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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
04004133-b6df-4c48-a8b8-af6dedff6999 | end-to-end-adversarial-shape-learning-for | 1910.06474 | null | https://arxiv.org/abs/1910.06474v1 | https://arxiv.org/pdf/1910.06474v1.pdf | End-to-End Adversarial Shape Learning for Abdomen Organ Deep Segmentation | Automatic segmentation of abdomen organs using medical imaging has many potential applications in clinical workflows. Recently, the state-of-the-art performance for organ segmentation has been achieved by deep learning models, i.e., convolutional neural network (CNN). However, it is challenging to train the conventional CNN-based segmentation models that aware of the shape and topology of organs. In this work, we tackle this problem by introducing a novel end-to-end shape learning architecture -- organ point-network. It takes deep learning features as inputs and generates organ shape representations as points that located on organ surface. We later present a novel adversarial shape learning objective function to optimize the point-network to capture shape information better. We train the point-network together with a CNN-based segmentation model in a multi-task fashion so that the shared network parameters can benefit from both shape learning and segmentation tasks. We demonstrate our method with three challenging abdomen organs including liver, spleen, and pancreas. The point-network generates surface points with fine-grained details and it is found critical for improving organ segmentation. Consequently, the deep segmentation model is improved by the introduced shape learning as significantly better Dice scores are observed for spleen and pancreas segmentation. | ['Jinzheng Cai', 'Dong Yang', 'Holger Roth', 'Lin Yang', 'Daguang Xu', 'Yingda Xia'] | 2019-10-15 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [-5.58087081e-02 4.41653430e-01 1.28481358e-01 -5.25264919e-01
-7.15744495e-01 -6.85258150e-01 1.31411314e-01 4.75140661e-01
-2.03626454e-01 3.59542668e-01 4.82288115e-02 -3.09438556e-01
1.47015154e-01 -8.64321172e-01 -9.07043993e-01 -7.25486696e-01
-1.21799745e-01 7.26839125e-01 1.56510770e-01 -1.76588565e-01
-7.15418011e-02 7.87565947e-01 -2.91375756e-01 1.80096686e-01
1.09832931e+00 9.81688917e-01 -3.53312880e-01 7.25511253e-01
-2.00969368e-01 3.94917727e-01 -2.80011445e-01 -3.90244395e-01
5.71446002e-01 -4.94903088e-01 -8.67392123e-01 6.90410584e-02
1.97051823e-01 -2.10829854e-01 -3.61140817e-03 8.38408470e-01
7.37495959e-01 -1.51021838e-01 8.63825917e-01 -7.82771349e-01
-6.21961653e-01 6.67882144e-01 -4.34107840e-01 -2.24468708e-02
-4.26958613e-02 4.73565489e-01 4.43345189e-01 -5.66044807e-01
2.92720407e-01 1.16486180e+00 1.30911529e+00 4.53700900e-01
-1.16796434e+00 -4.16172355e-01 -3.06172371e-01 -7.43888199e-01
-1.01667821e+00 1.17703058e-01 7.38731086e-01 -5.64140975e-01
3.91613841e-01 9.34031382e-02 8.27596962e-01 7.06671000e-01
5.35736084e-01 9.01355088e-01 7.61063337e-01 -1.35019392e-01
-2.15560459e-02 -2.90810019e-01 -7.26831332e-02 8.77171755e-01
3.25872391e-01 1.36150762e-01 4.30904776e-01 -9.15545002e-02
1.37824905e+00 3.52817893e-01 -2.05213010e-01 -7.66561508e-01
-1.49500048e+00 8.69980514e-01 1.20129621e+00 2.71300107e-01
-6.60199225e-01 4.22772259e-01 5.40361822e-01 -6.81742728e-02
3.28902900e-01 6.80753231e-01 -5.20564497e-01 3.73778462e-01
-8.93081069e-01 1.61302567e-01 8.76040876e-01 7.08740413e-01
3.44430864e-01 3.46902683e-02 -6.06005967e-01 6.45328760e-01
5.51671267e-01 3.73748720e-01 2.78548688e-01 -6.06382728e-01
2.96811670e-01 9.27346826e-01 -2.98972338e-01 -7.60981739e-01
-1.05038023e+00 -6.44058585e-01 -1.14802361e+00 1.83267251e-01
7.18374610e-01 -4.52250987e-01 -1.34483099e+00 1.32342064e+00
6.55022562e-01 2.28518501e-01 4.70224619e-02 1.12322724e+00
1.29113019e+00 3.91960621e-01 3.32293898e-01 4.59525228e-01
1.40231240e+00 -1.11358094e+00 -4.27182287e-01 1.57000124e-01
7.34236181e-01 -6.49521589e-01 6.98040783e-01 -4.49878015e-02
-1.36222982e+00 -4.07945156e-01 -7.82317042e-01 7.08715469e-02
-1.51683226e-01 6.83723241e-02 5.14603257e-01 5.54744601e-01
-9.69608963e-01 7.58428931e-01 -1.25110734e+00 -3.31945986e-01
9.93133068e-01 6.51720881e-01 -2.04485103e-01 7.34326392e-02
-7.65445054e-01 8.37130964e-01 3.93154085e-01 4.18532118e-02
-8.32078874e-01 -1.17859352e+00 -1.06241310e+00 2.16241598e-01
1.15803378e-02 -1.29221976e+00 1.14112949e+00 -9.59396362e-01
-1.72502291e+00 1.00977862e+00 5.76543808e-01 -5.30949652e-01
8.31557989e-01 2.05862120e-01 2.97188640e-01 3.33615899e-01
-1.94577798e-01 8.44350934e-01 6.74005032e-01 -1.42796195e+00
6.95031509e-03 -3.88059497e-01 -3.49288404e-01 7.32967257e-02
1.50596529e-01 -2.01190010e-01 -3.50990534e-01 -8.67101073e-01
2.87701607e-01 -1.12606573e+00 -5.57160258e-01 5.70362806e-01
-5.62539637e-01 1.22391529e-01 4.91102576e-01 -7.43486345e-01
6.31902277e-01 -1.88761997e+00 -5.04971929e-02 4.33621615e-01
3.95340174e-01 4.38038886e-01 -1.32130831e-01 -4.80689891e-02
-4.35392819e-02 1.29972622e-01 -6.43873036e-01 -2.47089058e-01
-2.53870070e-01 2.25301921e-01 3.58702391e-01 7.13445425e-01
3.55493426e-01 1.58321774e+00 -9.77937400e-01 -6.58681273e-01
4.78568703e-01 6.60362303e-01 -6.98240221e-01 4.28904563e-01
-9.87378880e-02 1.09506321e+00 -5.19380391e-01 8.13183725e-01
9.65620875e-01 -5.25062621e-01 -5.61313406e-02 -3.09666693e-01
1.65178984e-01 -3.51423293e-01 -5.97148061e-01 1.99176669e+00
-3.71332914e-01 -2.16197409e-02 2.63310492e-01 -9.13666964e-01
1.00741017e+00 5.41473806e-01 9.76264298e-01 -4.23716605e-01
4.61794138e-01 3.74708921e-01 2.31021926e-01 -5.26691973e-01
-1.48559675e-01 -5.60569227e-01 -5.63895591e-02 1.37853891e-01
1.72423586e-01 -4.50747281e-01 -2.17306554e-01 -2.91386664e-01
7.46897995e-01 1.84249029e-01 2.75448680e-01 -5.50431907e-01
5.44182837e-01 1.46492481e-01 3.25689137e-01 4.22667980e-01
-4.30967242e-01 1.17806089e+00 3.73630852e-01 -8.30048323e-01
-1.13038504e+00 -1.19394112e+00 -4.58700001e-01 8.18065763e-01
1.66564226e-01 3.65997195e-01 -9.05377209e-01 -9.94211614e-01
2.74629563e-01 2.14579344e-01 -7.42426395e-01 -1.87566578e-02
-8.72783542e-01 -8.14810157e-01 9.14991200e-01 7.83328056e-01
2.87502706e-01 -1.32809305e+00 -6.34261489e-01 3.84666800e-01
-4.29986306e-02 -8.98861587e-01 -6.55536532e-01 1.46458954e-01
-1.27715933e+00 -1.25484717e+00 -1.39740908e+00 -8.46341789e-01
1.02311695e+00 -4.34767216e-01 1.33298290e+00 3.23977202e-01
-4.68154907e-01 2.57904798e-01 -2.18768880e-01 -5.79129875e-01
-7.38126457e-01 1.65613458e-01 -5.18921018e-01 -9.99013260e-02
-1.91059574e-01 -2.89518178e-01 -1.11180985e+00 2.15053439e-01
-9.90957856e-01 2.13153362e-02 7.96446681e-01 8.87906849e-01
7.72192478e-01 -7.02139914e-01 4.61684734e-01 -1.01067781e+00
4.79353577e-01 -5.29090703e-01 -4.29329097e-01 2.61400878e-01
-2.11979419e-01 6.02203645e-02 6.40834510e-01 -2.18677670e-01
-6.73408568e-01 3.04280281e-01 -4.92189080e-01 -5.03296018e-01
-1.73140377e-01 4.42605704e-01 4.13254142e-01 -4.93818969e-01
7.18186915e-01 9.67145264e-02 4.02161658e-01 -2.91429013e-01
2.21754134e-01 2.50542611e-02 4.47914004e-01 -5.37957132e-01
5.75000048e-01 5.03365934e-01 5.44707894e-01 -3.43299687e-01
-6.41278565e-01 -5.38940728e-01 -9.49531376e-01 -4.78918850e-02
1.23661363e+00 -7.60679722e-01 -7.01385677e-01 4.10520911e-01
-1.10515058e+00 -5.65266311e-01 -4.45252210e-01 4.18067187e-01
-6.24546111e-01 4.51060206e-01 -8.37763429e-01 -2.27369756e-01
-1.04411399e+00 -1.34649360e+00 1.16255605e+00 3.20796013e-01
-5.94577715e-02 -1.59514523e+00 2.06688598e-01 3.48343179e-02
7.90167212e-01 1.11835229e+00 8.05460095e-01 -1.07628131e+00
-4.09530103e-01 -2.03314900e-01 -4.77332056e-01 2.29470640e-01
5.30466177e-02 -2.25951046e-01 -5.90896845e-01 -4.59138960e-01
-6.09937795e-02 -2.13588312e-01 6.69105232e-01 1.02415860e+00
1.50096405e+00 -2.32480377e-01 -3.06743234e-01 1.31586361e+00
1.50504661e+00 -2.97224037e-02 4.64300662e-01 8.08489546e-02
9.21014726e-01 3.07435483e-01 1.14535064e-01 3.63172352e-01
3.55870962e-01 2.84775257e-01 7.20201313e-01 -9.57337916e-01
-2.17461288e-01 -7.64089227e-02 -3.36834282e-01 6.71750009e-01
4.26935777e-03 -3.91782075e-02 -1.29714191e+00 5.09359241e-01
-1.71897733e+00 -2.96144664e-01 -3.34396899e-01 1.83907616e+00
8.43960762e-01 -2.71166533e-01 1.75382242e-01 -3.84919554e-01
5.02084792e-01 -2.81888306e-01 -5.80586791e-01 -3.18689853e-01
3.59574944e-01 3.61353546e-01 5.37877738e-01 2.70635307e-01
-1.40567446e+00 5.16182959e-01 6.13447189e+00 3.39561582e-01
-1.40029442e+00 -2.17335746e-02 9.43049908e-01 3.48470837e-01
-1.09114148e-01 -3.96637499e-01 -2.85936207e-01 4.49879140e-01
4.60088611e-01 3.51547927e-01 1.75129306e-02 7.23565280e-01
7.58352652e-02 3.14944535e-01 -1.17616749e+00 8.74329269e-01
-1.84640251e-02 -1.29799926e+00 2.95648389e-02 -1.21324807e-01
9.13655996e-01 2.46479720e-01 -7.01596811e-02 3.30417484e-01
2.18428910e-01 -1.40985751e+00 3.27453941e-01 5.66768289e-01
8.07095408e-01 -5.30016065e-01 1.06387210e+00 2.87782788e-01
-1.05008709e+00 1.87111542e-01 -2.35963419e-01 5.86166203e-01
1.91526756e-01 4.94436115e-01 -1.10559738e+00 6.79358304e-01
3.49322140e-01 4.95254159e-01 -5.21066546e-01 1.64478970e+00
7.64166415e-02 4.09036994e-01 -4.67442632e-01 2.47399718e-01
5.54881871e-01 -3.88980269e-01 4.01812017e-01 1.54419553e+00
4.06959355e-01 1.28741711e-01 4.84426439e-01 1.25389946e+00
-3.17718804e-01 2.79459238e-01 -4.35257822e-01 3.05869460e-01
-2.13910803e-01 1.56687939e+00 -1.16438293e+00 -3.31261784e-01
-1.53762519e-01 7.24421501e-01 8.01516883e-03 3.55213694e-02
-8.76576304e-01 -1.37342095e-01 1.90401539e-01 1.63975000e-01
1.31011382e-01 1.21632300e-01 -7.72226691e-01 -1.16183305e+00
-4.31430697e-01 -5.30611217e-01 5.66889763e-01 -3.64355564e-01
-1.58043897e+00 4.92014557e-01 -2.31463522e-01 -1.07760620e+00
1.06046326e-01 -6.20875597e-01 -1.07686961e+00 9.91012096e-01
-1.78685379e+00 -1.52032518e+00 -6.46152437e-01 5.20508945e-01
2.87927479e-01 2.42748380e-01 8.55831325e-01 1.66746050e-01
-1.50020763e-01 6.25685692e-01 3.32655460e-02 8.04390728e-01
5.83507419e-01 -1.58253098e+00 3.94356877e-01 4.00487542e-01
-3.88429314e-01 5.15064955e-01 2.09977999e-01 -6.44934058e-01
-1.30293632e+00 -1.46141851e+00 3.08141470e-01 -5.66970944e-01
1.41112387e-01 6.64086565e-02 -8.49764049e-01 6.60923839e-01
2.18865201e-01 8.73246551e-01 8.03331017e-01 -3.37539077e-01
1.76891565e-01 2.39217117e-01 -1.69944024e+00 2.68549681e-01
5.88622272e-01 1.25821590e-01 -5.67841709e-01 4.15191293e-01
5.69873631e-01 -9.76689517e-01 -1.40681171e+00 7.99253047e-01
4.36974615e-01 -6.69938028e-01 1.38835108e+00 -5.85313618e-01
5.79154670e-01 -1.39147460e-01 5.90687752e-01 -1.56748438e+00
-2.31006965e-01 -4.50158805e-01 -8.07069466e-02 6.59688652e-01
3.18949342e-01 -5.18179834e-01 8.70425165e-01 5.88100851e-01
-4.90193814e-01 -1.21548879e+00 -6.80003643e-01 -4.21694726e-01
7.80327439e-01 1.71139196e-01 6.69310510e-01 1.01474512e+00
-3.58078182e-01 -4.82567027e-03 2.25071073e-01 -3.37106772e-02
7.69594491e-01 2.12156162e-01 6.82665229e-01 -1.28568089e+00
5.86410165e-02 -6.14710689e-01 -1.26290575e-01 -9.57175076e-01
-3.09176952e-01 -1.34467101e+00 7.79562816e-02 -1.93244720e+00
1.48300007e-01 -7.09113181e-01 -4.25201237e-01 5.63146472e-01
-4.04126823e-01 2.85261422e-01 3.64881396e-01 1.05299063e-01
-5.42169094e-01 4.21155810e-01 2.21114492e+00 -1.46843538e-01
-3.24355304e-01 3.73922795e-01 -4.81614381e-01 7.28925645e-01
8.24538708e-01 -5.24698734e-01 9.93997529e-02 -3.61818850e-01
-1.71171322e-01 4.46163565e-01 5.21738827e-01 -9.42672551e-01
2.04449058e-01 3.04721475e-01 9.01994824e-01 -6.26766741e-01
-1.19332269e-01 -1.14006090e+00 -3.95351797e-02 9.92226303e-01
-3.21231097e-01 -8.44849423e-02 3.16949040e-01 2.56946772e-01
-9.58496705e-02 -1.42294943e-01 1.26331460e+00 -5.31441152e-01
8.36885497e-02 8.41821790e-01 1.64923873e-02 2.35112861e-01
1.15934098e+00 -1.80078223e-01 2.52154648e-01 3.88320610e-02
-9.64536369e-01 4.08427805e-01 3.09326321e-01 -3.41900028e-02
6.54547334e-01 -1.32950461e+00 -9.63021100e-01 2.96241522e-01
-2.57603154e-02 9.10610616e-01 1.36591166e-01 1.27779138e+00
-1.32510519e+00 2.74392754e-01 -3.93987536e-01 -1.05708683e+00
-8.15773785e-01 4.87657338e-01 8.46113384e-01 -7.19643950e-01
-8.71805310e-01 7.60005236e-01 3.87122840e-01 -1.11486864e+00
9.53634828e-03 -9.40690637e-01 -2.10640505e-01 -3.84468377e-01
-5.33652045e-02 1.15731739e-01 1.44192772e-02 -4.11559314e-01
-2.30289653e-01 7.91644335e-01 3.07796508e-01 5.84829092e-01
1.48985016e+00 1.84528530e-01 -1.64329425e-01 -1.47307277e-01
1.25292623e+00 -3.82757455e-01 -1.49845600e+00 -1.80245191e-01
-1.95868716e-01 -2.36276507e-01 -1.10643037e-01 -1.04622757e+00
-1.48969603e+00 1.02087665e+00 6.75265968e-01 7.76041895e-02
8.40284109e-01 -2.47977033e-01 1.25321651e+00 2.64973938e-02
1.21151000e-01 -5.18056631e-01 2.36741632e-01 5.41501045e-01
8.64689529e-01 -1.59859717e+00 -6.24781996e-02 -4.80101794e-01
-5.87976992e-01 1.46481621e+00 4.53700751e-01 -5.80584586e-01
8.31751287e-01 3.39637935e-01 4.45603579e-01 -4.38023508e-01
1.24459371e-01 2.53727939e-02 6.85484707e-01 4.67301279e-01
5.89690924e-01 2.68880546e-01 -1.10206395e-01 7.35879660e-01
2.27434859e-01 8.56548622e-02 3.76511440e-02 7.68579781e-01
-1.63972110e-01 -9.44232881e-01 -5.54226816e-01 4.31218296e-01
-9.52872753e-01 -2.19657738e-02 -4.56149094e-02 7.39718139e-01
4.15229806e-06 1.70377597e-01 -8.65460336e-02 3.49224657e-01
