 3.46550822e-01 3.68519306e-01 1.08790308e-01
7.02280223e-01 -5.89504540e-01 1.09173963e-02 1.42072365e-01
7.02291965e-01 -1.38541281e+00 1.24135241e-01 -4.73591506e-01
-4.86822873e-01 7.45299757e-01 6.80202901e-01 8.65907222e-02
4.59482014e-01 2.28676811e-01 8.62808973e-02 -5.19184172e-01
-4.69079524e-01 -5.21622598e-01 7.56314874e-01 1.09352112e+00
3.66864711e-01 -1.04803838e-01 -3.53613824e-01 2.47901604e-01
-5.24362326e-02 2.56939441e-01 2.43386224e-01 8.97743464e-01
-2.62546808e-01 -1.25691295e+00 -8.96219134e-01 2.94961542e-01
-3.87451351e-01 3.67249638e-01 -3.35626483e-01 1.31092274e+00
3.43233138e-01 9.42260146e-01 -1.66474953e-01 -1.84681445e-01
5.71575463e-01 7.57407676e-03 8.79160285e-01 -4.38958198e-01
-8.18764448e-01 -1.48316950e-01 1.93600915e-02 -7.72644877e-01
-9.68535602e-01 -1.04861903e+00 -6.69428527e-01 -2.00679645e-01
-4.03278857e-01 -1.28932804e-01 5.24970055e-01 9.06940997e-01
1.62432075e-01 6.07480705e-01 -3.33915412e-01 -1.09162307e+00
-3.67906243e-01 -7.34486103e-01 -3.43331069e-01 5.06629944e-01
4.16144073e-01 -1.46569538e+00 -9.43136495e-03 1.91102117e-01] | [6.35005521774292, -1.9932475090026855] |
a394f753-5951-4569-9df0-20c7264c235b | building-competitive-direct-acoustics-to-word | 1712.03133 | null | http://arxiv.org/abs/1712.03133v1 | http://arxiv.org/pdf/1712.03133v1.pdf | Building competitive direct acoustics-to-word models for English conversational speech recognition | Direct acoustics-to-word (A2W) models in the end-to-end paradigm have
received increasing attention compared to conventional sub-word based automatic
speech recognition models using phones, characters, or context-dependent hidden
Markov model states. This is because A2W models recognize words from speech
without any decoder, pronunciation lexicon, or externally-trained language
model, making training and decoding with such models simple. Prior work has
shown that A2W models require orders of magnitude more training data in order
to perform comparably to conventional models. Our work also showed this
accuracy gap when using the English Switchboard-Fisher data set. This paper
describes a recipe to train an A2W model that closes this gap and is at-par
with state-of-the-art sub-word based models. We achieve a word error rate of
8.8%/13.9% on the Hub5-2000 Switchboard/CallHome test sets without any decoder
or language model. We find that model initialization, training data order, and
regularization have the most impact on the A2W model performance. Next, we
present a joint word-character A2W model that learns to first spell the word
and then recognize it. This model provides a rich output to the user instead of
simple word hypotheses, making it especially useful in the case of words unseen
or rarely-seen during training. | ['Michael Picheny', 'George Saon', 'Bhuvana Ramabhadran', 'Brian Kingsbury', 'Kartik Audhkhasi'] | 2017-12-08 | null | null | null | null | ['english-conversational-speech-recognition'] | ['speech'] | [ 2.31162041e-01 1.82527915e-01 -1.07474610e-01 -4.47119176e-01
-1.34345484e+00 -5.26156068e-01 5.21937430e-01 -1.27790794e-01
-7.63501287e-01 4.02045101e-01 3.00052047e-01 -1.11027610e+00
3.79356086e-01 -2.21117169e-01 -4.25735742e-01 -5.08296728e-01
9.99346673e-02 6.36218429e-01 4.59945887e-01 -3.02120328e-01
-9.90476087e-02 4.01297882e-02 -1.05917823e+00 2.70739704e-01
3.12281489e-01 5.64010203e-01 7.46454775e-01 1.12072659e+00
-2.85929829e-01 4.07337159e-01 -7.64381230e-01 -2.34373033e-01
2.12886497e-01 -5.63543081e-01 -6.02683783e-01 -2.35983897e-02
2.59245217e-01 -2.14368448e-01 -3.19034994e-01 6.13095880e-01
8.50098014e-01 2.94796973e-01 4.97456223e-01 -6.72929466e-01
-5.58897197e-01 6.81755960e-01 1.18111931e-01 3.79942775e-01
2.18029812e-01 3.66588086e-01 1.16388440e+00 -1.26817417e+00
2.24069968e-01 1.26958942e+00 6.85839355e-01 8.43541861e-01
-1.29180849e+00 -6.43349528e-01 2.79269218e-01 4.85155061e-02
-1.60468924e+00 -1.06197500e+00 2.88296402e-01 -2.27391720e-01
1.91639495e+00 3.64039898e-01 4.20793295e-01 1.33482742e+00
-1.53790817e-01 7.48643637e-01 8.48184645e-01 -9.96381342e-01
1.58512875e-01 1.40694961e-01 3.84494722e-01 5.67479253e-01
-1.63669884e-01 4.87080254e-02 -7.05644011e-01 -8.62333477e-02
2.92050809e-01 -4.09754544e-01 -4.17320758e-01 4.03677911e-01
-9.91345108e-01 8.18260670e-01 -2.47168645e-01 3.03311944e-01
-1.45837158e-01 2.66781271e-01 2.62022763e-01 3.19153190e-01
4.15383756e-01 3.54524195e-01 -7.86031961e-01 -6.32005036e-01
-1.32757235e+00 -3.02810729e-01 8.77704203e-01 9.45544541e-01
3.55779141e-01 5.48035443e-01 -1.64591432e-01 1.13431764e+00
6.59826279e-01 7.21571684e-01 8.85888219e-01 -3.72982532e-01
4.95423734e-01 -2.35321835e-01 -1.42746270e-01 -1.90430313e-01
-1.68071240e-01 -4.36364055e-01 -2.47081697e-01 -2.68958975e-02
3.24067444e-01 -1.18309416e-01 -1.51874292e+00 1.65769219e+00
-3.50681305e-01 2.07573995e-01 1.94985703e-01 5.28499901e-01
6.92822397e-01 1.16088986e+00 2.75786296e-02 -9.29689258e-02
1.36624861e+00 -1.06497025e+00 -8.96417201e-01 -7.28229046e-01
8.34483147e-01 -1.08536279e+00 1.43862653e+00 3.82030755e-01
-1.25488377e+00 -6.19539917e-01 -1.00573194e+00 -9.46423877e-03
-5.44604599e-01 3.44563499e-02 9.17883739e-02 1.09613073e+00
-1.36104548e+00 1.03880018e-01 -8.54727507e-01 -4.90019053e-01
-1.56886578e-01 3.36317927e-01 -8.24752375e-02 7.29786837e-03
-1.32182741e+00 1.24472344e+00 5.46139836e-01 -6.54757917e-02
-9.75014806e-01 -4.75006700e-01 -8.56559396e-01 8.09049904e-02
8.82198513e-02 -7.04077035e-02 1.70025110e+00 -7.25887001e-01
-1.88417220e+00 7.63001680e-01 -5.66019356e-01 -5.00432670e-01
-4.58216220e-02 -1.14530273e-01 -8.81114244e-01 -3.12246680e-01
-5.11690557e-01 4.94156212e-01 8.39898527e-01 -9.69109118e-01
-4.46390301e-01 2.09925979e-01 -5.19824982e-01 -2.60347966e-03
-3.58304709e-01 1.97152063e-01 -8.02122533e-01 -7.35957384e-01
-1.79559037e-01 -8.30451071e-01 9.02869031e-02 -6.06432199e-01
-3.86914849e-01 -4.35796291e-01 4.25267577e-01 -8.59323800e-01
1.54659522e+00 -2.32587528e+00 -2.71757871e-01 3.14526916e-01
-2.85503268e-01 8.03293884e-01 -5.30778050e-01 7.54566967e-01
-1.24431171e-01 3.69788915e-01 -8.81423280e-02 -8.00207019e-01
9.20432881e-02 4.71694678e-01 -4.61549968e-01 2.00174302e-01
3.00191551e-01 7.51332581e-01 -5.31083584e-01 -1.10110097e-01
2.45176718e-01 6.41125381e-01 -6.19902730e-01 2.82965511e-01
3.34869176e-02 -1.47225916e-01 2.70170629e-01 3.06596011e-01
2.78345704e-01 2.41411448e-01 5.60570173e-02 3.34263444e-01
-1.47261411e-01 1.07761168e+00 -8.67913246e-01 1.58122897e+00
-9.11241531e-01 9.19563532e-01 6.30508140e-02 -8.10583949e-01
9.95513916e-01 6.92296267e-01 -3.13860476e-01 -7.00536191e-01
-5.35269454e-03 6.45255327e-01 4.06215429e-01 -1.96877062e-01
2.93577135e-01 -4.19481814e-01 2.42208064e-01 1.47603348e-01
4.78851944e-01 -3.82456988e-01 -1.54099369e-03 -6.19646311e-02
1.06597328e+00 -3.22189063e-01 1.44866392e-01 -2.76461542e-01
3.39832664e-01 -2.34638140e-01 3.95752639e-01 9.79361832e-01
2.49583144e-02 9.72084701e-01 2.14561988e-02 -1.43224709e-02
-9.16006327e-01 -1.12602568e+00 -1.05944641e-01 1.23222983e+00
-4.10044938e-01 -5.88047862e-01 -6.26733959e-01 -2.53657162e-01
-3.45499933e-01 1.50434351e+00 -1.19817508e-02 -2.76220679e-01
-6.94545209e-01 -4.85546052e-01 8.77053320e-01 4.23699319e-01
-3.30494493e-02 -1.00586236e+00 -1.02513008e-01 4.99621600e-01
-1.57338381e-01 -1.32281601e+00 -7.75808334e-01 7.65730441e-01
-5.80978155e-01 -3.92172545e-01 -7.48536110e-01 -1.12375188e+00
3.85243863e-01 1.24936014e-01 1.14008415e+00 4.98958379e-02
-2.23474042e-03 3.07623357e-01 -6.31335795e-01 -5.74704707e-01
-7.66615093e-01 2.47199908e-01 2.66914308e-01 -1.62189543e-01
6.58918619e-01 -2.14362204e-01 -2.00546950e-01 2.95648634e-01
-7.55856752e-01 -6.90511391e-02 6.22960031e-01 9.21181381e-01
2.60696828e-01 -3.26623112e-01 4.30195272e-01 -5.23151696e-01
7.10903883e-01 -2.76678205e-01 -4.74181652e-01 2.12044641e-01
-6.60467386e-01 5.91966771e-02 5.67816317e-01 -5.81400394e-01
-6.80905402e-01 -3.49971168e-02 -1.06499887e+00 -1.09536573e-01
-1.41278714e-01 4.97710347e-01 -1.75041303e-01 3.82274479e-01
5.16988337e-01 4.63782310e-01 8.87786970e-02 -7.40340889e-01
3.42217326e-01 1.13886952e+00 3.98725390e-01 -1.47330835e-01
7.76430368e-01 -1.84142590e-01 -8.37646365e-01 -1.26657844e+00
-4.76695806e-01 -8.05839002e-01 -3.40137482e-01 1.88247815e-01
7.93955088e-01 -9.23644364e-01 -2.28571102e-01 3.21014792e-01
-1.41957712e+00 -7.68102646e-01 -3.37491006e-01 8.71077001e-01
-2.99788862e-01 1.40132725e-01 -6.65567636e-01 -1.12104511e+00
-2.05383703e-01 -1.16083729e+00 8.58575940e-01 -2.64971137e-01
-5.96119761e-01 -1.08664405e+00 1.62275493e-01 1.90752178e-01
8.31970453e-01 -8.19061577e-01 9.43940461e-01 -1.18488920e+00
2.09575966e-02 -3.97635996e-01 2.31856406e-01 7.31319606e-01
-4.18231636e-03 -2.51947939e-01 -1.20042944e+00 -3.33725393e-01
-7.42189437e-02 -1.89249754e-01 8.86018991e-01 3.17901820e-01
8.59115362e-01 -2.04929903e-01 -9.94200036e-02 4.67426419e-01
1.09842896e+00 4.56156850e-01 5.80817521e-01 -5.28406464e-02
3.26335013e-01 3.07563841e-01 -2.81099454e-02 -1.85020380e-02
2.50482798e-01 7.96915174e-01 9.24737230e-02 -1.76251560e-01
-5.43609619e-01 -3.06606412e-01 9.90708768e-01 1.72186363e+00
4.15327877e-01 -7.82646120e-01 -1.34502649e+00 7.09428132e-01
-1.26616049e+00 -7.18728304e-01 -7.49350116e-02 2.32611799e+00
9.39039886e-01 5.70394099e-01 1.00009479e-02 1.02100730e-01
5.27085304e-01 1.93190947e-01 -5.86411618e-02 -8.66336942e-01
-1.08555727e-01 6.13216162e-01 6.47524536e-01 1.23633409e+00
-6.26090825e-01 1.26754546e+00 6.92781782e+00 1.29012823e+00
-1.04759216e+00 4.31636751e-01 4.40527171e-01 -1.66869700e-01
-4.43273991e-01 -4.14104313e-02 -1.22920310e+00 3.59283000e-01
1.70077240e+00 2.50551820e-01 5.64676821e-01 5.95409274e-01
3.57257217e-01 -2.34924182e-02 -1.04114997e+00 1.08401549e+00
2.77022451e-01 -8.95629764e-01 -3.86008061e-02 -8.23803619e-02
2.08033934e-01 5.53169966e-01 7.93892145e-02 7.17584789e-01
2.92607605e-01 -1.25055718e+00 8.36809576e-01 8.29309672e-02
1.00614786e+00 -4.55105096e-01 6.74059093e-01 4.93705004e-01
-9.83101010e-01 2.37526998e-01 -2.72567391e-01 -6.72496408e-02
4.65578496e-01 4.18955058e-01 -1.20481658e+00 -3.72986197e-02
2.13440597e-01 -1.08305238e-01 -3.69764745e-01 9.54634964e-01
-2.28159219e-01 1.45600724e+00 -5.42542875e-01 -3.34419429e-01
5.49086154e-01 2.02431127e-01 5.05742610e-01 1.94122422e+00
4.75898892e-01 3.76223512e-02 5.00165820e-02 2.90785432e-01
4.94828373e-02 2.00878322e-01 -5.03193259e-01 -3.55704159e-01
5.13946772e-01 7.13919282e-01 -6.38575196e-01 -4.24722284e-01
-7.48000205e-01 1.17111135e+00 1.91424921e-01 6.18636489e-01
-5.79989612e-01 -5.67198455e-01 8.36631417e-01 1.78217709e-01
5.05607903e-01 -8.52180600e-01 -8.46188422e-03 -9.36136007e-01
-2.40091011e-01 -9.68315363e-01 -9.95676368e-02 -5.80391705e-01
-1.11664975e+00 9.07030106e-01 -1.44247621e-01 -8.26595068e-01
-5.58247864e-01 -1.14642155e+00 -7.17336535e-01 1.25225556e+00
-1.67638409e+00 -7.36793101e-01 3.33012313e-01 2.45854482e-01
1.12332106e+00 -4.06342864e-01 1.29349113e+00 2.37716615e-01
-4.99226242e-01 7.91668475e-01 3.06639880e-01 3.36066991e-01
6.37460470e-01 -1.14446020e+00 1.10591555e+00 1.03271151e+00
6.05868757e-01 7.52286017e-01 6.28575921e-01 -4.42447841e-01
-1.37955976e+00 -7.43802309e-01 1.39870369e+00 -6.39125109e-01
6.27244771e-01 -9.13287461e-01 -8.32520425e-01 6.29816353e-01
4.98667568e-01 -1.01922877e-01 9.73856091e-01 3.89208049e-01
-3.49017143e-01 5.97337037e-02 -5.05120814e-01 7.31237710e-01
9.10220861e-01 -9.04415250e-01 -6.58942342e-01 2.78077006e-01
1.00456691e+00 -1.20456189e-01 -1.60247982e-01 4.06364799e-02
4.07342315e-01 -5.57246506e-01 8.41010094e-01 -6.65077925e-01
-3.32781225e-01 -1.54143162e-02 -5.55586755e-01 -1.66484010e+00
-3.36037517e-01 -9.21522439e-01 -5.28096072e-02 1.14113200e+00
1.11357367e+00 -7.36674547e-01 4.21123743e-01 5.78389525e-01
-4.79275703e-01 -4.84396100e-01 -1.14008689e+00 -1.30832613e+00
3.25443864e-01 -1.20998108e+00 1.86730266e-01 6.07448339e-01
-7.07068294e-02 4.79398102e-01 -2.33675003e-01 2.65004132e-02
-3.07905935e-02 -6.45747602e-01 4.29881513e-01 -6.83516085e-01
-5.60629845e-01 -6.55903578e-01 -3.57346565e-01 -1.55689561e+00
1.81709856e-01 -1.08264399e+00 6.58314466e-01 -1.42610002e+00
-3.58292073e-01 -2.93877959e-01 -2.85237193e-01 6.52516127e-01
-1.55846491e-01 2.51700550e-01 2.60636270e-01 -6.04072884e-02
-1.78346217e-01 5.50181091e-01 5.12830436e-01 -2.36958712e-02
-2.56385952e-01 5.77157550e-02 -3.71270210e-01 6.04418039e-01
1.05083144e+00 -5.20080805e-01 -4.46250588e-01 -8.27570140e-01
-8.64196122e-02 -2.87682265e-01 -6.46441197e-03 -9.40912127e-01
3.59744400e-01 2.27881908e-01 -9.21948776e-02 -3.96015882e-01
7.05614030e-01 -6.45108461e-01 -3.08006763e-01 4.99914706e-01
-2.27026045e-01 2.20618874e-01 4.07280385e-01 2.59256005e-01
-2.22065240e-01 -6.46830440e-01 7.14730918e-01 -4.77808975e-02
-7.68509626e-01 -2.17955038e-01 -1.08349586e+00 2.68928766e-01
5.02879024e-01 -2.01206565e-01 8.28799456e-02 -6.87913954e-01
-5.78580141e-01 -6.50981665e-02 -1.40012085e-01 5.39435625e-01
6.62069499e-01 -1.09891129e+00 -9.66555715e-01 6.37771845e-01
1.41688973e-01 -4.35580581e-01 -2.93348521e-01 6.85516238e-01
-3.77885044e-01 6.40111804e-01 4.93757933e-01 -3.44246864e-01
-1.15992677e+00 2.08595648e-01 2.84150749e-01 3.29410508e-02
-3.98091435e-01 1.36137855e+00 -1.47676766e-01 -5.97614944e-01
6.49850845e-01 -5.30606449e-01 3.03264737e-01 2.91467439e-02
5.76750457e-01 4.54714661e-03 3.78619045e-01 -6.14642799e-01
-5.68271935e-01 3.67449433e-01 -1.55205593e-01 -7.86542654e-01
9.14582253e-01 -5.22630252e-02 5.28307676e-01 8.82408679e-01
1.31877851e+00 2.89444178e-01 -7.22106814e-01 -2.55843997e-01
1.82194605e-01 -1.20721668e-01 2.73429900e-01 -8.12281549e-01
-5.29524624e-01 1.28164661e+00 5.44671953e-01 3.77882808e-01
7.39912093e-01 8.39683320e-03 1.08863795e+00 5.02240181e-01
-5.11640646e-02 -1.38052571e+00 -1.69808432e-01 1.10579252e+00
6.98765516e-01 -1.23411095e+00 -5.89169323e-01 -1.17779389e-01
-6.50495231e-01 1.16435277e+00 3.02510470e-01 2.00664997e-01
9.15948868e-01 6.38531148e-01 4.95174527e-01 2.59159267e-01
-8.91207397e-01 -5.53963363e-01 3.36104155e-01 5.29522598e-01
7.13318110e-01 8.07318240e-02 -1.40157357e-01 5.80854177e-01
-3.82777065e-01 -4.87524301e-01 1.77807555e-01 7.39261806e-01
-7.19488025e-01 -1.30423546e+00 -3.65493238e-01 2.28579387e-01
-3.62518460e-01 -7.95948982e-01 -3.35880548e-01 6.12608969e-01
-2.82081962e-01 1.34009635e+00 1.51746005e-01 -4.63763595e-01
4.32907641e-01 9.53908682e-01 -1.48345763e-02 -1.00636172e+00
-5.59265077e-01 3.28503370e-01 3.60471845e-01 -2.68632352e-01
1.55461073e-01 -3.83654326e-01 -1.26601267e+00 -1.04186565e-01
-6.67495012e-01 2.83038110e-01 9.64619935e-01 1.01878166e+00
2.18724146e-01 5.03369331e-01 3.16484779e-01 -7.86145389e-01
-6.54297411e-01 -1.27634203e+00 -3.35488290e-01 1.59594119e-02
5.42988122e-01 -4.55949455e-02 -5.74001253e-01 3.34907416e-03] | [14.362090110778809, 6.833063125610352] |
3b83c04b-7901-452f-8d75-3e924db1c722 | a-comprehensive-survey-on-deep-learning-for | 2306.02051 | null | https://arxiv.org/abs/2306.02051v2 | https://arxiv.org/pdf/2306.02051v2.pdf | A Comprehensive Survey on Deep Learning for Relation Extraction: Recent Advances and New Frontiers | Relation extraction (RE) involves identifying the relations between entities from unstructured texts. RE serves as the foundation for many natural language processing (NLP) applications, such as knowledge graph completion, question answering, and information retrieval. In recent years, deep neural networks have dominated the field of RE and made noticeable progress. Subsequently, the large pre-trained language models (PLMs) have taken the state-of-the-art of RE to a new level. This survey provides a comprehensive review of existing deep learning techniques for RE. First, we introduce RE resources, including RE datasets and evaluation metrics. Second, we propose a new taxonomy to categorize existing works from three perspectives (text representation, context encoding, and triplet prediction). Third, we discuss several important challenges faced by RE and summarize potential techniques to tackle these challenges. Finally, we outline some promising future directions and prospects in this field. This survey is expected to facilitate researchers' collaborative efforts to tackle the challenges of real-life RE systems. | ['Ruifeng Xu', 'Ying Shen', 'Wai Lam', 'Hong Cheng', 'Rui Zhang', 'Lingzhi Wang', 'Min Yang', 'Yang Deng', 'Xiaoyan Zhao'] | 2023-06-03 | null | null | null | null | ['knowledge-graph-completion', 'relation-extraction', 'information-retrieval'] | ['knowledge-base', 'natural-language-processing', 'natural-language-processing'] | [ 1.41807303e-01 3.34174484e-01 -4.93458182e-01 -2.72124976e-01
-5.95851541e-01 -3.43993604e-01 6.38857365e-01 5.91629088e-01
-4.34750050e-01 7.92090654e-01 3.24822336e-01 -3.03402454e-01
-1.82393208e-01 -1.17208040e+00 -4.79995996e-01 -2.45482311e-01
-3.69917661e-01 7.80747592e-01 -2.29582503e-01 -3.64612788e-01
8.36116523e-02 6.63030148e-01 -1.29786527e+00 4.08366919e-01