3.25657964e-01 -6.15638457e-02 5.03070712e-01 -7.11513102e-01
-1.43067849e+00 1.02146454e-02 -6.29019916e-01 -5.18019855e-01
-2.37638727e-01 -3.00337821e-01 -1.61436057e+00 -1.26043200e-01
-1.16294354e-01 -2.02687830e-01 7.99189687e-01 7.07560897e-01
1.92967474e-01 9.63992536e-01 4.27128613e-01 -1.15564036e+00
-8.87128174e-01 -8.23756576e-01 -1.49768189e-01 8.64545107e-01
3.62666994e-01 -2.56773025e-01 1.13951251e-01 4.72453460e-02] | [14.482056617736816, -2.5530683994293213] |
ed317689-b186-46db-8278-f4bec91728fa | two-stream-convolutional-neural-network-for | 1612.07978 | null | http://arxiv.org/abs/1612.07978v1 | http://arxiv.org/pdf/1612.07978v1.pdf | Two-stream convolutional neural network for accurate RGB-D fingertip detection using depth and edge information | Accurate detection of fingertips in depth image is critical for
human-computer interaction. In this paper, we present a novel two-stream
convolutional neural network (CNN) for RGB-D fingertip detection. Firstly edge
image is extracted from raw depth image using random forest. Then the edge
information is combined with depth information in our CNN structure. We study
several fusion approaches and suggest a slow fusion strategy as a promising way
of fingertip detection. As shown in our experiments, our real-time algorithm
outperforms state-of-the-art fingertip detection methods on the public dataset
HandNet with an average 3D error of 9.9mm, and shows comparable accuracy of
fingertip estimation on NYU hand dataset. | ['Xinghao Chen', 'Hengkai Guo', 'Guijin Wang'] | 2016-12-23 | null | null | null | null | ['fingertip-detection'] | ['computer-vision'] | [-4.31493111e-02 -5.07471561e-01 -2.84131840e-02 -1.03307135e-01
-1.81445479e-01 -6.04346156e-01 5.22263497e-02 -4.51210976e-01
-6.99007094e-01 2.38545150e-01 -1.83043316e-01 -1.67747185e-01
7.20555112e-02 -8.04095149e-01 -4.42404956e-01 -3.76370996e-01
-3.39533016e-02 1.86639100e-01 6.08876467e-01 -7.54818469e-02
4.07180220e-01 1.10018921e+00 -1.49341679e+00 2.77951539e-01
1.65431812e-01 1.50148451e+00 1.84310216e-03 1.07714355e+00
8.08284506e-02 1.94509178e-01 -3.79818290e-01 -3.09502810e-01
7.68988073e-01 3.27600986e-01 -3.28684330e-01 -2.31578097e-01
6.75235987e-01 -1.15157104e+00 -7.54771709e-01 6.21132553e-01
1.36849344e+00 -5.14202714e-01 5.81664741e-01 -1.08024013e+00
-9.66680720e-02 -2.77549401e-02 -7.56555021e-01 6.01247363e-02
3.51757318e-01 1.77305058e-01 3.71203870e-01 -1.18156302e+00
6.82412982e-01 1.15184724e+00 5.88133693e-01 6.22332215e-01
-6.76804066e-01 -6.70421124e-01 -1.92742616e-01 7.08975643e-02
-1.46801507e+00 -2.60401759e-02 9.35676694e-01 -1.30962282e-01
1.06487060e+00 -2.22211424e-03 6.88892007e-01 1.01284885e+00
6.06252134e-01 1.21909428e+00 9.09411311e-01 -5.23088276e-01
-1.91099018e-01 -2.08488792e-01 6.32425472e-02 1.01023555e+00
4.34846431e-01 3.83786768e-01 -5.11003077e-01 1.03191053e-02
1.62954128e+00 1.06707849e-01 2.37463593e-01 -3.44707333e-02
-1.19927979e+00 1.95853442e-01 7.26005137e-01 7.27499500e-02
-7.41628528e-01 4.12858784e-01 2.32713982e-01 -3.52498032e-02
2.90329933e-01 -1.37625948e-01 -4.63509649e-01 -2.11947113e-01
-7.47675657e-01 4.69965845e-01 7.99682498e-01 9.77182031e-01
1.22907013e-01 -6.83102667e-01 -3.12982380e-01 3.85609239e-01
4.99072254e-01 6.60064399e-01 -3.25559437e-01 -8.79788339e-01
6.19686365e-01 6.09114885e-01 4.14374381e-01 -8.59572709e-01
-5.18229544e-01 -1.01177149e-01 -8.01829219e-01 8.96569431e-01
1.01145625e+00 -4.26796436e-01 -1.04126084e+00 8.15451622e-01
3.34343575e-02 -2.99940705e-01 -6.93244040e-01 1.15939891e+00
8.61966789e-01 -2.29513720e-01 -2.99171031e-01 6.41325951e-01
1.42293751e+00 -6.18084133e-01 -4.91231412e-01 -1.63103864e-01
-3.06428403e-01 -9.62181628e-01 7.84509242e-01 8.73204112e-01
-1.04275346e+00 -7.58315623e-01 -8.70742798e-01 -2.52085388e-01
-5.03644347e-01 7.30879188e-01 8.84825587e-01 9.20121491e-01
-7.77286708e-01 4.80069607e-01 -8.59198451e-01 -2.76597321e-01
6.14083827e-01 6.01575434e-01 -2.16611028e-01 2.81753708e-02
-8.45756590e-01 4.98587400e-01 5.69363609e-02 3.79672110e-01
-4.35373047e-03 -3.27786863e-01 -2.00118542e-01 -2.32920080e-01
6.80229664e-02 -7.53087938e-01 1.21948159e+00 2.85353363e-02
-1.50699627e+00 9.41647232e-01 -4.50236946e-01 7.45045952e-03
9.98269200e-01 -5.51100314e-01 -1.18622988e-01 2.29668140e-01
-3.90495181e-01 7.16657996e-01 1.12317157e+00 -1.15452266e+00
-9.65171695e-01 -8.80554438e-01 5.15140705e-02 -2.36268163e-01
-1.08439766e-01 1.85778782e-01 -7.53390610e-01 -4.54139829e-01
4.61369038e-01 -8.48662674e-01 -3.40764411e-02 4.72279489e-01
-7.01106429e-01 -4.29073364e-01 6.36193693e-01 -7.68472314e-01
1.18625021e+00 -1.57887566e+00 -2.76521534e-01 5.87108433e-01
6.71624601e-01 2.36238107e-01 1.56307429e-01 1.74940377e-01
3.90567631e-01 -3.40520054e-01 3.51350993e-01 -5.77648997e-01
8.16443563e-02 -2.33402535e-01 -6.83148503e-02 1.44135177e-01
1.07400917e-01 1.07045019e+00 -8.78960788e-01 -7.36982465e-01
6.11944079e-01 9.59919155e-01 -2.05588609e-01 1.19384296e-01
3.25428993e-01 2.50715166e-01 -4.12696153e-01 1.53957593e+00
9.71202672e-01 1.21665031e-01 -2.64485598e-01 -7.35614419e-01
-2.93668628e-01 -4.83299866e-02 -1.29575336e+00 1.72270918e+00
-2.32697532e-01 7.34334826e-01 -2.42672721e-03 1.50478363e-01
8.62333477e-01 1.30977944e-01 4.24700290e-01 -3.73252243e-01
6.40013516e-01 3.25397611e-01 -2.59260610e-02 -4.91209149e-01
3.86609197e-01 6.29006922e-01 2.65906155e-01 4.26254451e-01
-1.19276285e-01 2.42874011e-01 -2.19102964e-01 -1.46842092e-01
9.79839504e-01 1.91223204e-01 -1.90807939e-01 2.33814657e-01
1.60068721e-01 -2.78226882e-01 -7.50789046e-02 1.24566424e+00
-6.41971707e-01 7.97774613e-01 3.58173341e-01 -6.23757124e-01
-1.01003885e+00 -1.17797124e+00 -3.20710748e-01 7.66294360e-01
2.98982002e-02 -1.02225825e-01 -8.47135484e-01 -5.01331091e-01
5.68006158e-01 -4.29695278e-01 -7.58989155e-01 4.84011710e-01
-7.43379712e-01 -3.89512300e-01 7.40573227e-01 1.16502416e+00
1.07185197e+00 -1.00955987e+00 -1.14708948e+00 2.27201834e-01
2.31331050e-01 -1.21374714e+00 -3.35882574e-01 -8.50278046e-03
-7.58647144e-01 -1.24215662e+00 -1.23157811e+00 -7.85001278e-01
5.24891376e-01 2.31014147e-01 8.84891450e-01 -1.18075281e-01
-9.36436892e-01 3.80908549e-01 -2.38900125e-01 -8.67010653e-01
4.56926733e-01 3.69663596e-01 1.27653986e-01 -2.91230708e-01
8.98643792e-01 -2.64622986e-01 -1.35202217e+00 1.27524570e-01
-2.35825792e-01 -3.33012015e-01 8.38844657e-01 4.74865377e-01
4.64532346e-01 -2.69107759e-01 -3.06187198e-02 3.60697694e-02
9.80007410e-01 2.22687736e-01 -3.99459958e-01 2.38551095e-01
-3.71588081e-01 -1.81229621e-01 -6.32966980e-02 -3.24406713e-01
-8.19571674e-01 6.91866100e-01 -3.75143021e-01 -4.17441338e-01
-3.54829907e-01 -1.66469634e-01 1.88061222e-01 -6.03563070e-01
7.63779104e-01 -4.61640917e-02 1.45636827e-01 -7.53422320e-01
2.37912953e-01 1.36226606e+00 7.93831408e-01 -3.56381208e-01
3.13175619e-01 7.27582753e-01 3.30121875e-01 -7.41714060e-01
-2.77427971e-01 -5.00212371e-01 -1.35811722e+00 -5.64952374e-01
6.99784279e-01 -3.54835898e-01 -1.47641015e+00 1.27594852e+00
-1.49518442e+00 6.59084925e-03 4.21405852e-01 2.94925958e-01
-1.04266480e-01 4.05011684e-01 -7.32182622e-01 -1.23381126e+00
-8.32334936e-01 -8.07048023e-01 1.43754411e+00 3.99780840e-01
-1.63532421e-02 -6.40752435e-01 -2.58695900e-01 -2.38057330e-01
3.93279552e-01 3.70962948e-01 2.83440024e-01 1.70745894e-01
-5.37805974e-01 -1.20845938e+00 -9.34435785e-01 1.50620028e-01
5.15567899e-01 2.17908740e-01 -1.20358229e+00 -1.76648483e-01
-7.27459431e-01 3.75427231e-02 1.03365791e+00 7.48012781e-01
1.51614678e+00 5.94246209e-01 -5.46156049e-01 3.54771495e-01
1.28953505e+00 -5.63333705e-02 7.58836925e-01 2.23982513e-01
6.88566208e-01 4.10086483e-01 5.35502315e-01 8.43142390e-01
2.55327493e-01 3.75312269e-01 6.12302661e-01 -2.77201235e-01
-2.91132420e-01 6.33789226e-02 -3.25928003e-01 -5.51484451e-02
-9.56782281e-01 -6.13863952e-02 -9.50273991e-01 2.50110745e-01
-1.68979025e+00 -6.17818654e-01 -1.44187152e-01 2.16481113e+00
4.63424891e-01 3.74600619e-01 6.83505714e-01 5.42079568e-01
7.01665878e-01 -2.96531886e-01 -6.83861315e-01 7.02595785e-02
1.46496832e-01 8.09031546e-01 1.08346510e+00 2.17645779e-01
-1.39699137e+00 9.43666041e-01 6.94674492e+00 2.63780765e-02
-1.10509062e+00 -3.70325536e-01 9.81706530e-02 -1.67608678e-01
6.29866064e-01 -7.37117946e-01 -1.09521353e+00 3.29849988e-01
1.79888643e-02 5.96736610e-01 3.87362033e-01 7.65186906e-01
-6.42179623e-02 -2.58666992e-01 -9.25637007e-01 1.41103768e+00
-8.80032107e-02 -9.94774938e-01 -3.83269876e-01 1.00225590e-01
8.56641233e-02 1.90316468e-01 -1.97726116e-02 -4.19159323e-01
-1.99585065e-01 -9.34917510e-01 2.90767789e-01 9.92723227e-01
1.09306347e+00 -9.20090139e-01 8.60390961e-01 1.78047881e-01
-1.46020031e+00 -3.36663984e-02 -4.66306180e-01 -2.10399464e-01
-3.69432196e-03 5.85619926e-01 -7.64669418e-01 1.47257736e-02
1.01410806e+00 3.71520221e-01 -5.85043252e-01 1.18244708e+00
-1.52888805e-01 3.71009018e-03 -5.35844445e-01 -4.81690079e-01
-1.16611719e-01 3.82625341e-01 3.37803662e-01 1.45359910e+00
6.71417862e-02 4.22693752e-02 -2.22913086e-01 9.70460951e-01
4.46894728e-02 -4.76347268e-01 -3.65722716e-01 -6.03790544e-02
6.34708166e-01 1.37112451e+00 -7.96726286e-01 -2.67533600e-01
-2.61809111e-01 1.36907971e+00 -1.93943098e-01 2.32480168e-01
-3.94220173e-01 -1.51176453e+00 7.15106308e-01 2.88056314e-01
1.72342390e-01 -6.95365608e-01 -7.96831012e-01 -9.00869071e-01
3.58489960e-01 -3.13202143e-02 -8.99834335e-02 -6.82209134e-01
-1.17507958e+00 3.85867059e-01 -4.06607717e-01 -1.28205240e+00
-2.10647836e-01 -1.66677308e+00 -4.66741264e-01 1.17673123e+00
-1.49768257e+00 -1.23239446e+00 -1.23070931e+00 8.43836248e-01
1.75093874e-01 -3.35804045e-01 7.77353585e-01 2.94881523e-01
-3.78010005e-01 9.41181242e-01 -2.70017773e-01 8.63921046e-01
9.44479406e-01 -1.34813607e+00 1.24873412e+00 5.16476572e-01
-4.49158132e-01 8.74457181e-01 2.46513918e-01 -8.22085261e-01
-1.74912977e+00 -7.00270772e-01 8.07077050e-01 -6.99199021e-01
1.79255366e-01 -2.70212919e-01 -2.66183794e-01 4.25767064e-01
-1.38888240e-01 1.94476321e-01 2.63669342e-01 -2.14272235e-02
-4.47625369e-01 1.06059812e-01 -1.57855594e+00 5.04674315e-01
1.39015830e+00 -6.75980568e-01 -3.35423052e-01 1.31083906e-01
1.58335626e-01 -8.05759788e-01 -1.03159547e+00 4.66691434e-01
1.76640594e+00 -1.05975008e+00 1.43292212e+00 -3.27176690e-01
2.38506690e-01 -1.25905499e-03 -3.41099612e-02 -6.24245524e-01
-5.09135723e-01 -2.28351906e-01 -5.84047675e-01 5.18845320e-01
-1.78508341e-01 -5.99274099e-01 1.36382949e+00 6.02172256e-01
3.81835818e-01 -8.57707381e-01 -7.30347812e-01 -7.17935145e-01
-1.88127309e-01 -6.92616999e-01 4.13446307e-01 8.61539170e-02
-1.03550151e-01 -3.54410827e-01 -7.69563019e-02 5.69016971e-02
1.30239999e+00 -6.86050802e-02 7.75113761e-01 -1.46216488e+00
1.11661348e-02 -5.89019239e-01 -6.18967831e-01 -1.27504802e+00
-4.35642511e-01 -1.93847597e-01 -2.30293557e-01 -1.60551250e+00
-1.46539956e-01 -1.25971422e-01 -4.28907692e-01 5.87162673e-01
2.19010022e-02 7.44129360e-01 -4.93513457e-02 6.15635887e-02
6.07048981e-02 -5.00439763e-01 1.38434589e+00 -5.00584655e-02
-4.50879723e-01 3.67270648e-01 -1.54979393e-01 5.59519827e-01
9.08181369e-01 1.39906839e-01 2.77174681e-01 -3.21851224e-01
6.38631731e-02 -1.66192099e-01 9.27159131e-01 -1.07941234e+00
4.24302936e-01 2.18475506e-01 1.10572672e+00 -1.33405507e+00
3.53961170e-01 -5.72216272e-01 -8.36726129e-01 8.49842846e-01
1.61502048e-01 -2.17547387e-01 1.70910835e-01 -3.32843289e-02
3.56047273e-01 1.17672749e-01 4.97429639e-01 -3.39798093e-01
-7.34940290e-01 4.92594898e-01 -1.88636154e-01 -6.84652746e-01
4.40282643e-01 -7.20146596e-01 -2.04744205e-01 -3.02389786e-02
-6.08740151e-01 -1.90410584e-01 2.98370458e-02 7.63653696e-01
1.01839125e+00 -1.38970065e+00 -6.60751104e-01 4.44768310e-01
1.03473775e-01 -2.09084526e-01 -2.14873776e-01 5.77986419e-01
-7.48008311e-01 8.48846018e-01 -6.49460912e-01 -6.46499693e-01
-1.34316730e+00 4.67390567e-03 5.15290499e-01 4.20769095e-01
-5.38409770e-01 1.28903830e+00 -7.46104419e-01 -1.21791042e-01
8.59974861e-01 -7.40352690e-01 1.40379086e-01 -1.81860894e-01
7.66125739e-01 7.97324419e-01 3.47279370e-01 8.91988911e-03
-6.48590088e-01 9.73632157e-01 -9.40294117e-02 -7.37091303e-02
8.51463616e-01 1.86415181e-01 -1.36088774e-01 -1.20640290e-03
1.06121111e+00 -1.61116868e-01 -1.39358842e+00 -9.43754688e-02
-3.23118865e-01 -9.10624862e-01 2.70499289e-01 -1.07412934e+00
-8.21575403e-01 1.04870188e+00 1.39470124e+00 -4.52382714e-02
1.01761270e+00 -5.29191568e-02 1.11153698e+00 7.63267934e-01
7.43744016e-01 -1.35917735e+00 -1.45306624e-02 5.19142151e-01
1.03313565e+00 -1.50873518e+00 -5.65863512e-02 -5.19181788e-01
1.71401665e-01 1.60236526e+00 5.36038995e-01 -5.17557144e-01
9.83030856e-01 5.73093235e-01 1.89489331e-02 -1.99839935e-01
1.48117244e-01 -6.20448053e-01 5.40365100e-01 7.54419744e-01
5.01774490e-01 8.66896063e-02 -8.34623352e-02 5.33834040e-01
4.84524993e-03 5.80762386e-01 -1.92436710e-01 1.44996250e+00
-5.16851068e-01 -1.14862251e+00 -4.08465952e-01 6.25680864e-01
-2.47874141e-01 4.61915974e-03 -8.86865377e-01 7.11836100e-01
2.96730548e-01 6.24180198e-01 2.39501432e-01 -1.00255573e+00
5.77118635e-01 1.34626761e-01 1.35915244e+00 -4.19761986e-02
-5.63298583e-01 5.39238891e-03 -3.26820701e-01 -7.48393834e-01
-9.50239003e-02 -4.13148642e-01 -1.05024302e+00 -4.77711499e-01
-2.76284367e-01 -8.67459297e-01 1.24560905e+00 7.07977057e-01
5.31564891e-01 4.34273183e-01 5.21841049e-01 -1.76908207e+00
-4.31308538e-01 -9.60638821e-01 -7.14083970e-01 -3.21625650e-01
3.57608706e-01 -7.17145026e-01 -1.59230959e-02 -3.04686278e-01] | [6.495484352111816, -0.4573672413825989] |
94887cb5-9dc9-45c9-9712-bf45bced2d53 | implicit-acoustic-echo-cancellation-for | 2111.10639 | null | https://arxiv.org/abs/2111.10639v4 | https://arxiv.org/pdf/2111.10639v4.pdf | Implicit Acoustic Echo Cancellation for Keyword Spotting and Device-Directed Speech Detection | In many speech-enabled human-machine interaction scenarios, user speech can overlap with the device playback audio. In these instances, the performance of tasks such as keyword-spotting (KWS) and device-directed speech detection (DDD) can degrade significantly. To address this problem, we propose an implicit acoustic echo cancellation (iAEC) framework where a neural network is trained to exploit the additional information from a reference microphone channel to learn to ignore the interfering signal and improve detection performance. We study this framework for the tasks of KWS and DDD on, respectively, an augmented version of Google Speech Commands v2 and a real-world Alexa device dataset. Notably, we show a 56% reduction in false-reject rate for the DDD task during device playback conditions. We also show comparable or superior performance over a strong end-to-end neural echo cancellation + KWS baseline for the KWS task with an order of magnitude less computational requirements. | ['Thibaud Sénéchal', 'Thomas Balestri', 'Samuele Cornell'] | 2021-11-20 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 4.63204354e-01 -1.05276860e-01 2.49797747e-01 -2.12404072e-01
-1.37804008e+00 -5.90134501e-01 4.84176636e-01 -1.02104343e-01
-4.76337433e-01 7.70882592e-02 5.27248204e-01 -6.55321538e-01
2.77406096e-01 2.82332629e-01 -7.19706535e-01 -2.99718678e-01
7.69670382e-02 -3.10253590e-01 3.44482630e-01 4.30564098e-02
-5.11840135e-02 2.44501114e-01 -1.52811873e+00 3.62594515e-01
4.24852461e-01 1.14470315e+00 3.69080871e-01 1.33482671e+00
2.18355045e-01 6.40146434e-01 -9.76367235e-01 -5.85166477e-02
1.68280497e-01 6.88636750e-02 -1.68392181e-01 -2.37798616e-01
4.33598101e-01 -6.82854176e-01 -8.14210594e-01 7.87703156e-01
1.33081257e+00 1.81661293e-01 2.97388285e-01 -1.04098475e+00
4.43306938e-02 4.34684515e-01 -3.34148109e-01 3.92632246e-01
6.08128309e-01 2.76018083e-01 7.43089080e-01 -1.41126096e+00
-3.63989472e-02 1.01805139e+00 8.33788216e-01 5.13522923e-01
-1.19577789e+00 -8.67127240e-01 -1.41420299e-02 -1.09246217e-01
-1.41754270e+00 -1.37865090e+00 5.97131550e-01 -2.69815594e-01
1.41345406e+00 6.23180866e-01 8.25763419e-02 1.41447258e+00
-3.10137123e-01 1.15127146e+00 5.16912758e-01 -4.00905997e-01
1.72345534e-01 1.41255692e-01 1.14373185e-01 -4.20872634e-03
-4.78837699e-01 3.33483160e-01 -1.09852338e+00 -3.88729185e-01
2.52886087e-01 -3.12456101e-01 -6.56478643e-01 2.00259462e-01
-1.00785685e+00 8.55857506e-02 -2.69032687e-01 1.53334469e-01
-3.60356003e-01 1.05677418e-01 6.30714834e-01 3.68419170e-01
5.65391660e-01 3.78796935e-01 -6.96690559e-01 -8.61455083e-01
-1.08812881e+00 9.04510543e-02 1.06073034e+00 9.95202541e-01
-1.93365440e-02 3.66148382e-01 -4.26192760e-01 1.15846062e+00
7.92794451e-02 7.82927930e-01 3.30650479e-01 -7.49659777e-01
7.55876660e-01 -3.76207739e-01 2.54220575e-01 -5.72585881e-01
-2.41379365e-01 -5.57847381e-01 -5.45735598e-01 -7.12265447e-02
9.39255953e-02 -3.94407511e-01 -7.81266510e-01 1.78196502e+00
2.51516312e-01 6.92515612e-01 -8.15446824e-02 9.17693138e-01
6.26198530e-01 7.18418837e-01 -9.40636396e-02 -2.22783223e-01
1.07645869e+00 -8.51886809e-01 -1.01170635e+00 -3.88147056e-01
4.94234443e-01 -9.32512939e-01 1.27641332e+00 6.04457378e-01
-1.08272183e+00 -5.51935554e-01 -1.17706168e+00 1.50104672e-01
5.89923235e-03 3.42094928e-01 1.23295344e-01 8.97802234e-01
-1.12366426e+00 2.37634391e-01 -7.79550850e-01 1.78545285e-02
-7.65140653e-02 5.41063607e-01 -1.77492604e-01 3.44629377e-01
-1.07697439e+00 1.69478104e-01 -1.17358968e-01 5.27131595e-02
-9.05054986e-01 -1.04481149e+00 -5.78991532e-01 2.41112813e-01
6.44088805e-01 -1.35358512e-01 1.87560320e+00 -7.13030756e-01
-1.82718563e+00 5.63913524e-01 -2.64772713e-01 -4.56073910e-01
4.22607124e-01 -7.20721364e-01 -9.36017394e-01 -2.31016353e-01
-1.77632317e-01 6.93728914e-03 1.24906600e+00 -8.74003947e-01
-7.28784919e-01 -1.36749759e-01 -2.38589615e-01 2.52024710e-01
-5.51533401e-01 2.19829112e-01 -8.63994062e-01 -1.01215184e+00
-1.72896594e-01 -8.76876831e-01 2.66458958e-01 -3.06837380e-01
-6.50425911e-01 -6.65723830e-02 9.30496216e-01 -1.12049568e+00
1.51262021e+00 -2.90220261e+00 -2.47310817e-01 8.70967805e-02
2.85918206e-01 7.60032713e-01 -3.44877511e-01 3.36324155e-01
-1.22654036e-01 -1.46438554e-01 4.53791879e-02 -9.24991190e-01
-5.64240990e-03 -1.96682483e-01 -4.19552594e-01 3.06429833e-01
2.55790614e-02 5.72322667e-01 -7.38749623e-01 2.76457191e-01
3.05832773e-01 6.35527372e-01 -6.16134703e-01 6.39976323e-01
2.96341270e-01 3.11524481e-01 4.58240993e-02 3.43676358e-01
6.49821639e-01 2.00670376e-01 6.22740909e-02 -1.13990948e-01
-6.66704625e-02 8.83431435e-01 -1.11400151e+00 1.76115513e+00
-9.95678663e-01 1.10985494e+00 7.69930780e-01 -4.07033175e-01
3.44499677e-01 6.88093364e-01 8.55447501e-02 -1.00470030e+00
-8.00698623e-02 3.39291424e-01 9.38530732e-03 -3.52568656e-01
5.44016361e-01 4.00081992e-01 1.33306310e-01 1.71835333e-01
1.55052869e-02 1.16357751e-01 -6.94625378e-01 3.65466446e-01
1.65694737e+00 -6.17815197e-01 -1.35798171e-01 -5.21065444e-02
3.97453070e-01 -8.37139845e-01 1.29824281e-01 1.18174148e+00
-1.25888363e-01 6.88168764e-01 1.52434349e-01 3.61376911e-01
-7.76175857e-01 -1.24745464e+00 1.28754571e-01 1.44583857e+00
-1.41803995e-01 -6.76992357e-01 -9.81210351e-01 -3.42204124e-01
-4.07298096e-02 8.64464998e-01 8.90671555e-03 -3.24607462e-01
-5.45672774e-01 -1.88940868e-01 1.06594253e+00 4.83304471e-01
1.51777491e-01 -6.28911674e-01 -1.93786263e-01 3.89227062e-01
-1.47234485e-01 -1.63767886e+00 -1.01833487e+00 3.42254877e-01
-1.93259567e-01 -4.43686426e-01 -8.33750665e-01 -5.52431464e-01
8.65541250e-02 5.41166902e-01 9.42633331e-01 -2.36082390e-01
-9.32408944e-02 7.36450970e-01 -2.88305372e-01 -3.82151157e-01
-5.34616649e-01 1.12389430e-01 5.03743529e-01 9.66846198e-02
1.61292568e-01 -6.57031178e-01 -5.73040843e-01 3.54745358e-01
-6.41860008e-01 -1.84207916e-01 4.29740787e-01 7.88898885e-01
-2.36522732e-03 -3.60325277e-02 6.56403542e-01 -4.08653259e-01
1.16727912e+00 -2.57350236e-01 -3.82470340e-01 -1.09735467e-01
-3.55321527e-01 -2.49253795e-01 5.31385243e-01 -8.71789277e-01
-1.00840187e+00 1.29350172e-02 -8.53422642e-01 -5.86357832e-01
-1.36840880e-01 1.98280066e-01 -3.87861609e-01 5.87696433e-02
4.67508525e-01 3.53102446e-01 -4.78112161e-01 -7.53580987e-01
1.48180202e-01 1.66889429e+00 9.34803545e-01 -6.56834990e-02
4.88085955e-01 2.47306805e-02 -6.99988723e-01 -1.28944194e+00
-1.51568010e-01 -9.62304652e-01 6.20142668e-02 1.26649849e-02
4.71631855e-01 -1.30765426e+00 -7.92462826e-01 7.24433303e-01
-1.46527612e+00 -3.79465848e-01 -1.78831853e-02 5.53683162e-01
-2.70249337e-01 3.21311355e-01 -5.51823318e-01 -1.38973343e+00
-5.76326966e-01 -1.20776939e+00 1.50298369e+00 -2.43355617e-01
-3.68674755e-01 -2.57682145e-01 -1.40238509e-01 1.22985430e-01
6.70100391e-01 -5.26849151e-01 5.49978137e-01 -9.29422081e-01
-3.93007696e-02 -4.43029255e-01 3.61232832e-02 6.14924252e-01
1.72629312e-01 -4.53419119e-01 -1.64993548e+00 -3.67300391e-01
2.33722448e-01 6.06296547e-02 5.37480354e-01 2.60568500e-01
1.04324639e+00 -3.19352865e-01 -2.42957175e-01 3.59704763e-01
6.66022182e-01 4.52726334e-01 4.00013179e-01 -4.31562304e-01
6.01693153e-01 2.94250518e-01 3.01526308e-01 5.79898536e-01
-6.02538846e-02 1.17643142e+00 -2.49343505e-03 -1.79741472e-01
-5.32168150e-01 -3.76462579e-01 6.63878918e-01 1.05060399e+00
6.36165321e-01 -7.69992888e-01 -7.58625448e-01 5.54463744e-01
-1.56446087e+00 -5.54035962e-01 -1.23797823e-02 2.53446364e+00
5.87734044e-01 3.20568591e-01 1.69928193e-01 4.24610525e-01
7.17870414e-01 1.49767548e-01 -7.13316083e-01 -3.54443610e-01
-2.88118664e-02 3.63346368e-01 4.76545841e-01 6.78248763e-01
-9.65255439e-01 7.21412480e-01 6.14807463e+00 1.16063190e+00
-1.23090827e+00 5.55182159e-01 3.92089128e-01 -4.37115580e-01
1.04647130e-01 -7.01104462e-01 -4.24631715e-01 4.92472112e-01
1.50949550e+00 2.35763788e-01 7.31763124e-01 7.16270268e-01
5.56536555e-01 -1.01796016e-02 -1.31513214e+00 1.54136288e+00
-9.27363187e-02 -9.58969712e-01 -5.82264245e-01 6.37100264e-02
1.81502342e-01 4.15223956e-01 2.48463571e-01 4.37138975e-01
-2.44533226e-01 -8.33214760e-01 7.80849040e-01 7.82904252e-02
1.28043330e+00 -5.92505693e-01 2.71322191e-01 4.96333838e-01
-1.32336831e+00 -4.42992225e-02 3.06250274e-01 -1.59633309e-02
2.85890192e-01 5.86055577e-01 -1.12452674e+00 1.03341006e-01
7.12620974e-01 4.95057255e-02 -1.44757628e-01 9.68961298e-01
-2.31540501e-02 1.06578827e+00 -6.09775484e-01 4.00466509e-02
-6.11133762e-02 5.05298555e-01 1.04912734e+00 1.47414255e+00
2.71614701e-01 1.13249451e-01 -3.45747799e-01 4.36475486e-01
-3.16522509e-01 -1.98297650e-01 -5.27828217e-01 -1.94015503e-01
7.93527007e-01 8.42331409e-01 -1.39506862e-01 -2.65112191e-01
-4.08139050e-01 1.55237162e+00 -1.21914610e-01 6.46795750e-01
-8.78121376e-01 -6.59765959e-01 1.07890737e+00 1.50414199e-01
2.14552194e-01 -3.31136197e-01 8.67578387e-02 -1.00361764e+00
3.84956956e-01 -1.08856845e+00 -3.06717575e-01 -6.64481401e-01
-8.86031270e-01 6.10249281e-01 -4.59354639e-01 -9.69220996e-01
-5.44109583e-01 -4.77495462e-01 -6.29496336e-01 1.13728249e+00
-8.86125267e-01 -4.28553045e-01 1.39618546e-01 5.02802730e-01
9.48235154e-01 -1.15475170e-01 7.97955334e-01 9.79576886e-01
-4.23237085e-01 1.15818596e+00 3.97514492e-01 1.15374744e-01
6.84721410e-01 -1.07244992e+00 1.13539684e+00 1.03519249e+00
1.83351964e-01 5.70759892e-01 9.42040741e-01 -5.52547634e-01
-1.61371970e+00 -9.19309258e-01 6.75247252e-01 -3.96141529e-01
4.61040497e-01 -1.23352528e+00 -7.96603739e-01 2.49943599e-01
9.29531753e-02 -6.68332272e-04 6.00612700e-01 1.38746500e-01
-3.27764302e-01 -9.79940500e-03 -7.08749413e-01 6.96843147e-01
1.22709465e+00 -1.29214537e+00 -1.99353233e-01 2.66056359e-01
1.11077273e+00 -5.83850384e-01 -3.45153153e-01 2.15654954e-01
7.46094823e-01 -8.35048735e-01 1.17657888e+00 -3.81398171e-01
-3.38777512e-01 -1.63236544e-01 -3.09168160e-01 -1.26470196e+00
1.55499294e-01 -1.40076232e+00 -4.80131447e-01 1.47395003e+00
4.93369579e-01 -4.23764199e-01 5.75423896e-01 7.29996681e-01
-4.11990225e-01 -3.84515256e-01 -1.24062467e+00 -8.97721410e-01
-5.35917759e-01 -1.16598010e+00 3.01884979e-01 2.54999369e-01
-7.82238618e-02 4.68958259e-01 -7.88225591e-01 5.01641512e-01
2.51582116e-01 -8.49643767e-01 8.09994280e-01 -6.84491277e-01
-7.21771240e-01 -9.09868553e-02 -2.25144878e-01 -1.76366103e+00
3.35195996e-02 -3.55113745e-01 4.65823412e-01 -8.04238856e-01
-4.78361368e-01 -7.06652254e-02 -3.07730108e-01 -7.58290663e-02
-2.18939468e-01 -5.24805412e-02 3.24587464e-01 -2.24519018e-02
-6.15485370e-01 5.35251021e-01 2.99845845e-01 -1.83687642e-01
-5.40674150e-01 4.12409484e-01 -4.63875204e-01 5.31033158e-01
4.33259875e-01 -5.36056936e-01 -4.55109030e-01 -6.03449821e-01
1.64990779e-02 3.49038273e-01 2.11038798e-01 -1.19453847e+00
4.86286908e-01 6.01222575e-01 -7.90910795e-02 -3.99013221e-01
7.70051062e-01 -8.46693516e-01 -3.88918519e-02 2.45081648e-01
-4.88523513e-01 -2.66863167e-01 5.71478724e-01 7.90531993e-01
-2.38864738e-02 2.65944719e-01 4.09114212e-01 6.50438190e-01
-4.39269006e-01 -9.90404561e-02 -7.84187019e-01 2.70638429e-03
3.93623978e-01 -2.42980719e-02 2.43226308e-02 -8.18174183e-01
-5.95524549e-01 4.24857065e-03 -2.65671104e-01 7.82703340e-01
6.92339122e-01 -1.07830846e+00 -4.22718763e-01 3.24785858e-01
1.09863982e-01 -3.34311664e-01 3.32273692e-01 8.68689716e-01
2.26955101e-01 5.22049487e-01 7.14712501e-01 -6.28880978e-01
-1.49155903e+00 2.91484118e-01 5.25844336e-01 1.45085096e-01
-4.63871092e-01 1.18532372e+00 4.72753167e-01 -4.35522974e-01
1.08031380e+00 -3.90724093e-01 3.17461491e-01 -4.40684855e-01
8.44326735e-01 5.58350682e-01 5.72826684e-01 -5.12968123e-01
-5.50281584e-01 -6.83001522e-03 -1.70263555e-02 -5.15226662e-01
8.62419724e-01 -4.52541858e-01 6.73579454e-01 4.28628832e-01
1.46317804e+00 5.41029572e-01 -1.12915480e+00 -3.60050112e-01
-2.46537589e-02 -3.07610899e-01 4.24221426e-01 -1.09875882e+00
-6.58739686e-01 9.14209664e-01 1.09924674e+00 2.09061995e-01
1.28128958e+00 -9.41522345e-02 1.07694244e+00 3.46339285e-01
9.27615389e-02 -1.01501215e+00 -9.73108262e-02 5.38762093e-01
8.65786076e-01 -1.07827163e+00 -8.87644708e-01 -3.82771462e-01
-4.49818760e-01 7.39459455e-01 2.91984022e-01 3.50212872e-01
6.88990474e-01 9.06853020e-01 1.38722062e-01 1.71279669e-01
-6.13069475e-01 -5.10696582e-02 2.79446185e-01 6.09886646e-01
2.69205064e-01 2.18789913e-02 2.53885716e-01 9.05084133e-01
-7.87928626e-02 -2.37257883e-01 9.96689405e-03 7.65771568e-01
-9.15929377e-02 -7.64928401e-01 -3.12716216e-01 4.27473545e-01
-6.44565523e-01 -6.63967907e-01 -6.26393080e-01 1.39637232e-01
-2.86238402e-01 1.63095701e+00 6.72622919e-02 -9.37806487e-01
7.96412945e-01 1.10965244e-01 -1.07647702e-01 -4.39578295e-01
-8.92273009e-01 4.92875248e-01 2.18658403e-01 -7.70275831e-01
1.61102071e-01 -3.69498998e-01 -1.05155826e+00 -2.49257326e-01
-6.12456441e-01 1.05890594e-01 9.63526249e-01 8.85493279e-01