8.40404809e-01 9.20876443e-01 -7.56790563e-02 5.27086437e-01
-5.17761707e-01 1.23560727e+00 -6.84286475e-01 -8.55549753e-01
-2.57922709e-01 -1.63061023e-01 -1.43129325e+00 -2.56205618e-01
2.51254976e-01 2.11141277e-02 -7.61005282e-01 8.11087072e-01
5.34645975e-01 6.82742238e-01 5.87207198e-01 -1.19931662e+00
-1.05555677e+00 8.32893550e-01 -4.09915268e-01 3.91538829e-01
5.11397481e-01 -7.80906558e-01 1.54591572e+00 -1.15363228e+00
8.47510874e-01 9.43445206e-01 5.41193902e-01 2.61929810e-01
-7.31047392e-01 -4.17695284e-01 3.27357769e-01 7.97336459e-01
-1.57172859e+00 -4.43739504e-01 6.78934991e-01 -2.06258714e-01
1.89019930e+00 3.57248455e-01 4.33998704e-01 7.49018550e-01
-4.11684625e-02 1.02821815e+00 3.53564143e-01 -6.50379002e-01
-2.45580718e-01 -2.50432462e-01 3.30646694e-01 7.77163863e-01
3.53916228e-01 -4.89845455e-01 -5.56696057e-01 -1.32308409e-01
4.98172492e-01 -2.28885561e-01 -1.78150803e-01 -1.92153946e-01
-1.17107725e+00 7.47232199e-01 4.64084804e-01 7.32573390e-01
-3.10366452e-01 -2.98877545e-02 4.54373896e-01 2.85627425e-01
6.11766040e-01 6.58500791e-01 -6.53836906e-01 1.05799146e-01
-5.51151037e-01 2.90705144e-01 1.13980401e+00 1.07394791e+00
6.42143190e-01 -1.19065531e-01 -5.15416622e-01 1.14811409e+00
4.84490693e-02 7.64058605e-02 2.02584222e-01 -4.46110904e-01
8.40923011e-01 7.73988783e-01 -1.25961810e-01 -1.26242590e+00
-6.68274760e-01 -4.10026550e-01 -1.07083821e+00 -8.15027416e-01
-6.61122650e-02 -3.02776605e-01 -6.79455936e-01 1.35246921e+00
1.93416983e-01 2.15615854e-01 2.52195448e-01 4.27634180e-01
1.75907230e+00 7.05774069e-01 2.39867568e-01 -1.64856225e-01
1.45496142e+00 -1.19382811e+00 -8.94905031e-01 -3.08002055e-01
8.24485421e-01 -6.24980927e-01 4.77496058e-01 -9.78585705e-02
-1.06045043e+00 -2.45401472e-01 -8.09763670e-01 -5.95315754e-01
-8.15656066e-01 5.44100404e-01 1.00618815e+00 2.74373055e-01
-1.04071915e+00 5.32465577e-01 -7.18852818e-01 -6.39201939e-01
3.26678693e-01 6.11946464e-01 -4.35665041e-01 5.74945025e-02
-1.76995742e+00 1.18779266e+00 5.12437940e-01 4.92473811e-01
-4.08585489e-01 -3.22432280e-01 -1.11176741e+00 5.20222545e-01
5.23970902e-01 -8.99613202e-01 1.20230210e+00 -2.32736632e-01
-1.29897499e+00 1.19140613e+00 -4.06270534e-01 -4.79909539e-01
-2.14724153e-01 -4.71894294e-01 -7.68318832e-01 -2.67122742e-02
-7.30316862e-02 2.99632758e-01 8.30563381e-02 -1.01536059e+00
-5.34129620e-01 -1.78339988e-01 2.43548959e-01 3.79275948e-01
-3.09695244e-01 4.88941669e-01 -5.48623919e-01 -5.61334670e-01
-1.76378936e-01 -6.70416474e-01 -2.52737939e-01 -7.42920697e-01
-7.89618433e-01 -9.02853668e-01 2.32914165e-01 -5.04541695e-01
1.54741287e+00 -1.70372808e+00 1.68928206e-01 5.46650626e-02
6.81474030e-01 6.64455295e-01 -2.10139215e-01 1.20609474e+00
-4.21229511e-01 1.73210785e-01 -1.00752003e-02 -4.74221706e-01
-2.78604366e-02 -7.34734610e-02 -4.37252849e-01 1.84634939e-01
2.60486275e-01 1.54964674e+00 -9.96925294e-01 -4.05067742e-01
9.55212414e-02 5.72960734e-01 5.18335328e-02 2.84983337e-01
-2.59778202e-01 1.53956100e-01 -6.02281749e-01 4.38563317e-01
3.14102918e-01 -7.33945489e-01 5.78174949e-01 -2.55232215e-01
1.00515388e-01 8.75951767e-01 -8.83301854e-01 1.31087780e+00
-3.89438570e-01 8.47627997e-01 -4.45253879e-01 -1.50469339e+00
1.16811192e+00 5.47832608e-01 7.23036468e-01 -4.82541084e-01
1.56225190e-01 7.76285976e-02 -1.27701461e-01 -6.31720841e-01
9.41514492e-01 2.02042475e-01 2.46737137e-01 3.03822696e-01
2.01562554e-01 2.58860767e-01 5.17740428e-01 2.71694243e-01
1.06823123e+00 1.53042178e-03 8.44597280e-01 2.51924515e-01
7.74136126e-01 -8.58997181e-02 3.86232644e-01 4.68513072e-01
7.08003193e-02 1.88850239e-01 2.99445927e-01 -6.64828002e-01
-5.53942382e-01 -6.36819422e-01 6.08428717e-02 1.07543182e+00
3.47297899e-02 -7.94601679e-01 -3.48308325e-01 -7.59370863e-01
-2.31915101e-01 5.42370975e-01 -5.76763332e-01 -6.18875921e-02
-8.05987418e-01 -9.23950553e-01 7.26624608e-01 7.82769561e-01
7.83447564e-01 -1.39302492e+00 2.53909111e-01 3.21677178e-01
-6.80126250e-01 -1.47780240e+00 1.42204789e-02 6.70417994e-02
-7.68132925e-01 -1.16139293e+00 -6.48994803e-01 -1.40108550e+00
3.82916689e-01 3.84244263e-01 1.65588164e+00 3.06038380e-01
-8.34559798e-02 4.18062568e-01 -5.93582511e-01 -2.56771296e-01
1.04237862e-01 5.18003345e-01 -1.74764290e-01 -3.70261669e-01
1.15224779e+00 -6.17279947e-01 -3.97292644e-01 -5.96978739e-02
-6.21084332e-01 -2.25200444e-01 4.68050271e-01 6.59375846e-01
5.96077323e-01 6.89964741e-02 7.99292266e-01 -1.06942797e+00
9.60023463e-01 -6.33964956e-01 -1.92583695e-01 8.89689922e-01
-6.61963105e-01 -1.14428133e-01 3.43642086e-01 -1.19911671e-01
-1.09183562e+00 -3.72303784e-01 -3.25995952e-01 1.56022355e-01
-1.33016720e-01 1.08645844e+00 -6.99734166e-02 -4.64326655e-03
6.21338129e-01 1.35330334e-01 -7.48405039e-01 -5.31866312e-01
5.58073580e-01 6.41283035e-01 2.26778373e-01 -6.17647231e-01
3.51086706e-01 2.28950888e-01 6.97315559e-02 -9.05877054e-01
-1.29690254e+00 -6.70671105e-01 -5.29301703e-01 2.39183292e-01
8.69692802e-01 -8.05022538e-01 -7.86344051e-01 2.39688739e-01
-1.43833768e+00 -1.07754685e-01 -2.51043260e-01 3.35245430e-01
-2.60391057e-01 4.66510713e-01 -1.03069365e+00 -5.67870975e-01
-8.14522147e-01 -8.17184210e-01 9.25586998e-01 4.60781932e-01
-2.01660648e-01 -1.34654844e+00 3.38156402e-01 5.18086016e-01
1.90644965e-01 1.36977537e-02 9.20631170e-01 -1.10154521e+00
-4.53870326e-01 -2.37881646e-01 -3.87424856e-01 1.67291276e-02
2.04633951e-01 -8.23637545e-02 -9.38334644e-01 2.09489036e-02
-5.57007492e-01 -3.66490006e-01 9.68962789e-01 1.91737115e-01
1.14242065e+00 -6.64832368e-02 -9.09378827e-01 5.29038906e-01
1.00032270e+00 1.86575457e-01 6.28788590e-01 3.21767509e-01
9.44417715e-01 6.66202724e-01 4.85568017e-01 8.45100731e-02
1.10751259e+00 4.79272693e-01 5.23249432e-02 -1.09381370e-01
-1.58997580e-01 -3.15198839e-01 -1.15928359e-01 1.15003347e+00
-5.42546391e-01 -8.33291650e-01 -1.06304860e+00 5.42753458e-01
-1.97012198e+00 -8.93677711e-01 -1.58805117e-01 1.78548145e+00
7.17677355e-01 -2.81724989e-01 -3.07507038e-01 -4.83144373e-02
9.09877539e-01 3.07854533e-01 -3.21044534e-01 -3.33537728e-01
-1.77368119e-01 3.84392560e-01 7.51156881e-02 4.04789925e-01
-1.41664159e+00 1.40905058e+00 6.58113909e+00 6.33166313e-01
-7.47737586e-01 -2.24969208e-01 5.07710278e-01 5.08650362e-01
-2.89420098e-01 3.12619796e-03 -9.97501910e-01 -3.06969583e-01
8.47864628e-01 -3.32825541e-01 3.71511936e-01 6.75372422e-01
-4.40921724e-01 1.88636467e-01 -1.17644167e+00 1.15281868e+00
3.57977390e-01 -1.69708037e+00 9.12946388e-02 -1.39981627e-01
6.97802782e-01 4.00845826e-01 -2.33040422e-01 7.52918541e-01
5.19533932e-01 -1.22626436e+00 -3.18000883e-01 5.63614786e-01
9.39887762e-01 -6.94750488e-01 9.37084615e-01 9.32105929e-02
-1.68164217e+00 2.15512425e-01 -3.95630717e-01 -1.70537710e-01
2.21916184e-01 7.24874318e-01 -8.54244232e-01 1.11603522e+00
6.19673073e-01 9.16223586e-01 -3.46972048e-01 9.14029300e-01
-6.07640088e-01 3.07932198e-01 -4.64685000e-02 -1.65943742e-01
-1.08629435e-01 -2.08098203e-01 3.80161762e-01 1.31234908e+00
-6.25197664e-02 2.39239722e-01 1.88422963e-01 4.62006599e-01
-7.67991245e-01 3.34081441e-01 -5.23680985e-01 -6.21485233e-01
5.93784273e-01 1.42539847e+00 -6.40325069e-01 -3.80875885e-01
-7.06632137e-01 7.42492378e-01 9.81605589e-01 6.66819155e-01
-3.08701694e-01 -8.13886642e-01 4.68056053e-01 -4.59109306e-01
5.89700677e-02 -5.11529863e-01 5.15971333e-02 -1.58681405e+00
-5.84382787e-02 -6.87057257e-01 7.08381772e-01 -4.95998323e-01
-1.36185682e+00 9.54053223e-01 -1.49774412e-02 -8.34346950e-01
-3.16499144e-01 -5.72831988e-01 -2.95982271e-01 6.98838234e-01
-1.65525174e+00 -1.10934877e+00 -2.16320619e-01 3.48933488e-01
3.46982241e-01 -2.72795320e-01 1.27276492e+00 6.13047957e-01
-7.66806543e-01 6.25908315e-01 -1.01909284e-02 7.57254899e-01
6.78070366e-01 -1.16072106e+00 9.80297208e-01 6.18528783e-01
7.39612818e-01 8.47154260e-01 1.22351833e-01 -6.95290148e-01
-1.20490682e+00 -9.05732095e-01 1.88117933e+00 -3.89910519e-01
8.79230499e-01 -2.99147069e-01 -8.15198004e-01 1.19751537e+00
2.54665762e-01 1.15371816e-01 1.09104323e+00 7.76811421e-01
-2.26264164e-01 -2.63773967e-02 -8.14876974e-01 7.51987576e-01
1.10021567e+00 -6.83188319e-01 -5.75800419e-01 5.07095873e-01
7.08550036e-01 -5.04815519e-01 -1.15753448e+00 5.79629421e-01
3.19288731e-01 -4.85063076e-01 9.51968729e-01 -9.84845936e-01
2.74085879e-01 1.03606589e-01 7.01876953e-02 -1.22627294e+00
-4.71039563e-01 -7.73460150e-01 -7.67829776e-01 1.27998900e+00
4.54029113e-01 -5.46236753e-01 7.98037946e-01 6.49702251e-01
-4.32160161e-02 -1.15420747e+00 -5.01888514e-01 -3.46437603e-01
1.56750947e-01 -4.27777469e-01 7.30536103e-01 1.21265554e+00
3.00195247e-01 1.02906668e+00 -1.75242037e-01 -7.48928711e-02
4.59237536e-03 3.61617357e-01 5.90030134e-01 -1.24754941e+00
2.94509120e-02 -4.12397951e-01 -2.50851989e-01 -1.51021075e+00
5.30548394e-01 -1.26013768e+00 -3.62970650e-01 -2.23423791e+00
2.49217033e-01 -3.75615627e-01 -3.39008540e-01 5.41371942e-01
-4.65877295e-01 1.15685491e-02 -3.22884060e-02 7.16127753e-02
-9.21325386e-01 6.76897645e-01 1.27637076e+00 -2.92300969e-01
-5.83043136e-02 5.46851642e-02 -8.61751139e-01 5.85562170e-01
9.52158332e-01 -2.48095319e-01 -3.86741370e-01 -7.12224603e-01
9.54420090e-01 7.79747264e-03 -1.71602219e-01 -4.40135300e-01
4.77179706e-01 -1.77979752e-01 2.73927987e-01 -6.82483137e-01
2.82739550e-01 -5.96622288e-01 -4.75247681e-01 -2.43393242e-01
-4.66325730e-01 1.11851133e-01 6.26217872e-02 3.17424655e-01
-5.42800903e-01 -1.60881966e-01 1.47252142e-01 -6.14536330e-02
-7.99819708e-01 6.67737067e-01 -2.41463795e-01 3.90201211e-01
7.00731695e-01 3.53671342e-01 -5.42014837e-01 -5.19574344e-01
-7.75457978e-01 5.35874784e-01 -4.63521957e-01 4.97263908e-01
6.31569803e-01 -1.09477162e+00 -8.18120956e-01 -2.90761441e-01
2.23588645e-01 2.04558745e-01 1.67829692e-01 5.94426215e-01
-3.44224066e-01 8.53376567e-01 3.87059450e-01 5.85457757e-02
-1.15211618e+00 7.38891959e-01 3.80894482e-01 -8.99754286e-01
-6.14371240e-01 1.19082355e+00 -7.14329332e-02 -6.24343336e-01
3.76343668e-01 -7.30972067e-02 -1.01977015e+00 -6.16787300e-02
6.05833948e-01 4.22624737e-01 4.78448391e-01 -5.93218744e-01
-5.13324201e-01 3.50567967e-01 -3.66065204e-01 3.37651551e-01
1.41323376e+00 -1.19715177e-01 -4.52580273e-01 3.19977432e-01
9.98442233e-01 1.47439167e-01 -3.01112384e-01 -6.78075433e-01
5.28239667e-01 2.02531219e-01 -1.14684433e-01 -7.45600820e-01
-1.04459417e+00 8.65216851e-01 -2.48068452e-01 4.50295895e-01
1.04262161e+00 2.20550328e-01 1.06204963e+00 1.15127277e+00
2.43496060e-01 -8.96397173e-01 -1.87393382e-01 1.05012631e+00
7.85743773e-01 -1.30163670e+00 2.36477181e-01 -7.33267665e-01
-5.99624813e-01 1.12152052e+00 5.40863574e-01 -1.08442716e-01
8.62256527e-01 1.96123436e-01 -1.22593462e-01 -4.23406899e-01
-7.31155932e-01 -6.88447833e-01 7.05186486e-01 7.58841217e-01
1.18896580e+00 -3.20351683e-02 -4.28442061e-01 7.41615832e-01
-3.11933666e-01 -6.85610101e-02 8.97265673e-02 7.43406236e-01
-1.55144289e-01 -1.60864913e+00 2.56399363e-01 6.13100708e-01
-6.78544044e-01 -5.85166037e-01 -7.17155397e-01 5.68349779e-01
-1.71339586e-01 1.29528809e+00 -3.87476827e-03 -5.05548596e-01
3.13787401e-01 4.40794230e-02 5.19109607e-01 -9.79283512e-01
-5.85674286e-01 -3.72336239e-01 7.28641808e-01 -2.61196822e-01
-5.61609328e-01 -1.72903225e-01 -1.33852267e+00 -2.91988909e-01
-5.01515865e-01 5.28541148e-01 3.46093118e-01 1.17305923e+00
6.01849139e-01 6.52094603e-01 2.23876655e-01 -3.68313193e-01
-5.85140958e-02 -1.09929883e+00 -3.30115110e-01 -5.67601882e-02
1.98077992e-01 -6.90092564e-01 3.98357175e-02 -1.47181749e-01] | [9.268342018127441, 8.635517120361328] |
31072593-89fe-4b64-b969-f4b04ccad19c | semantics-as-a-foreign-language | null | null | https://aclanthology.org/D18-1263 | https://aclanthology.org/D18-1263.pdf | Semantics as a Foreign Language | We propose a novel approach to semantic dependency parsing (SDP) by casting the task as an instance of multi-lingual machine translation, where each semantic representation is a different foreign dialect. To that end, we first generalize syntactic linearization techniques to account for the richer semantic dependency graph structure. Following, we design a neural sequence-to-sequence framework which can effectively recover our graph linearizations, performing almost on-par with previous SDP state-of-the-art while requiring less parallel training annotations. Beyond SDP, our linearization technique opens the door to integration of graph-based semantic representations as features in neural models for downstream applications. | ['Ido Dagan', 'Gabriel Stanovsky'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [ 2.20304593e-01 6.70171797e-01 -5.99411488e-01 -8.04841459e-01
-1.21920133e+00 -9.64270532e-01 5.85299730e-01 2.11054951e-01
-1.95287317e-01 7.61140764e-01 5.74183166e-01 -8.10798049e-01
4.61778730e-01 -7.67032266e-01 -1.07695925e+00 -1.81546062e-01
1.39809385e-01 6.55101418e-01 7.55825266e-02 -4.91763324e-01
-5.44560552e-02 2.00517014e-01 -8.00058663e-01 4.91096795e-01
8.37178826e-01 2.87313282e-01 1.90050378e-01 5.71522415e-01
-4.78687465e-01 7.70730674e-01 -1.85036510e-01 -8.87146115e-01
7.55068734e-02 -6.46327615e-01 -1.18468559e+00 -3.47709835e-01
4.10823256e-01 1.76871642e-01 -1.28143862e-01 9.72482145e-01
5.36691286e-02 -1.31286487e-01 4.40182656e-01 -8.57358158e-01
-1.24605393e+00 8.75922561e-01 -1.79018974e-01 1.06676370e-02
3.82580042e-01 -3.72608095e-01 1.48107243e+00 -6.34707987e-01
1.04319799e+00 1.49033701e+00 6.39415383e-01 6.78461254e-01
-1.38010764e+00 -2.18560010e-01 4.72743303e-01 3.54307070e-02
-6.94689989e-01 -4.81985897e-01 6.66824758e-01 -1.17838636e-01
1.52666223e+00 -2.20504224e-01 2.97520161e-01 1.01441038e+00
1.61450267e-01 9.13632393e-01 9.24005687e-01 -8.39646101e-01
3.58073115e-02 -1.40210539e-01 5.15020788e-01 1.00166416e+00
1.56724125e-01 -1.17150009e-01 -4.76029932e-01 1.48130402e-01
4.38913018e-01 -5.63831449e-01 9.96745974e-02 -3.66017848e-01
-7.88363695e-01 1.12759864e+00 4.11974341e-01 2.25281730e-01
6.51360005e-02 3.35459024e-01 7.42232263e-01 5.79684496e-01
7.83482194e-01 1.62244886e-01 -1.02542794e+00 2.07864583e-01
-2.87535071e-01 1.02881901e-01 1.14323330e+00 1.15971816e+00
8.11436355e-01 4.74981107e-02 1.03076734e-01 8.58180285e-01
1.54560238e-01 1.90781429e-01 3.53788257e-01 -8.70624661e-01
8.08112085e-01 3.65878850e-01 -3.98203135e-01 -2.14959905e-01
-3.29077184e-01 -1.62844107e-01 -1.26381323e-01 -2.64632821e-01
2.52288520e-01 -1.72886670e-01 -8.13245296e-01 2.20204782e+00
2.00508609e-01 -4.06464487e-02 5.08754373e-01 4.98944134e-01
5.60111046e-01 6.28240526e-01 5.19757628e-01 -3.68537307e-02
1.52109087e+00 -1.32374060e+00 -3.82763177e-01 -7.10889697e-01
1.28129292e+00 -4.68585670e-01 1.10412526e+00 -1.62049890e-01
-1.24389327e+00 -2.42530674e-01 -9.13065255e-01 -6.33455396e-01
-6.09979093e-01 -2.28459522e-01 9.79155779e-01 5.50604403e-01
-1.54086053e+00 7.92064786e-01 -8.89885902e-01 -6.08419240e-01
1.59881994e-01 2.58677930e-01 -6.12397075e-01 -3.29412699e-01
-1.28344774e+00 1.12781882e+00 4.71418083e-01 -2.21628159e-01
-2.87551522e-01 -7.90464818e-01 -1.47633517e+00 5.75620048e-02
2.91691087e-02 -1.01114118e+00 1.60528839e+00 -1.15843785e+00
-1.59156168e+00 1.34577072e+00 -5.64015269e-01 -5.56492090e-01
-1.92266386e-02 -3.31886828e-01 -1.85889035e-01 -6.77811652e-02
3.57389271e-01 6.57013893e-01 4.21837211e-01 -1.02144098e+00
-3.77260327e-01 -5.95674157e-01 3.77487898e-01 3.64814281e-01
1.54478341e-01 6.39452159e-01 -4.30107862e-01 -6.23186171e-01
-3.91058289e-02 -9.41223025e-01 -3.37647408e-01 -4.65200603e-01
-3.70065153e-01 -3.07604522e-01 3.52453560e-01 -7.88342118e-01
8.12500536e-01 -1.66374767e+00 6.68721199e-01 -3.10312092e-01
-1.63372189e-01 3.28443915e-01 -4.88203973e-01 7.40934193e-01
-4.08906460e-01 2.32600674e-01 -5.19981861e-01 -7.53553450e-01
-3.93319838e-02 6.53351426e-01 -2.46055260e-01 1.94434881e-01
4.95178282e-01 1.49186766e+00 -9.49141741e-01 -2.52755493e-01
-1.11747816e-01 2.93028295e-01 -7.25156367e-01 1.89453796e-01
-5.01128674e-01 4.65181559e-01 -5.11122048e-01 5.48307955e-01
5.22147179e-01 -1.22496724e-01 8.04255605e-01 9.12263021e-02
-8.80891085e-02 1.00718737e+00 -2.39325732e-01 2.29008532e+00
-8.88108790e-01 2.85339385e-01 2.40420401e-01 -1.51106000e+00
7.17779219e-01 1.82403803e-01 -8.96681175e-02 -6.02234721e-01
3.05329524e-02 3.55446279e-01 -3.00114363e-01 -1.62906930e-01
3.01405877e-01 -5.38004875e-01 -5.07656455e-01 5.21041572e-01
6.07968152e-01 -1.24158092e-01 4.70278971e-02 3.10017854e-01
9.59839821e-01 7.98190057e-01 4.90195483e-01 -4.49288517e-01
4.16157812e-01 2.49354050e-01 4.53618050e-01 4.71572429e-01
1.29042864e-01 2.34024629e-01 7.59944797e-01 -3.65116000e-01
-1.02456522e+00 -1.18879652e+00 1.79627091e-01 1.55501521e+00
-3.03768903e-01 -4.23848003e-01 -1.20834482e+00 -1.15098834e+00
-1.28713951e-01 8.52528214e-01 -4.03358996e-01 -7.93185234e-02
-1.18769634e+00 -7.47379243e-01 7.82240629e-01 7.66747415e-01
2.01068148e-02 -9.90437150e-01 1.02539927e-01 2.70379454e-01
-1.26302660e-01 -1.48631418e+00 -3.71428847e-01 5.86755633e-01
-9.95993555e-01 -8.02107155e-01 -3.59559387e-01 -1.50112557e+00
5.18190861e-01 5.11984564e-02 1.62502456e+00 -1.67983603e-02
1.31283075e-01 9.49079245e-02 -4.05127257e-01 -1.53912798e-01
-8.90711904e-01 4.76066887e-01 -2.03075588e-01 -5.83150566e-01
4.19098586e-01 -7.93271482e-01 -7.77283087e-02 -4.25240725e-01
-6.87957883e-01 2.33010530e-01 2.53423452e-01 8.15115929e-01
4.99907285e-01 -8.23037982e-01 7.49084949e-01 -1.66088867e+00
6.65149868e-01 -6.22996449e-01 -6.38126969e-01 3.49368811e-01
-3.75739843e-01 5.17020524e-01 1.25935960e+00 7.33083487e-02
-1.23417008e+00 2.41506428e-01 -4.69377100e-01 8.75304416e-02
-1.92115173e-01 6.11095011e-01 -3.48236442e-01 -3.91841233e-02
3.74922365e-01 4.89226989e-02 -6.60263151e-02 -6.90962791e-01
1.16008818e+00 3.30920875e-01 5.99948823e-01 -8.10593247e-01
6.44121230e-01 1.42871022e-01 5.98563366e-02 -4.75986421e-01
-1.25609124e+00 -1.66113988e-01 -1.15816474e+00 7.08042502e-01
1.16738331e+00 -1.13658082e+00 -2.35802621e-01 2.53267109e-01
-1.53053808e+00 -6.45521343e-01 -3.08023226e-02 1.00588568e-01
-7.32693970e-01 4.77334291e-01 -1.24120355e+00 -1.98866487e-01
-2.99868107e-01 -9.89228427e-01 1.28238750e+00 -2.16022298e-01
-9.21654552e-02 -1.56909740e+00 2.82927066e-01 3.49161506e-01
1.38552293e-01 -1.43644139e-01 1.56699228e+00 -9.44736540e-01
-5.70326269e-01 -1.67647433e-02 -3.83971483e-01 4.13345158e-01
-1.17213782e-02 -6.84994876e-01 -8.12867224e-01 -1.82717871e-02
-6.50115609e-02 -3.66666675e-01 8.92889082e-01 3.04414779e-01
5.22362769e-01 -3.59130621e-01 -2.82620788e-01 9.67849731e-01
1.51630712e+00 -2.42457107e-01 3.54877830e-01 3.19499552e-01
1.02550185e+00 9.19744670e-01 3.41410249e-01 -3.23610604e-01
8.16029012e-01 5.83058119e-01 1.15005508e-01 -2.72437721e-01
-3.93365502e-01 -6.59245670e-01 6.55466378e-01 1.19447613e+00
3.46640736e-01 -2.26998612e-01 -7.49558747e-01 6.57754421e-01
-1.81760573e+00 -3.31080168e-01 -1.96508452e-01 1.83966088e+00
7.80223370e-01 -1.59380794e-01 2.48323777e-04 -5.23285091e-01
6.12513006e-01 3.14048946e-01 -2.29090631e-01 -1.09659576e+00
-4.55136932e-02 7.15890229e-01 8.60673070e-01 1.01140571e+00
-9.43151653e-01 1.96293449e+00 7.16531610e+00 3.51563185e-01
-7.73867190e-01 5.02873898e-01 4.85557556e-01 4.82090116e-01
-6.28148317e-01 5.56790590e-01 -9.40961778e-01 -3.72281857e-02
1.35727525e+00 -5.98190576e-02 7.63135135e-01 7.67063975e-01
-1.34479284e-01 2.79409230e-01 -1.36849880e+00 3.94392461e-01
1.76454321e-01 -1.41237593e+00 2.01902628e-01 -2.92453289e-01
4.11956370e-01 5.27817249e-01 -3.34821939e-01 3.71315479e-01
1.13730156e+00 -9.01219368e-01 5.05835235e-01 -3.31735849e-01
9.24101889e-01 -5.51313877e-01 4.02689010e-01 1.39527440e-01
-1.30913651e+00 1.06027082e-01 -3.61101836e-01 -2.00980306e-01
6.51986659e-01 2.47258037e-01 -6.39155090e-01 9.64486659e-01
2.13524967e-01 8.98478389e-01 -3.51208568e-01 -1.30741764e-02
-9.81089950e-01 5.28183937e-01 4.60007973e-02 1.45195499e-01
4.05922294e-01 -4.16847348e-01 3.59828711e-01 1.70433486e+00
1.25015676e-01 5.54406941e-02 1.67849556e-01 6.00104570e-01
-3.86123598e-01 3.31489503e-01 -1.00564373e+00 -7.47534558e-02
2.67441154e-01 9.12916958e-01 -6.08523607e-01 -5.15072167e-01
-9.47632790e-01 1.51579797e+00 1.17365408e+00 3.80756855e-01
-5.67287564e-01 -7.96877593e-02 8.22580278e-01 -3.32183182e-01
3.47155899e-01 -4.31553245e-01 -4.07888412e-01 -1.64852870e+00
1.81317255e-02 -6.88392222e-01 5.04148424e-01 -6.22799516e-01
-1.31692410e+00 5.63797951e-01 -6.24667071e-02 -5.12768447e-01
-6.24801040e-01 -9.41489697e-01 -5.18606484e-01 1.09340990e+00
-1.89136410e+00 -1.68048120e+00 6.03807390e-01 3.03986520e-01
7.73140132e-01 -1.54199544e-02 1.27819455e+00 4.13958877e-02
-4.23785090e-01 5.26128471e-01 -1.72486398e-02 2.37906486e-01
5.78867853e-01 -1.32303536e+00 1.47572756e+00 1.08645630e+00
3.07647884e-01 6.83596551e-01 3.81219178e-01 -7.01973736e-01
-1.62276185e+00 -1.20139074e+00 1.55830550e+00 -7.07398415e-01
9.63749647e-01 -8.33186269e-01 -9.55654204e-01 1.38379109e+00
2.99392074e-01 -6.33865595e-02 5.87303042e-01 6.28461957e-01
-7.88880348e-01 3.46249044e-01 -7.96958327e-01 7.19163001e-01
1.54236472e+00 -9.07789409e-01 -8.25504541e-01 5.45014501e-01
1.43041348e+00 -4.02568370e-01 -7.44494140e-01 6.94571063e-02
1.57739252e-01 -5.68193734e-01 1.02408004e+00 -1.19181085e+00
6.42426610e-01 1.68604851e-01 -3.48912150e-01 -1.62550533e+00
-3.91846329e-01 -5.76203883e-01 1.86271608e-01 1.28547966e+00
7.92682469e-01 -8.25909793e-01 6.01438820e-01 4.81349319e-01
-7.13857234e-01 -4.70163137e-01 -9.10817981e-01 -7.28475332e-01
7.47247994e-01 -5.06402493e-01 3.66175324e-01 1.09090650e+00
4.38676894e-01 8.90111148e-01 -2.08106846e-01 6.98789433e-02
5.03707945e-01 2.13198185e-01 4.07207489e-01 -9.81712699e-01
-6.11698747e-01 -2.53301144e-01 -1.47956952e-01 -1.21199739e+00
1.09898472e+00 -1.73116767e+00 5.45831285e-02 -1.75637734e+00
2.54362933e-02 -3.00045580e-01 -1.58789903e-01 7.44160652e-01
-1.87349901e-01 1.93486601e-01 2.04015002e-01 8.11433196e-02
-4.87944514e-01 2.88571477e-01 8.45072985e-01 2.73136407e-01
-1.27047509e-01 -3.50424320e-01 -1.13382149e+00 6.54612243e-01
7.51511753e-01 -7.02689469e-01 -3.26319695e-01 -1.10016060e+00
1.95260435e-01 3.22238624e-01 -1.92833636e-02 -2.04838142e-01
-5.31588830e-02 -7.11089894e-02 -2.41972387e-01 1.06400155e-01
1.78292796e-01 -2.26808399e-01 -3.31767142e-01 2.21984372e-01
-4.99005973e-01 2.41070434e-01 2.88580835e-01 3.80605131e-01
-2.26449013e-01 -2.22784758e-01 6.53826177e-01 -5.56376457e-01
-6.88067973e-01 7.61914998e-02 -3.53750676e-01 3.18741918e-01
5.29997170e-01 2.51915365e-01 -4.51085359e-01 -1.28486335e-01
-7.20608711e-01 1.28043383e-01 6.88948989e-01 6.33745372e-01
5.51721267e-02 -9.12353158e-01 -5.86093307e-01 2.05617830e-01
6.07648194e-02 -2.18139708e-01 -1.95652753e-01 3.64636153e-01
-4.30844784e-01 6.94618702e-01 2.70586815e-02 -1.09266400e-01
-1.05735075e+00 7.28997827e-01 1.50045425e-01 -4.70516354e-01
-6.97497785e-01 7.99798548e-01 5.57308555e-01 -1.09439504e+00
-3.68849099e-01 -2.07222417e-01 -2.63240747e-02 -3.47685099e-01
1.79692671e-01 -2.52771527e-01 2.87361622e-01 -6.91744506e-01
-4.04866815e-01 6.25479639e-01 -1.19871460e-01 -2.22810328e-01
1.41118348e+00 -2.74045020e-01 -4.32659268e-01 2.99392521e-01
1.48652947e+00 1.00180306e-01 -8.59637022e-01 -2.50538975e-01
4.63407785e-01 5.34168147e-02 -3.96999598e-01 -8.12940776e-01
-7.99034178e-01 9.06291485e-01 -1.96049839e-01 -3.38531017e-01
8.31651747e-01 5.41445732e-01 1.02444530e+00 4.92718101e-01
3.89470220e-01 -8.52199078e-01 -4.90679830e-01 9.69771206e-01
3.71409029e-01 -1.06616139e+00 -3.89968812e-01 -8.73262584e-01
-7.36353517e-01 9.57271099e-01 2.14605555e-01 -4.79608297e-01
5.04145980e-01 4.29962337e-01 2.85922170e-01 2.47247331e-02
-7.90360093e-01 -3.23156118e-01 3.32345106e-02 8.26738775e-01
9.09830987e-01 2.43274048e-01 -5.86425185e-01 6.97701871e-01
-2.29180560e-01 -1.64636299e-02 3.53452146e-01 8.71757746e-01
-2.40025118e-01 -1.96356177e+00 3.15420896e-01 -5.70076928e-02
-8.32132876e-01 -8.45646262e-01 -3.04373384e-01 7.33701050e-01
-3.67220610e-01 7.68250108e-01 -9.23169777e-02 -5.94113804e-02
3.41261297e-01 5.85112154e-01 5.66017807e-01 -1.16827643e+00
-5.42727590e-01 -2.04787791e-01 7.22993970e-01 -8.13227594e-01
-2.77443618e-01 -5.70430636e-01 -1.46061003e+00 1.90025438e-02
4.71288450e-02 1.57083884e-01 8.38710785e-01 1.21029007e+00
5.13998032e-01 3.42306763e-01 3.25117797e-01 -7.28721142e-01
-3.71392578e-01 -6.21941864e-01 -2.26433516e-01 4.66200620e-01
1.79771721e-01 -2.50425637e-01 -2.01442670e-02 2.01893240e-01] | [10.45916748046875, 9.445958137512207] |
9f19357a-53c2-40fc-a1d6-72ae1fdaa037 | channel-estimation-for-reconfigurable-4 | 1912.03619 | null | https://arxiv.org/abs/1912.03619v2 | https://arxiv.org/pdf/1912.03619v2.pdf | Channel Estimation for Reconfigurable Intelligent Surface Aided Multi-User mmWave MIMO Systems | Channel acquisition is one of the main challenges for the deployment of reconfigurable intelligent surface (RIS) aided communication systems. This is because an RIS has a large number of reflective elements, which are passive devices with no active transmitting/receiving abilities. In this paper, we study the channel estimation problem for the RIS aided multi-user millimeter-wave (mmWave) multi-input multi-output (MIMO) system. Specifically, we propose a novel channel estimation protocol for the above system to estimate the cascaded channels, which are the products of the channels from the base station (BS) to the RIS and from the RIS to the users. Further, since the cascaded channels are typically sparse, this allows us to formulate the channel estimation problem as a sparse recovery problem using compressive sensing (CS) techniques, thereby allowing the channels to be estimated with less training overhead. Moreover, the sparse channel matrices of the cascaded channels of all users have a common block sparsity structure due to the common channel between the BS and the RIS. To take advantage of the common sparsity pattern, we propose a two-step multi-user joint channel estimation procedure. In the first step, we make use of the common column-block sparsity and project the received signals onto the common column subspace. In the second step, we make use of the row-block sparsity of the projected signals and propose a multi-user joint sparse matrix recovery algorithm that takes into account the common channel between the BS and the RIS. | ['Wei Yu', 'Hei Victor Cheng', 'Ying-Chang Liang', 'Jie Chen'] | 2019-12-08 | null | null | null | null | ['compressive-sensing'] | ['computer-vision'] | [ 4.85081971e-01 2.23334804e-01 1.96121112e-01 3.49235564e-01
-3.96317601e-01 -3.13026547e-01 -5.79085052e-02 -6.29707992e-01
2.25378945e-01 6.47804081e-01 2.56011099e-01 -2.86968678e-01
-3.16264510e-01 -8.27886224e-01 -6.82891428e-01 -1.28916454e+00
-1.69616655e-01 -1.02844670e-01 -2.86060810e-01 -1.83663413e-01
-1.84015751e-01 2.99685895e-01 -1.22896743e+00 -5.41483983e-02
7.02969968e-01 1.11230147e+00 2.14259982e-01 4.69386578e-01
2.36010864e-01 5.86281538e-01 -4.04755354e-01 1.24343373e-01
2.58376509e-01 -5.57732880e-01 1.84675455e-01 2.77299166e-01
-1.91930085e-01 -3.98061067e-01 -7.74286747e-01 8.35123658e-01
4.81700003e-01 -2.57064313e-01 4.97756153e-01 -1.06075275e+00
1.22040279e-01 4.43938196e-01 -8.80815983e-01 -5.36261559e-01
4.35532600e-01 -5.35484254e-01 6.54248238e-01 -8.13864231e-01
3.06065172e-01 9.16757703e-01 4.28157181e-01 -1.64232731e-01
-6.97498202e-01 -9.74764109e-01 -2.60291755e-01 -2.02626675e-01
-1.69162714e+00 -7.49754131e-01 7.44366109e-01 -3.88038546e-01
3.04433256e-01 3.64736736e-01 8.60312641e-01 6.76721334e-01
1.95855036e-01 3.22601050e-01 8.94903719e-01 -6.53109670e-01
2.53875196e-01 7.45382756e-02 -2.38870867e-02 7.96832323e-01
8.89461339e-01 3.08178775e-02 -4.69935060e-01 -3.67540002e-01
1.04222405e+00 1.14053480e-01 -5.44043541e-01 -2.01314703e-01
-1.11767685e+00 6.26420617e-01 2.10726216e-01 5.90353608e-01
-5.50145209e-01 6.72415793e-02 -4.28488314e-01 2.85937876e-01
-1.32208895e-02 1.10386856e-01 1.64929792e-01 4.22620535e-01
-7.51471341e-01 -3.49633694e-01 1.25222778e+00 1.34759569e+00
9.76944506e-01 1.23748243e-01 7.86478668e-02 5.76647043e-01
9.57303166e-01 1.19229066e+00 -1.18195206e-01 -3.52810681e-01
6.41642928e-01 2.53749162e-01 1.05460070e-01 -1.25726569e+00
-3.84484410e-01 -1.32447088e+00 -1.32685888e+00 -3.24332476e-01
1.05018504e-01 -7.17145681e-01 -5.34018278e-01 1.40238047e+00
2.47102931e-01 3.88729066e-01 6.25217199e-01 6.58282042e-01
5.39936125e-01 7.63984025e-01 -5.79460025e-01 -6.41613662e-01
1.22313023e+00 -5.31840265e-01 -7.01412022e-01 -4.63868678e-01
5.80464125e-01 -8.84970427e-01 6.50714412e-02 3.80800992e-01
-9.09259975e-01 -6.63740784e-02 -1.47785473e+00 5.55483758e-01
4.50545609e-01 7.55816340e-01 5.09527087e-01 7.12101758e-01
-6.31737828e-01 -2.15581849e-01 -6.51955307e-01 2.48145536e-02
3.95702384e-02 3.46144706e-01 -1.30492985e-01 -5.32018721e-01
-8.97785962e-01 2.60500550e-01 -5.15899919e-02 4.38275695e-01
-4.85959470e-01 -2.65395790e-01 -6.99323833e-01 1.23289496e-01
3.68871540e-01 -7.31572092e-01 5.51797569e-01 -7.72602558e-01
-1.51462162e+00 -1.94800254e-02 -3.23355496e-01 1.01504996e-01
7.57113844e-02 -5.27555030e-03 -4.78197247e-01 6.44910187e-02
-2.45380729e-01 -5.94762087e-01 7.95310080e-01 -1.38537800e+00
-3.94114405e-01 -3.69243771e-01 -2.11747080e-01 1.50389627e-01
-3.34996730e-01 -6.02640569e-01 -5.27072787e-01 -3.76065582e-01
8.58859479e-01 -1.24250090e+00 -5.19264698e-01 -5.93184829e-01
-7.58863926e-01 8.66355956e-01 7.02626705e-01 -5.96243560e-01
1.29632199e+00 -2.72336793e+00 4.31673646e-01 1.02460384e+00
1.72623962e-01 -1.84617653e-01 -9.15751904e-02 8.83104801e-01
1.53069139e-01 -7.21334279e-01 -9.99320149e-02 -2.17594326e-01
-5.04810691e-01 -6.00381456e-02 -1.77047968e-01 6.09492600e-01
-6.04852140e-01 3.00582230e-01 -5.56992590e-01 1.04468502e-01
-1.95919544e-01 4.48823035e-01 -6.01349533e-01 3.24101657e-01
3.73989075e-01 9.42993999e-01 -1.01674068e+00 5.04648387e-01
9.17496026e-01 -3.37924331e-01 5.48124611e-01 -5.14580429e-01
-2.62734532e-01 -2.10801587e-01 -1.67300010e+00 1.46306837e+00
-7.72854269e-01 1.85231313e-01 5.76394141e-01 -8.03040445e-01
9.42556500e-01 5.51734686e-01 8.07089150e-01 -7.79091775e-01
1.67913318e-01 6.42463744e-01 3.67652103e-02 -4.28755432e-01
1.03464045e-01 -1.69401914e-01 -9.61168334e-02 5.07502496e-01
-6.70900643e-02 1.35250628e-01 -1.88725621e-01 4.10509288e-01
1.28262293e+00 -5.33986628e-01 4.90898788e-01 -5.70585318e-02
6.87189460e-01 -4.18639570e-01 5.75504661e-01 5.34259856e-01
8.64876032e-01 1.66030645e-01 1.98550820e-01 2.84036279e-01
-6.23474300e-01 -7.15348721e-01 3.13334793e-01 2.10642666e-01
4.98987794e-01 -2.87063867e-01 -2.88577110e-01 8.59232843e-02
-6.88558295e-02 2.31914654e-01 -1.80540942e-02 -1.65949658e-01
-2.93729603e-01 -8.05377603e-01 1.58219844e-01 -3.86453979e-02
6.12713456e-01 -7.21179098e-02 -3.16448838e-01 4.42866385e-01
-2.57938147e-01 -1.26143634e+00 6.92355037e-02 7.44958147e-02
-7.33349562e-01 -1.04406536e+00 -7.33016312e-01 -6.35893166e-01
1.02757955e+00 8.06684792e-01 2.83449471e-01 7.56050348e-02
-7.73020983e-02 7.28930354e-01 -4.33621347e-01 -3.13905448e-01
4.80802283e-02 -2.65560299e-01 -1.47793805e-02 6.66400671e-01
-4.14500505e-01 -7.60555208e-01 -5.37904978e-01 5.35914779e-01
-5.37064493e-01 3.76042932e-01 1.29259586e+00 5.79012334e-01
4.74530369e-01 2.83016115e-01 3.87498021e-01 -6.14812016e-01
4.69548851e-02 -7.33999491e-01 -6.92857862e-01 1.66610032e-01
1.67226389e-01 -9.26510990e-02 3.88786435e-01 -1.02073796e-01
-1.07384908e+00 5.38492203e-01 -7.63064399e-02 3.66184749e-02
4.41705704e-01 8.35454822e-01 -8.14387500e-01 -7.20976651e-01
2.59857148e-01 2.21741155e-01 -1.18872724e-01 -2.43435368e-01
1.71934083e-01 9.92260814e-01 3.30921076e-02 -1.37180820e-01
1.21190703e+00 6.20938480e-01 5.03947496e-01 -1.56351471e+00
-7.19137967e-01 -7.87041605e-01 -9.38647687e-02 -3.35703135e-01
3.76359731e-01 -1.47898149e+00 -4.77676392e-01 6.72145724e-01
-8.92464638e-01 -2.24530876e-01 2.35254526e-01 1.10015178e+00
-2.22419634e-01 4.17218238e-01 -2.66275525e-01 -1.02648270e+00
-2.33885333e-01 -9.20642316e-01 8.67091358e-01 1.73695385e-01
1.51721567e-01 -7.12707162e-01 -1.26631051e-01 2.34651580e-01
5.88478386e-01 1.56084463e-01 7.62108743e-01 -2.38194782e-02
-1.02510631e+00 -3.63308609e-01 -1.16133124e-01 1.75655670e-02
1.25752568e-01 -6.77573562e-01 -6.13393843e-01 -7.22534716e-01
2.54579484e-01 4.07667369e-01 5.21506786e-01 5.96071720e-01
6.68444037e-01 -2.58662939e-01 -6.86111808e-01 7.20704496e-01
1.73262632e+00 6.06273338e-02 9.86530662e-01 -1.38672560e-01
7.61012435e-01 3.72890413e-01 5.89550734e-01 8.32022846e-01
2.48078987e-01 5.33145905e-01 3.36927682e-01 -2.75228560e-01
4.85439114e-02 1.19710781e-01 4.62898612e-01 1.22510648e+00
-1.55194560e-02 -4.55245703e-01 -2.66846120e-01 1.27133399e-01
-1.70756423e+00 -5.92914164e-01 -6.11679971e-01 2.56449366e+00
-1.03059215e-02 -3.21396023e-01 -3.86779785e-01 2.22598001e-01
5.45387447e-01 -8.50217268e-02 -2.63980955e-01 3.40612054e-01
-1.20427266e-01 1.39719054e-01 9.69827414e-01 4.55761582e-01
-6.48723960e-01 1.81218520e-01 5.17715502e+00 4.84368742e-01
-1.08639383e+00 8.46782816e-04 -1.01967856e-01 1.83091089e-01
-6.12791240e-01 1.94456294e-01 -7.63669729e-01 1.48315474e-01
6.32418931e-01 5.36596105e-02 4.00372177e-01 4.02245104e-01
3.97953615e-02 -4.43910480e-01 -8.45012307e-01 1.12064695e+00
1.44411370e-01 -9.02970910e-01 -1.71750292e-01 4.80829418e-01
8.68819058e-01 -3.18039805e-01 -1.08185776e-01 -6.28333986e-02
-1.62406325e-01 -4.69205707e-01 4.99649137e-01 6.39902174e-01
7.96989202e-01 -5.54713249e-01 5.69265783e-01 4.22260344e-01
-1.34498298e+00 -4.83477145e-01 -1.36605173e-01 -1.56388327e-01
3.42551887e-01 1.57820058e+00 -4.92343068e-01 1.14556682e+00
9.74346995e-02 6.05855107e-01 2.28336111e-01 1.21668065e+00
-3.32681835e-01 5.76869667e-01 -3.87839466e-01 5.88024817e-02
-1.34761348e-01 -5.70103467e-01 7.91799486e-01 7.41213679e-01
1.25295115e+00 5.49819291e-01 1.86987028e-01 2.35798523e-01
8.17215666e-02 2.37016842e-01 -7.06332803e-01 5.85072264e-02