8.24737728e-01 1.01716375e+00 5.88826299e-01 -9.43251610e-01
-6.87085509e-01 -9.67959046e-01 -6.60059035e-01 1.16658524e-01
8.82508636e-01 -2.32038498e-01 -5.49337268e-01 -3.91003370e-01] | [14.715478897094727, 6.129934787750244] |
10913376-3659-4e93-961c-cf3762ce5fde | can-bert-do-it-controller-area-network | 2210.09439 | null | https://arxiv.org/abs/2210.09439v1 | https://arxiv.org/pdf/2210.09439v1.pdf | CAN-BERT do it? Controller Area Network Intrusion Detection System based on BERT Language Model | Due to the rising number of sophisticated customer functionalities, electronic control units (ECUs) are increasingly integrated into modern automotive systems. However, the high connectivity between the in-vehicle and the external networks paves the way for hackers who could exploit in-vehicle network protocols' vulnerabilities. Among these protocols, the Controller Area Network (CAN), known as the most widely used in-vehicle networking technology, lacks encryption and authentication mechanisms, making the communications delivered by distributed ECUs insecure. Inspired by the outstanding performance of bidirectional encoder representations from transformers (BERT) for improving many natural language processing tasks, we propose in this paper ``CAN-BERT", a deep learning based network intrusion detection system, to detect cyber attacks on CAN bus protocol. We show that the BERT model can learn the sequence of arbitration identifiers (IDs) in the CAN bus for anomaly detection using the ``masked language model" unsupervised training objective. The experimental results on the ``Car Hacking: Attack \& Defense Challenge 2020" dataset show that ``CAN-BERT" outperforms state-of-the-art approaches. In addition to being able to identify in-vehicle intrusions in real-time within 0.8 ms to 3 ms w.r.t CAN ID sequence length, it can also detect a wide variety of cyberattacks with an F1-score of between 0.81 and 0.99. | ['Jean-Luc Danger', 'Hadi Ghauch', 'Maria Mushtaq', 'Natasha Alkhatib'] | 2022-10-17 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [-2.36915186e-01 9.25998669e-04 -4.31661248e-01 -1.81156337e-01
-6.61714435e-01 -7.02253163e-01 4.86060917e-01 -3.41159068e-02
-4.60089058e-01 3.87214005e-01 -6.39307797e-01 -1.17853045e+00
1.51683658e-01 -9.01270092e-01 -8.93810511e-01 -5.56219161e-01
-3.68803710e-01 3.30353111e-01 6.06231809e-01 -3.70266289e-01
-8.69798362e-02 9.51113701e-01 -1.39854884e+00 3.92400175e-01
1.92232922e-01 1.43693113e+00 -4.84354436e-01 4.35706526e-01
-2.34087259e-01 7.82818973e-01 -7.70923972e-01 -6.58372760e-01
2.04725504e-01 -6.23163627e-03 -4.40336913e-01 -7.91495264e-01
4.58203517e-02 -5.46166599e-01 -9.55178082e-01 1.43829751e+00
-9.82486084e-02 -2.71402448e-01 1.33766487e-01 -2.01064563e+00
-6.06998010e-03 6.24996483e-01 -3.48946959e-01 3.39977443e-01
-1.26672620e-02 5.53911150e-01 7.85649359e-01 -3.54265183e-01
6.53381109e-01 1.23866785e+00 4.95611012e-01 6.59326971e-01
-1.01177168e+00 -1.57547402e+00 -2.24543661e-02 8.37486327e-01
-1.35408139e+00 -3.42146486e-01 7.63621449e-01 -1.92431301e-01
1.11757517e+00 2.65327320e-02 2.51698587e-02 1.45920706e+00
9.69805062e-01 4.53650296e-01 5.26780069e-01 1.59611672e-01
2.62961566e-01 9.75161493e-02 3.37470084e-01 7.13357985e-01
4.13352430e-01 6.04796112e-01 -9.24084559e-02 -1.61211640e-01
7.19136745e-02 -1.10425754e-02 2.44402573e-01 2.84147888e-01
-8.29079568e-01 1.05750549e+00 4.04872566e-01 1.83876842e-01
-3.54519933e-01 5.63778341e-01 1.34692574e+00 7.59923398e-01
-2.61387378e-01 2.90550590e-01 -5.57278514e-01 -4.26232547e-01
-2.04102680e-01 -1.01671338e-01 8.51926267e-01 8.99356663e-01
5.78104317e-01 5.31615078e-01 6.20596826e-01 9.41316709e-02
4.72139567e-01 6.17635012e-01 1.29631460e-01 -5.91315925e-01
3.37460071e-01 2.98757195e-01 -4.34549749e-01 -9.06979680e-01
-3.64021748e-01 -1.66767105e-01 -9.69417036e-01 5.21814764e-01
3.17184329e-01 -1.90420821e-01 -6.56232655e-01 1.62596369e+00
7.09199160e-02 2.27629334e-01 9.18126181e-02 4.14706975e-01
2.92991489e-01 7.44155884e-01 1.35955140e-01 -2.82915626e-02
1.60830534e+00 -4.51112658e-01 -1.01651573e+00 -1.40046731e-01
7.09483266e-01 -5.81079781e-01 3.17874491e-01 6.17770731e-01
-4.14710641e-01 -5.77275455e-01 -1.42426586e+00 5.24864078e-01
-8.10963631e-01 -6.77701950e-01 4.60204929e-01 9.96523738e-01
-5.02631783e-01 1.38761565e-01 -8.13554823e-01 2.24057347e-01
5.48956215e-01 5.54763794e-01 -6.94232643e-01 3.76331583e-02
-1.87091422e+00 9.61170793e-01 4.83507752e-01 1.65909186e-01
-1.24396098e+00 -5.72301507e-01 -8.21695864e-01 -6.67869747e-02
7.02131450e-01 1.21126927e-01 1.33042359e+00 -7.52806962e-01
-1.51567054e+00 3.49523991e-01 1.02171063e-01 -1.00522745e+00
1.58013776e-01 -1.73405141e-01 -1.34182489e+00 2.85984486e-01
-1.67172581e-01 2.60434598e-01 8.30207229e-01 -8.64112794e-01
-7.29455709e-01 -4.32378799e-02 1.60094593e-02 -1.17426753e+00
4.09510396e-02 5.77479005e-01 7.54168555e-02 -2.61351168e-01
-4.71646398e-01 -1.09286177e+00 -4.01804112e-02 -1.68983892e-01
-5.15326917e-01 -3.16377163e-01 1.52126825e+00 -6.13238275e-01
1.27339149e+00 -2.33516145e+00 -5.74614286e-01 7.31753469e-01
2.16104820e-01 9.02283967e-01 -5.96155897e-02 5.49178362e-01
-4.52966660e-01 1.83967412e-01 2.17357650e-01 1.84554651e-01
3.66540909e-01 2.27170646e-01 -6.82997048e-01 6.47747397e-01
3.42551649e-01 8.96248519e-01 -7.70688355e-01 -2.18267478e-02
2.62696415e-01 3.02194297e-01 -2.96843320e-01 8.83435011e-02
-3.43338192e-01 2.05930933e-01 -4.60447311e-01 5.49681246e-01
6.95625544e-01 2.36284003e-01 3.32664847e-02 -3.98972221e-02
-7.87418261e-02 3.41721594e-01 -7.72066534e-01 8.85457039e-01
-3.48022133e-01 1.14676917e+00 2.42210835e-01 -9.51654434e-01
9.54943419e-01 5.56158245e-01 3.29763472e-01 -1.08834314e+00
6.11536443e-01 2.69683957e-01 5.31433523e-01 -3.40812743e-01
-7.56982267e-02 5.36703644e-03 -4.68804806e-01 2.30200842e-01
-1.20265961e-01 4.33359504e-01 6.00225218e-02 2.00107932e-01
1.45726085e+00 -7.38301754e-01 -3.08168884e-02 1.07467316e-01
9.51084018e-01 -2.33765334e-01 6.75829947e-01 3.94895226e-01
-3.33317131e-01 -4.28797603e-01 1.02269280e+00 -5.89260876e-01
-9.71097291e-01 -1.26093268e+00 -1.54481456e-01 4.88267541e-01
5.27325179e-03 -6.21751308e-01 -5.41784823e-01 -9.60937500e-01
2.34539688e-01 9.99669313e-01 -1.63045391e-01 -8.47068846e-01
-7.75643408e-01 2.06705276e-02 1.38798904e+00 5.97705722e-01
6.93359911e-01 -8.48495007e-01 -2.70762712e-01 3.81181419e-01
2.02363178e-01 -1.76477492e+00 -1.88213751e-01 1.75255805e-01
-1.44071519e-01 -1.36268127e+00 6.01493716e-01 -5.81041276e-01
1.30659193e-01 -5.30589782e-02 5.73836625e-01 -5.10890149e-02
-4.18924838e-01 8.74910504e-02 -1.77836478e-01 -5.24996161e-01
-1.18100321e+00 -1.68076545e-01 4.11837965e-01 3.13304484e-01
8.94318342e-01 -3.20957363e-01 -1.29262015e-01 5.12993038e-01
-1.10092115e+00 -5.57600677e-01 7.12011039e-01 5.25459766e-01
1.44061610e-01 7.45646775e-01 9.24827993e-01 -5.69188774e-01
2.16096327e-01 -7.48323441e-01 -1.17248595e+00 -3.69235754e-01
-4.56660777e-01 1.46992477e-02 1.04657531e+00 -6.16403937e-01
-5.43915212e-01 -3.60740423e-01 -7.33486235e-01 -6.59111261e-01
-3.89633477e-01 3.76732171e-01 -6.89433873e-01 -1.78864047e-01
1.67345479e-01 -2.87076365e-02 4.27078843e-01 -8.65398198e-02
1.78612694e-01 6.77244365e-01 5.94511867e-01 -2.76659369e-01
1.24368560e+00 3.03454638e-01 1.28758669e-01 -1.01132309e+00
-2.33389959e-02 -1.83624163e-01 1.76304709e-02 -3.80008638e-01
9.03461754e-01 -6.67278886e-01 -1.54353404e+00 6.28031552e-01
-1.57029629e+00 1.72377918e-02 8.20233375e-02 5.28382182e-01
-2.02124357e-01 4.35825795e-01 -9.53361273e-01 -7.47304022e-01
-1.72080681e-01 -1.40550375e+00 6.02918863e-01 -1.96238443e-01
-2.81511813e-01 -5.49659252e-01 -2.34226078e-01 2.16160551e-01
6.09936774e-01 3.58971477e-01 1.37153745e+00 -1.23503494e+00
-8.49662602e-01 -9.78736579e-01 -3.27535629e-01 9.11926925e-01
-1.47954032e-01 -2.80708428e-02 -9.13646519e-01 -3.39804202e-01
1.41332597e-01 7.82970935e-02 3.25472027e-01 -3.03123653e-01
1.16790271e+00 -3.68484557e-01 -3.35741907e-01 3.89015287e-01
1.20555604e+00 8.21986139e-01 1.10836589e+00 3.73962909e-01
5.45212507e-01 4.12861407e-01 3.38449925e-01 1.22634731e-01
-1.18119083e-01 6.90836191e-01 9.41274405e-01 4.73484159e-01
2.78881967e-01 -1.68915853e-01 9.16329920e-01 5.07161200e-01
5.35067141e-01 -1.74313694e-01 -8.89770746e-01 1.27904890e-02
-1.27659094e+00 -1.10417140e+00 -2.65357465e-01 1.99440706e+00
3.80621731e-01 7.89267004e-01 -5.38443401e-02 3.86973739e-01
6.13299966e-01 4.43503149e-02 -5.29820979e-01 -9.45342302e-01
1.99247107e-01 5.05715787e-01 9.14683044e-01 3.09377581e-01
-1.03460205e+00 8.65591824e-01 4.67114735e+00 9.64249432e-01
-1.22335470e+00 1.17802478e-01 1.74879044e-01 2.06237197e-01
1.90455735e-01 -1.16591454e-01 -1.02457988e+00 8.94128680e-01
1.97252905e+00 -5.64809283e-03 2.13639319e-01 8.63683820e-01
8.45955610e-02 4.18614298e-01 -1.17017663e+00 7.29034662e-01
-2.37548932e-01 -1.24522698e+00 -1.50469631e-01 3.40828031e-01
6.24819957e-02 2.09117830e-01 1.61108151e-02 6.07757390e-01
1.89607888e-01 -1.09848881e+00 5.69093347e-01 2.35449895e-01
7.46824265e-01 -1.22135687e+00 1.14845216e+00 1.02144353e-01
-1.14135253e+00 -4.12323952e-01 6.20627441e-02 3.89918089e-01
3.90377790e-01 4.03393030e-01 -5.45313478e-01 5.16785860e-01
5.24802744e-01 1.17334761e-01 -1.75821677e-01 5.95841825e-01
-2.66202807e-01 8.39423418e-01 -3.70291770e-01 1.58355385e-02
5.95997870e-01 9.97016430e-02 6.14832342e-01 1.01864386e+00
7.75046647e-02 -2.54492253e-01 -3.29150595e-02 6.95512235e-01
-9.49386433e-02 -5.56997836e-01 -9.52475190e-01 -3.95457566e-01
6.01078689e-01 9.19791520e-01 -3.55192751e-01 -7.68928975e-03
-6.01049125e-01 3.67237598e-01 -4.27396476e-01 3.30307543e-01
-1.36558843e+00 -8.21128726e-01 1.40909648e+00 1.47883132e-01
4.60774601e-01 -5.60699344e-01 2.82054156e-01 -4.42172050e-01
-1.06890373e-01 -1.40570605e+00 2.56291509e-01 -2.52954781e-01
-1.28774226e+00 8.40063989e-01 -9.52216610e-02 -1.34790599e+00
-3.01194757e-01 -1.17082536e+00 -5.15851080e-01 4.02699798e-01
-1.58003044e+00 -9.34345126e-01 3.23846370e-01 7.03393519e-01
2.68936217e-01 -6.93991005e-01 7.71978498e-01 7.45057106e-01
-7.50011683e-01 9.20453548e-01 3.67401578e-02 7.00374246e-01
4.75738525e-01 -5.27269840e-01 7.65763879e-01 8.42900097e-01
-2.04459235e-01 5.76506972e-01 5.96881986e-01 -6.22050524e-01
-2.05066466e+00 -1.04419756e+00 6.28965020e-01 -1.21412061e-01
1.37851346e+00 -5.21419406e-01 -9.80692029e-01 9.32344139e-01
3.01627487e-01 2.37655477e-03 7.13473856e-01 -3.81945521e-01
-9.93578672e-01 -5.10544062e-01 -9.64183092e-01 4.96923625e-01
3.91380996e-01 -8.93597960e-01 -3.92227411e-01 6.93280250e-02
9.42331791e-01 2.87720487e-02 -6.56245947e-01 3.80468756e-01
4.83571231e-01 -7.16637373e-01 9.26581919e-01 -1.03821206e+00
-5.20017855e-02 -3.28595310e-01 -1.71748072e-01 -7.82305658e-01
5.85240088e-02 -1.05328441e+00 -3.18815053e-01 1.01621222e+00
2.41103217e-01 -1.28753567e+00 6.28679872e-01 3.06807846e-01
-2.18933284e-01 -3.55074108e-01 -1.63574660e+00 -1.05343068e+00
6.10049106e-02 -1.17754531e+00 7.93485165e-01 5.98748565e-01
-1.65233482e-02 1.99721083e-01 -1.52966604e-01 7.01813161e-01
5.55415273e-01 -5.08516729e-01 7.61996150e-01 -1.27266145e+00
2.83097457e-02 -7.37523019e-01 -1.10587859e+00 -8.04557443e-01
7.72891760e-01 -7.78045893e-01 3.47828604e-02 -3.56826931e-01
-6.32225752e-01 -2.48919606e-01 -5.83095551e-01 3.17751795e-01
7.35894263e-01 -1.74563751e-02 6.14391342e-02 -4.08629745e-01
-3.76634508e-01 2.92622805e-01 3.62629473e-01 -5.18694520e-01
4.22372222e-01 1.23224184e-01 -1.20266855e-01 6.57796323e-01
8.62230003e-01 -5.70185959e-01 -1.58809960e-01 1.43106906e-02
1.71342239e-01 2.75505371e-02 3.74700010e-01 -9.90240633e-01
3.85004729e-01 1.09220453e-01 -2.75002778e-01 -5.89532614e-01
1.80413708e-01 -1.38343060e+00 1.43024966e-01 7.93884397e-01
-4.85096499e-02 5.31074405e-01 6.14614546e-01 7.21243024e-01
-3.47025633e-01 4.61815782e-02 8.35594058e-01 4.89226669e-01
-8.23867559e-01 4.06079501e-01 -1.18390298e+00 -2.80502200e-01
1.23913002e+00 6.09865636e-02 -5.30718446e-01 -3.24242413e-01
-2.89728165e-01 3.52603048e-01 5.00909388e-02 9.12825763e-01
8.18550766e-01 -1.27708673e+00 -3.86156321e-01 8.52677405e-01
1.81007311e-01 -5.04316688e-01 1.83268473e-01 7.47757614e-01
-5.47146320e-01 7.68788934e-01 -2.30247125e-01 -2.46052042e-01
-1.15997767e+00 8.80723774e-01 3.05285901e-01 -2.21145719e-01
-6.24965131e-01 3.53686392e-01 -1.27404228e-01 -1.01470351e-01
4.86733049e-01 -2.48470426e-01 2.00456202e-01 -2.20910832e-01
8.92999530e-01 5.44362307e-01 1.83648333e-01 -7.56761849e-01
-5.73679209e-01 5.75653836e-02 -3.58659446e-01 2.00595915e-01
7.69693136e-01 2.70695597e-01 -3.45455259e-01 1.61678150e-01
1.71854842e+00 -2.42786318e-01 -6.15143538e-01 -1.62718102e-01
4.52268690e-01 -1.65016979e-01 1.17244653e-01 -4.38145369e-01
-1.34530413e+00 9.53886986e-01 5.96244812e-01 4.34840083e-01
7.50767589e-01 -1.66030675e-01 1.48315704e+00 7.66572893e-01
6.35545909e-01 -7.01205909e-01 -3.62817235e-02 8.33816051e-01
3.34627807e-01 -9.20865595e-01 -6.70193195e-01 -2.69961715e-01
-3.04472327e-01 1.31453598e+00 4.62272942e-01 -4.03843932e-02
8.98363173e-01 8.19037914e-01 9.98306572e-02 -2.72631228e-01
-7.72812307e-01 3.67840320e-01 -2.30085686e-01 4.82941628e-01
-3.23187560e-01 -1.02483064e-01 2.49479920e-01 5.49880266e-01
-6.24971502e-02 -2.80801952e-01 3.77521187e-01 7.11428761e-01
-3.35558891e-01 -1.29446661e+00 -2.87654102e-01 2.49865070e-01