5.67031860e-01 1.27393115e+00 -2.47600883e-01 -1.30025819e-01
-8.00465107e-01 6.84584677e-01 -4.20986563e-01 7.64577150e-01
-3.55222464e-01 -4.83929843e-01 5.88466287e-01 8.46070424e-02
2.87348419e-01 -8.34128261e-01 -3.10980558e-01 -1.11859822e+00
6.30376711e-02 -7.30441511e-01 -7.90823326e-02 -3.28859508e-01
-6.77108228e-01 1.97747171e-01 -6.17458582e-01 -1.70048916e+00
2.09844951e-02 -5.81850000e-02 -3.55391949e-01 8.96456838e-01
-1.71708870e+00 -1.20758820e+00 -5.79901516e-01 7.13335276e-01
-4.00409818e-01 -1.86085463e-01 7.76423335e-01 6.18038893e-01
-4.98766124e-01 2.83190280e-01 4.58072931e-01 -1.97272878e-02
3.40245306e-01 -3.07021439e-01 -5.43859780e-01 9.65392649e-01
-3.25808108e-01 7.26228416e-01 5.26159942e-01 -7.18334556e-01
-2.19857216e+00 -1.05111825e+00 2.33750984e-01 6.69826150e-01
3.89346629e-01 -3.87951314e-01 -2.49980032e-01 7.00203001e-01
3.81094497e-03 -4.19348955e-01 1.28347278e+00 -5.65296598e-02
4.09849919e-02 -1.30677044e-01 -7.69092917e-01 3.88242900e-01
8.69515836e-01 -2.06789255e-01 2.37713903e-01 3.75785261e-01
6.84655830e-02 -4.30390894e-01 -6.55033469e-01 3.63153100e-01
8.37266982e-01 -4.55544800e-01 9.49473739e-01 4.77193236e-01
1.68542087e-01 -5.49473882e-01 -5.53379953e-01 -1.39492142e+00
-4.91760910e-01 -6.17413580e-01 -2.16709629e-01 1.04495084e+00
3.60617369e-01 -9.86578107e-01 6.09232664e-01 -8.19791630e-02
-2.42620096e-01 -2.32745796e-01 -8.39937687e-01 -5.74407279e-01
-8.08958173e-01 -3.99882123e-02 3.54048252e-01 5.61813235e-01
6.13461360e-02 6.07979476e-01 -8.85190010e-01 1.02094936e+00
9.86439645e-01 1.70314923e-01 1.03606212e+00 -1.36938131e+00
-8.80322933e-01 2.70119578e-01 -2.75572300e-01 -1.53656185e+00
-2.42005318e-01 -6.81561470e-01 1.34186253e-01 -1.65455437e+00
5.12920544e-02 -1.04440331e+00 2.79475730e-02 7.41539747e-02
2.34682024e-01 1.64105147e-01 -1.44318402e-01 3.07324082e-01
-1.88951775e-01 6.00001872e-01 1.30419910e+00 2.00655125e-02
-5.29778779e-01 4.10367578e-01 -7.87233472e-01 3.07256550e-01
3.32178622e-01 -8.32323804e-02 -3.46198320e-01 -6.30328000e-01
4.75369096e-01 6.64128840e-01 2.83801351e-02 -1.42981744e+00
4.76567000e-01 -4.08251956e-02 1.18707590e-01 -3.93271983e-01
6.30189896e-01 -1.27326035e+00 8.00850391e-01 6.95648313e-01
4.61400747e-01 -8.98961604e-01 -1.25553593e-01 8.80035281e-01
-1.30503923e-01 9.17686075e-02 5.70686698e-01 1.32026777e-01
-2.41972238e-01 2.68198550e-01 -6.24451518e-01 -8.13298762e-01
1.22186601e+00 -6.58161491e-02 -1.20947227e-01 -6.83562636e-01
-5.21994054e-01 2.24774301e-01 1.51872948e-01 -2.55576998e-01
5.93002319e-01 -1.03412867e+00 -6.20152891e-01 6.21547341e-01
3.06140408e-02 -1.58127069e-01 5.89872539e-01 1.24531698e+00
-1.20850675e-01 4.49226528e-01 -4.13795114e-02 -4.34649169e-01
-1.22567570e+00 -1.61287040e-01 5.03859036e-02 -2.07213555e-02
-3.45127016e-01 6.50401592e-01 2.93064892e-01 1.35201961e-01
-1.01087503e-01 1.67110682e-01 -3.67338777e-01 -2.09523831e-02
5.60853541e-01 4.10691023e-01 1.64913222e-01 -6.83522999e-01
-1.02514870e-01 9.36875403e-01 4.69095975e-01 -1.96159363e-01
1.43727612e+00 -4.11704212e-01 -3.19673777e-01 -1.20128589e-02
1.13378704e+00 8.61319602e-01 -4.65565205e-01 -3.93113971e-01
-6.55712783e-01 -5.87078631e-01 1.80872738e-01 -3.20271283e-01
-1.27413392e+00 4.68100011e-01 3.72860968e-01 -6.94697723e-02
1.26924765e+00 -1.05710097e-01 7.47530818e-01 5.88081717e-01
9.61151719e-01 -6.39794469e-01 4.53484803e-02 4.16141778e-01
6.39501274e-01 -3.63818765e-01 1.56849414e-01 -1.17598784e+00
-1.05071086e-02 1.13323116e+00 8.97900015e-02 6.63363859e-02
9.07570064e-01 3.82355243e-01 -6.50736541e-02 -3.34243536e-01
-1.83505073e-01 -2.30399266e-01 -6.81301057e-02 3.84414405e-01
2.94595748e-01 2.57843822e-01 -3.84131968e-01 6.81358099e-01
4.54470627e-02 -1.20026790e-01 9.13662612e-01 9.30054724e-01
-6.30944550e-01 -1.39267564e+00 -6.81707501e-01 6.03922904e-01
5.41604683e-02 1.61875784e-01 5.30554876e-02 1.23891510e-01
-2.43190840e-01 1.36924851e+00 -5.14008701e-01 -4.73434240e-01
5.59522927e-01 -4.64212924e-01 4.65996802e-01 -7.11254895e-01
3.33178580e-01 5.36689222e-01 2.17932731e-01 -5.04768908e-01
-2.42908373e-01 -6.24677122e-01 -1.17427599e+00 1.61095306e-01
-6.71139419e-01 4.45910007e-01 9.18015063e-01 1.05265570e+00
4.93318588e-01 7.68593013e-01 1.25031161e+00 -6.86712146e-01
-1.57345787e-01 -4.87329304e-01 -1.27274907e+00 -2.10090697e-01
2.46191606e-01 -6.82365060e-01 -3.81284922e-01 -4.16676044e-01] | [6.318902492523193, 1.2848496437072754] |
f1e6b9fc-beb8-45dd-8b29-30d7105c00ba | green-runner-a-tool-for-efficient-model | 2305.16849 | null | https://arxiv.org/abs/2305.16849v1 | https://arxiv.org/pdf/2305.16849v1.pdf | Green Runner: A tool for efficient model selection from model repositories | Deep learning models have become essential in software engineering, enabling intelligent features like image captioning and document generation. However, their popularity raises concerns about environmental impact and inefficient model selection. This paper introduces GreenRunnerGPT, a novel tool for efficiently selecting deep learning models based on specific use cases. It employs a large language model to suggest weights for quality indicators, optimizing resource utilization. The tool utilizes a multi-armed bandit framework to evaluate models against target datasets, considering tradeoffs. We demonstrate that GreenRunnerGPT is able to identify a model suited to a target use case without wasteful computations that would occur under a brute-force approach to model selection. | ['Luis Cruz', 'Taylan Selvi', 'Anj Simmons', 'Scott Barnett', 'Jai Kannan'] | 2023-05-26 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [ 1.17597673e-02 -2.27585132e-03 -7.43088186e-01 -2.12256551e-01
-1.19443238e+00 -5.30871272e-01 4.37583208e-01 -1.67670935e-01
-3.61955464e-01 6.33230805e-01 -9.61097181e-02 -7.51435637e-01
-5.25615454e-01 -7.32773185e-01 -7.62897789e-01 -3.97742689e-01
1.39372736e-01 7.04600692e-01 -5.19573689e-01 2.19173685e-01
5.06244302e-01 5.61395884e-01 -1.46749616e+00 4.59148496e-01
6.81302845e-01 1.00696707e+00 4.34328258e-01 9.05421972e-01
-4.74435896e-01 8.90816391e-01 -7.41691947e-01 -4.86444443e-01
3.75604868e-01 5.96653484e-02 -8.57474923e-01 -5.48275635e-02
2.45081395e-01 -3.59935701e-01 6.31542727e-02 7.99761593e-01
4.83209044e-01 -8.67469758e-02 6.97461784e-01 -1.50171447e+00
-6.28370702e-01 8.51447046e-01 -6.88325644e-01 1.59841925e-01
5.90428070e-04 4.52079594e-01 1.36436987e+00 -6.16966367e-01
3.25920403e-01 1.39274788e+00 3.65572482e-01 4.52053726e-01
-1.46666110e+00 -7.04215884e-01 3.63957822e-01 2.74780989e-01
-1.14562714e+00 -4.18522090e-01 7.27481663e-01 -4.57292259e-01
1.48212397e+00 5.35120964e-01 9.13313925e-01 9.56787229e-01
3.69324982e-01 1.13857651e+00 1.01493299e+00 -8.33514392e-01
5.79077780e-01 1.37056515e-01 7.07575306e-02 7.70047963e-01
2.79712021e-01 2.14143664e-01 -7.53335834e-01 -6.68083429e-01
4.16068524e-01 -1.53823018e-01 2.11257800e-01 -2.71507084e-01
-8.40353906e-01 9.20567691e-01 8.74138251e-02 1.49251685e-01
-7.63178945e-01 8.88323903e-01 2.47451067e-01 2.05946818e-01
7.12791741e-01 9.86023843e-01 -9.16565597e-01 -4.50617760e-01
-1.23549926e+00 4.83022720e-01 7.27407038e-01 1.11974871e+00
7.53500164e-01 2.06239998e-01 -4.48629171e-01 9.23180163e-01
5.72169781e-01 5.06216288e-01 4.45381552e-01 -1.09267318e+00
4.59142864e-01 4.06023443e-01 3.01897407e-01 -6.33037269e-01
-2.18438849e-01 -6.22138619e-01 -1.82827771e-01 1.19017474e-01
-2.84314156e-03 -3.95980060e-01 -9.57242668e-01 1.47214735e+00
-1.59917697e-01 -9.24376473e-02 -1.45602494e-01 4.26508427e-01
1.52873039e-01 8.25602591e-01 2.57808864e-01 -1.89858720e-01
9.12890792e-01 -9.14976180e-01 -4.89720374e-01 -7.22107291e-01
8.28073382e-01 -6.21847868e-01 1.23025441e+00 8.38744521e-01
-1.13511086e+00 -1.27893567e-01 -8.54294717e-01 4.35641974e-01
-2.56625026e-01 1.55446261e-01 1.09988630e+00 1.10869944e+00
-1.13962173e+00 3.76272053e-01 -8.13672185e-01 2.34026425e-02
6.49086475e-01 3.78440350e-01 3.53809267e-01 -2.07298413e-01
-7.83691287e-01 1.19322765e+00 3.44563425e-01 -1.05650097e-01
-1.25395703e+00 -7.77074218e-01 -5.83909512e-01 3.33852947e-01
4.09150749e-01 -7.05103993e-01 1.55150044e+00 -9.01541770e-01
-1.44910383e+00 5.09511888e-01 9.74085778e-02 -6.43104792e-01
4.34464633e-01 -3.94430697e-01 -1.57696083e-01 -4.87314403e-01
-1.98018357e-01 4.82916445e-01 8.04240167e-01 -1.15868115e+00
-5.94411790e-01 -7.12921321e-02 2.03627676e-01 -3.00286748e-02
-6.92455769e-01 3.28702599e-01 -5.88310599e-01 -4.53211427e-01
-4.97374058e-01 -1.01192975e+00 -7.98599362e-01 -3.49963188e-01
-5.06002426e-01 -5.23802638e-02 4.16042149e-01 -3.66271615e-01
1.71813548e+00 -1.42852366e+00 -8.69760141e-02 4.41600800e-01
-1.22398380e-02 1.10682556e-02 -5.13508201e-01 2.53029913e-01
9.05967690e-03 6.80856347e-01 1.53860286e-01 4.02955934e-02
3.97667289e-01 -4.15881276e-02 -5.29848412e-02 2.67254174e-01
3.52333009e-01 8.43227029e-01 -6.16602421e-01 -6.93547308e-01
2.23545939e-01 1.11649558e-01 -8.08740497e-01 2.72894263e-01
-6.85075164e-01 -6.02901578e-01 -5.16217411e-01 9.58329797e-01
2.99658835e-01 -4.20065165e-01 2.21080586e-01 1.76418349e-02
5.58795445e-02 4.79612984e-02 -8.96050215e-01 1.32734299e+00
-1.21030879e+00 6.96478248e-01 -2.06663460e-01 -9.03408051e-01
9.13065076e-01 -1.33382604e-01 5.14314115e-01 -6.45425797e-01
1.41268954e-01 9.76156816e-02 -4.34802920e-02 -6.69500947e-01
2.77803719e-01 9.73753855e-02 -1.70122564e-01 8.37666810e-01
-9.64740962e-02 -3.31411719e-01 1.74539313e-01 -1.40725866e-01
1.23609352e+00 2.84429461e-01 2.87008524e-01 -4.76039052e-01
-8.86401162e-02 1.08254269e-01 3.00594568e-01 1.05384076e+00
2.39270687e-01 2.88460970e-01 4.67496455e-01 -8.17344546e-01
-9.39500034e-01 -4.95356232e-01 3.50891501e-01 1.44312894e+00
-4.60410893e-01 -2.66885579e-01 -9.15407717e-01 -8.52932334e-01
-1.03849946e-02 1.19535077e+00 -6.06744289e-01 -3.64706606e-01
-2.95479804e-01 -1.05075514e+00 1.97340399e-01 3.72649848e-01
-1.44518241e-01 -1.09493029e+00 -1.14793575e+00 3.39726776e-01
1.22882791e-01 -5.34338295e-01 -3.18844825e-01 5.46544969e-01
-6.90652549e-01 -8.44897270e-01 -8.28141451e-01 -1.85160846e-01
5.15328586e-01 -7.57002905e-02 1.81393003e+00 1.92947298e-01
-4.31795627e-01 4.17073965e-01 -5.12024090e-02 -9.92065191e-01
-3.79375398e-01 2.91461766e-01 -1.95693374e-01 -4.11734879e-01
6.46197796e-01 2.45355181e-02 -5.17595708e-01 -2.19941582e-03
-8.58906865e-01 1.89095736e-01 8.75119388e-01 1.03026593e+00
4.63818699e-01 -5.40903397e-02 4.28048134e-01 -9.46484566e-01
1.25827932e+00 -4.99899864e-01 -1.01525521e+00 8.55125248e-01
-1.17207229e+00 4.03285325e-01 2.51698494e-01 -7.60628700e-01
-9.26596701e-01 8.08298662e-02 1.94972634e-01 -4.14597780e-01
2.75891036e-01 9.77987528e-01 -1.35069087e-01 -1.18176006e-01
8.81173372e-01 -3.06587778e-02 -3.85255098e-01 -3.33337814e-01
3.83086056e-01 7.64336765e-01 -2.77446836e-01 -1.01946425e+00
5.35963237e-01 -2.08685011e-01 -3.62158388e-01 -3.80824000e-01
-6.82561636e-01 2.25590356e-02 -3.83480825e-02 -4.55535680e-01
2.19257638e-01 -6.02314293e-01 -4.48100477e-01 2.48246957e-02
-1.25522757e+00 -4.96394843e-01 -2.55279094e-01 2.46786609e-01
-8.76863658e-01 -2.45900035e-01 -1.17072769e-01 -1.16210580e+00
-5.71940243e-01 -1.48776615e+00 8.24875534e-01 -1.10032056e-02
-4.76650685e-01 -7.27471292e-01 -1.17319465e-01 4.50611621e-01
6.19648755e-01 -1.05476398e-02 1.21153116e+00 -6.66859925e-01
-7.01186478e-01 -4.85159606e-01 -8.12345296e-02 2.93296486e-01
-2.26716131e-01 3.45682293e-01 -9.08108532e-01 -2.03859836e-01
-4.80482012e-01 -4.40171182e-01 3.74959826e-01 7.49103427e-01
1.70489430e+00 -6.63482010e-01 -2.21391618e-01 3.75923365e-01
1.51366949e+00 6.64472282e-01 3.95318568e-01 8.34286988e-01
3.94256294e-01 5.61549485e-01 8.59740198e-01 7.57085264e-01
-4.15869951e-02 4.63077754e-01 6.04733348e-01 -1.43537194e-01
3.47608864e-01 9.32379067e-02 8.93296227e-02 2.22991630e-01
1.51902676e-01 -6.01620317e-01 -1.59732008e+00 8.93252015e-01
-1.77364457e+00 -3.47922206e-01 5.56352437e-01 1.72304678e+00
7.63223290e-01 4.97891128e-01 9.26543474e-02 -1.18556492e-01
4.79045808e-01 9.98431351e-03 -8.80597591e-01 -8.81379366e-01
1.48264930e-01 3.25104564e-01 9.70003426e-01 1.80997044e-01
-5.76776922e-01 9.38514948e-01 7.73201132e+00 1.04952776e+00
-9.15658951e-01 7.50201046e-02 1.20882666e+00 -6.02931738e-01
-7.31193125e-01 -1.96675315e-01 -6.69534981e-01 3.29629093e-01
1.26082623e+00 -8.35915029e-01 5.30458629e-01 1.29358554e+00
4.58489776e-01 5.11143245e-02 -1.17877972e+00 9.79971945e-01
-1.18159913e-01 -1.89302576e+00 1.67044923e-01 1.69986740e-01
8.84683371e-01 9.73839313e-02 3.60626549e-01 3.46662015e-01
9.27501380e-01 -1.14887559e+00 8.97854030e-01 5.63895047e-01
5.38424313e-01 -1.28538823e+00 6.08212709e-01 3.21287632e-01
-4.17509407e-01 -5.38196087e-01 -2.57084668e-01 9.72310007e-02
-3.04305047e-01 5.33097386e-01 -1.21999741e+00 -1.64887652e-01
6.57557368e-01 -5.36524616e-02 -3.78728390e-01 1.16247976e+00
3.93691450e-01 7.54850745e-01 -1.98495194e-01 -6.31595135e-01
4.82285768e-01 2.32687056e-01 1.69971570e-01 1.33751965e+00
7.16757834e-01 -5.26316524e-01 1.14299454e-01 9.60772634e-01
-7.74460509e-02 2.90797293e-01 -5.93241513e-01 -3.08679730e-01
6.60585999e-01 8.17370236e-01 -5.37970245e-01 -3.20420086e-01
-2.68001437e-01 2.66330004e-01 1.69177670e-02 4.41606671e-01
-7.63593137e-01 -1.36132121e-01 7.09816158e-01 8.90099332e-02
9.79483053e-02 9.66507420e-02 -6.03580654e-01 -5.60647249e-01
-2.52967417e-01 -1.37436020e+00 2.48246759e-01 -9.04250741e-01
-8.80988419e-01 6.31779194e-01 3.82383853e-01 -9.92364168e-01
-4.86686319e-01 -5.91034174e-01 -3.46211582e-01 1.00737238e+00
-1.37230587e+00 -1.20093548e+00 1.91511825e-01 2.74416991e-02
7.97645211e-01 -5.61335146e-01 6.82307005e-01 -1.30363658e-01
-7.69856513e-01 6.68231905e-01 2.58510381e-01 -5.10664105e-01
2.74191916e-01 -1.10649443e+00 5.53516328e-01 7.15545833e-01
2.41554320e-01 6.14503086e-01 1.01255095e+00 -3.55549484e-01
-1.43993902e+00 -9.76273894e-01 5.20597577e-01 -1.36369795e-01
6.36676848e-01 -1.44124299e-01 -3.11712176e-01 4.53125834e-01
3.17846298e-01 -2.72257656e-01 8.35513711e-01 4.60797876e-01
-2.98505664e-01 -1.43418133e-01 -1.30501330e+00 7.00458050e-01
5.75841963e-01 -4.18414176e-01 1.25626504e-01 5.68715274e-01
7.54507661e-01 -2.02202946e-01 -9.16428268e-01 1.75288945e-01
7.23531485e-01 -6.14667833e-01 8.83104503e-01 -1.06924570e+00
5.43356657e-01 1.82745874e-01 -1.98916584e-01 -1.44191980e+00
-4.65100884e-01 -8.83007109e-01 -5.37242711e-01 9.67079163e-01
1.00290298e+00 -1.37630746e-01 1.17159665e+00 1.31843495e+00
1.44714549e-01 -1.02982426e+00 -8.37906718e-01 -7.53041506e-01
2.68008299e-02 -8.83317828e-01 1.17332530e+00 7.03210294e-01
-3.97330314e-01 -6.99878186e-02 -6.06841087e-01 -1.43894970e-01
5.74276567e-01 1.67666346e-01 6.11180484e-01 -1.02141726e+00
-4.73097771e-01 -7.27730989e-01 -3.12266108e-02 -6.57973230e-01
2.61801332e-01 -3.85967284e-01 8.54315236e-02 -1.47471535e+00
2.35760063e-01 -4.97844875e-01 -8.28979075e-01 6.93759203e-01
5.98899387e-02 -1.45989150e-01 3.40953320e-01 -1.33658648e-01
-6.60934210e-01 1.30892530e-01 7.46663749e-01 -6.58981740e-01
-1.62447110e-01 2.43916601e-01 -1.18618536e+00 7.53502488e-01
9.99239504e-01 -6.72016442e-01 -3.78512025e-01 -7.08966553e-01
6.65079236e-01 1.48778707e-01 -1.43340575e-02 -8.79436135e-01
4.16451395e-02 -7.94026613e-01 2.27621362e-01 -4.77813780e-01
1.36728808e-01 -9.25934196e-01 4.78432924e-01 5.72899818e-01
-1.03749192e+00 3.02563757e-01 5.32030523e-01 5.13552785e-01
1.98922917e-01 -7.61742651e-01 6.71525717e-01 -3.83550346e-01
-8.06954563e-01 2.11342081e-01 -6.96404755e-01 -3.10349584e-01
1.02561378e+00 -8.02395195e-02 -2.55700890e-02 -3.24435145e-01
-4.46015626e-01 2.43986487e-01 2.65489012e-01 5.34310400e-01
6.77328348e-01 -1.21970415e+00 -7.44520605e-01 -1.61835134e-01
3.19452047e-01 -3.03131521e-01 1.60360873e-01 1.99108720e-02
-7.07112193e-01 4.99127418e-01 -9.94033813e-02 -3.59039873e-01
-1.42184055e+00 4.84890699e-01 4.95026618e-01 -6.87350869e-01
5.73266447e-02 1.03242552e+00 -1.27473757e-01 -1.86758950e-01
3.69390041e-01 -1.77758843e-01 6.31781342e-03 2.02723648e-02
4.46716487e-01 4.80643779e-01 3.48179102e-01 1.80380076e-01
-1.50903568e-01 -8.57819840e-02 -2.89132506e-01 -1.02934301e-01
1.64472926e+00 9.27694440e-02 5.08644804e-02 2.09122568e-01
9.22240615e-01 -4.45494652e-01 -1.17779100e+00 -2.07546026e-01
4.92981672e-01 -6.57844663e-01 8.31016839e-01 -1.23466623e+00
-1.20655286e+00 7.64954746e-01 8.79822671e-01 3.74277771e-01
1.24952364e+00 -2.75160939e-01 4.67606694e-01 7.64212787e-01
4.89445895e-01 -1.40746725e+00 6.69758543e-02 1.30734444e-01
9.88645494e-01 -1.05413580e+00 8.13694969e-02 3.23800743e-01
-5.94363987e-01 1.00951135e+00 8.29230964e-01 3.07112634e-01
6.13394439e-01 6.53927565e-01 6.32689670e-02 -1.57345101e-01
-1.33614540e+00 1.26041517e-01 1.30687803e-01 7.61932015e-01
3.99943978e-01 2.76706457e-01 9.81657878e-02 3.83729905e-01
8.63716826e-02 6.42268434e-02 4.01029587e-01 1.02541971e+00
-3.26181144e-01 -1.16858292e+00 -5.98390043e-01 9.88428414e-01
-7.01346993e-01 -4.32330310e-01 -2.83618838e-01 5.53999841e-01