-6.11866176e-01 2.78157830e-01 -7.40376413e-02 9.94568527e-01
6.07018173e-02 1.14016008e+00 2.41186157e-01 -8.61751020e-01
4.49478686e-01 4.11270827e-01 -2.98938066e-01 -5.10164648e-02
-4.28149194e-01 -4.73397404e-01 2.70950466e-01 -1.01426971e+00
4.09546167e-01 -3.58460993e-01 -1.47160184e+00 -9.11335051e-01
-2.66834497e-01 3.73276860e-01 7.48540282e-01 9.92356122e-01
3.52801055e-01 8.25222731e-01 8.50226879e-01 -2.58413941e-01
-7.50270247e-01 -5.42189300e-01 -5.94093204e-01 1.48197457e-01
6.11697912e-01 -6.65141761e-01 -5.36436439e-01 -4.64947522e-01] | [5.250732898712158, 7.347848415374756] |
c54ff77c-3d60-45d1-8844-bb40ac501a58 | devis-making-deformable-transformers-work-for | 2207.11103 | null | https://arxiv.org/abs/2207.11103v1 | https://arxiv.org/pdf/2207.11103v1.pdf | DeVIS: Making Deformable Transformers Work for Video Instance Segmentation | Video Instance Segmentation (VIS) jointly tackles multi-object detection, tracking, and segmentation in video sequences. In the past, VIS methods mirrored the fragmentation of these subtasks in their architectural design, hence missing out on a joint solution. Transformers recently allowed to cast the entire VIS task as a single set-prediction problem. Nevertheless, the quadratic complexity of existing Transformer-based methods requires long training times, high memory requirements, and processing of low-single-scale feature maps. Deformable attention provides a more efficient alternative but its application to the temporal domain or the segmentation task have not yet been explored. In this work, we present Deformable VIS (DeVIS), a VIS method which capitalizes on the efficiency and performance of deformable Transformers. To reason about all VIS subtasks jointly over multiple frames, we present temporal multi-scale deformable attention with instance-aware object queries. We further introduce a new image and video instance mask head with multi-scale features, and perform near-online video processing with multi-cue clip tracking. DeVIS reduces memory as well as training time requirements, and achieves state-of-the-art results on the YouTube-VIS 2021, as well as the challenging OVIS dataset. Code is available at https://github.com/acaelles97/DeVIS. | ['Laura Leal-Taixé', 'Guillem Brasó', 'Tim Meinhardt', 'Adrià Caelles'] | 2022-07-22 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.30211607e-01 -2.34137297e-01 -1.17188752e-01 -1.12997189e-01
-1.11918044e+00 -7.81370699e-01 2.48694167e-01 -2.55467892e-01
-4.69877243e-01 3.98926020e-01 -1.41268611e-01 5.19571230e-02
1.86664723e-02 -3.22232276e-01 -9.72665846e-01 -6.27326965e-01
5.87627143e-02 4.81654018e-01 9.81929779e-01 -4.93563637e-02
-1.05136484e-02 3.18146646e-01 -1.73301005e+00 5.52441180e-01
5.33364177e-01 1.30777943e+00 5.24588346e-01 7.04658210e-01
8.92862584e-03 6.46359742e-01 -2.97701180e-01 -5.53183198e-01
5.45739412e-01 -8.18568319e-02 -8.72560501e-01 1.96276948e-01
1.07635295e+00 -4.83414471e-01 -4.25753474e-01 8.38261724e-01
3.92454922e-01 2.45133430e-01 3.58559877e-01 -1.29921627e+00
-5.31853557e-01 2.76991278e-01 -6.15158796e-01 5.72143912e-01
1.56519353e-01 3.09324890e-01 9.22007322e-01 -1.07012713e+00
7.58311450e-01 1.01960325e+00 6.99451625e-01 6.62618637e-01
-9.66153979e-01 -3.84899199e-01 4.71416026e-01 4.63504374e-01
-1.30045056e+00 -4.73332047e-01 3.67346317e-01 -6.14392161e-01
9.67793584e-01 4.96859193e-01 9.64862108e-01 9.56355035e-01
-1.75460358e-03 1.35639858e+00 7.99965799e-01 2.29671467e-02
-9.84357074e-02 -1.48843944e-01 5.06988242e-02 8.39591324e-01
-5.71543202e-02 -1.36840343e-01 -5.17342091e-01 2.43761301e-01
7.25506544e-01 2.26469472e-01 -3.78425896e-01 -3.93824428e-01
-1.11683154e+00 4.79097188e-01 2.49936447e-01 3.64851713e-01
-3.03709686e-01 3.53354663e-01 4.91988420e-01 1.71599314e-01
7.55181253e-01 -2.88366284e-02 -6.68057978e-01 -3.11179996e-01
-1.27379680e+00 2.89070815e-01 4.71800625e-01 1.07913911e+00
5.68497300e-01 -1.28682442e-02 -5.00365913e-01 5.92181444e-01
1.17318906e-01 5.61810732e-01 3.11632931e-01 -1.09094346e+00
4.29180294e-01 4.62994933e-01 -5.88908643e-02 -7.20500708e-01
-3.79500210e-01 -1.63067192e-01 -3.37584496e-01 9.47075114e-02
6.78284466e-01 2.43769232e-02 -1.13858306e+00 1.46317327e+00
6.87692523e-01 6.45619452e-01 -3.85760248e-01 1.24611115e+00
9.38209951e-01 5.85964501e-01 5.57637885e-02 -1.98377952e-01
1.41765249e+00 -1.36749220e+00 -6.27466083e-01 -1.81076810e-01
4.91819441e-01 -6.40097678e-01 9.21751320e-01 4.04015303e-01
-1.59545827e+00 -5.64643145e-01 -7.00935006e-01 -3.89435828e-01
-3.30360383e-01 4.91453428e-03 5.51532686e-01 5.17076969e-01
-1.24881792e+00 5.36762834e-01 -1.15767837e+00 -1.80688739e-01
7.63632834e-01 5.15053689e-01 -2.07982332e-01 -3.61425010e-03
-8.57194960e-01 5.87502420e-01 2.16027796e-01 1.58986092e-01
-9.71812725e-01 -1.00937402e+00 -7.59202600e-01 -1.50265656e-02
1.01528239e+00 -6.45602405e-01 1.16182733e+00 -1.07075214e+00
-1.33785176e+00 9.40864086e-01 -2.73085088e-01 -5.14308512e-01
7.77740419e-01 -5.34121096e-01 -6.81308955e-02 3.72922570e-01
8.80409181e-02 8.13517809e-01 1.03701019e+00 -9.08230722e-01
-7.50459611e-01 -3.10413688e-01 2.08203703e-01 2.64174551e-01
-3.82133454e-01 2.37737074e-01 -1.29624748e+00 -7.08524942e-01
-2.02776715e-01 -1.08144808e+00 -7.46624023e-02 2.59164214e-01
-2.00015992e-01 -2.46048138e-01 9.75387633e-01 -7.37683177e-01
1.29604220e+00 -2.27333403e+00 5.30381978e-01 -3.13817650e-01
3.72919053e-01 5.88611662e-01 -2.79705137e-01 2.05827691e-02
1.31754413e-01 -5.30187860e-02 -2.56005347e-01 -6.51086807e-01
-4.75628413e-02 1.62343740e-01 -5.66460229e-02 4.84488130e-01
1.16230667e-01 1.23892426e+00 -7.37529635e-01 -7.45445013e-01
3.85385752e-01 4.31692362e-01 -7.76954532e-01 2.54113860e-02
-4.23165470e-01 3.67571831e-01 -4.44580257e-01 9.70452070e-01
6.52570367e-01 -3.17844242e-01 -1.35018229e-01 -4.98907626e-01
-2.12409660e-01 -2.93718725e-01 -1.11765385e+00 1.95096421e+00
1.71757154e-02 7.25645959e-01 2.68008560e-01 -1.01938665e+00
1.25256076e-01 2.15341866e-01 1.07672942e+00 -7.68429279e-01
1.32067397e-01 3.08965117e-01 -3.70790243e-01 -8.32828104e-01
5.62422216e-01 2.54151970e-01 1.85843959e-01 -4.52694017e-03
1.70788735e-01 4.07254435e-02 6.28995240e-01 1.88478798e-01
1.02638197e+00 6.20970309e-01 -5.58180213e-02 -1.23123311e-01
4.31896210e-01 4.40343469e-02 6.23077035e-01 6.21915936e-01
-2.93311834e-01 8.24128509e-01 1.68323740e-01 -5.04697144e-01
-7.53625095e-01 -9.05083537e-01 -1.46894068e-01 1.25586808e+00
4.10189658e-01 -4.74686623e-01 -9.31166232e-01 -8.74403656e-01
1.02151543e-01 1.42503873e-01 -6.56690419e-01 3.24160159e-01
-8.81730318e-01 -5.29435992e-01 3.10910672e-01 5.90411723e-01
2.44695783e-01 -1.02894676e+00 -8.93747091e-01 3.22414845e-01
-3.33399057e-01 -1.54408598e+00 -8.76627028e-01 -1.60274446e-01
-6.26041412e-01 -1.31961882e+00 -1.05403507e+00 -6.35782361e-01
2.73167312e-01 3.77451986e-01 9.98532891e-01 1.40722528e-01
-7.42865682e-01 7.22971499e-01 -3.61922652e-01 -2.92187691e-01
8.66915584e-02 4.16187681e-02 -6.26270548e-02 2.06325695e-01
2.37116516e-02 -2.90936708e-01 -7.15483189e-01 5.17848313e-01
-1.00782120e+00 2.23267689e-01 3.26730222e-01 4.89099622e-01
8.65512550e-01 -5.34994483e-01 2.16972813e-01 -6.48352802e-01
-1.78391874e-01 -3.79861057e-01 -7.96740651e-01 3.65995914e-01
-1.43451706e-01 -3.42664152e-01 2.06816196e-01 -5.48438430e-01
-7.95558155e-01 2.73238122e-01 -1.05312139e-01 -1.07774520e+00
9.29318890e-02 2.25171536e-01 7.55817220e-02 -2.45800361e-01
2.57779509e-01 3.00025076e-01 -4.89019975e-02 -5.21671832e-01
2.40687221e-01 3.39036316e-01 3.86275887e-01 -4.85318363e-01
6.52661920e-01 6.10604942e-01 -1.52797908e-01 -7.25581646e-01
-1.04134536e+00 -7.05712259e-01 -7.61922359e-01 -5.02909124e-01
1.29567659e+00 -1.03840983e+00 -7.10510969e-01 5.70082724e-01
-1.13792360e+00 -7.28322804e-01 -4.09521282e-01 2.36494377e-01
-7.67673731e-01 3.98070127e-01 -5.68250060e-01 -5.59092224e-01
-2.45374233e-01 -1.29863489e+00 1.37779820e+00 -4.06719260e-02
7.59195834e-02 -8.07662487e-01 -1.66126311e-01 7.82928348e-01
3.10871571e-01 3.80398482e-01 1.60254925e-01 -4.56456751e-01
-1.13053036e+00 -2.32977886e-02 -3.35031658e-01 2.33657286e-01
-2.47644052e-01 1.97598222e-03 -8.45871687e-01 -4.33427185e-01
-1.92478701e-01 -2.69483298e-01 1.05452144e+00 6.48339212e-01
1.39350104e+00 -1.52352035e-01 -3.70547533e-01 9.14137602e-01
1.44095731e+00 2.19875827e-01 3.77925396e-01 1.86694950e-01
1.07096982e+00 4.26356882e-01 8.46370876e-01 3.14873278e-01
4.08150911e-01 1.09019196e+00 5.40452778e-01 -1.37339547e-01
-4.50368315e-01 3.59727949e-01 4.96599257e-01 5.70689440e-01
-4.71149892e-01 -4.86891419e-01 -8.31344128e-01 7.26945817e-01
-2.09046698e+00 -1.17586482e+00 -3.18271309e-01 1.83805048e+00
5.26691616e-01 -8.54788870e-02 4.21571583e-01 -1.25879809e-01
6.39331520e-01 2.11542919e-01 -7.05935717e-01 6.27004355e-02
-9.10295993e-02 1.00997746e-01 5.15100956e-01 3.49060833e-01
-1.37435269e+00 1.09590030e+00 5.25356007e+00 1.02816987e+00
-1.02702844e+00 5.71728408e-01 4.56106365e-01 -7.47383416e-01
3.69380265e-02 -3.03156912e-01 -9.42836344e-01 5.74956834e-01
6.61938667e-01 3.72807533e-02 5.05298972e-01 6.39933050e-01
1.59403279e-01 -1.19039036e-01 -1.11955917e+00 1.12870121e+00
1.94075987e-01 -1.52028799e+00 -3.56366932e-02 -2.06571277e-02
6.59523189e-01 5.00697196e-01 1.51095137e-01 2.58132219e-01
-2.90577322e-01 -6.71468675e-01 1.18765545e+00 4.93715346e-01
8.66509020e-01 -2.71793038e-01 3.92419875e-01 3.69655825e-02
-1.60307825e+00 -2.01317638e-01 -6.70316592e-02 3.23812604e-01
4.82543170e-01 2.78567821e-01 -1.57951936e-01 6.18389547e-01
1.13876629e+00 1.02965701e+00 -5.36118567e-01 1.24189675e+00
4.39541519e-01 4.77421969e-01 -4.79073107e-01 2.96814680e-01
4.62159187e-01 -7.37533495e-02 7.73104846e-01 1.44556928e+00
2.95046121e-01 2.17861339e-01 5.18345773e-01 5.60136616e-01
-1.24504473e-02 6.88096359e-02 -1.50268450e-01 -3.52028422e-02
1.46837831e-02 1.31477571e+00 -8.55430365e-01 -5.35957336e-01
-6.13977492e-01 1.07300556e+00 2.79366970e-01 3.37361664e-01
-1.29803920e+00 1.75561905e-01 7.32883990e-01 4.77179676e-01
7.79161274e-01 -2.20621079e-01 4.54834811e-02 -1.35309327e+00
3.24385256e-01 -7.06791341e-01 4.93084580e-01 -6.36665344e-01
-8.97117078e-01 4.72888887e-01 7.10431710e-02 -1.33120739e+00
1.15071744e-01 -5.80159187e-01 -2.31482610e-01 3.46794546e-01
-1.62933815e+00 -1.30529249e+00 -4.04235989e-01 9.47479784e-01
1.15270984e+00 1.25436306e-01 3.50828171e-01 7.87695587e-01
-7.28996694e-01 4.66608405e-01 1.99480616e-02 7.57523552e-02
6.35891497e-01 -1.08048308e+00 3.03332031e-01 8.59317243e-01
1.91003025e-01 1.21698804e-01 4.82134014e-01 -5.85561454e-01
-1.51586640e+00 -1.24979734e+00 3.70490342e-01 -6.72876954e-01
7.18281031e-01 -3.26697558e-01 -9.57642078e-01 7.73501873e-01
9.10143703e-02 5.73414326e-01 2.96278298e-01 -3.32809448e-01
-1.85055193e-02 8.19434300e-02 -9.91014063e-01 4.18392688e-01
1.50311732e+00 -2.33772814e-01 -1.25587374e-01 5.72793245e-01
6.88004911e-01 -9.54643726e-01 -9.47943032e-01 4.12446916e-01
5.38661301e-01 -1.04331100e+00 1.06790817e+00 -3.60937744e-01
3.20397764e-01 -3.58866602e-01 -1.03216290e-01 -6.69343412e-01
-3.27053130e-01 -7.12032914e-01 -5.23666263e-01 1.05513966e+00
1.76259473e-01 -3.32686692e-01 8.46241057e-01 6.53594732e-01
-5.44396341e-01 -1.05926967e+00 -1.13567102e+00 -8.12423289e-01
-3.31513524e-01 -6.00717247e-01 3.04685622e-01 6.88194215e-01
-4.38792855e-01 -1.47158876e-01 -3.15496802e-01 1.40851349e-01
6.63241684e-01 2.21219853e-01 6.43639326e-01 -9.90689874e-01
-4.96319622e-01 -3.97832304e-01 -5.17257512e-01 -1.22486734e+00
5.25853485e-02 -8.79066229e-01 -6.94463104e-02 -1.44598925e+00
1.99684173e-01 -2.45916456e-01 -1.99594319e-01 7.02597678e-01
-2.83400953e-01 7.12285578e-01 6.56957388e-01 3.04867029e-01
-1.25666201e+00 3.72272581e-01 1.34362209e+00 -1.69059217e-01
-1.42361701e-01 -3.99533957e-02 -2.22677231e-01 7.31688261e-01
4.92716372e-01 -4.01584834e-01 -3.56330752e-01 -7.16028988e-01
-7.95381814e-02 2.88960487e-01 6.12321317e-01 -1.08455670e+00
3.19622606e-01 -1.60234183e-01 1.72134116e-01 -6.88985527e-01
7.43573189e-01 -7.92979658e-01 2.18017593e-01 3.11801493e-01
4.73152995e-02 1.78451054e-02 4.26650822e-01 6.21195614e-01
-1.81603178e-01 3.30763943e-02 7.49543011e-01 -1.47456944e-01
-1.14735389e+00 8.67632210e-01 -1.76183477e-01 3.59267354e-01
1.44022954e+00 -6.31992996e-01 -1.76775083e-02 2.38694578e-01
-1.08866060e+00 4.75762010e-01 3.67042542e-01 7.31104434e-01
5.93737483e-01 -9.16309834e-01 -6.75646544e-01 1.39807556e-02
-9.16863754e-02 2.01213226e-01 7.84077942e-01 1.27332354e+00
-3.98829222e-01 3.57570648e-01 -1.26666978e-01 -9.64604974e-01
-1.67330837e+00 6.14457250e-01 4.20930684e-01 -1.56599984e-01
-7.90805757e-01 1.04502451e+00 3.66535366e-01 1.00385129e-01
2.42250770e-01 -3.58965814e-01 -1.53617024e-01 3.62315685e-01
5.07765889e-01 3.94257128e-01 9.06452239e-02 -6.81645691e-01
-3.91105294e-01 1.02254260e+00 -1.43880635e-01 1.56252772e-01
1.47178268e+00 -2.72652537e-01 1.45709917e-01 3.37592840e-01
1.11454427e+00 -2.70694375e-01 -1.77846050e+00 -2.34907046e-01
-2.48923719e-01 -5.91354549e-01 2.08602529e-02 -5.56830525e-01
-1.48679507e+00 6.58822477e-01 6.57235384e-01 2.69730777e-01
1.07357824e+00 1.40030548e-01 1.14099765e+00 3.48150432e-02
3.84320915e-01 -1.13995004e+00 1.32854804e-01 3.88946116e-01
8.73839378e-01 -1.28444803e+00 -6.77320734e-02 -5.98986030e-01
-6.69486046e-01 1.02812910e+00 7.42360294e-01 1.78689677e-02
6.33354664e-01 4.01053995e-01 -1.72080815e-01 -2.47572348e-01
-6.54946566e-01 -4.01025891e-01 5.85072458e-01 3.14783573e-01
2.05547512e-01 -2.54002064e-01 -3.57830487e-02 3.30417812e-01
2.97721952e-01 2.96417117e-01 2.19430953e-01 8.45740139e-01
-2.89556473e-01 -7.74215221e-01 -1.77540377e-01 4.20543909e-01
-7.01126993e-01 -5.44612259e-02 1.04069173e-01 6.55996442e-01
3.27908128e-01 6.46480083e-01 1.17390305e-01 -9.70457569e-02
3.94523501e-01 -1.36892319e-01 5.92608750e-01 -5.05733788e-01
-8.05742621e-01 1.47156775e-01 -2.77819019e-02 -1.12560034e+00
-7.25345552e-01 -9.60593402e-01 -1.23832011e+00 -1.84390396e-01
-2.60933101e-01 -2.34290957e-01 4.54881161e-01 1.00082266e+00
5.14050424e-01 6.52814209e-01 1.40168875e-01 -1.41914463e+00
-2.12585479e-01 -5.19092143e-01 -2.50554174e-01 3.86415601e-01
4.88872975e-01 -6.97976828e-01 -4.53882217e-02 2.92369395e-01] | [9.11889362335205, -0.06587432324886322] |
b96da0a4-0744-4f20-9b24-f8cff4bc70be | mipt-nsu-utmn-at-semeval-2021-task-5 | 2104.04739 | null | https://arxiv.org/abs/2104.04739v1 | https://arxiv.org/pdf/2104.04739v1.pdf | MIPT-NSU-UTMN at SemEval-2021 Task 5: Ensembling Learning with Pre-trained Language Models for Toxic Spans Detection | This paper describes our system for SemEval-2021 Task 5 on Toxic Spans Detection. We developed ensemble models using BERT-based neural architectures and post-processing to combine tokens into spans. We evaluated several pre-trained language models using various ensemble techniques for toxic span identification and achieved sizable improvements over our baseline fine-tuned BERT models. Finally, our system obtained a F1-score of 67.55% on test data. | ['Dmitry Morozov', 'Anna Glazkova', 'Mikhail Kotyushev'] | 2021-04-10 | null | https://aclanthology.org/2021.semeval-1.124 | https://aclanthology.org/2021.semeval-1.124.pdf | semeval-2021 | ['toxic-spans-detection'] | ['natural-language-processing'] | [-4.83529896e-01 -4.01176691e-01 -1.77749082e-01 -1.69346869e-01
-1.40666068e+00 -4.33753222e-01 5.42140484e-01 4.05433744e-01
-6.88946724e-01 1.15560639e+00 5.43370426e-01 -1.96255818e-01
1.55104369e-01 -8.34030628e-01 -4.09506410e-01 -2.07747549e-01
-7.12636769e-01 1.47769466e-01 2.10137805e-03 -2.79773980e-01
4.00504678e-01 6.03494644e-01 -6.08063757e-01 6.16268933e-01
7.99117804e-01 1.09487867e+00 -5.11437356e-01 1.16255069e+00
1.42319039e-01 1.24144983e+00 -1.32435203e+00 -4.36334372e-01
-1.29459485e-01 -2.44652480e-01 -9.17855978e-01 -5.19403994e-01
5.55635691e-01 -1.82418913e-01 -9.80141580e-01 6.32502437e-01
8.13192844e-01 5.55415377e-02 8.99940133e-01 -8.94345105e-01
-9.61500704e-01 1.30230784e+00 -3.69921476e-01 8.79117668e-01
3.61653596e-01 4.21077907e-01 9.41989124e-01 -8.66120279e-01
2.89876580e-01 1.01404881e+00 1.07838523e+00 6.25767410e-01
-1.21342790e+00 -7.90611029e-01 -9.74348411e-02 5.11810482e-01
-1.62049854e+00 -5.54674804e-01 5.47398031e-01 -4.21331346e-01
1.63928843e+00 2.38674283e-01 3.33212733e-01 1.43847656e+00
7.83774257e-01 8.28911602e-01 1.02672935e+00 -2.06026122e-01
3.69603559e-02 -3.31831217e-01 5.07680058e-01 4.97298360e-01
1.06369354e-01 1.03978783e-01 -1.14409292e+00 -1.64359316e-01
2.12129489e-01 -4.68537718e-01 1.52222410e-01 1.14275312e+00
-9.38950777e-01 7.85393596e-01 4.00991499e-01 4.32340086e-01
-5.40881336e-01 6.91854596e-01 1.16264379e+00 2.86566734e-01
7.86390722e-01 4.52581644e-01 -1.65658206e-01 -2.02500641e-01
-1.37553370e+00 3.68072093e-01 6.01378143e-01 7.64756083e-01
-2.44114637e-01 4.65822548e-01 -1.08159602e+00 6.13950193e-01
1.15334816e-01 6.35892272e-01 1.58229604e-01 -5.16925752e-01
7.37992167e-01 1.62756592e-01 -4.99000251e-02 -3.70509982e-01
-5.59938312e-01 -5.54040372e-01 -7.61203468e-01 -2.68732291e-02
1.17359400e-01 -4.36054111e-01 -1.25122654e+00 1.56280982e+00
-4.95589197e-01 1.48904389e-02 -2.39072919e-01 2.76558250e-01
8.13089728e-01 4.09470052e-01 9.00011361e-01 2.23144051e-02
1.03044426e+00 -9.40344274e-01 -6.63847148e-01 8.44720155e-02
6.82555854e-01 -3.81064862e-01 7.03295529e-01 6.94920242e-01
-1.20603108e+00 -3.85124922e-01 -1.05005288e+00 1.84869140e-01
-5.64181328e-01 -1.31241724e-01 2.50172079e-01 8.47282648e-01
-1.10261655e+00 9.09245253e-01 -3.25392962e-01 -6.76113069e-02
5.75044811e-01 2.22892299e-01 -1.05224796e-01 4.13509965e-01
-1.73144937e+00 1.08461976e+00 5.67753971e-01 -1.18228726e-01
-1.53690183e+00 -6.07156754e-01 -9.29823458e-01 1.10282518e-01
-9.35641900e-02 -5.42924941e-01 1.33209133e+00 2.15172336e-01
-6.32851303e-01 6.79577529e-01 1.51572883e-01 -1.09166849e+00
3.97361845e-01 -7.04363823e-01 -1.00787508e+00 2.05833316e-01
8.79677944e-03 3.04952741e-01 9.62154791e-02 -6.08076632e-01
-2.62980670e-01 5.68727963e-02 -2.08568975e-01 -1.22013152e-01
-5.60970962e-01 7.00902700e-01 1.50609598e-01 -6.89804792e-01
-1.02551055e+00 -4.63558406e-01 -2.21677557e-01 -4.98897374e-01
-8.16935718e-01 -6.67306900e-01 4.85332489e-01 -1.35041738e+00
1.98296416e+00 -1.38808489e+00 -3.95891458e-01 7.80116841e-02
1.38531387e-01 1.30518720e-01 -3.84562939e-01 6.84286237e-01
-6.84813783e-02 4.69773173e-01 -7.46482164e-02 -6.15860343e-01
1.62989572e-01 -3.90750349e-01 -3.33738387e-01 3.42438132e-01
1.31057218e-01 8.70996654e-01 -7.82937765e-01 -8.12194884e-01
5.29756323e-02 2.10638821e-01 -6.38557598e-02 1.60125479e-01
-2.97835886e-01 1.22584671e-01 -8.62707496e-02 1.09609354e+00
3.29251915e-01 4.02046740e-01 -5.15274763e-01 1.90595984e-01
-6.15520179e-02 5.39657116e-01 -2.73623109e-01 1.83104527e+00
-4.44819778e-01 5.80777407e-01 -5.27293921e-01 -4.61356163e-01
8.79107952e-01 3.55984181e-01 5.78771591e-01 -6.14303052e-01
2.26536214e-01 2.94316541e-02 -3.04678530e-01 -5.08706629e-01
8.61982584e-01 -2.10619837e-01 -9.49548423e-01 3.18560243e-01
2.43567377e-01 1.76634952e-01 6.29927039e-01 4.16008681e-01
1.74375796e+00 -1.11881047e-01 -3.39812115e-02 -2.75676131e-01
5.03126383e-01 -1.36563033e-01 4.00907248e-01 6.75032377e-01
-4.44571018e-01 5.85082531e-01 2.61405230e-01 -2.67894834e-01
-9.10169303e-01 -1.39194083e+00 -1.78930759e-02 1.16728759e+00
-6.00992382e-01 -6.45207405e-01 -6.56746328e-01 -1.14980721e+00
-1.42159648e-02 1.48479247e+00 -7.60181487e-01 -1.90274507e-01
-5.97407758e-01 -5.64344287e-01 1.73017430e+00 1.14125514e+00
5.77718794e-01 -1.53502393e+00 -2.63693959e-01 5.02401590e-01
-1.98535457e-01 -7.11836100e-01 -8.48487496e-01 4.35687423e-01
-4.55319166e-01 -7.38302767e-01 -5.96307755e-01 -4.08093631e-01
-3.60301793e-01 -5.29176652e-01 1.49084294e+00 -7.58957267e-02
-3.57295960e-01 -2.00465798e-01 -1.36675924e-01 -8.38655889e-01
-1.78930134e-01 2.11816952e-01 3.52785051e-01 -7.91872025e-01
5.27213097e-01 -4.37517434e-01 -6.07218146e-01 -1.25735566e-01
-5.45434773e-01 -6.40881717e-01 1.99429706e-01 5.41262805e-01
1.49137691e-01 -3.08353454e-01 1.35862160e+00 -3.88049543e-01
1.06010950e+00 -6.71139657e-01 7.47859627e-02 3.67909372e-01
-5.94744503e-01 -6.23064153e-02 5.98052084e-01 -3.22937846e-01
-8.86051893e-01 -2.64369458e-01 -6.17521405e-01 -4.85366404e-01
-6.55588135e-02 6.04599655e-01 -6.56251237e-02 3.25531274e-01
9.40044880e-01 1.91208959e-01 -5.63335180e-01 -6.76868677e-01
4.60861444e-01 1.05190718e+00 8.59436691e-01 -7.91255891e-01
3.50314349e-01 -2.58889288e-01 -3.81725430e-01 -4.35808510e-01
-9.07918215e-01 -5.40047526e-01 -5.47442913e-01 -5.26754737e-01
9.41009700e-01 -1.08999836e+00 -3.80134374e-01 9.08727765e-01
-1.30333197e+00 -3.91620010e-01 -2.67629951e-01 1.75865635e-01
-5.18245041e-01 3.39358121e-01 -1.29934859e+00 -1.14575112e+00
-1.44511282e+00 -3.89960200e-01 7.94357657e-01 2.60710716e-01
-6.94860399e-01 -1.04210031e+00 6.50407791e-01 8.95187035e-02
6.14636242e-01 6.12397254e-01 7.50736952e-01 -9.98739898e-01
3.02911311e-01 -3.42777669e-01 -1.48515999e-01 2.74655849e-01
-3.81752223e-01 -2.28016734e-01 -1.02135897e+00 -4.87664491e-02
-6.57396317e-01 -8.20802152e-01 1.30798817e+00 1.77199915e-01
1.61113834e+00 -1.79176956e-01 -1.84976280e-01 7.79278874e-02
1.40482569e+00 1.46336749e-01 7.66986132e-01 4.19375420e-01
2.97881126e-01 2.93011451e-03 3.43607932e-01 7.28288412e-01
8.22376609e-02 2.35125661e-01 3.28575253e-01 3.69052202e-01
-3.89309525e-01 -2.89008915e-01 8.95815194e-01 4.90778655e-01
-7.67955482e-02 -7.02768803e-01 -9.38456118e-01 1.03011513e+00
-1.46503150e+00 -1.45838034e+00 -2.44697198e-01 1.62240899e+00
9.39817011e-01 5.70177794e-01 7.09516108e-01 2.18140364e-01
5.26098788e-01 3.82614583e-01 -6.03302538e-01 -1.17953396e+00
-1.05214551e-01 5.78146994e-01 6.27405524e-01 1.63019076e-02
-1.40892243e+00 8.15894008e-01 8.22828388e+00 1.33246303e+00
-6.67987466e-01 2.62605846e-01 7.18997836e-01 -6.13124132e-01
-2.53427863e-01 -7.29450464e-01 -7.97869742e-01 6.32769048e-01
1.91211271e+00 -6.32686377e-01 5.80306426e-02 6.50631845e-01
2.93655664e-01 2.12893385e-04 -1.04350007e+00 5.90525866e-01
4.36760277e-01 -1.31122029e+00 -2.73990810e-01 1.11438580e-01
5.46656668e-01 5.17355144e-01 -2.32226606e-02 1.01169479e+00
6.49214923e-01 -1.54740894e+00 1.03614902e+00 9.58476782e-01
1.20255744e+00 -1.10537052e+00 6.80888236e-01 3.29174399e-01
-1.07468069e+00 -9.75319594e-02 -1.96725219e-01 1.19770057e-01
5.12355983e-01 8.63016486e-01 -7.32520998e-01 5.91210485e-01
6.29070163e-01 4.78937030e-01 -8.47093999e-01 1.51692355e+00
-2.47696862e-01 1.08598113e+00 -4.18505907e-01 -1.69842467e-01
2.54114211e-01 6.63856685e-01 5.84131181e-01 1.91817749e+00
4.11665142e-01 -3.43920141e-01 -8.56616423e-02 7.05627084e-01
-5.33405364e-01 -1.07988358e-01 -6.01584554e-01 -1.21461242e-01
5.28182387e-01 1.11067677e+00 -2.71771789e-01 -5.70443034e-01
1.02048621e-01 7.62104571e-01 4.29800183e-01 -8.38038474e-02
-1.34544492e+00 -7.36012876e-01 2.87901998e-01 -1.32007971e-01
-1.22194752e-01 -3.48371029e-01 -8.69030237e-01 -7.15381145e-01
-4.35606241e-01 -5.61363995e-01 8.78501236e-01 -7.09847689e-01
-1.48185074e+00 9.66626942e-01 2.19261214e-01 -9.27915156e-01
-2.33635917e-01 -4.53815430e-01 -1.45742464e+00 9.20869768e-01
-8.17932785e-01 -1.29088950e+00 1.94528520e-01 2.39618912e-01
9.17428851e-01 -2.67848819e-01 1.03568137e+00 3.10576677e-01
-7.56540060e-01 1.10010314e+00 3.80148627e-02 3.59229952e-01
8.64654362e-01 -1.40857744e+00 6.44595981e-01 1.07051933e+00
-8.56558606e-03 3.16678941e-01 6.77081287e-01 -9.61054385e-01
-4.28117424e-01 -1.59319580e+00 1.34280217e+00 -6.71372116e-01
9.40168500e-01 -1.81901619e-01 -6.38400316e-01 5.61028779e-01
9.58558142e-01 -5.17995000e-01 7.48812675e-01 4.01916087e-01
-5.18953800e-01 9.02531371e-02 -1.40471435e+00 2.80431777e-01
1.17716706e+00 -5.79554796e-01 -7.22734571e-01 4.25303102e-01
5.15723944e-01 5.76445647e-02 -1.31161702e+00 3.07993829e-01
4.22958702e-01 -5.82366467e-01 9.24066842e-01 -8.87114108e-01
7.32134640e-01 1.96937606e-01 -4.18579206e-02 -1.45721638e+00
-4.59067553e-01 -2.52795279e-01 -6.50127649e-01 1.41384423e+00
6.27391875e-01 1.08712539e-01 6.36123598e-01 6.55293226e-01
-4.40606833e-01 -1.11151671e+00 -9.94432032e-01 -1.13670182e+00
7.29210138e-01 -6.86995804e-01 4.51233327e-01 3.22536469e-01
5.04582107e-01 4.06095862e-01 -5.94000518e-01 -3.37429494e-01
7.64108777e-01 -2.50238180e-01 -7.85432458e-02 -5.98298609e-01
-4.50617783e-02 -5.42399108e-01 -1.73959121e-01 -2.59253353e-01
3.88982922e-01 -1.01392210e+00 2.01247036e-01 -1.63150430e+00
5.74745834e-01 -1.48652107e-01 -1.11269784e+00 8.20295632e-01
-7.81123415e-02 6.36408150e-01 2.52734959e-01 -2.08132938e-01
-1.09395647e+00 4.96046662e-01 4.03084964e-01 -5.91744721e-01
1.04877040e-01 -1.94331661e-01 -7.94548213e-01 4.38397557e-01
1.64777303e+00 -6.66419983e-01 1.72400370e-01 -3.07875454e-01
-4.50982898e-02 5.56552708e-02 6.65757135e-02 -1.42072058e+00
2.24363990e-02 -1.67963505e-01 7.87143171e-01 -1.31038165e+00
1.74113631e-01 8.09192136e-02 -1.14234677e-02 6.60441995e-01
-7.18046725e-01 2.15040460e-01 5.00917554e-01 3.97797257e-01
3.28804642e-01 -4.68743205e-01 4.76397097e-01 -3.72872874e-02
-6.71943009e-01 1.51064992e-01 -9.64864731e-01 -5.07578067e-02
1.05323160e+00 2.90176302e-01 -5.72391331e-01 -2.88472891e-01
-9.05590713e-01 5.34580588e-01 1.95884500e-02 3.42162400e-01
7.70506799e-01 -1.43384254e+00 -1.25492430e+00 -5.99406838e-01
9.72347185e-02 -6.58751011e-01 3.37091029e-01 7.18048275e-01
-4.22071576e-01 3.62258941e-01 -2.53950387e-01 1.44012958e-01
-1.50930285e+00 5.36494434e-01 3.62368405e-01 -1.08746159e+00
-3.22980434e-01 9.34048772e-01 -7.02656388e-01 -8.01286474e-02
2.58674949e-01 6.99463412e-02 -3.46467495e-01 3.87254879e-02
8.53627026e-01 9.19046402e-01 9.33161750e-02 -5.20672202e-01
-6.14011586e-01 -1.83735993e-02 -1.33252621e-01 -1.37170777e-01
1.25364506e+00 1.46128401e-01 2.00759903e-01 4.10787731e-01
1.03761470e+00 7.94106275e-02 -5.01605213e-01 8.73878375e-02
3.79850358e-01 2.93479502e-01 1.34284154e-03 -1.58137417e+00
-6.68837190e-01 9.09599364e-01 4.42948967e-01 -1.20605215e-01
1.09174562e+00 -1.51999548e-01 1.21678817e+00 3.65232438e-01
3.17462951e-01 -1.34266424e+00 1.95359349e-01 8.38477910e-01
1.16410863e+00 -6.53716385e-01 8.76098350e-02 9.50348601e-02
-6.18979275e-01 1.15357304e+00 6.69303298e-01 -5.38072169e-01
3.52579653e-01 5.05576849e-01 -1.50032103e-01 -2.57703781e-01
-1.12103057e+00 -1.65248603e-01 2.42021576e-01 5.04658580e-01
5.00804424e-01 1.87580690e-01 -6.95694208e-01 9.58992898e-01
-1.94564730e-01 1.99440092e-01 3.37889820e-01 6.20866716e-01
-7.53621221e-01 -7.07202256e-01 -4.33426738e-01 8.03570509e-01
-1.08514583e+00 -2.80357689e-01 -5.59506953e-01 5.21565259e-01