-1.73725244e-02 7.91347384e-01 -1.91613302e-01 -5.46430886e-01
2.63877600e-01 -1.34920284e-01 3.27050805e-01 -6.30651832e-01
-6.85973585e-01 3.08320642e-01 4.73020583e-01 -4.98966217e-01
-1.27737239e-01 -5.63618064e-01 -3.96119386e-01 -9.89878401e-02
-6.34729266e-01 1.94862485e-01 9.95459735e-01 7.28893101e-01
4.43251729e-01 7.18624711e-01 5.80710173e-01 -8.36502492e-01
-7.11935043e-01 -8.23217213e-01 -3.37856472e-01 -2.67741114e-01
9.37987864e-02 -5.81474781e-01 5.75716719e-02 -9.68795270e-02] | [8.339669227600098, 3.721388578414917] |
ddf0b2e3-28ff-4ac1-9338-ac6479b175f0 | event-causality-extraction-with-event-1 | 2301.11621 | null | https://arxiv.org/abs/2301.11621v1 | https://arxiv.org/pdf/2301.11621v1.pdf | Event Causality Extraction with Event Argument Correlations | Event Causality Identification (ECI), which aims to detect whether a causality relation exists between two given textual events, is an important task for event causality understanding. However, the ECI task ignores crucial event structure and cause-effect causality component information, making it struggle for downstream applications. In this paper, we explore a novel task, namely Event Causality Extraction (ECE), aiming to extract the cause-effect event causality pairs with their structured event information from plain texts. The ECE task is more challenging since each event can contain multiple event arguments, posing fine-grained correlations between events to decide the causeeffect event pair. Hence, we propose a method with a dual grid tagging scheme to capture the intra- and inter-event argument correlations for ECE. Further, we devise a event type-enhanced model architecture to realize the dual grid tagging scheme. Experiments demonstrate the effectiveness of our method, and extensive analyses point out several future directions for ECE. | ['Jinqiao Shi', 'Tingwen Liu', 'Quangang Li', 'Xin Cong', 'Jiawei Sheng', 'Shiyao Cui'] | 2023-01-27 | event-causality-extraction-with-event | https://aclanthology.org/2022.coling-1.201 | https://aclanthology.org/2022.coling-1.201.pdf | coling-2022-10 | ['event-causality-identification'] | ['natural-language-processing'] | [ 2.55650461e-01 -1.01645283e-01 -1.77329347e-01 -3.37880105e-01
-5.77633262e-01 -7.82970726e-01 8.75079393e-01 6.25146568e-01
-2.62436599e-01 7.20859587e-01 7.03551769e-01 -6.08436584e-01
-3.44550729e-01 -9.25307930e-01 -5.24183035e-01 -3.58304381e-01
-3.97681147e-01 7.01098219e-02 6.11599803e-01 2.16737032e-01
1.59001827e-01 3.68496895e-01 -1.20811391e+00 5.43599606e-01
5.71090817e-01 6.70138299e-01 1.56292215e-01 5.07535875e-01
-3.81547958e-01 1.36135876e+00 -7.55487204e-01 -2.92647213e-01
-4.51125085e-01 -7.32277632e-01 -1.04199862e+00 -6.68344975e-01
-4.30604637e-01 -1.28052965e-01 -1.42917514e-01 8.13660204e-01
2.78979719e-01 -1.58988312e-02 7.88730323e-01 -1.54948866e+00
8.98246169e-02 9.60465074e-01 -4.66356844e-01 7.49024391e-01
3.95451337e-01 -2.88069993e-01 1.38068366e+00 -8.81352723e-01
6.04081869e-01 1.21271312e+00 3.86292428e-01 1.50945589e-01
-5.52845418e-01 -8.95393014e-01 3.66888642e-01 5.39605319e-01
-1.15744603e+00 -7.94383958e-02 6.86976671e-01 -2.95507669e-01
9.81891811e-01 4.25728470e-01 3.92583787e-01 1.04870939e+00
3.48973602e-01 7.76920795e-01 6.04718566e-01 -3.29704821e-01
2.06540212e-01 -5.36737859e-01 1.59656346e-01 2.67822236e-01
9.98372138e-02 4.73652259e-02 -6.88879848e-01 -4.06260788e-01
5.97167253e-01 -2.40933336e-02 -1.60086602e-01 4.83497590e-01
-1.46470809e+00 7.01813340e-01 2.00354099e-01 6.37317955e-01
-4.94977146e-01 4.96485047e-02 7.65746593e-01 1.04635144e-02
2.89992899e-01 1.44133002e-01 -6.88309252e-01 -1.36224344e-01
-4.59297329e-01 3.64118934e-01 8.45806956e-01 6.61562145e-01
2.79488266e-01 -5.90456784e-01 -5.86823285e-01 5.96696317e-01
3.37981164e-01 -2.23828908e-02 2.21414313e-01 -2.75500238e-01
5.55245817e-01 7.06551731e-01 2.05003664e-01 -1.23187876e+00
-7.66886830e-01 -1.55519366e-01 -7.29670882e-01 -3.56583416e-01
4.26828623e-01 -3.26299816e-01 -2.92818636e-01 1.85956824e+00
6.90231025e-01 7.74251997e-01 1.23565868e-02 8.64371717e-01
9.33357477e-01 9.24230695e-01 6.94598854e-01 -5.88493526e-01
1.85002518e+00 -3.44063491e-01 -1.04377711e+00 -1.24197297e-01
5.60630023e-01 -7.79594898e-01 6.31302714e-01 3.17152217e-02
-7.15028763e-01 -9.08400416e-02 -6.56207919e-01 4.53528389e-02
-1.79078385e-01 2.95320321e-02 9.32718933e-01 3.19467448e-02
-1.90890923e-01 2.50212967e-01 -8.07220161e-01 -1.13495298e-01
1.08172044e-01 1.15029529e-01 -2.52080828e-01 3.06617230e-01
-1.95529735e+00 5.86311877e-01 1.00505090e+00 2.77234405e-01
-6.38089597e-01 -8.63436282e-01 -8.16724539e-01 3.76220226e-01
7.56530285e-01 -2.72524506e-01 1.36019862e+00 -3.39136302e-01
-8.73578906e-01 4.33727771e-01 -4.99347270e-01 -3.28440160e-01
1.23814180e-01 -7.08772019e-02 -1.03586853e+00 2.68879570e-02
2.60469407e-01 5.83842807e-02 3.46481800e-01 -7.80118227e-01
-1.25083828e+00 -1.89932778e-01 -1.42695727e-02 9.41977128e-02
-3.05850625e-01 7.97655046e-01 -5.79937160e-01 -1.07581556e+00
1.02957210e-03 -4.25996304e-01 -1.29305854e-01 -5.84696472e-01
-4.07913208e-01 -9.22262311e-01 7.59803951e-01 -5.41966975e-01
1.95299304e+00 -2.15941834e+00 -3.71436059e-01 7.28729963e-02
3.22307497e-01 -1.03901654e-01 2.65849441e-01 6.34877205e-01
-4.69062269e-01 4.14235480e-02 -1.27206683e-01 2.95074463e-01
2.32679918e-02 3.00364643e-01 -7.30328023e-01 -5.09137735e-02
7.12399483e-01 8.28405380e-01 -1.21167481e+00 -6.80547178e-01
-1.78569809e-01 1.79632276e-01 -2.49385685e-01 5.02343357e-01
-2.80941904e-01 5.34216583e-01 -7.50882149e-01 2.63090283e-01
3.70058119e-01 -3.93980861e-01 5.41697681e-01 -3.42556566e-01
-3.68524402e-01 8.59401226e-01 -1.39577913e+00 1.28194547e+00
-2.66736567e-01 2.91872352e-01 -2.40826190e-01 -9.31388319e-01
7.12913454e-01 8.16027224e-01 5.70328414e-01 -6.70692921e-01
-5.27671678e-03 2.40821689e-01 6.41316995e-02 -6.15097404e-01
2.63833046e-01 -3.03443789e-01 -5.43649673e-01 4.37297910e-01
-1.80072203e-01 7.23510027e-01 5.21600425e-01 3.75126272e-01
1.57822740e+00 -5.84822744e-02 7.26633549e-01 -1.43219084e-01
5.20181656e-01 -4.88703896e-04 1.20019698e+00 5.22610664e-01
-9.97021049e-02 4.26643461e-01 1.03554416e+00 -4.82448161e-01
-4.20010686e-01 -8.05085480e-01 -1.01910383e-01 7.77721286e-01
2.55837828e-01 -8.18055749e-01 -4.37105477e-01 -1.04502678e+00
-3.49857837e-01 7.20933259e-01 -5.74243665e-01 3.03412769e-02
-1.01455605e+00 -1.30391288e+00 5.80723166e-01 8.48473907e-01
3.79184067e-01 -1.11113727e+00 -5.56023002e-01 6.96486175e-01
-8.00936282e-01 -1.11406779e+00 -6.09115303e-01 2.53655404e-01
-3.93574715e-01 -1.34981418e+00 -1.09263651e-01 -6.35809898e-01
1.05511928e-02 -1.80082515e-01 1.24786127e+00 -2.68674999e-01
-1.54006043e-02 -3.63950163e-01 -5.74392319e-01 -5.47075450e-01
-3.17151010e-01 -1.10585183e-01 -2.71282762e-01 7.22280070e-02
5.70456564e-01 -5.77628732e-01 -6.73916876e-01 4.74600703e-01
-9.96142507e-01 9.63741615e-02 5.39856851e-01 6.03753507e-01
5.52218556e-01 7.97249198e-01 9.26127017e-01 -9.63871658e-01
4.29921299e-01 -7.81711459e-01 -5.00962973e-01 2.51805037e-01
-3.37557793e-01 1.57860927e-02 7.66457736e-01 -3.36145073e-01
-1.55746233e+00 -1.83689028e-01 -3.25932443e-01 2.22852945e-01
-3.45717221e-01 8.13422859e-01 -7.84375250e-01 8.46206129e-01
2.57369667e-01 -1.13402553e-01 -9.61631000e-01 -3.35244417e-01
5.87937534e-02 4.05078530e-01 7.46431351e-01 -5.73901832e-01
3.94456804e-01 4.24795061e-01 1.51073467e-02 -6.33942187e-02
-9.43152130e-01 -6.12909555e-01 -4.06815797e-01 3.33637111e-02
1.00744295e+00 -8.04382503e-01 -8.97138357e-01 1.20593764e-01
-1.60820341e+00 -6.77219182e-02 5.76285310e-02 7.19199419e-01
-5.14401905e-02 1.62699446e-01 -6.82901859e-01 -6.06053054e-01
-1.92462340e-01 -6.73941016e-01 9.20006037e-01 1.83684856e-01
-6.54902160e-01 -9.37899828e-01 2.21179023e-01 -2.17120185e-01
-3.27622861e-01 2.07734317e-01 1.11495018e+00 -9.48698878e-01
-1.81627154e-01 -7.99584389e-02 -5.07788897e-01 -6.28909826e-01
1.04069896e-01 -8.27297047e-02 -6.32800937e-01 3.46463203e-01
-1.21290110e-01 2.44562089e-01 6.41273320e-01 1.90955311e-01
1.13572228e+00 -4.51237530e-01 -6.89123452e-01 6.88100308e-02
1.03055322e+00 5.07982492e-01 6.01043761e-01 1.72305644e-01
7.84946799e-01 7.56459892e-01 8.68658364e-01 5.91325164e-01
6.73224628e-01 6.62451684e-01 1.87506899e-01 -1.79806024e-01
-1.99596167e-01 -5.28493464e-01 1.28600195e-01 5.65695107e-01
-1.32922279e-02 -6.29005432e-01 -8.99208128e-01 7.37008929e-01
-1.95990288e+00 -1.08548248e+00 -8.69961143e-01 1.85995579e+00
1.08225679e+00 2.35135853e-01 3.91760021e-02 4.66153413e-01
8.45321953e-01 5.93264028e-02 -1.20836750e-01 -8.50187391e-02
-4.89448979e-02 5.00900112e-02 1.92001954e-01 2.34821394e-01
-1.21782362e+00 7.26653457e-01 5.58632851e+00 9.58205163e-01
-8.69714320e-01 2.46450022e-01 4.77538705e-01 4.24486756e-01
-4.06105965e-01 2.78008342e-01 -9.29771960e-01 7.39397764e-01
9.59365666e-01 -8.35594758e-02 -2.83859760e-01 2.70131260e-01
5.85558474e-01 -1.28683716e-01 -1.07077086e+00 6.29602194e-01
-5.44908226e-01 -1.27096868e+00 -6.08338006e-02 -1.88937277e-01
2.81913251e-01 -4.16957110e-01 -6.84390426e-01 4.03590836e-02
3.72562975e-01 -5.92638195e-01 8.08165133e-01 1.75762564e-01
5.54137647e-01 -6.57254517e-01 6.17864609e-01 2.03344345e-01
-1.72915602e+00 -1.33164339e-02 1.91704273e-01 -1.68273091e-01
5.74660182e-01 1.09495139e+00 -8.43647301e-01 1.00471139e+00
7.67764449e-01 7.38490224e-01 -1.28246099e-01 8.98030579e-01
-7.21366107e-01 1.21651471e+00 -3.45268548e-01 -4.59415130e-02
-8.45346376e-02 3.24986935e-01 5.47844887e-01 1.70167220e+00
2.70717919e-01 6.34513140e-01 5.00317626e-02 7.86780715e-01
1.55363260e-02 3.30223851e-02 -1.09982371e-01 -5.40325902e-02
6.67375267e-01 1.22280586e+00 -1.29586267e+00 -4.15947855e-01
-4.87882286e-01 8.53809059e-01 2.22306296e-01 2.26940081e-01
-1.05817580e+00 -5.40126264e-01 3.14542294e-01 -1.71348706e-01
1.89072236e-01 2.63307486e-02 -3.93500596e-01 -1.24550831e+00
1.46662831e-01 -5.33187449e-01 1.12636030e+00 -6.15437448e-01
-1.34843957e+00 3.51142138e-01 1.45659849e-01 -1.07549798e+00
-2.93218136e-01 -3.02200347e-01 -1.09574676e+00 7.41979480e-01
-1.53167248e+00 -9.83303130e-01 2.69581676e-02 5.35916090e-01
4.39081669e-01 4.98389035e-01 5.68780780e-01 6.84534252e-01
-7.76354849e-01 3.21788043e-01 -6.21341586e-01 4.29088175e-01
7.59153426e-01 -1.22722161e+00 3.25004101e-01 1.15695941e+00
2.05812633e-01 5.26106894e-01 6.64652109e-01 -1.00375926e+00
-1.05527163e+00 -1.30590355e+00 1.84773135e+00 -3.53022844e-01
8.75897467e-01 -3.36646348e-01 -1.09011471e+00 6.45767510e-01
-2.22804900e-02 -5.61416009e-03 6.82263017e-01 3.68266255e-01
-4.23610806e-01 -8.56081117e-03 -6.25007153e-01 5.92205703e-01
1.21275473e+00 -2.94982016e-01 -8.16900313e-01 1.05459057e-01
8.33914340e-01 -2.26989210e-01 -8.96917164e-01 6.36592150e-01
4.10063237e-01 -6.74385667e-01 9.08066571e-01 -4.21278030e-01
6.48866832e-01 -7.35669792e-01 2.14295238e-01 -9.07714367e-01
-3.60029072e-01 -6.13533616e-01 -2.49039963e-01 1.88317072e+00
4.71232086e-01 -4.40802068e-01 2.43733525e-01 4.41485941e-01
-9.59936753e-02 -2.14495495e-01 -7.03837514e-01 -5.71154833e-01
-4.30775732e-01 -8.60476315e-01 9.59524810e-01 1.28404939e+00
3.88651788e-01 7.85668135e-01 -1.44312426e-01 6.72559321e-01
1.56000480e-01 5.14953554e-01 1.80421457e-01 -1.19448364e+00
-3.30145687e-01 -4.60549563e-01 3.40594128e-02 -5.86325347e-01
1.32992268e-01 -7.96180487e-01 1.36503085e-01 -1.52157283e+00
3.37245673e-01 -5.10957003e-01 -4.20542657e-01 7.33539760e-01
-9.27647650e-01 -1.23227634e-01 -2.71807671e-01 1.35973677e-01
-7.20147312e-01 3.14935923e-01 8.39236856e-01 1.65558472e-01
-1.94224939e-01 -3.21741700e-02 -3.99761975e-01 8.64707649e-01
6.14106059e-01 -7.37077951e-01 -4.62908596e-01 -2.40950763e-01
5.38999140e-01 4.12040859e-01 6.37367964e-01 -4.35643971e-01
3.64278972e-01 -4.98355091e-01 2.54118323e-01 -7.50975788e-01
-4.49468195e-01 -7.51798689e-01 4.13641334e-01 2.00834751e-01
-3.88207018e-01 9.31564495e-02 2.28850141e-01 7.03072608e-01
-5.72478771e-01 -1.54484063e-01 7.45448694e-02 1.58750743e-01
-9.58505511e-01 1.47018895e-01 -4.81029660e-01 1.69569686e-01
8.79673660e-01 5.23593724e-01 -1.37314916e-01 1.03844672e-01
-6.54688895e-01 1.73249289e-01 -2.86584765e-01 4.38983351e-01
3.71128827e-01 -1.42719400e+00 -7.88853824e-01 -1.40966102e-01
2.87409246e-01 4.42295671e-02 2.83784628e-01 8.56726229e-01
1.03222512e-01 3.60391945e-01 4.05913353e-01 -8.06915388e-02
-1.25133908e+00 6.00470483e-01 -1.68031603e-01 -9.11394656e-01
-6.69100046e-01 7.85534620e-01 7.20215440e-01 -1.18615277e-01
-8.09603781e-02 -3.92991036e-01 -6.26891315e-01 1.87387004e-01
9.09794748e-01 2.60610431e-01 1.77508593e-01 -2.57943779e-01
-5.10266364e-01 -5.97279985e-03 1.05176196e-01 -1.84321687e-01
1.11790109e+00 -9.03989822e-02 -4.93637562e-01 2.97342807e-01
9.45270896e-01 -4.76472490e-02 -1.02952921e+00 -1.12494543e-01
6.10235095e-01 -2.18858868e-01 1.81862197e-04 -9.72805083e-01
-7.08442628e-01 7.12510526e-01 -1.47088990e-01 4.33210880e-01
1.44963980e+00 3.28974068e-01 1.01483083e+00 -2.91746438e-01
4.23983373e-02 -6.49724960e-01 -3.36314976e-01 5.64075053e-01
8.42778981e-01 -9.63664293e-01 -3.99717778e-01 -9.13849473e-01
-3.68335217e-01 9.96204019e-01 4.34157938e-01 4.57096726e-01
5.71523964e-01 6.42113209e-01 -3.24651867e-01 -4.37124670e-01
-8.76844645e-01 -3.38180393e-01 3.41927499e-01 8.36589336e-02
6.42841995e-01 1.02053605e-01 -9.26567614e-01 1.13553059e+00
1.51269823e-01 1.99593917e-01 -6.44474626e-02 7.86265850e-01
6.21275939e-02 -1.32136583e+00 -4.88947868e-01 1.75854638e-01
-9.18942511e-01 -2.58504152e-01 -2.52605975e-01 5.44891953e-01
3.93224388e-01 1.21316743e+00 1.99168965e-01 -1.50776565e-01
4.89790648e-01 -4.22031321e-02 5.67995608e-02 -3.97398293e-01
-6.74180567e-01 3.67156088e-01 3.11844409e-01 -5.30780375e-01
-4.52908903e-01 -8.43464792e-01 -1.90103674e+00 6.70663342e-02
-1.41622514e-01 3.98238569e-01 1.77921548e-01 1.24032927e+00
4.37535375e-01 9.36654210e-01 5.93417764e-01 -9.28999216e-04
9.94708985e-02 -6.68349445e-01 -3.85886669e-01 5.12809694e-01
1.18302979e-01 -6.56478405e-01 3.02841514e-02 4.62862879e-01] | [9.062359809875488, 9.134879112243652] |
01d13095-86af-4495-ab3a-b525e65ac9e1 | a-new-finsler-minimal-path-model-with | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Chen_A_New_Finsler_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Chen_A_New_Finsler_CVPR_2016_paper.pdf | A New Finsler Minimal Path Model With Curvature Penalization for Image Segmentation and Closed Contour Detection | In this paper, we propose a new curvature penalized minimal path model for image segmentation via closed contour detection based on the weighted Euler elastica curves, firstly introduced to the field of computer vision in [22]. Our image segmentation method extracts a collection of curvature penalized minimal geodesics, concatenated to form a closed contour, by connecting a set of user-specified points. Globally optimal minimal paths can be computed by solving an Eikonal equation. This first order PDE is traditionally regarded as unable to penalize curvature, which is related to the path acceleration in active contour models. We introduce here a new approach that enables finding a global minimum of the geodesic energy including a curvature term. We achieve this through the use of a novel Finsler metric adding to the image domain the orientation as an extra space dimension. This metric is non-Riemannian and asymmetric, defined on an orientation lifted space, incorporating the curvature penalty in the geodesic energy. Experiments show that the proposed Finsler minimal path model indeed outperforms state-of-the-art minimal path models in both synthetic and real images. | ['Jean-Marie Mirebeau', 'Laurent D. Cohen', 'Da Chen'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['contour-detection'] | ['computer-vision'] | [ 2.23762378e-01 5.26799858e-01 1.17531307e-01 -3.42907578e-01
-4.94878858e-01 -7.14691579e-01 5.49357295e-01 3.49458516e-01
-8.39464188e-01 1.79941788e-01 -3.31684202e-01 -1.73628032e-01
-4.52422172e-01 -6.28208876e-01 -5.83621681e-01 -6.78243279e-01
-2.31347620e-01 2.91175932e-01 5.85895479e-01 -3.57537001e-01
6.09545827e-01 8.29634309e-01 -1.04325271e+00 -6.89283848e-01
1.10112643e+00 6.09102011e-01 -1.35755949e-02 7.10782468e-01
2.11627539e-02 -1.81034952e-01 -8.57767474e-04 -5.57282925e-01
2.67298728e-01 -4.83498424e-01 -1.06630850e+00 4.10303861e-01
1.88731730e-01 3.53021696e-02 4.66262907e-01 1.20666003e+00
-8.53024945e-02 3.39107543e-01 8.63494039e-01 -8.98301184e-01
-3.11506063e-01 1.00858636e-01 -5.44381797e-01 -7.48319179e-02
3.74810249e-02 -2.10261643e-01 1.03955936e+00 -9.48876441e-01
1.01477063e+00 8.95494401e-01 6.22605681e-01 3.82390350e-01
-1.44735599e+00 2.34657869e-01 -2.65281409e-01 -4.81549352e-02
-1.27034128e+00 5.06531224e-02 1.05517256e+00 -6.70979202e-01
3.50179613e-01 4.79269594e-01 7.90781140e-01 4.10673529e-01
2.82089591e-01 4.21364039e-01 7.24771619e-01 -5.66725314e-01
4.23309386e-01 1.21916663e-02 5.20046234e-01 9.86105800e-01