2.06884384e-01 1.25097299e+00 -7.61282742e-02 -9.63534176e-01
3.16287547e-01 4.74471718e-01 4.46564972e-01 -4.06555593e-01
-1.24836600e+00 -1.06232598e-01 1.02193546e+00 -3.38145256e-01
1.40991464e-01 -5.39955139e-01 -1.29543149e+00 -5.35481870e-01
-2.65718848e-01 3.73623073e-01 2.70450592e-01 6.15005255e-01
1.18773431e-01 8.17229390e-01 6.03376448e-01 -6.12039566e-01
-9.21589971e-01 -1.65443969e+00 -8.08469594e-01 1.86185747e-01
-5.21535240e-02 -1.83117285e-01 -1.82658285e-01 1.90731287e-02] | [8.967960357666016, 10.652958869934082] |
846202d8-bde7-4516-812b-9a2ad536594a | isolation-forest | null | null | https://ieeexplore.ieee.org/abstract/document/4781136 | https://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf?q=isolation-forest | Isolation forest | Most existing model-based approaches to anomaly detection construct a profile of normal instances, then identify instances that do not conform to the normal profile as anomalies. This paper proposes a fundamentally different model-based method that explicitly isolates anomalies instead of profiles normal points. To our best knowledge, the concept of isolation has not been explored in current literature. The use of isolation enables the proposed method, iForest, to exploit sub-sampling to an extent that is not feasible in existing methods, creating an algorithm which has a linear time complexity with a low constant and a low memory requirement. Our empirical evaluation shows that iForest performs favourably to ORCA, a near-linear time complexity distance-based method, LOF and Random Forests in terms of AUC and processing time, and especially in large data sets. iForest also works well in high dimensional problems which have a large number of irrelevant attributes, and in situations where training set does not contain any anomalies. | ['Zhi-Hua Zhou', 'Kai Ming Ting', 'Fei Tony Liu'] | 2008-12-15 | null | null | null | null | ['unsupervised-anomaly-detection-with-specified-5', 'unsupervised-anomaly-detection-with-specified-4', 'unsupervised-anomaly-detection-with-specified-7', 'unsupervised-anomaly-detection-with-specified-6', 'unsupervised-anomaly-detection-with-specified'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 2.25155726e-01 -1.08013069e-02 -1.00037642e-01 -2.38857046e-01
-4.22818989e-01 -4.36421573e-01 5.84732115e-01 5.48016667e-01
-1.75900981e-01 4.84066516e-01 -4.91335362e-01 -5.54465592e-01
-5.61141610e-01 -9.36397910e-01 -1.00521244e-01 -7.32687116e-01
-4.13777858e-01 9.69296277e-01 6.66931927e-01 7.28342608e-02
6.18052244e-01 7.44398534e-01 -1.89477539e+00 1.68519050e-01
9.23172176e-01 1.00766706e+00 -6.81225002e-01 6.06050014e-01
-4.03644711e-01 4.87928957e-01 -5.66933393e-01 -3.00584495e-01
5.03625512e-01 -3.71343881e-01 -7.50854492e-01 1.86663046e-01
4.20220166e-01 -4.93448488e-02 2.71091491e-01 7.94983864e-01
1.65034503e-01 1.03708997e-01 8.21111619e-01 -1.42141807e+00
-1.35219589e-01 1.17870174e-01 -7.67466903e-01 5.46849191e-01
3.80302697e-01 -1.95007026e-01 7.39222288e-01 -5.83839715e-01
2.34055981e-01 9.83192146e-01 8.76170814e-01 2.91751504e-01
-1.49878061e+00 -3.99429560e-01 3.61519635e-01 2.58420467e-01
-1.33377159e+00 -1.41957998e-01 6.33350074e-01 -4.15911198e-01
1.14233720e+00 7.72258937e-01 5.24810970e-01 7.77943254e-01
3.51071432e-02 4.49249893e-01 1.18484998e+00 -6.82396352e-01
6.63693011e-01 5.72111383e-02 3.53587359e-01 4.55411077e-01
7.07652628e-01 9.15804431e-02 -4.23708297e-02 -9.12203670e-01
1.78520560e-01 3.27186853e-01 2.10009038e-01 -6.12964988e-01
-6.50754094e-01 8.43466163e-01 -3.41984153e-01 4.75637138e-01
-4.42086458e-01 -3.96495074e-01 5.50106823e-01 6.81281149e-01
5.44438004e-01 5.34292459e-01 -5.78842938e-01 -9.55534354e-02
-1.11270249e+00 3.23420614e-01 9.39716935e-01 6.73129857e-01
5.39586723e-01 1.81546926e-01 2.41505891e-01 7.32560754e-01
2.32719302e-01 1.92298576e-01 5.07732093e-01 -5.21086574e-01
1.70827448e-01 9.89311397e-01 -5.41486852e-02 -9.55133796e-01
-3.77205342e-01 -3.01754713e-01 -7.39402533e-01 4.94947910e-01
6.23209953e-01 2.51691729e-01 -8.99599850e-01 1.33138192e+00
4.60311413e-01 3.27416003e-01 8.39454606e-02 2.73300141e-01
-5.21031804e-02 1.60370931e-01 1.95882339e-02 -5.34526289e-01
9.24747884e-01 -4.61092293e-01 -4.43119258e-01 -6.48207264e-03
7.52103686e-01 -7.92536318e-01 9.67619896e-01 9.54550505e-01
-6.80831492e-01 -2.13576742e-02 -9.87309635e-01 6.95969105e-01
-6.60340786e-01 -3.87832969e-01 7.05669284e-01 1.03716826e+00
-8.07875037e-01 6.06627047e-01 -8.88854802e-01 -7.24089801e-01
4.18645948e-01 2.94713557e-01 -3.15192223e-01 3.42675857e-02
-7.00284183e-01 7.84725785e-01 4.59073603e-01 -2.67020047e-01
-3.90231103e-01 -4.15980309e-01 -5.01796901e-01 -4.60138619e-02
5.71265459e-01 -1.84422299e-01 9.20755446e-01 -7.79614627e-01
-8.27388942e-01 6.57734990e-01 -3.79850209e-01 -8.07008743e-01
7.19786823e-01 -1.99546888e-01 -9.51053023e-01 -9.09716412e-02
-2.68298085e-03 -2.86949068e-01 8.16270947e-01 -1.07480896e+00
-8.28658879e-01 -7.11950898e-01 -2.40960449e-01 -3.01380306e-01
-2.99079239e-01 -7.84329548e-02 1.56407692e-02 -4.93627697e-01
4.14735317e-01 -8.79539311e-01 -5.75973988e-01 -2.43655667e-01
-4.50409710e-01 -3.06054682e-01 1.07941914e+00 -1.97961926e-01
1.61704624e+00 -1.91171324e+00 -5.27753115e-01 1.03395224e+00
1.79675236e-01 6.03798628e-01 1.77273154e-01 7.50902414e-01
-2.89423078e-01 1.60564139e-01 -7.23547220e-01 1.89612955e-01
-1.66965291e-01 4.71709460e-01 -2.90493488e-01 5.31691730e-01
1.45341948e-01 6.36099502e-02 -7.31261969e-01 -4.04280365e-01
1.43170267e-01 -1.65248767e-01 -6.11529946e-01 7.25043789e-02
-1.01364166e-01 4.51879203e-02 -5.19553006e-01 8.53030801e-01
7.62265086e-01 2.93145236e-02 1.08595230e-01 5.97415924e-01
-5.10257743e-02 -5.64880110e-03 -1.50662482e+00 9.23568785e-01
-9.39926356e-02 3.09532791e-01 -4.28580374e-01 -1.42062092e+00
1.24269652e+00 3.29953492e-01 7.14998722e-01 -3.99648845e-01
-2.72368312e-01 5.43896854e-01 1.31760135e-01 -5.32128274e-01
5.74461222e-02 1.51701467e-02 1.86250746e-01 4.13465202e-01
-2.57256508e-01 5.52181423e-01 3.18832874e-01 -3.57404165e-02
1.64858723e+00 -5.15826568e-02 8.15878630e-01 -3.33906144e-01
9.40048695e-01 -7.21792504e-02 5.21825552e-01 1.01655817e+00
-8.18367749e-02 4.73989636e-01 7.64101028e-01 -7.40828395e-01
-1.00360060e+00 -1.12413573e+00 -4.42004174e-01 8.82215500e-01
-2.67045259e-01 -6.04727089e-01 -5.67988694e-01 -9.54679251e-01
1.99100062e-01 8.24427903e-01 -6.50407135e-01 -2.01317780e-02
-4.43433344e-01 -1.16411877e+00 5.27999341e-01 1.08817227e-01
1.49324745e-01 -9.18714046e-01 -5.40317416e-01 3.49740952e-01
3.36280912e-01 -7.92587757e-01 2.37190679e-01 2.54922658e-01
-1.35062599e+00 -1.39027929e+00 2.15444174e-02 -1.81927100e-01
8.14431727e-01 -1.75644726e-01 1.21649861e+00 2.23488450e-01
-4.35165316e-01 1.42808154e-01 -4.85453725e-01 -7.11101592e-01
-3.13612252e-01 7.33235429e-05 4.91340049e-02 2.22977981e-01
1.22314239e+00 -9.31240916e-01 -2.05718577e-01 3.62929821e-01
-1.02189183e+00 -7.86237419e-01 5.85752964e-01 7.54689574e-01
6.35774374e-01 3.04873794e-01 7.66947150e-01 -1.51292157e+00
6.22777402e-01 -8.76163661e-01 -5.67825317e-01 9.05131474e-02
-1.44906867e+00 -7.25775165e-03 7.81747341e-01 -2.57373661e-01
-7.38686800e-01 1.47968620e-01 -1.93909243e-01 -3.47454786e-01
-7.81721950e-01 6.78748190e-02 -4.58350480e-02 1.07912771e-01
8.98981154e-01 3.92627388e-01 1.39849558e-01 -8.08634520e-01
-2.23701790e-01 6.96840048e-01 2.66336262e-01 -4.14863735e-01
7.93750346e-01 5.34830809e-01 3.63785744e-01 -9.35382187e-01
-5.13422608e-01 -8.37208390e-01 -8.58392000e-01 2.00278327e-01
2.58529723e-01 -3.80043566e-01 -2.50710964e-01 2.53396481e-01
-5.60985684e-01 3.45811278e-01 -4.18697149e-01 2.47679025e-01
-4.15440232e-01 5.81403375e-01 -1.04659021e-01 -1.29877889e+00
-2.67759264e-01 -6.04625702e-01 5.92122436e-01 -1.49079457e-01
-4.79351133e-01 -9.39937651e-01 3.25355053e-01 -8.75772461e-02
2.82277316e-01 8.28551829e-01 1.01640868e+00 -1.62187064e+00
-1.70179799e-01 -8.37826133e-01 1.91462293e-01 2.55836695e-01
1.80647776e-01 2.06905618e-01 -1.03341079e+00 -2.92875022e-01
-5.38872480e-02 3.26286554e-01 6.81879878e-01 7.32399672e-02
1.44155133e+00 -3.89627755e-01 -3.88428509e-01 3.56868774e-01
1.47826898e+00 5.97483218e-01 7.24712253e-01 6.88738465e-01
3.03733498e-01 7.06057370e-01 7.17385352e-01 5.93133926e-01
-5.52753918e-02 5.26567817e-01 4.82469112e-01 -1.20786309e-01
4.87939507e-01 1.24413215e-01 7.04652965e-02 4.62223887e-01
-2.07132518e-01 -5.70621192e-02 -1.16985893e+00 7.66736984e-01
-1.97801256e+00 -1.37122750e+00 -5.84279299e-01 2.52186537e+00
2.05510214e-01 4.68016922e-01 6.24529660e-01 8.51104259e-01
5.34380972e-01 -1.42443657e-01 -3.60073239e-01 -1.03524625e+00
2.36133501e-01 2.53068596e-01 2.49826983e-01 1.92552581e-01
-1.15419757e+00 3.64295244e-01 6.63733959e+00 6.34723485e-01
-7.80462980e-01 -1.22511365e-01 4.25247967e-01 -1.30192280e-01
-8.60754550e-02 1.79739609e-01 -5.06610572e-01 6.76665008e-01
1.22095525e+00 -4.95695323e-02 6.39941841e-02 1.22155654e+00
-6.98815240e-03 -2.13557154e-01 -1.11104202e+00 6.77286744e-01
-3.38702090e-02 -7.87457168e-01 -2.85889041e-02 3.92359853e-01
4.24978048e-01 -1.13476127e-01 -2.28601620e-01 8.21746439e-02
-2.05562892e-03 -8.63284886e-01 9.28406194e-02 4.75118607e-01
3.13780785e-01 -1.12410057e+00 1.00147378e+00 3.64037752e-01
-9.90658402e-01 -2.96154559e-01 -3.63655418e-01 -1.80149511e-01
-2.01038375e-01 9.11025286e-01 -9.64471579e-01 6.40754104e-01
8.90303612e-01 4.25706089e-01 -8.19951653e-01 1.35327554e+00
3.98818612e-01 9.40799117e-01 -5.80959976e-01 1.96692914e-01
3.32552254e-01 -2.77914703e-01 7.48602808e-01 1.38189042e+00
4.24816102e-01 -2.61597216e-01 3.18652868e-01 2.57309407e-01
6.88807547e-01 4.12860841e-01 -9.61216807e-01 1.26965731e-01
4.85164165e-01 9.63088036e-01 -7.83181906e-01 -3.52939397e-01
-7.43982196e-01 6.11944795e-01 -3.44598405e-02 1.44553065e-01
-3.43055010e-01 -6.30786002e-01 6.06359422e-01 5.41354835e-01
2.18788937e-01 1.90132648e-01 -4.26081479e-01 -9.46888506e-01
3.56633931e-01 -1.02874708e+00 9.18369114e-01 1.78330153e-01
-1.66908312e+00 8.70910943e-01 2.44936034e-01 -1.63721991e+00
-6.24264717e-01 -5.66213489e-01 -8.25668812e-01 7.44564891e-01
-1.05124784e+00 -6.96059287e-01 -1.35396495e-01 7.36265123e-01
2.94510335e-01 -4.98527914e-01 1.22958660e+00 1.91066369e-01
-5.46266139e-01 5.85673988e-01 2.25848660e-01 -1.06920794e-01
4.85968441e-01 -1.52495706e+00 4.55878198e-01 1.12458253e+00
5.23131080e-02 6.09390318e-01 9.22883153e-01 -5.54967523e-01
-6.72483385e-01 -1.01812840e+00 8.65523219e-01 -5.49792647e-01
5.15182495e-01 -1.95432171e-01 -1.39261270e+00 5.56399107e-01
-2.63629794e-01 2.14595839e-01 1.16791022e+00 4.97147113e-01
-4.04076070e-01 -2.10911900e-01 -1.61774766e+00 3.19965333e-01
9.55937624e-01 -1.86842650e-01 -6.88884914e-01 2.40452304e-01
-1.49694443e-01 8.13438669e-02 -8.43774140e-01 5.44139087e-01
5.58096886e-01 -1.37566733e+00 8.30259800e-01 -8.36495101e-01
-1.92764759e-01 -3.63212138e-01 -1.97804928e-01 -9.50474560e-01
-1.83193937e-01 -4.82458025e-01 -5.33164442e-01 1.25682902e+00
4.37223017e-01 -1.28726006e+00 8.43672514e-01 4.20624793e-01
1.07354291e-01 -1.03783917e+00 -1.02782989e+00 -1.18076956e+00
-1.28259853e-01 -5.77229202e-01 9.51367736e-01 1.02695191e+00
-1.73170775e-01 -2.69599527e-01 -2.82666564e-01 2.45962679e-01
7.87536800e-01 1.65933177e-01 7.29674816e-01 -2.06804442e+00
-2.22884938e-01 -3.83290768e-01 -9.36315417e-01 -1.11090809e-01
-1.85297579e-01 -5.69688737e-01 -3.99108887e-01 -1.00152898e+00
-1.29426077e-01 -7.44093835e-01 -6.79838657e-01 5.26606202e-01
-3.91387008e-02 1.13393269e-01 -4.31028813e-01 3.78735781e-01
-4.44872051e-01 -1.21612232e-02 4.87728655e-01 3.03596318e-01
-2.77925074e-01 3.87415141e-01 -6.41008496e-01 1.08462286e+00
9.98447359e-01 -8.66771579e-01 -5.17140269e-01 4.01591718e-01
-1.62905961e-01 -3.03026021e-01 6.76419139e-02 -1.28957248e+00
-2.34239012e-01 -4.18076664e-01 6.77550733e-01 -5.24026275e-01
-9.10216868e-02 -1.09655952e+00 2.54930705e-01 5.90462387e-01
-2.02797323e-01 6.44162774e-01 -5.11033349e-02 7.37470031e-01
-2.94052035e-01 -4.30969834e-01 8.52004349e-01 -1.50385007e-01
-7.05598116e-01 2.90630221e-01 -5.06685495e-01 -5.89129180e-02
1.44558024e+00 -5.64063370e-01 3.96318771e-02 -5.52588403e-02
-7.96369553e-01 -6.01921566e-02 6.62447751e-01 3.60542446e-01
3.79324973e-01 -1.18501663e+00 -4.55530554e-01 5.54253042e-01
4.99440104e-01 -1.83124155e-01 7.59036280e-03 8.64000916e-01
-6.14695966e-01 2.97899097e-01 -2.38554180e-01 -8.00686777e-01
-1.43556738e+00 8.32783818e-01 1.44030347e-01 -4.68490630e-01
-9.36664760e-01 2.48271257e-01 -1.68965846e-01 -4.81916010e-01
7.00554177e-02 1.09914660e-01 -3.31754655e-01 1.10079326e-01
7.01168120e-01 6.72314048e-01 5.10444641e-01 -4.74199206e-01
-5.39342940e-01 3.97615820e-01 -3.39366436e-01 1.91227898e-01
1.16314447e+00 1.05954908e-01 -2.40460932e-01 5.98064005e-01
7.70285368e-01 1.10423334e-01 -7.23802686e-01 -2.72034079e-01
8.76608849e-01 -9.50313926e-01 -3.34940255e-01 -7.19034910e-01
-7.83196032e-01 6.75307751e-01 9.03374970e-01 8.59307349e-01
1.53142810e+00 -2.74141043e-01 3.40082765e-01 4.96997863e-01
4.37761992e-01 -1.05576432e+00 -2.18859211e-01 3.19228441e-01
5.22440434e-01 -1.03025341e+00 -5.42675778e-02 -2.85113841e-01
-4.30570126e-01 1.15179384e+00 9.05468941e-01 -3.22530091e-01
6.90498590e-01 3.33672285e-01 9.13323760e-02 -1.40715633e-02