4.24881995e-01 -2.26593271e-01 2.56915689e-01 -2.37993538e-01
8.65543604e-01 -3.20600957e-01 -3.03602546e-01 -9.37306941e-01
-1.15841472e+00 1.05940259e+00 4.57194865e-01 5.37223518e-01
-1.07430227e-01 -1.35625392e-01 -1.79375857e-01 -2.31176838e-01
5.68197131e-01 3.84741127e-01 -1.82176724e-01 1.16836444e-01
-7.69430995e-01 1.51408941e-01 8.61729145e-01 5.71675181e-01
1.15235889e+00 -2.67643452e-01 1.84751332e-01 4.92525578e-01
7.64533281e-01 1.61605969e-01 7.46069327e-02 -1.18414760e+00
6.42215088e-02 1.08165681e+00 1.74098119e-01 -1.34900677e+00
-6.99737549e-01 -1.98188350e-01 -6.22097909e-01 5.50732434e-01
8.06348443e-01 3.42229120e-02 -4.42720711e-01 1.49107862e+00
6.95114017e-01 2.50906646e-01 -1.05075680e-01 1.21175861e+00
6.38833717e-02 1.70704260e-01 -1.07822262e-01 -3.05043936e-01
1.10686648e+00 -7.75109589e-01 -4.65211958e-01 2.65796423e-01
8.21770728e-01 -7.43512750e-01 9.41777170e-01 3.52737278e-01
-1.19224524e+00 2.90847197e-02 -1.28143728e+00 -3.41557860e-02
-4.10270512e-01 1.53257668e-01 2.67270148e-01 8.32285404e-01
-1.10428965e+00 1.13698137e+00 -1.02343023e+00 -2.88345158e-01
-8.30939785e-02 2.71923691e-01 -3.00736189e-01 6.56701267e-01
-8.38265598e-01 7.46077061e-01 2.39067778e-01 3.17341655e-01
-9.48782042e-02 -2.48125434e-01 -8.45301569e-01 -3.40564251e-01
3.21113169e-01 -3.33861351e-01 6.22027576e-01 -9.17200983e-01
-1.65665185e+00 1.04622602e+00 2.12026060e-01 -4.52548593e-01
8.82206440e-01 -1.88615486e-01 -1.45086959e-01 5.54980218e-01
-2.96931267e-01 4.61339593e-01 8.09650958e-01 -1.23754299e+00
4.54193540e-03 -4.12581772e-01 7.35925063e-02 5.73625527e-02
-3.98004740e-01 -1.63355589e-01 -4.30336714e-01 -6.11167073e-01
4.10374850e-01 -1.18742180e+00 -5.05534649e-01 1.02850690e-01
-4.54036534e-01 -3.42721313e-01 9.15832937e-01 -6.37145638e-01
1.11238170e+00 -1.83725679e+00 5.80450594e-01 5.67914009e-01
2.54879683e-01 2.50085443e-01 3.40552703e-02 8.06381404e-02
1.40950382e-01 3.76833886e-01 -1.25948846e+00 -3.58993113e-01
-2.43169039e-01 5.23562059e-02 1.92271382e-01 9.82953846e-01
4.20864522e-01 5.44147134e-01 -9.93594229e-01 -6.96306944e-01
5.60021251e-02 7.32603729e-01 -4.54853654e-01 -1.22425221e-01
-1.06196634e-01 6.44146264e-01 -4.54374313e-01 -1.33633032e-01
8.96922648e-01 1.22476101e-01 -8.45778435e-02 4.53750454e-02
-6.95826411e-01 -2.85085022e-01 -1.41883898e+00 1.93045330e+00
1.83249429e-01 2.84904420e-01 1.37823790e-01 -1.13183796e+00
1.20548785e+00 7.65073523e-02 7.78970420e-01 -1.27622128e-01
3.46337169e-01 4.45656478e-01 -1.68628290e-01 -3.13066095e-01
5.63263893e-01 2.76748955e-01 2.58257419e-01 3.83010387e-01
-1.22206420e-01 -2.04673484e-01 4.10885364e-01 1.82022974e-01
7.18732297e-01 3.44123185e-01 -2.09979326e-01 -8.51869464e-01
1.17059767e+00 -3.78420129e-02 4.17209953e-01 6.92777559e-02
-9.24449135e-03 8.79459321e-01 6.19538069e-01 -2.86318660e-01
-1.02185297e+00 -9.35289741e-01 -4.67555612e-01 2.69452214e-01
3.94589245e-01 -3.15941989e-01 -1.63323081e+00 -7.81780243e-01
-3.73536974e-01 5.07180274e-01 -5.25083780e-01 -1.80531219e-01
-9.22803462e-01 -7.61801183e-01 2.91330397e-01 -1.42027900e-01
5.86616755e-01 -7.69550025e-01 -1.03374565e+00 2.24795714e-01
-1.87420659e-02 -8.73852849e-01 -7.26114333e-01 -4.99623239e-01
-1.27390242e+00 -1.31498742e+00 -1.03147697e+00 -5.45673251e-01
8.22757542e-01 1.56806856e-02 7.00955927e-01 2.60965914e-01
-5.50210238e-01 6.46624386e-01 -3.30388129e-01 4.49743727e-03
-4.37962770e-01 4.83019501e-02 -2.43954062e-01 7.54429758e-01
3.81770078e-03 -2.39405602e-01 -9.75219429e-01 5.23057461e-01
-1.00904906e+00 -5.54448590e-02 2.51359433e-01 2.64663845e-01
7.76267648e-01 -3.41299236e-01 5.65697476e-02 -5.40074229e-01
3.79039168e-01 -2.95105845e-01 -8.81700039e-01 6.86015040e-02
-7.30693698e-01 2.92976409e-01 2.85719961e-01 -3.45421016e-01
-8.30371678e-01 4.16499108e-01 -1.01882055e-01 4.96906787e-02
1.16296157e-01 2.35872313e-01 1.39991418e-01 -7.24027812e-01
5.10774016e-01 -5.23530506e-02 -1.48521340e-03 -7.59544551e-01
6.17066383e-01 2.27190048e-01 2.39054471e-01 -3.29991937e-01
7.57514775e-01 8.82769644e-01 6.74165189e-01 -1.13846886e+00
-3.20132524e-01 -5.72979212e-01 -1.27577925e+00 -4.94242132e-01
1.12238216e+00 1.40180007e-01 -5.50135374e-01 4.87722069e-01
-1.48568499e+00 -3.49033237e-01 -3.42140019e-01 5.37895203e-01
-9.13586974e-01 9.38460410e-01 -4.00619149e-01 -1.02050519e+00
-2.31741339e-01 -1.11281848e+00 9.22793448e-01 2.09111020e-01
-3.80680561e-02 -1.39135671e+00 6.06079936e-01 -9.91956517e-02
4.86214757e-02 7.54824281e-01 7.69644201e-01 -3.14830095e-01
-4.52109843e-01 -1.58458516e-01 1.59642641e-02 4.42873508e-01
-2.43584082e-01 2.73939878e-01 -2.55972087e-01 -8.94661024e-02
4.05799717e-01 5.65696359e-01 9.80001450e-01 2.60422409e-01
6.57695949e-01 -2.62469113e-01 -1.29075304e-01 5.63291848e-01
1.65168011e+00 2.27744263e-02 5.80737710e-01 6.65007979e-02
8.84647489e-01 1.07206333e+00 5.21682203e-01 -2.80344933e-02
2.95614481e-01 7.19774365e-01 5.39641738e-01 -2.20545590e-01
1.40692428e-01 2.79791325e-01 2.37527221e-01 9.88731146e-01
-6.46731615e-01 3.02049726e-01 -9.10633266e-01 6.29338026e-01
-1.93289113e+00 -4.46303904e-01 -1.08259714e+00 2.81099796e+00
6.47134960e-01 2.26837039e-01 2.44664729e-01 3.46550643e-01
6.96892262e-01 -2.35085040e-01 -2.85194367e-01 -6.40354097e-01
-1.35167956e-01 5.80109805e-02 6.98606908e-01 1.21263433e+00
-1.03835261e+00 7.28577852e-01 4.93919754e+00 5.12167931e-01
-1.06540632e+00 3.31675142e-01 7.47950822e-02 6.86346173e-01
-3.06922942e-01 3.18615884e-01 -5.56193173e-01 2.74496406e-01
8.61180663e-01 4.33773361e-03 -6.36047050e-02 6.59636855e-01
2.47097164e-01 -1.98755637e-01 -5.87064385e-01 4.61238265e-01
2.52997931e-02 -1.02268362e+00 -1.75976366e-01 4.46244776e-01
6.93699777e-01 -3.59923750e-01 -1.61989272e-01 -7.06225336e-01
-2.84165591e-01 -5.46629488e-01 7.57195711e-01 8.61382961e-01
3.19696039e-01 -7.35592246e-01 2.63291180e-01 3.39733273e-01
-1.37149715e+00 3.59837323e-01 -1.26152977e-01 3.23843181e-01
3.55314225e-01 4.88838345e-01 -4.18086201e-01 4.85700607e-01
8.06485713e-02 5.55691004e-01 -6.22898519e-01 1.31642091e+00
8.04323936e-04 4.44373846e-01 -4.59997922e-01 -6.32900968e-02
4.86441493e-01 -1.26890707e+00 1.14860272e+00 1.31626320e+00
2.15289593e-01 1.16730303e-01 -1.79848552e-01 1.01870441e+00
2.87620813e-01 7.71490574e-01 -5.09241700e-01 1.93224087e-01
-3.50402474e-01 1.58638132e+00 -1.53323996e+00 4.19423804e-02
-2.67672211e-01 1.18153703e+00 7.97578320e-02 2.80402213e-01
-4.97770756e-01 -5.45443237e-01 3.45234662e-01 2.23494262e-01
-1.24081522e-01 -8.80198479e-01 -3.30007911e-01 -9.85404015e-01
1.20606951e-01 1.92970306e-01 1.13993771e-01 -2.03197703e-01
-5.93919516e-01 4.12555009e-01 1.62427891e-02 -1.02540171e+00
4.64609936e-02 -7.03748107e-01 -6.67534351e-01 7.05858231e-01
-1.23750234e+00 -9.03386712e-01 -7.00972527e-02 6.05300903e-01
3.69964033e-01 5.03708601e-01 8.45600784e-01 6.35984316e-02
-4.38199371e-01 1.50072291e-01 8.53152722e-02 6.69155344e-02
2.22156838e-01 -1.69753516e+00 4.84647989e-01 1.01860034e+00
8.81500368e-04 6.61258519e-01 7.35433161e-01 -6.45067751e-01
-1.29438531e+00 -8.86249065e-01 7.76113093e-01 -3.59770060e-01
5.93541324e-01 -2.78109759e-01 -1.19409084e+00 5.38022578e-01
4.21694778e-02 -1.05813399e-01 2.56832957e-01 -6.41006410e-01
-1.51031893e-02 2.39334792e-01 -1.30068827e+00 6.85347199e-01
9.08320367e-01 8.46562311e-02 -2.43703008e-01 2.60008931e-01
7.41255403e-01 4.49041128e-02 -8.90228987e-01 4.34904844e-01
3.28137398e-01 -9.57739234e-01 8.55633438e-01 -2.99384832e-01
-1.03561141e-01 -2.03439370e-01 3.47208709e-01 -9.11610305e-01
6.16401574e-03 -1.25212228e+00 -2.65851039e-02 9.87913668e-01
4.37180877e-01 -7.15260923e-01 7.64644086e-01 5.53510010e-01
-1.45958781e-01 -9.44235802e-01 -9.55816984e-01 -8.51651728e-01
3.83006990e-01 -4.47895020e-01 -2.05806154e-03 7.48855233e-01
7.80842155e-02 -3.84323597e-02 5.46305850e-02 -9.85221341e-02
1.22860014e+00 -3.97965103e-01 3.03827047e-01 -1.76897895e+00
1.03188992e-01 -8.14393938e-01 -4.35419083e-01 -9.66156900e-01
-6.12509958e-02 -9.42300916e-01 -4.33660448e-02 -1.22141385e+00
-3.48373145e-01 -5.33290684e-01 4.82430272e-02 -1.60392776e-01
2.42869169e-01 4.53649521e-01 1.03787556e-01 2.95803845e-01
-3.64435941e-01 4.08103228e-01 1.29424691e+00 2.57715464e-01
-4.94099349e-01 6.05679713e-02 -1.76145524e-01 1.25354886e+00
7.83091128e-01 -3.46411169e-01 -2.20922649e-01 -2.57444698e-02
4.82974172e-01 -2.28085518e-01 4.30345565e-01 -8.06010365e-01
2.69200355e-01 5.76748624e-02 -3.86421382e-01 -2.84055501e-01
2.95863926e-01 -8.38269949e-01 -1.43764719e-01 5.23467064e-01
-1.77138150e-01 -1.71421185e-01 -1.73832029e-01 5.59069037e-01
2.25816786e-01 -7.68834054e-01 1.14361572e+00 1.56590324e-02
-4.07603264e-01 2.02968046e-01 -2.63197243e-01 1.58600938e-02
1.21151805e+00 -3.98359954e-01 2.52886236e-01 1.02689222e-01
-9.31559265e-01 -1.32111505e-01 6.49666071e-01 2.04004928e-01
6.95120692e-01 -1.17000031e+00 -6.22899652e-01 1.03204086e-01
-1.96600109e-01 -7.06952289e-02 1.92743286e-01 1.30995595e+00
-9.75500882e-01 1.54332265e-01 1.07421372e-02 -6.92454040e-01
-1.18478537e+00 3.71831000e-01 5.92826068e-01 -2.15915248e-01
-7.51912713e-01 5.33872426e-01 -1.16826944e-01 -4.12661344e-01
-1.23729520e-01 -3.64118785e-01 -4.15745735e-01 1.25094682e-01
1.59263387e-01 8.82382274e-01 6.97204620e-02 -1.28238344e+00
-2.95310378e-01 1.44524741e+00 3.92610997e-01 -6.03238642e-01
9.51070666e-01 -2.35557035e-01 -2.60891259e-01 4.75711793e-01
1.61844277e+00 8.09035227e-02 -1.36225057e+00 1.42304108e-01
5.67298353e-01 -1.56296179e-01 -5.51661998e-02 -2.65967846e-01
-9.17688131e-01 9.07381773e-01 6.97150707e-01 6.45520091e-01
7.38272250e-01 -1.78758159e-01 8.09711993e-01 -8.75349790e-02
2.25166559e-01 -1.25766754e+00 1.12373941e-01 4.01665956e-01
1.11388910e+00 -9.68383729e-01 -3.33391160e-01 -7.83984721e-01
-2.68820047e-01 1.39434457e+00 3.80959846e-02 -6.43039048e-01
1.04284430e+00 -1.25607803e-01 -1.43607631e-01 -1.20442495e-01
3.03881198e-01 -3.30582529e-01 5.41903615e-01 2.09974349e-01
3.54455858e-01 5.02070896e-02 -1.39958441e+00 2.05673143e-01
-4.97809462e-02 -3.00706744e-01 7.51833379e-01 7.34668732e-01
-6.67193711e-01 -1.25328040e+00 -3.49718392e-01 -2.75783300e-01
-4.81306672e-01 3.53778511e-01 -5.30249417e-01 6.75056279e-01
2.12489083e-01 9.71397102e-01 1.96998939e-01 -6.04598522e-02
4.46813256e-01 1.21125460e-01 4.42093462e-01 -2.79012024e-01
-2.84986407e-01 3.82959433e-02 -3.91633242e-01 -5.19183695e-01
-5.87083340e-01 -7.30829597e-01 -1.65748870e+00 1.99494302e-01
-3.85705382e-01 3.88102204e-01 1.30873537e+00 9.42964137e-01
2.07097083e-01 -2.62338463e-02 6.53854549e-01 -1.00527978e+00
-3.35767776e-01 -7.98503518e-01 -6.31061554e-01 6.33747578e-01
3.20652604e-01 -5.90286434e-01 -6.31013453e-01 2.59843647e-01] | [7.3459038734436035, 3.938511610031128] |
8aa82778-1871-4707-8e41-3a0d66105877 | why-topological-data-analysis-detects | 2304.06877 | null | https://arxiv.org/abs/2304.06877v1 | https://arxiv.org/pdf/2304.06877v1.pdf | Why Topological Data Analysis Detects Financial Bubbles? | We present a heuristic argument for the propensity of Topological Data Analysis (TDA) to detect early warning signals of critical transitions in financial time series. Our argument is based on the Log-Periodic Power Law Singularity (LPPLS) model, which characterizes financial bubbles as super-exponential growth (or decay) of an asset price superimposed with oscillations increasing in frequency and decreasing in amplitude when approaching a critical transition (tipping point). We show that whenever the LPPLS model is fitting with the data, TDA generates early warning signals. As an application, we illustrate this approach on a sample of positive and negative bubbles in the Bitcoin historical price. | ['Vahid Nateghi', 'Matteo Manzi', 'Marian Gidea', 'Samuel W. Akingbade'] | 2023-04-14 | null | null | null | null | ['topological-data-analysis'] | ['graphs'] | [-4.36747164e-01 4.56120148e-02 -1.69361606e-01 4.73163754e-01
-1.79270983e-01 -1.04804051e+00 9.25068855e-01 4.61289465e-01
2.88918942e-01 5.00674367e-01 1.50536031e-01 -1.07251990e+00
-2.33110949e-01 -8.33991766e-01 -5.52146196e-01 -5.70940673e-01
-1.36391509e+00 2.16753468e-01 6.63376510e-01 -2.58292764e-01
6.01540387e-01 2.68773079e-01 -8.39712083e-01 1.59131303e-01
6.59900546e-01 1.24247408e+00 -4.31600869e-01 7.59146988e-01
1.29149809e-01 9.05070782e-01 -6.38827145e-01 -4.85939652e-01
5.54885447e-01 -5.49764633e-01 -5.75612962e-01 -7.30627358e-01
-2.72011548e-01 1.83929931e-02 -3.43931317e-02 1.37263715e+00
3.69855054e-02 -6.00645602e-01 4.67614383e-01 -1.40196288e+00
-5.94280958e-01 1.00040531e+00 -6.70971870e-01 1.19519162e+00
3.74042988e-01 3.99800725e-02 1.24133718e+00 -1.02029502e+00
6.34688735e-01 6.83556199e-01 1.10112679e+00 -3.27586114e-01
-1.43552625e+00 -2.31384486e-01 -6.52985692e-01 2.88072377e-02
-7.00161457e-01 -1.02623962e-01 1.09301996e+00 -8.27096462e-01
6.62991285e-01 3.61416340e-01 1.15367770e+00 7.42115974e-01
7.82986641e-01 3.02050561e-01 1.26288927e+00 -2.02844366e-01
5.52564323e-01 -4.25932258e-01 2.68966168e-01 2.68054962e-01
6.64327919e-01 7.78885663e-01 -5.90434313e-01 -8.20290506e-01
7.86786616e-01 -4.44075495e-01 -1.33762762e-01 -1.08015381e-01
-1.07063591e+00 1.04965401e+00 2.68978477e-01 6.54917657e-01
-5.98074138e-01 8.14392641e-02 4.42662507e-01 9.76715684e-01
6.93388581e-01 5.59705675e-01 -4.36694562e-01 -4.48676586e-01
-9.86226380e-01 1.36323258e-01 1.05227292e+00 3.08141768e-01
6.68497458e-02 -1.02139367e-02 3.04984391e-01 -2.61670619e-01
2.50122488e-01 4.34480101e-01 7.14612484e-01 -8.20896983e-01
3.07780474e-01 2.32869267e-01 3.03417772e-01 -1.12283337e+00
-5.08375227e-01 -4.75536555e-01 -5.51640928e-01 5.54881513e-01
1.04266214e+00 -1.02436773e-01 6.56603947e-02 1.30018866e+00
-4.89752963e-02 2.21041650e-01 2.57906765e-01 4.44047958e-01
5.23570515e-02 7.50334799e-01 -4.67855483e-01 -9.68044400e-01
1.43777859e+00 -2.48724192e-01 -6.14964664e-01 4.40229297e-01
3.48238468e-01 -5.62430263e-01 7.82666683e-01 6.39505088e-01
-1.19639337e+00 3.06433882e-03 -1.03928673e+00 7.36663580e-01
-2.54643232e-01 -9.27757025e-01 1.31442621e-01 4.13362026e-01
-8.86227608e-01 8.81398976e-01 -7.83402443e-01 1.00231327e-01
6.38806969e-02 -3.20205182e-01 3.64718646e-01 1.22955894e+00
-1.40106702e+00 5.59274137e-01 2.50318021e-01 -1.02667183e-01
-6.45217061e-01 -1.01898074e+00 -2.30121270e-01 1.57310456e-01
-1.25354484e-01 8.04206058e-02 1.19722176e+00 -9.20905888e-01
-1.14797843e+00 5.08525491e-01 5.03200889e-01 -1.23748922e+00
9.04769957e-01 1.30366877e-01 -6.53533220e-01 5.33793032e-01
-8.42828602e-02 -5.37577152e-01 9.91448283e-01 -9.49436605e-01
-3.03927898e-01 8.53987876e-03 -4.55921113e-01 -7.98353076e-01
-1.30547453e-02 3.64071518e-01 1.27152789e+00 -9.84040439e-01
2.38905430e-01 -3.17317963e-01 1.24857761e-01 -5.17266273e-01
-1.75592422e-01 -3.39147747e-01 7.34421730e-01 -9.54133213e-01
1.64661741e+00 -1.99212575e+00 -6.22264326e-01 6.07543707e-01
3.06950480e-01 -2.03699134e-02 3.69214207e-01 7.59997845e-01
-5.38412154e-01 4.95739460e-01 -2.31253013e-01 5.80781579e-01
4.01584879e-02 -3.04516137e-01 -1.14354873e+00 7.91141510e-01
4.99503724e-02 9.25587654e-01 -1.17635155e+00 -1.69241756e-01
-4.43919122e-01 -1.51084453e-01 -2.29440913e-01 -1.86923102e-01
-2.76430875e-01 2.66360134e-01 -6.51008859e-02 7.66483426e-01
6.18848503e-01 -3.21919590e-01 -4.95490618e-02 2.46639967e-01
-8.81673634e-01 5.64285517e-01 -5.20168185e-01 5.51680565e-01
4.43965107e-01 1.01675916e+00 -2.22655073e-01 -1.11584163e+00
1.03115213e+00 5.68494856e-01 7.53060520e-01 -8.67685199e-01
-3.06621399e-02 6.98215604e-01 2.70564258e-01 -5.02096653e-01
1.72813535e-01 -3.61930549e-01 -1.63786352e-01 9.30230200e-01
-2.86982983e-01 2.66354233e-01 3.15197051e-01 8.51751044e-02
1.26441562e+00 -2.20052436e-01 2.82504797e-01 -9.03955221e-01
2.03633726e-01 9.73327085e-02 4.73707259e-01 4.48024988e-01
-5.60802579e-01 1.33624658e-01 1.53226459e+00 -7.24541485e-01
-1.43855751e+00 -1.28414416e+00 -4.09924060e-01 4.14930910e-01
-1.67260364e-01 -4.92963314e-01 -5.40423870e-01 -5.15597761e-01
2.45494753e-01 7.06903636e-01 -7.57802904e-01 4.00884561e-02
-7.26678133e-01 -1.02317488e+00 3.58978838e-01 7.68787786e-02
3.73220146e-01 -1.07212555e+00 -1.46292579e+00 2.98040569e-01
-8.28326270e-02 -4.98374134e-01 -2.04067409e-01 1.69190988e-01
-1.06278813e+00 -1.23483872e+00 -6.23235345e-01 -3.93097848e-01
3.01023722e-01 -3.76356930e-01 1.23127949e+00 -8.11332911e-02
-6.72678053e-02 1.42249569e-01 -3.19172978e-01 -1.60373867e-01
-9.53127682e-01 -6.93673134e-01 9.93612185e-02 1.04416739e-02
3.62617522e-02 -8.57796252e-01 -8.59585702e-01 1.68884650e-01
-6.89799666e-01 -6.64668798e-01 -1.54884204e-01 5.81039369e-01
-1.83034196e-01 3.85193497e-01 8.42052221e-01 7.60448501e-02
1.15471995e+00 -8.79966259e-01 -1.28144205e+00 2.55907625e-01
-1.04284310e+00 -1.05947234e-01 6.25673771e-01 -6.97445869e-01
-3.34813952e-01 -6.64734840e-01 6.89541578e-01 -4.23694909e-01
7.17128873e-01 1.03028989e+00 7.76729107e-01 -9.87123698e-02
6.66757941e-01 1.91482618e-01 2.64544021e-02 -5.21181107e-01
1.82831325e-02 2.08331943e-01 7.53279686e-01 -2.85653710e-01
7.77284801e-01 4.13299859e-01 1.60891652e-01 -6.12460136e-01
-5.21900281e-02 1.64693780e-03 -5.42331815e-01 -5.54338396e-01
3.84399593e-01 -2.75110722e-01 -9.30254638e-01 5.28806031e-01
-1.12551045e+00 -3.14083010e-01 -7.23399162e-01 3.52221847e-01
-5.72292447e-01 3.31546664e-01 -7.91472554e-01 -1.29176259e+00
-1.26730293e-01 -8.42817649e-02 -8.98896977e-02 2.16191590e-01
-2.76874065e-01 -1.33460283e+00 8.42227101e-01 -5.55274308e-01
5.50719261e-01 1.02353799e+00 8.77118766e-01 -1.12462449e+00
-2.23360956e-01 -2.03320637e-01 2.18936205e-01 -3.78788263e-02
-1.11072846e-01 5.73459148e-01 -4.52182949e-01 -3.28889340e-01
8.25951934e-01 3.52255166e-01 5.26494086e-01 4.12386358e-01
-3.78510915e-02 -1.01865101e+00 4.93109226e-04 2.76601315e-01
1.37858641e+00 4.58822876e-01 1.11169457e-01 5.49269855e-01
-4.37955290e-01 6.77464783e-01 4.28029358e-01 7.75022805e-01
-1.77745372e-01 1.40901387e-01 4.27181929e-01 2.99027830e-01
3.31684351e-01 -5.01157999e-01 9.19453144e-01 8.62761199e-01
-1.37951389e-01 3.78250808e-01 -1.42555463e+00 6.95577443e-01
-1.63819230e+00 -1.40352118e+00 -6.37001336e-01 2.33137822e+00
7.62000501e-01 6.29491389e-01 1.03479791e+00 2.33570069e-01
7.43152499e-01 1.78798795e-01 -3.69820505e-01 -4.24602330e-01
-3.74958724e-01 -5.12462258e-02 6.15658045e-01 6.01970673e-01
-7.24732637e-01 1.52994856e-01 7.82233763e+00 1.84964135e-01
-1.23208344e+00 2.16456205e-01 7.66074717e-01 4.59111124e-01
-5.06778419e-01 3.72795939e-01 9.52490419e-02 9.79411244e-01
1.25325751e+00 -1.05246770e+00 2.45009601e-01 4.05785233e-01
3.34902316e-01 -2.54392594e-01 -6.96639240e-01 5.97096086e-01
-6.24247313e-01 -1.45776224e+00 -5.08495331e-01 3.62127930e-01
4.37868237e-01 -2.63696797e-02 1.51470929e-01 -4.13675517e-01
5.83764017e-01 -4.42619324e-01 1.11259592e+00 6.26911461e-01
3.63201827e-01 -5.84029675e-01 4.71766740e-01 2.24465400e-01
-1.17971730e+00 -5.59332192e-01 -8.54856372e-02 -3.20179671e-01
5.25854409e-01 8.97122264e-01 -6.55262530e-01 3.41073573e-01
6.57751858e-01 5.39815247e-01 -5.99983752e-01 1.23864579e+00
-1.64272115e-01 1.00172043e+00 -7.47546196e-01 4.40374054e-02
-2.95346603e-03 -6.58077538e-01 1.12279081e+00 7.13667810e-01
7.81739235e-01 -2.17538942e-02 -4.74800378e-01 1.29565084e+00
5.23463011e-01 -2.63905704e-01 -6.81167245e-01 -4.10665274e-01
5.21612048e-01 6.87523365e-01 -1.23172295e+00 -2.91318774e-01
-4.03336078e-01 3.02174896e-01 -3.71310204e-01 2.66099513e-01
-5.77945769e-01 -3.86572838e-01 2.04326451e-01 4.03631687e-01
3.75743747e-01 -4.17486817e-01 -6.93747520e-01 -1.08166778e+00
1.42062500e-01 -3.37315321e-01 7.70142913e-01 -5.84004462e-01
-1.22361195e+00 4.07964408e-01 -5.76698333e-02 -1.48368669e+00
-4.73439127e-01 -3.18704605e-01 -1.50278962e+00 3.87100577e-01
-9.03581023e-01 -2.72863239e-01 5.25748610e-01 5.75522125e-01
-2.69713551e-01 -1.78014114e-01 2.11348653e-01 -2.45957285e-01
-1.76052347e-01 -1.87480241e-01 3.20845425e-01 2.15379611e-01
-4.95719165e-02 -1.53846681e+00 8.16892922e-01 1.18963480e+00
1.72210976e-01 3.22135240e-01 1.58334124e+00 -9.79745448e-01
-8.62193584e-01 -1.70813113e-01 1.27035546e+00 -4.57541615e-01
1.83506286e+00 -3.37692499e-01 -7.81756103e-01 4.93259013e-01
4.72771615e-01 2.01081917e-01 4.07633930e-01 -5.20403504e-01
-2.62707680e-01 4.97007817e-02 -1.13836026e+00 3.37684989e-01
2.66937464e-01 -3.07839632e-01 -1.09160542e+00 2.76153505e-01
6.29927576e-01 4.79747772e-01 -1.14130843e+00 4.32489574e-01
6.08839214e-01 -1.49704671e+00 6.18148565e-01 -4.09907609e-01
1.28650814e-01 -2.56252866e-02 3.14668357e-01 -1.05749977e+00
-2.75145262e-01 -1.80083013e+00 -5.59502304e-01 9.55863953e-01
2.57224590e-01 -1.20421505e+00 3.61352980e-01 1.56653166e-01
4.46791798e-01 -4.33041871e-01 -1.62696147e+00 -1.19536936e+00
5.40134251e-01 -6.82588667e-02 4.16093290e-01 1.12696052e+00
1.21229017e+00 -2.56125033e-01 8.82592350e-02 -7.62012675e-02
8.33298862e-01 3.87679547e-01 1.04995646e-01 -1.56183434e+00
-2.84268782e-02 -9.63315606e-01 -3.39645714e-01 -4.75868106e-01
-6.38931215e-01 -5.42468011e-01 -3.27315003e-01 -3.78040314e-01
-2.83993006e-01 -1.41340420e-01 -5.13344765e-01 -1.05189875e-01
5.93142748e-01 -3.26967128e-02 2.12767288e-01 7.75599599e-01
-1.32652922e-02 2.10618660e-01 6.20004833e-01 4.04081732e-01
-2.67997801e-01 2.05112308e-01 -1.66227952e-01 6.93817496e-01
8.72591317e-01 -5.11008501e-01 -2.32850970e-03 8.40899527e-01
1.08801043e+00 6.24012589e-01 8.03876460e-01 -6.66186690e-01
2.66983420e-01 7.88682140e-03 -2.11972043e-01 -8.05919349e-01
-6.58366442e-01 -5.14984012e-01 1.80986986e-01 1.42355633e+00
-3.65650803e-02 8.94971907e-01 -1.27922401e-01 4.29138750e-01
-3.37279409e-01 -2.31923163e-01 5.82014024e-01 5.73175400e-02
1.43795460e-01 -7.52707720e-02 -5.92424929e-01 3.29755753e-01
1.17237616e+00 1.83317568e-02 -8.56118083e-01 -3.36697310e-01
-8.09266508e-01 1.51861802e-01 4.27564144e-01 2.40040138e-01
4.66796309e-01 -1.57608938e+00 -9.19697881e-01 4.28974450e-01
-5.14609993e-01 -7.36630738e-01 -4.98050272e-01 1.56758344e+00
-8.89425576e-01 4.51380700e-01 -1.97882280e-01 -3.92358780e-01
-6.95230305e-01 1.10845566e+00 5.65879762e-01 -3.62033546e-01
-8.84907246e-01 3.23238015e-01 -3.51806939e-01 5.70246935e-01
-1.25515327e-01 -5.45283556e-01 -6.85307905e-02 7.16901243e-01
5.95528126e-01 7.22657204e-01 -4.17879581e-01 -4.22479808e-01
-2.97640532e-01 3.57378811e-01 5.97803652e-01 -5.73905349e-01
1.36266053e+00 -4.23068665e-02 -5.68667233e-01 1.10401452e+00
8.17344546e-01 3.31180513e-01 -1.34460115e+00 -5.25651239e-02
1.18819475e+00 -2.17623532e-01 -5.04134655e-01 -5.25421262e-01
-7.79550493e-01 4.76312667e-01 3.89056593e-01 1.96440077e+00
7.29359925e-01 1.79732382e-01 7.00843573e-01 -2.83222884e-01
3.25768501e-01 -1.13594246e+00 2.31771722e-01 3.28134000e-01
1.16733837e+00 -5.27558386e-01 -3.55021060e-01 8.82202238e-02
-1.13673352e-01 1.59529638e+00 -5.21341622e-01 -8.33490252e-01
1.28057647e+00 5.16742945e-01 -1.11957505e-01 -5.84733009e-01
-9.10588264e-01 5.05961061e-01 -2.06257582e-01 4.15741771e-01
-6.09722361e-02 1.07920133e-01 -8.05394351e-01 6.60574913e-01
-7.01397121e-01 -1.79733634e-01 7.18691051e-01 8.74207497e-01
-6.91641510e-01 -4.80302691e-01 -6.94932044e-01 -7.08348379e-02
-8.07196319e-01 -2.37400711e-01 -5.53181052e-01 7.59467840e-01
-3.54103595e-01 6.11599445e-01 5.71635842e-01 -2.85551935e-01
-1.24334164e-01 3.92647982e-01 -1.14036947e-01 4.42147166e-01
-5.62771797e-01 3.44218075e-01 -1.67535782e-01 -5.09252369e-01
-2.68656522e-01 -1.28028727e+00 -9.13000882e-01 -4.02270973e-01
-1.50499448e-01 6.20596588e-01 2.60046870e-01 5.09949088e-01
-1.67896654e-02 -9.71893780e-03 1.17852592e+00 -3.63129944e-01
-9.47056830e-01 -1.02648389e+00 -9.42428768e-01 1.09483920e-01
1.01927209e+00 -2.27022976e-01 -1.23574579e+00 8.54600146e-02] | [4.77943229675293, 4.124654293060303] |
38b35eee-aac6-45dc-bc86-7b9109d0983c | knowledge-augmented-deep-neural-networks-for | null | null | http://proceedings.neurips.cc/paper/2020/hash/a51fb975227d6640e4fe47854476d133-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/a51fb975227d6640e4fe47854476d133-Paper.pdf | Knowledge Augmented Deep Neural Networks for Joint Facial Expression and Action Unit Recognition | Facial expression and action units (AUs) represent two levels of descriptions of the facial behavior. Due to the underlying facial anatomy and the need to form a meaningful coherent expression, they are strongly correlated. This paper proposes to systematically capture their dependencies and incorporate them into a deep learning framework for joint facial expression recognition and action unit detection. Specifically, we first propose a constraint optimization method to encode the generic knowledge on expression-AUs probabilistic dependencies into a Bayesian Network (BN). The BN is then integrated into a deep learning framework as a weak supervision for an AU detection model. A data-driven facial expression recognition(FER) model is then constructed from data. Finally, the FER model and AU detection model are trained jointly to refine their learning. Evaluations on benchmark datasets demonstrate the effectiveness of the proposed knowledge integration in improving the performance of both the FER model and the AU detection model. The proposed AU detection model is demonstrated to be able to achieve competitive performance without AU annotations. Furthermore, the proposed Bayesian Network capturing the generic knowledge is demonstrated to generalize well to different datasets. | ['Qiang Ji', 'Yuru Wang', 'Tengfei Song', 'Zijun Cui'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['action-unit-detection'] | ['computer-vision'] | [ 2.94647038e-01 3.02726984e-01 -3.27957660e-01 -8.91017437e-01
-8.16474319e-01 -2.64408495e-02 5.42757034e-01 -3.64861459e-01
-1.97893813e-01 3.91198128e-01 3.27669159e-02 4.88024205e-01
1.34745374e-01 -3.85495871e-01 -6.48733497e-01 -7.94292450e-01
-3.85732204e-02 9.66369584e-02 -3.59377712e-02 6.64620399e-02
-3.35199267e-01 6.88415945e-01 -1.65024388e+00 2.54333675e-01
4.81620550e-01 1.62716401e+00 -2.28312146e-02 4.24171776e-01
2.96016097e-01 1.21759319e+00 -4.59728658e-01 -5.28389156e-01
-8.34951084e-03 -2.26342723e-01 -5.16292691e-01 4.81609017e-01
3.59136879e-01 -5.45603514e-01 -9.77254882e-02 1.08842719e+00
2.20815420e-01 4.55404110e-02 8.83815408e-01 -1.41690338e+00
-1.03853300e-01 2.85295755e-01 -7.31401563e-01 -3.69782239e-01
1.48366615e-01 -1.61981553e-01 1.07168555e+00 -9.58806157e-01
5.93651235e-01 1.29408824e+00 4.07210290e-01 7.43938625e-01
-1.06711936e+00 -8.11422050e-01 3.44256908e-01 1.78544685e-01
-1.55812728e+00 -7.42961228e-01 8.83090973e-01 -4.30513918e-01
8.44559193e-01 -1.76626965e-01 5.85574746e-01 1.13754904e+00
-3.28247726e-01 1.10727942e+00 9.42588568e-01 -3.72223973e-01
2.55732834e-01 1.73225015e-01 2.09097371e-01 1.17860138e+00
-2.58562416e-01 -1.64750844e-01 -7.48743176e-01 -2.27095306e-01
6.77690089e-01 -1.66992843e-01 4.05219719e-02 -2.68387590e-02
-1.62186727e-01 6.18640184e-01 3.82072031e-01 1.76209286e-01
-6.41665459e-01 4.54182088e-01 4.85728383e-01 -2.24611238e-01
4.69225109e-01 -2.75379568e-01 -1.17246263e-01 -1.08025260e-01
-8.51588011e-01 1.20150417e-01 8.22919250e-01 9.03031349e-01
8.87697160e-01 6.90733716e-02 -2.89162308e-01 1.12474060e+00
5.81632793e-01 3.22100282e-01 3.59812826e-02 -1.06851268e+00
-1.02840781e-01 6.53314173e-01 4.49742340e-02 -1.04467285e+00
-3.52667153e-01 -2.24457622e-01 -5.95442355e-01 2.84270942e-01
6.97515160e-02 -1.76782623e-01 -8.85418892e-01 2.26129174e+00
3.12956512e-01 4.95576084e-01 2.76789106e-02 8.48419249e-01
7.36781418e-01 5.75758517e-01 4.42788780e-01 -2.22320631e-01
1.48488641e+00 -8.79828811e-01 -8.78946602e-01 -1.37005717e-01
8.03957939e-01 -3.55067700e-01 3.40215474e-01 4.94816422e-01
-9.64727998e-01 -6.25042140e-01 -1.08141041e+00 2.08548363e-02
1.24090657e-01 8.17999244e-01 6.95212483e-01 4.55511212e-01
-7.90905654e-01 1.62495464e-01 -1.17709041e+00 -2.35463217e-01
8.22968543e-01 6.08663619e-01 -6.92929149e-01 1.77590325e-02
-1.13155854e+00 8.66384208e-01 3.83746594e-01 5.61961353e-01
-1.23262501e+00 -1.44980669e-01 -1.07633579e+00 8.24579746e-02
1.53244615e-01 -4.07075047e-01 1.31495833e+00 -1.48449492e+00
-1.93868327e+00 8.84122193e-01 -3.17353427e-01 -2.05425858e-01
2.09782779e-01 -2.91835904e-01 -2.64992625e-01 3.89611751e-01
-2.57974744e-01 7.76324213e-01 1.00263548e+00 -1.28644812e+00
-7.42488384e-01 -6.02536082e-01 5.67212962e-02 1.81265727e-01
-4.82288390e-01 3.14401984e-01 -9.13998604e-01 -3.06328356e-01
-1.03441425e-01 -9.64486897e-01 -2.98444033e-02 2.60000855e-01
-2.73647994e-01 -5.62565863e-01 8.92258346e-01 -5.42950749e-01
1.22355890e+00 -2.22381306e+00 3.20930809e-01 4.05821800e-01
1.10794067e-01 2.11045563e-01 -2.52194226e-01 -1.70446619e-01
-1.35207279e-02 -2.93257087e-01 -1.18481539e-01 -9.35362041e-01
8.44277665e-02 6.75420463e-01 1.20461337e-01 6.39123738e-01
7.32078791e-01 7.65336096e-01 -5.73937833e-01 -4.66235965e-01
1.91272143e-02 7.12376416e-01 -5.27985871e-01 7.10399449e-01
-3.42418104e-01 2.58786291e-01 -6.12389624e-01 8.74802589e-01
6.49079919e-01 -6.08491749e-02 5.40577471e-01 -5.83302259e-01
3.82572055e-01 -1.38782158e-01 -9.87876952e-01 1.50921571e+00
-5.52892923e-01 3.69231254e-01 5.45937181e-01 -1.24118304e+00
1.28336132e+00 3.40172887e-01 4.66567844e-01 -3.75903666e-01
4.32711393e-01 6.00933284e-02 -1.15045518e-01 -5.27272224e-01
-6.13221973e-02 -3.46633047e-01 -5.86192915e-03 1.92495376e-01
6.98917449e-01 1.95803761e-01 7.64399245e-02 -4.30148318e-02
8.85807514e-01 3.86494517e-01 3.54746372e-01 -6.86774477e-02
9.61690724e-01 -5.99006355e-01 9.97765243e-01 3.22493792e-01
-2.69394547e-01 1.94329709e-01 8.12765658e-01 -3.29228848e-01
-4.83060032e-01 -8.23391199e-01 -2.33059719e-01 1.40935791e+00
-1.01648957e-01 -5.43076158e-01 -8.32768381e-01 -8.66625130e-01
-1.23414896e-01 3.24428648e-01 -8.97702694e-01 -2.74001688e-01
-1.87707171e-01 -7.34471977e-01 7.10260570e-01 8.44596922e-01
4.63842332e-01 -8.77859771e-01 -3.73758465e-01 4.67468388e-02
-6.72956463e-03 -1.55684066e+00 8.40950087e-02 3.50441337e-01
-3.65284234e-01 -8.99209559e-01 -3.87241632e-01 -7.24952638e-01
6.78827405e-01 -3.93791437e-01 6.48069203e-01 5.85864894e-02
-1.59677297e-01 4.06268060e-01 -2.42097721e-01 -4.80599642e-01
-5.25756717e-01 -4.39359039e-01 2.41844624e-01 7.60790765e-01
6.56283379e-01 -6.86719060e-01 -3.61488312e-01 3.43780458e-01
-8.71217370e-01 -2.51383930e-01 7.22990870e-01 9.38119948e-01
5.89981556e-01 -1.91736996e-01 4.18626308e-01 -6.49195969e-01
3.96487385e-01 -3.25233221e-01 -6.32039130e-01 2.50319153e-01
-2.65516132e-01 -4.90089469e-02 2.04448700e-01 -2.93240815e-01
-1.52298021e+00 6.15029633e-01 -3.71300280e-01 -7.03585982e-01
-4.32560295e-01 6.34315193e-01 -4.89114672e-01 -7.07842037e-02
3.52354169e-01 -1.09778963e-01 1.77114025e-01 -4.03742760e-01
2.78189361e-01 7.84010470e-01 5.56444824e-01 -8.75216961e-01
1.68466359e-01 5.77147424e-01 1.52029082e-01 -8.52718294e-01
-1.15377724e+00 -4.31666166e-01 -8.56965780e-01 -3.58172566e-01
1.06807470e+00 -1.20222414e+00 -6.99944437e-01 5.56665242e-01
-1.35762703e+00 -1.45619884e-01 3.11083943e-01 4.29376990e-01
-7.68393874e-01 3.39673817e-01 -7.49087930e-01 -1.15187514e+00
-1.55678585e-01 -1.26295495e+00 1.45823014e+00 2.44483531e-01
-3.49080890e-01 -8.43097985e-01 5.34507781e-02 3.97719115e-01
-1.15455367e-01 4.61675763e-01 6.20958447e-01 -5.79315960e-01
-1.49451137e-01 -4.56682801e-01 -3.93198758e-01 8.94750178e-01
9.15133879e-02 3.86614949e-01 -1.38592613e+00 1.05191529e-01
-5.71636446e-02 -9.03763175e-01 9.10268366e-01 1.72884747e-01
1.10850072e+00 -1.69659406e-01 -2.84403026e-01 5.87508202e-01
1.10626018e+00 6.48610890e-02 4.95929509e-01 -2.12831110e-01
5.84882617e-01 8.87638569e-01 6.29361868e-01 7.69742608e-01
7.47851208e-02 9.92713153e-01 6.51971161e-01 -8.10223520e-02
9.55142602e-02 -1.30411029e-01 6.96189642e-01 5.67781210e-01
-2.89371550e-01 8.17638785e-02 -6.77022278e-01 1.16197772e-01
-2.14112520e+00 -8.12744439e-01 2.11368576e-01 1.58021152e+00
8.40763986e-01 -5.31829596e-02 3.06093153e-02 -2.67812401e-01
4.65591609e-01 1.42608672e-01 -4.18043822e-01 -5.67505300e-01
-4.45176512e-02 2.71403402e-01 -1.11517534e-01 3.89517814e-01
-1.24609685e+00 1.11365902e+00 5.83696127e+00 9.18036938e-01
-1.14767540e+00 1.80929571e-01 6.37288928e-01 2.24639460e-01
2.59518862e-01 -2.75980592e-01 -1.02259839e+00 -4.27146591e-02
8.26980829e-01 1.89556912e-01 -8.10868368e-02 1.19389927e+00
2.21121565e-01 -7.90252537e-02 -1.22095811e+00 1.10122657e+00
2.25708470e-01 -8.88772011e-01 1.57525331e-01 -9.78413448e-02
5.39707959e-01 -2.60207474e-01 -3.46003883e-02 4.39474374e-01
2.44177416e-01 -1.24126220e+00 6.41486883e-01 7.76783049e-01
7.28706419e-01 -8.52370858e-01 8.49306941e-01 2.38288000e-01
-1.05288935e+00 -9.71326977e-02 -3.51813376e-01 -2.94836536e-02
1.27466097e-01 2.55589753e-01 -7.29889750e-01 4.05520469e-01
5.53047061e-01 9.36490119e-01 -1.79715008e-01 3.12169760e-01
-6.51214540e-01 5.87251246e-01 -4.75821912e-01 1.15564696e-01
3.43949020e-01 -2.11119130e-01 2.20141411e-01 1.44720650e+00
4.76493388e-02 1.24498248e-01 1.55853599e-01 1.04213583e+00
-1.81873769e-01 1.40889153e-01 -3.25871646e-01 -2.20874518e-01
8.95564482e-02 1.71797204e+00 -2.21295550e-01 -2.94418514e-01
-6.11316681e-01 9.55988467e-01 8.40036154e-01 3.01714838e-01
-9.18335915e-01 1.82893172e-01 1.03849101e+00 -3.06658477e-01
3.15409511e-01 6.93349689e-02 2.44931653e-01 -1.02254879e+00
-3.43260095e-02 -7.65964985e-01 4.15869415e-01 -6.94038928e-01
-1.27287674e+00 7.37562954e-01 1.86582245e-02 -8.94451737e-01
-3.69338304e-01 -9.93085742e-01 -5.02419472e-01 8.62814486e-01
-1.47325313e+00 -1.54408240e+00 -4.87813205e-01 7.56825030e-01
2.09335670e-01 -1.33856460e-01 1.15505993e+00 2.17952937e-01
-9.14073825e-01 6.83634996e-01 -4.10852581e-01 3.86991352e-01
5.45943856e-01 -9.18833673e-01 -6.29636109e-01 6.63323104e-01
1.15223646e-01 5.94263494e-01 3.29171777e-01 -4.17639434e-01
-1.11427438e+00 -1.24282265e+00 2.46518180e-01 -1.67665362e-01
7.14821279e-01 -4.42581773e-01 -8.75469923e-01 7.05740929e-01
-7.67212957e-02 4.68409717e-01 9.95211720e-01 9.80266929e-02
-6.54708803e-01 -3.93828839e-01 -8.67181957e-01 2.08671540e-01
6.95943356e-01 -7.45377183e-01 -3.57817560e-01 5.28974235e-02
4.80504856e-02 -2.21050769e-01 -1.01689446e+00 7.27787554e-01
9.03047442e-01 -9.96355653e-01 4.73279715e-01 -7.55730629e-01
5.07434130e-01 -1.55056193e-01 -3.95492882e-01 -8.85229051e-01
-1.53401539e-01 -5.32314658e-01 -4.61924940e-01 1.30252779e+00
2.70190984e-01 -3.28901000e-02 8.57436478e-01 7.68832922e-01
9.21822637e-02 -9.02865469e-01 -9.89943683e-01 -5.21132052e-01
-3.30297410e-01 -6.78477764e-01 2.06715018e-01 7.96442509e-01
1.14233851e-01 4.12821800e-01 -5.58031142e-01 3.54605168e-01
5.52018046e-01 -5.32865673e-02 8.22202504e-01 -1.10009837e+00
-4.21464384e-01 -4.92015332e-01 -8.40772152e-01 -1.11268783e+00
7.34527886e-01 -8.18576455e-01 2.02066377e-01 -1.11841941e+00
3.97665322e-01 -2.58822255e-02 -5.97000420e-01 7.34399498e-01
-1.81855634e-01 2.63478428e-01 -1.75253935e-02 7.02874511e-02
-7.09663630e-01 9.56287205e-01 8.15772831e-01 1.14213480e-02
-6.07879944e-02 -5.12718782e-03 -3.82886499e-01 1.06949258e+00
5.12500823e-01 -2.70669788e-01 -2.93494016e-01 -1.34020716e-01
-8.57822597e-02 3.25956754e-02 3.38599980e-01 -7.21590042e-01
1.59036130e-01 1.13718398e-01 3.13652039e-01 -1.44911334e-01
7.14706481e-01 -8.78211677e-01 -2.73454338e-01 -3.89324874e-02
-3.24627727e-01 -5.54729581e-01 2.72271812e-01 6.23834968e-01
-5.28991103e-01 -1.27346233e-01 9.86755490e-01 1.88446730e-01
-9.09517467e-01 4.73366976e-01 -2.61003941e-01 -5.19362509e-01
1.20580685e+00 -5.22693172e-02 2.22505257e-01 -5.39821625e-01
-1.25171912e+00 2.02425614e-01 -7.23876432e-02 4.28736895e-01
6.25467420e-01 -1.33333063e+00 -5.32236874e-01 1.61089227e-01
4.48201418e-01 -1.27096102e-01 1.11895688e-01 1.00000405e+00
-1.80101320e-01 1.77634582e-01 -3.53585124e-01 -8.17843139e-01
-1.52725279e+00 1.39059722e-01 6.12792671e-01 -1.75460815e-01
-4.98376153e-02 1.01634908e+00 4.90911275e-01 -2.80836642e-01
5.54850101e-01 -7.44423866e-02 -3.85789603e-01 1.20327868e-01
5.66140354e-01 1.07967528e-02 -1.39610752e-01 -1.05380905e+00
-3.34551483e-01 5.20758033e-01 -1.23240925e-01 -1.37952305e-02
1.38059258e+00 1.22677209e-02 -3.54181439e-01 1.34286940e-01
1.41466939e+00 -3.35466266e-01 -1.47438860e+00 -3.72729868e-01
1.99572355e-01 -1.96592193e-02 2.26699769e-01 -4.90139753e-01
-1.13391495e+00 1.00357032e+00 5.05367100e-01 -5.10579884e-01
1.32052541e+00 1.94182813e-01 2.04445809e-01 4.24277842e-01
1.21374346e-01 -1.07342005e+00 2.00209603e-01 3.46439123e-01
9.03572738e-01 -1.19815171e+00 -1.63006037e-01 -5.34043968e-01
-6.86297834e-01 1.46108651e+00 9.37873721e-01 -1.02958806e-01
8.00898135e-01 5.89530647e-01 1.32790744e-01 -3.82312328e-01
-8.51798356e-01 -4.30997699e-01 4.00648773e-01 4.56416786e-01
4.19000208e-01 -1.03710219e-01 -1.13712177e-01 1.09578705e+00
3.47373188e-01 2.88777053e-01 -2.39483882e-02 7.64545202e-01
-2.39385486e-01 -9.96182084e-01 4.79576848e-02 2.06667725e-02
-5.08388877e-01 3.60474885e-01 -6.02744460e-01 7.58104205e-01
3.22005987e-01 7.00529277e-01 4.19282727e-02 -3.74984443e-01
2.85690278e-01 2.52589941e-01 5.77395558e-01 -6.86010003e-01
-1.81881502e-01 3.92984539e-01 2.37296805e-01 -9.02774692e-01
-7.97650933e-01 -4.58696961e-01 -1.27281034e+00 3.25106770e-01
-2.09842622e-01 1.40481927e-02 5.05433559e-01 1.08379292e+00
2.53250450e-01 2.34044254e-01 6.78416073e-01 -7.50197053e-01
-6.02052212e-01 -1.07372928e+00 -7.82576025e-01 5.49179375e-01
1.26451477e-01 -9.75442708e-01 -1.47760198e-01 1.87122315e-01] | [13.611403465270996, 1.6252211332321167] |
3499edc2-5bd4-4b29-bc8f-023149fc21c0 | entity-aware-syntax-tree-based-data | 2209.02267 | null | https://arxiv.org/abs/2209.02267v1 | https://arxiv.org/pdf/2209.02267v1.pdf | Entity Aware Syntax Tree Based Data Augmentation for Natural Language Understanding | Understanding the intention of the users and recognizing the semantic entities from their sentences, aka natural language understanding (NLU), is the upstream task of many natural language processing tasks. One of the main challenges is to collect a sufficient amount of annotated data to train a model. Existing research about text augmentation does not abundantly consider entity and thus performs badly for NLU tasks. To solve this problem, we propose a novel NLP data augmentation technique, Entity Aware Data Augmentation (EADA), which applies a tree structure, Entity Aware Syntax Tree (EAST), to represent sentences combined with attention on the entity. Our EADA technique automatically constructs an EAST from a small amount of annotated data, and then generates a large number of training instances for intent detection and slot filling. Experimental results on four datasets showed that the proposed technique significantly outperforms the existing data augmentation methods in terms of both accuracy and generalization ability. | ['Noboru Matsuda', 'Wenge Rong', 'Jiangneng Li', 'Jianbin Cui', 'Jiaxing Xu'] | 2022-09-06 | null | null | null | null | ['text-augmentation', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.71787006e-01 5.14914334e-01 -4.49454963e-01 -4.78515118e-01
-1.94328517e-01 -1.59499332e-01 5.65993965e-01 4.34456319e-01
-5.12599170e-01 8.51233602e-01 6.91429257e-01 -3.70083839e-01
5.12312472e-01 -8.18982720e-01 -4.43527460e-01 -1.20040044e-01
2.54051507e-01 5.70320606e-01 7.68205374e-02 -3.10423851e-01
8.49254206e-02 -1.67265326e-01 -1.29416728e+00 2.56262541e-01
1.25335729e+00 1.04481220e+00 1.87148005e-01 2.30230212e-01
-1.10117757e+00 1.08628368e+00 -3.92278492e-01 -3.17131817e-01
9.29080471e-02 -3.65573645e-01 -1.15175259e+00 1.43519975e-02
-8.32776278e-02 -3.48990887e-01 6.10797890e-02 8.41527522e-01
2.27900833e-01 2.86492735e-01 2.96702802e-01 -1.43752384e+00
-8.38098943e-01 8.13002467e-01 -4.31459963e-01 1.17667593e-01
2.39909843e-01 -4.30026233e-01 1.09019375e+00 -1.00417674e+00
4.70267296e-01 1.13802445e+00 6.04607642e-01 8.11474204e-01
-8.65293086e-01 -6.47507310e-01 3.52578372e-01 -4.25225534e-02
-1.03939307e+00 -2.71429926e-01 9.38462734e-01 -3.26268017e-01
1.03596103e+00 1.59362957e-01 2.63142914e-01 1.02103066e+00
-5.86982787e-01 1.17991328e+00 9.91464436e-01 -8.14415574e-01
1.77012444e-01 5.21257520e-01 9.26130056e-01 4.93396014e-01
2.63392359e-01 -4.63593066e-01 -1.71954438e-01 -2.77995348e-01
5.62036633e-01 5.54202944e-02 -6.44855723e-02 -1.30422845e-01
-1.12527347e+00 8.16003144e-01 3.39421988e-01 4.14165318e-01
-5.94201982e-01 -2.71531671e-01 7.01899171e-01 -5.71668185e-02
6.30449712e-01 6.30410552e-01 -9.58930314e-01 -1.62643850e-01
-2.05024928e-01 9.27048475e-02 7.85630941e-01 9.97335374e-01
8.20165813e-01 -4.46287915e-02 -3.80070776e-01 1.00018740e+00
2.70838171e-01 3.39022249e-01 9.53691244e-01 -8.16666111e-02
8.86987746e-01 1.52807081e+00 2.57168531e-01 -8.34801555e-01
-5.24681985e-01 -1.53946146e-01 -8.22384894e-01 -4.79462743e-01
1.68741316e-01 -5.51998496e-01 -1.13965356e+00 2.00501752e+00
4.75514114e-01 2.14295432e-01 6.18516862e-01 5.16332030e-01
1.06531453e+00 9.65109527e-01 9.11474288e-01 -2.98294783e-01
1.78784728e+00 -1.05234182e+00 -1.27416873e+00 -9.31489348e-01
1.29127216e+00 -4.48687583e-01 1.39939880e+00 -3.39438617e-01
-4.89881128e-01 -5.80119550e-01 -7.94824362e-01 -3.04434538e-01
-8.04576695e-01 5.19259334e-01 8.79019558e-01 6.56319439e-01
-3.63816708e-01 -1.34833664e-01 -4.91632074e-01 -4.18641865e-01
4.02409434e-01 3.29513550e-01 -5.50064862e-01 5.61602637e-02
-1.53686261e+00 6.89073145e-01 1.10635579e+00 -7.67960995e-02
-2.36281399e-02 -6.20395184e-01 -1.46091080e+00 1.57666117e-01
6.79694712e-01 -6.23313069e-01 1.37703228e+00 -1.03764451e+00
-9.44264948e-01 8.29361796e-01 -4.01803732e-01 -7.49065757e-01
-1.90026134e-01 -5.84239900e-01 -5.14616370e-01 -4.82518643e-01
1.57084852e-01 7.12229550e-01 5.16766965e-01 -1.22119999e+00
-6.16044164e-01 -5.78588605e-01 -6.65120110e-02 4.28421438e-01
-8.74493420e-01 1.35516763e-01 -1.50659725e-01 -7.67504990e-01
1.98336405e-04 -5.16395211e-01 -5.24082005e-01 -6.51047111e-01
-5.90360880e-01 -6.97508931e-01 1.13571429e+00 -9.12497222e-01
1.70629752e+00 -1.90844417e+00 -2.25319073e-01 -6.43265173e-02
1.38068989e-01 6.93212211e-01 -7.77205974e-02 2.90863901e-01
-2.86618322e-01 2.75888234e-01 -4.23273861e-01 -4.12900925e-01
-4.97976765e-02 4.08176810e-01 -5.26518464e-01 -5.45934618e-01
2.98876792e-01 1.08276868e+00 -8.56501222e-01 -4.30610120e-01
7.38189295e-02 1.27528146e-01 -5.29205322e-01 4.58409011e-01
-5.40662169e-01 2.80412972e-01 -9.25125122e-01 4.34391946e-01
5.23526311e-01 -4.33466941e-01 1.00328133e-01 -1.44126549e-01
4.58869003e-02 6.89638734e-01 -1.15271986e+00 1.39737117e+00
-5.67290783e-01 2.76725382e-01 -3.40148091e-01 -9.71632481e-01
1.07619357e+00 5.13708532e-01 4.04933006e-01 -6.49008989e-01
4.44733590e-01 1.15941152e-01 -1.81126326e-01 -6.89566553e-01
6.83706343e-01 -2.08612531e-01 -4.44916904e-01 4.24151450e-01
-6.60359785e-02 5.24385273e-01 1.73463300e-01 3.33853573e-01
9.48693454e-01 -2.20484614e-01 8.91673148e-01 -2.57230289e-02
8.56527269e-01 1.92649454e-01 7.52956212e-01 4.88083541e-01
-2.25426942e-01 9.47868749e-02 3.10965300e-01 -5.54218948e-01
-1.22630775e+00 -3.06439996e-01 3.89780700e-02 1.22457409e+00
-4.13153023e-02 -5.86925149e-01 -9.35961545e-01 -1.16677225e+00
-1.92340016e-01 9.71206725e-01 -6.49554312e-01 -1.88058376e-01
-5.06255090e-01 -8.38783324e-01 2.09742084e-01 9.33017790e-01
9.53307748e-01 -1.48672986e+00 -1.37159392e-01 2.44185567e-01
-5.65424860e-01 -1.70643437e+00 -3.00082982e-01 2.52396576e-02
-6.69115841e-01 -8.54654372e-01 -2.27410316e-01 -1.05160213e+00
9.13420379e-01 2.02302605e-01 9.86157477e-01 1.05327941e-01
4.65000831e-02 4.21135612e-02 -6.97910666e-01 -8.29488993e-01
-5.22495449e-01 1.90667838e-01 1.00038186e-01 1.60092637e-01
1.08649075e+00 -2.82945931e-01 -1.19933546e-01 4.27154005e-02
-9.32675421e-01 5.14373481e-01 7.75109351e-01 8.28858256e-01
4.16422278e-01 1.05613045e-01 9.91557240e-01 -1.57015824e+00
7.72094250e-01 -4.90613014e-01 -1.62713170e-01 3.06702852e-01
-5.10694087e-01 1.91879362e-01 7.28598714e-01 -2.80299097e-01
-1.36624742e+00 2.72258222e-01 -4.03135538e-01 1.75879642e-01
-5.25781810e-01 7.32404470e-01 -5.77061772e-01 4.34290916e-01
6.23297989e-01 3.27145457e-01 -2.40251228e-01 -6.80913091e-01
4.34832186e-01 1.01025903e+00 4.21936482e-01 -4.07689333e-01
7.59869754e-01 1.62209228e-01 -5.12247086e-01 -9.44529355e-01
-1.36388218e+00 -8.13988090e-01 -8.20572078e-01 3.45660388e-01
1.03471804e+00 -9.12866950e-01 -2.37934396e-01 4.04758036e-01
-1.22056198e+00 -7.21694827e-02 -5.60146570e-01 3.54983032e-01
-1.06642149e-01 3.12917203e-01 -2.73474932e-01 -1.10969472e+00
-7.82859147e-01 -5.44455409e-01 9.36056912e-01 5.05303681e-01
-2.74886936e-01 -9.14828479e-01 -9.12662819e-02 7.11017251e-01
2.42824599e-01 -3.08553334e-02 1.05093014e+00 -1.38640952e+00
-2.04344481e-01 -4.52205360e-01 -4.90738094e-01 3.56833428e-01
3.84886086e-01 -6.46963060e-01 -1.01646650e+00 2.05269396e-01
8.07990581e-02 -4.19383019e-01 5.42662799e-01 -5.86111145e-03
1.20311403e+00 -7.58284688e-01 -3.74707729e-01 -3.23468484e-02
1.09672630e+00 4.29620713e-01 6.64715946e-01 3.88201088e-01
8.37965608e-01 5.95704198e-01 8.25957119e-01 4.45487976e-01
6.11200154e-01 5.16987324e-01 1.71647564e-01 -3.39363754e-01
1.52907044e-01 -6.52730584e-01 7.86577761e-02 8.54104102e-01
2.25270376e-01 -3.02792639e-01 -1.02269828e+00 6.87372625e-01
-1.88297820e+00 -5.88780880e-01 -9.94925722e-02 1.90228438e+00
1.01656556e+00 1.52406663e-01 -4.36941423e-02 3.22773278e-01
6.21967435e-01 2.00466299e-03 -3.63848448e-01 -1.17969729e-01
3.43995020e-02 3.93828377e-03 3.26722771e-01 4.00260091e-01
-1.39084184e+00 1.29506552e+00 5.29472065e+00 5.42195201e-01
-6.34005606e-01 4.82541025e-02 6.35172367e-01 7.75889575e-01
-2.13979229e-01 7.54562169e-02 -1.16852665e+00 4.34972584e-01
8.63041103e-01 -2.15050906e-01 -9.62862745e-02 1.10090435e+00
2.46340260e-01 8.84143561e-02 -7.79209077e-01 7.38224506e-01
1.31186470e-01 -1.25230551e+00 3.61155808e-01 -1.54712394e-01
5.77177703e-01 -3.34681600e-01 -3.91440362e-01 7.55784333e-01
3.27661306e-01 -8.04721475e-01 1.74042527e-02 3.98210511e-02
3.86388153e-01 -4.81115907e-01 1.03362525e+00 7.30457127e-01
-1.15857899e+00 -2.27366716e-01 -2.17784986e-01 -3.49525362e-01
3.11310917e-01 3.65775853e-01 -1.25576162e+00 2.66494930e-01
4.47380811e-01 5.66415071e-01 -6.15465343e-01 7.11787343e-01
-2.67714679e-01 7.17712641e-01 -3.44313025e-01 -1.83321744e-01
2.53478050e-01 -5.48220649e-02 3.37925315e-01 1.04535520e+00
-3.51113863e-02 5.14892101e-01 4.43075508e-01 6.33977830e-01
-4.80464846e-01 6.89693749e-01 -7.03752279e-01 -4.90520388e-01
5.71826100e-01 1.25352788e+00 -4.58531767e-01 -6.08971179e-01
-7.26020336e-01 9.48550045e-01 4.13297087e-01 1.45918041e-01
-4.56800163e-01 -3.76870066e-01 3.67565155e-01 -8.17163065e-02
-1.84340522e-01 1.36522027e-02 -5.61445177e-01 -1.32654834e+00
2.55574316e-01 -6.05044484e-01 7.92857051e-01 -8.64082932e-01
-1.17705619e+00 6.64395034e-01 -1.86641484e-01 -1.08489001e+00
-1.86288834e-01 -5.49118936e-01 -5.74239314e-01 8.04145634e-01
-1.56480861e+00 -1.39795220e+00 -4.46240306e-01 3.47383738e-01
9.35861409e-01 -1.54857859e-01 1.03263974e+00 5.07029176e-01
-9.98985648e-01 4.31054235e-01 -4.37304616e-01 6.03736401e-01
3.15886259e-01 -1.30125332e+00 6.77718580e-01 9.24204946e-01
1.59108236e-01 6.87566817e-01 6.38956964e-01 -8.18267047e-01
-9.12384868e-01 -1.17821038e+00 1.49636471e+00 -5.85722208e-01
6.41746581e-01 -3.60110730e-01 -1.33673930e+00 1.10359848e+00
1.72819808e-01 7.49334414e-03 9.67794061e-01 3.78350466e-01
-1.49261788e-01 3.26838076e-01 -1.10096192e+00 5.34694135e-01
8.70882869e-01 -3.25142890e-01 -1.16172671e+00 3.28828365e-01
1.11774945e+00 -3.58639508e-01 -5.60293853e-01 5.43743193e-01
-5.55990152e-02 -2.16289043e-01 8.89276266e-01 -1.31271851e+00
3.28121752e-01 -2.82460004e-01 1.34877663e-03 -1.03787053e+00
4.62581171e-03 -2.92611480e-01 -4.22825515e-01 1.75089335e+00
5.81257939e-01 -5.11727035e-01 9.90062237e-01 1.08042109e+00
-1.83249265e-01 -5.38594186e-01 -5.07736802e-01 -4.71792847e-01
-3.29510927e-01 -3.95983398e-01 6.25910401e-01 1.33977246e+00
2.69268125e-01 1.13036931e+00 -5.46628118e-01 2.13627249e-01
1.60952345e-01 -6.13923185e-02 8.82429063e-01 -1.39909959e+00
2.81666905e-01 -7.17213284e-03 -2.29411066e-01 -1.18265140e+00
3.55856597e-01 -8.19491684e-01 -7.84220472e-02 -1.74920201e+00
2.69977987e-01 -6.33555293e-01 -7.21220449e-02 1.00286305e+00
-7.42873073e-01 -2.58185863e-01 -1.13444850e-01 7.00199679e-02
-5.08122623e-01 9.78792846e-01 9.09503698e-01 8.19414705e-02
-6.65803432e-01 1.40876919e-01 -1.09060216e+00 1.00703859e+00
8.54437470e-01 -2.31725276e-01 -5.50947368e-01 -2.54021406e-01
3.48022836e-03 -2.65139610e-01 5.73574454e-02 -8.15980732e-01
4.53657424e-03 -1.76779866e-01 9.93053988e-02 -6.54617548e-01
1.62015721e-01 -1.21718633e+00 -5.66243708e-01 2.02764884e-01
-5.41645408e-01 -6.16728105e-02 4.29524243e-01 5.02613544e-01
-3.41133416e-01 -4.06550914e-01 4.91215914e-01 2.32337248e-02
-1.34678400e+00 2.67721891e-01 -1.47913724e-01 3.28746706e-01
1.11598003e+00 -3.94405834e-02 -4.77303714e-01 -1.55124124e-02
-6.97631001e-01 5.51964581e-01 -1.01379953e-01 7.01393604e-01
3.58116806e-01 -1.43550014e+00 -3.54979008e-01 3.71464014e-01
5.14154851e-01 2.08342880e-01 2.93368757e-01 3.65901768e-01
-1.49550766e-01 5.94241381e-01 -6.41355710e-03 -5.41104078e-02
-1.20454156e+00 6.13975704e-01 4.24647816e-02 -6.63918316e-01
-6.59995079e-01 6.77961111e-01 3.27391863e-01 -7.29363382e-01
2.53924698e-01 -3.61973882e-01 -8.57444942e-01 -1.89652462e-02
8.39560866e-01 -1.34897262e-01 -1.65772159e-02 -7.64308929e-01
-4.04325351e-02 6.17931932e-02 -4.01651621e-01 2.83481389e-01
1.18428719e+00 -2.24226043e-01 -4.64292243e-02 2.54358590e-01
9.47955847e-01 -3.10000889e-02 -6.72233105e-01 -7.54685938e-01
4.58907962e-01 -3.00574273e-01 -3.39093171e-02 -8.23415697e-01
-6.19189084e-01 7.33897269e-01 3.75382036e-01 3.64117175e-01
1.07306039e+00 2.05685534e-02 1.28819788e+00 6.97289765e-01
-5.11458181e-02 -1.11694813e+00 -3.44303921e-02 7.42343903e-01
6.24108613e-01 -1.60829234e+00 -3.91292244e-01 -9.97484028e-01
-9.79547262e-01 7.02250123e-01 1.34054971e+00 3.95783961e-01
5.91173410e-01 1.05207734e-01 1.35697022e-01 -9.96706188e-02
-4.20300096e-01 -4.92201686e-01 2.99694777e-01 5.23040533e-01
5.27063191e-01 -1.65471196e-01 -4.88824815e-01 1.20609236e+00
-1.14559568e-01 1.06348405e-02 2.54421055e-01 1.07325017e+00
-7.15686619e-01 -1.30731130e+00 -8.92421901e-02 6.74708664e-01
-3.86548460e-01 -3.84012848e-01 -4.97084856e-01 7.37224579e-01
6.52582571e-02 5.79006255e-01 4.54516448e-02 -3.67314309e-01
5.34646928e-01 6.32101536e-01 -3.72905523e-01 -1.05992305e+00
-3.59150916e-01 -4.88537043e-01 3.29983205e-01 -5.79031706e-02
-4.52383429e-01 -4.55545992e-01 -1.67818654e+00 2.82369673e-01
-4.11070198e-01 5.66756368e-01 4.65705782e-01 1.35864615e+00
5.29332697e-01 4.77572143e-01 5.37028611e-01 -1.21413663e-01
-1.25805855e-01 -1.24187100e+00 -1.88397512e-01 7.82278121e-01
1.65122598e-01 -6.09248579e-01 -2.02631000e-02 1.35188103e-01] | [9.823588371276855, 9.151368141174316] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.