-9.49332178e-01 -3.30340236e-01 2.00534940e-01 8.67210686e-01
1.73013806e-01 2.80836709e-02 -6.40836656e-01 2.00240031e-01
-3.94764543e-02 -3.09368193e-01 5.03093123e-01 1.18008375e+00
-5.32692552e-01 -1.40293527e+00 -4.81315523e-01 1.06418037e+00
-8.27790558e-01 1.85751885e-01 -4.43613261e-01 1.03685284e+00
3.54705572e-01 9.05813098e-01 3.14448804e-01 -3.43942940e-01
6.20403051e-01 6.40072405e-01 -6.43818527e-02 -4.98990238e-01
-4.48346704e-01 -1.65333804e-02 1.31596789e-01 -7.28116274e-01
-1.18101791e-01 -8.30036700e-01 -9.06381369e-01 -2.90678024e-01
-3.10008317e-01 3.90622258e-01 2.71194011e-01 9.97914255e-01
3.06367993e-01 2.12954357e-01 7.55155265e-01 -2.53690630e-01
-5.34753442e-01 -8.65910947e-01 -7.87583351e-01 5.90518534e-01
4.43520814e-01 -7.37971723e-01 -4.65550423e-01 -3.89873236e-01] | [7.551881790161133, 2.783468008041382] |
09bb1f8f-31f3-4b07-8510-5c6430be120d | the-2015-sheffield-system-for-transcription | 1512.06643 | null | http://arxiv.org/abs/1512.06643v1 | http://arxiv.org/pdf/1512.06643v1.pdf | The 2015 Sheffield System for Transcription of Multi-Genre Broadcast Media | We describe the University of Sheffield system for participation in the 2015
Multi-Genre Broadcast (MGB) challenge task of transcribing multi-genre
broadcast shows. Transcription was one of four tasks proposed in the MGB
challenge, with the aim of advancing the state of the art of automatic speech
recognition, speaker diarisation and automatic alignment of subtitles for
broadcast media. Four topics are investigated in this work: Data selection
techniques for training with unreliable data, automatic speech segmentation of
broadcast media shows, acoustic modelling and adaptation in highly variable
environments, and language modelling of multi-genre shows. The final system
operates in multiple passes, using an initial unadapted decoding stage to
refine segmentation, followed by three adapted passes: a hybrid DNN pass with
input features normalised by speaker-based cepstral normalisation, another
hybrid stage with input features normalised by speaker feature-MLLR
transformations, and finally a bottleneck-based tandem stage with noise and
speaker factorisation. The combination of these three system outputs provides a
final error rate of 27.5% on the official development set, consisting of 47
multi-genre shows. | ['Yu-Lan Liu', 'Mortaza Doulaty', 'Rosanna Milner', 'Raymond W. M. Ng', 'Oscar Saz', 'Thomas Hain', 'Salil Deena', 'Madina Hasan'] | 2015-12-21 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 4.37214553e-01 1.15639322e-01 5.85972607e-01 -6.65622354e-01
-1.75542796e+00 -6.63612366e-01 6.59054160e-01 7.07101077e-02
-5.14268398e-01 3.26822758e-01 6.78593099e-01 -3.13834608e-01
6.24108873e-02 1.83431283e-01 -5.38120985e-01 -8.25169861e-01
1.64479658e-01 7.81102121e-01 1.89714849e-01 -4.57978845e-01
2.37514153e-02 3.05009454e-01 -1.45757973e+00 9.11403835e-01
2.81942874e-01 8.97716343e-01 1.63431123e-01 1.44513690e+00
8.62876177e-02 5.52314043e-01 -9.95203257e-01 -3.69108140e-01
-6.58183172e-03 -5.25407076e-01 -1.19825912e+00 1.17911734e-01
5.59967279e-01 2.91110631e-02 3.91779505e-02 4.98528689e-01
1.09166634e+00 3.60631555e-01 5.27040422e-01 -8.48498046e-01
5.25617711e-02 1.24148130e+00 -1.73227459e-01 7.74074793e-01
5.01580119e-01 -1.38317838e-01 8.22502077e-01 -9.21696186e-01
1.68008387e-01 1.49809849e+00 7.46926010e-01 5.40290594e-01
-1.51886666e+00 -5.54842770e-01 -1.37501061e-01 1.28238082e-01
-1.41906250e+00 -1.28197730e+00 3.45147878e-01 -5.25038600e-01
1.42181802e+00 7.26864994e-01 4.70353872e-01 1.26392138e+00
-4.16763783e-01 6.82042897e-01 9.34864342e-01 -6.94161177e-01
1.10840566e-01 1.35212228e-01 -7.48176686e-03 -2.92021215e-01
-9.25218046e-01 -2.72589266e-01 -7.24159598e-01 -7.45724365e-02
2.52938479e-01 -9.79481578e-01 -2.08864570e-01 5.49298465e-01
-1.31259799e+00 7.88922429e-01 -2.24849835e-01 5.22157490e-01
-4.31240082e-01 -1.51151076e-01 6.82931304e-01 6.42869473e-01
5.80327153e-01 3.51185560e-01 -5.82278609e-01 -4.72087145e-01
-1.41402209e+00 3.73103827e-01 9.49693799e-01 1.04228795e+00
1.91393450e-01 3.27398509e-01 -2.35503912e-01 1.63653648e+00
4.65201527e-01 6.61127269e-01 6.28682256e-01 -7.55191743e-01
7.17299521e-01 -1.24365240e-01 -2.50450015e-01 -3.28962356e-01
-5.84916651e-01 -1.83398828e-01 -6.47590518e-01 -5.48813701e-01
3.01352262e-01 -4.87500817e-01 -9.87570286e-01 1.57496500e+00
4.54814821e-01 -1.45217506e-02 2.90965527e-01 6.60083890e-01
9.57954049e-01 1.22683656e+00 -1.88595995e-01 -3.25531334e-01
1.37198794e+00 -1.01581740e+00 -6.98746324e-01 -7.55849332e-02
4.56293166e-01 -1.43054581e+00 7.68543363e-01 6.63501024e-01
-1.33840156e+00 -7.03117728e-01 -9.59501147e-01 -9.17931274e-02
-1.65280402e-01 1.31211558e-03 -2.19671518e-01 7.92522132e-01
-1.43251276e+00 -1.85697064e-01 -6.52683496e-01 -3.79698157e-01
-5.29692397e-02 5.29526055e-01 -1.53193489e-01 2.71541834e-01
-1.19558740e+00 8.09212327e-01 3.36023569e-01 1.49697796e-01
-8.84438992e-01 -5.51124454e-01 -8.04250598e-01 -1.48477271e-01
3.32678109e-02 -5.55017516e-02 1.83158576e+00 -1.20416939e+00
-1.81168747e+00 7.93208003e-01 -3.23286533e-01 -6.75442696e-01
4.34638888e-01 -1.29660398e-01 -7.92374074e-01 -8.62051025e-02
-1.52026638e-01 4.67653632e-01 9.11985874e-01 -9.49016511e-01
-1.05608952e+00 -2.08717078e-01 -5.60871422e-01 4.36882615e-01
9.98700708e-02 7.84526587e-01 -4.23891932e-01 -5.65721214e-01
1.52068704e-01 -7.97161937e-01 3.59115899e-02 -1.33429873e+00
-3.93763304e-01 -1.70341507e-01 5.67355156e-01 -1.40537715e+00
1.29535127e+00 -2.43396926e+00 4.37421143e-01 4.07365918e-01
-4.09004807e-01 -3.57758328e-02 -1.18998259e-01 4.52750117e-01
-8.47903714e-02 -1.20356418e-01 -7.26654753e-02 -6.95482671e-01
1.04239628e-01 -4.09386642e-02 -2.99701899e-01 3.17586929e-01
2.08092615e-01 3.69212121e-01 -4.68732178e-01 -3.11686099e-01
1.68808416e-01 5.81980348e-01 -2.76973635e-01 2.44184911e-01
1.04743488e-01 6.45527720e-01 3.63260180e-01 4.65422839e-01
4.59660143e-01 7.08139479e-01 -2.24298462e-02 4.85803410e-02
-5.28830349e-01 8.15548658e-01 -1.40157080e+00 1.57600498e+00
-4.49110687e-01 7.42315650e-01 7.63519108e-01 -8.72545481e-01
9.39573050e-01 8.71716797e-01 3.06818366e-01 -4.02054012e-01
2.42722109e-01 3.97473842e-01 2.02801481e-01 -4.33844507e-01
6.78708971e-01 -1.97020367e-01 -3.66338938e-01 1.49380118e-01
6.38785779e-01 -7.73619533e-01 3.08916688e-01 9.32649449e-02
8.07690144e-01 -4.20319825e-01 -5.78229316e-02 -2.45397821e-01
7.06211507e-01 -2.71127641e-01 3.84340525e-01 6.66060328e-01
-1.35484591e-01 1.10765409e+00 2.88565814e-01 5.36126196e-02
-1.15535390e+00 -6.89795017e-01 -2.81796932e-01 1.49701440e+00
-5.57189465e-01 -4.37972009e-01 -1.01953852e+00 -2.11082041e-01
-4.03062761e-01 9.88813996e-01 -3.72982174e-01 2.54015565e-01
-8.58869135e-01 -7.91179538e-01 1.05851972e+00 2.67806441e-01
8.34100097e-02 -1.10552740e+00 2.05215123e-02 4.62829798e-01
-6.53493047e-01 -1.08109796e+00 -6.59941792e-01 6.51143968e-01
-1.77348956e-01 -3.86702418e-01 -9.19247270e-01 -1.10348165e+00
-1.12537198e-01 -1.42086104e-01 8.32869291e-01 -3.24008167e-01
1.52874097e-01 2.02836618e-01 -4.89064932e-01 -4.87732440e-01
-1.26026750e+00 4.72243905e-01 5.44857159e-02 2.16838181e-01
2.36560285e-01 -3.57867628e-01 -1.37965024e-01 6.05566859e-01
-6.70852125e-01 1.05595544e-01 4.72854942e-01 8.23186815e-01
4.68148708e-01 -2.50945181e-01 6.77425563e-01 -4.65959579e-01
5.97773552e-01 -2.22084388e-01 -3.95929247e-01 2.42897779e-01
1.06272303e-01 -4.78706032e-01 3.26832324e-01 -4.85542089e-01
-1.15822363e+00 1.18811376e-01 -1.10201705e+00 1.38202623e-01
-4.27770048e-01 3.27977598e-01 -5.42571664e-01 2.34729886e-01
8.50181222e-01 2.88078547e-01 -1.23902120e-01 -5.74301541e-01
3.39711279e-01 1.60832584e+00 7.82351315e-01 -3.14803064e-01
3.19584966e-01 -3.60723794e-01 -8.16409171e-01 -1.35040772e+00
-6.28989220e-01 -9.59559798e-01 -8.71159196e-01 -3.91988218e-01
9.87637281e-01 -9.69758272e-01 -1.66369811e-01 8.39420974e-01
-1.08727133e+00 -5.67965627e-01 -2.33215839e-01 4.85198110e-01
-6.06908977e-01 -1.09618276e-01 -6.60387158e-01 -8.32370043e-01
-4.01299477e-01 -1.41013730e+00 1.24196923e+00 -1.63559821e-02
-5.64156890e-01 -5.54669082e-01 2.03796357e-01 8.94828439e-01
4.33697522e-01 -3.27939630e-01 6.59254253e-01 -1.33995068e+00
2.89305776e-01 -6.37716800e-02 4.26187187e-01 9.27101612e-01
-1.81721777e-01 -2.17854008e-02 -1.35342479e+00 -2.62263417e-01
8.54696110e-02 -2.63046473e-01 5.07711589e-01 4.94151086e-01
5.78155816e-01 -3.87811184e-01 1.35856524e-01 5.30216515e-01
6.66341484e-01 3.94901574e-01 4.45379585e-01 3.20919484e-01
3.81060719e-01 6.23522282e-01 3.35913271e-01 1.45760804e-01
4.08786625e-01 8.45133960e-01 -6.70277104e-02 1.83406081e-02
-1.41763330e-01 3.59774768e-01 6.33827627e-01 1.45145094e+00
1.08873472e-01 -2.84850895e-01 -1.28743982e+00 7.29598880e-01
-1.54719532e+00 -8.82079661e-01 -4.57722306e-01 2.15293193e+00
9.26003158e-01 2.77242959e-01 6.78767383e-01 5.27955711e-01
8.54579210e-01 6.01011850e-02 1.07516855e-01 -9.07903016e-01
-3.22969973e-01 -2.71751615e-03 4.50959206e-01 6.37507796e-01
-1.20455098e+00 9.00948584e-01 6.34907150e+00 9.48544264e-01
-1.07821643e+00 4.42722082e-01 6.31338298e-01 -3.48329395e-01
3.40987463e-03 -1.85979113e-01 -1.08094788e+00 4.42062914e-01
1.82222784e+00 2.20180213e-01 6.09348655e-01 3.97427708e-01
3.66890639e-01 3.28470990e-02 -9.48951662e-01 9.01930630e-01
4.38312501e-01 -1.02392685e+00 -3.36938322e-01 -1.92231134e-01
6.57845438e-01 5.65310597e-01 -2.30626166e-01 5.43971360e-01
4.18149270e-02 -7.48685420e-01 1.42615080e+00 1.80610254e-01
5.93521774e-01 -9.71200883e-01 8.95936608e-01 4.02832776e-01
-8.43582690e-01 -1.84710920e-01 7.75121525e-02 3.76672804e-01
6.20380223e-01 1.33097142e-01 -1.29887629e+00 4.92379457e-01
8.35338175e-01 9.12679806e-02 -3.40281874e-01 1.01225114e+00
9.27731916e-02 1.39725137e+00 -6.21381581e-01 1.90852702e-01
2.60575533e-01 2.98432082e-01 8.67703319e-01 1.92818272e+00
1.35151401e-01 -1.11548215e-01 1.80988405e-02 -2.82269064e-02
1.87683448e-01 2.84772456e-01 1.51014790e-01 2.98657179e-01
5.36908567e-01 1.20379853e+00 -6.81483626e-01 -3.93713742e-01
-1.68702856e-01 8.66109133e-01 -1.75108954e-01 2.70511329e-01
-7.66482890e-01 -3.06554615e-01 1.37011752e-01 7.77507573e-03
3.59444678e-01 -2.95311194e-02 -1.42206237e-01 -6.39918804e-01
-2.56459951e-01 -1.46798408e+00 3.79829735e-01 -5.76423526e-01
-8.73283744e-01 1.02051878e+00 8.44862685e-02 -6.08180344e-01
-5.77777565e-01 1.29779149e-02 -4.03908730e-01 1.23656416e+00
-8.43263090e-01 -1.42167222e+00 1.19317308e-01 3.45088720e-01
1.15275145e+00 -3.54165941e-01 9.57293749e-01 5.65345824e-01
-5.74066639e-01 6.82897925e-01 3.80945235e-01 -1.21067598e-01
7.88330734e-01 -1.43171489e+00 7.80146360e-01 7.86025882e-01
1.00835212e-01 5.01067750e-02 8.09489846e-01 -3.16754848e-01
-9.15857613e-01 -9.01635528e-01 1.19762206e+00 -4.20251787e-01
6.66513860e-01 -9.93937790e-01 -8.54474366e-01 7.28491366e-01
3.38928372e-01 -6.00349009e-01 8.27609181e-01 2.02576116e-01
1.71555072e-01 -2.28506625e-01 -8.40302467e-01 6.38852268e-02
3.79630983e-01 -3.96213382e-01 -7.51010299e-01 2.76194841e-01
6.67584300e-01 -7.42057502e-01 -9.21460986e-01 1.06377438e-01
5.70834279e-01 -4.54918206e-01 7.08136439e-01 -1.96770370e-01
-6.22438826e-02 -7.29177669e-02 -4.94813979e-01 -1.49336755e+00
-1.50289834e-01 -1.17939067e+00 4.34852004e-01 2.04281545e+00
9.18578029e-01 -3.04551393e-01 7.95717235e-04 2.10844398e-01
-6.57736242e-01 -2.17096046e-01 -1.48328781e+00 -4.08046216e-01
-1.35744572e-01 -7.72642136e-01 5.11004210e-01 5.06298959e-01
-8.75107422e-02 6.77952170e-01 -3.57792497e-01 1.27272710e-01
7.37495497e-02 -5.88505983e-01 9.44047213e-01 -8.27192008e-01
-3.54392737e-01 -5.47124803e-01 -3.55539203e-01 -1.05829859e+00
-1.37919486e-01 -7.47587621e-01 6.92148447e-01 -1.31103683e+00
-3.35760802e-01 -1.77612737e-01 4.61300015e-02 2.64570296e-01
2.01749936e-01 2.24811420e-01 2.77708489e-02 1.24157295e-01
-4.91495758e-01 2.39308476e-01 4.77454782e-01 -7.50487819e-02
-4.78609443e-01 6.27864718e-01 -4.71964866e-01 5.49179971e-01
6.55980885e-01 -4.92260635e-01 -7.04316869e-02 -5.36813736e-01
-1.17916748e-01 1.65525585e-01 -9.61213838e-03 -8.86814415e-01
1.83725446e-01 2.50309467e-01 1.26308560e-01 -8.71602178e-01
4.98165846e-01 -5.17849147e-01 2.38870293e-01 -5.99806197e-02
-4.83908623e-01 1.14658549e-01 5.83684325e-01 1.88797433e-02
-3.52969319e-01 -3.36602181e-01 8.65126193e-01 1.07326187e-01
-2.66102403e-01 -4.32973027e-01 -1.03526282e+00 1.39416888e-01
6.48730874e-01 -2.09181383e-01 2.57801801e-01 -4.87466186e-01
-9.84595537e-01 2.10178033e-01 -1.83362842e-01 5.30803859e-01
7.94574246e-02 -9.45778847e-01 -1.29219174e+00 4.75745410e-01
-3.01777929e-01 2.12577656e-01 3.72281700e-01 1.22373700e+00
-5.29650986e-01 3.08305591e-01 2.98149377e-01 -8.88951242e-01
-1.83974791e+00 -2.50717223e-01 4.01483655e-01 -4.03854474e-02
-3.02809447e-01 1.37696362e+00 -2.59248108e-01 -4.84123975e-01
5.42446196e-01 -1.78961679e-01 -2.94128686e-01 4.70400512e-01
5.90192080e-01 4.76414055e-01 7.56475270e-01 -1.46789384e+00
-4.23631966e-01 4.36499789e-02 -2.91817218e-01 -9.27271485e-01
1.39194775e+00 -5.12471318e-01 1.20486878e-02 9.57997978e-01
9.04073596e-01 7.30916113e-02 -1.01913488e+00 1.68618485e-02
2.39699304e-01 7.37933442e-02 1.97707340e-01 -1.11149788e+00
-3.68457556e-01 7.33441114e-01 4.00417089e-01 6.64146721e-01
1.14358592e+00 1.05273709e-01 7.67110348e-01 1.05003662e-01
-3.36431563e-01 -1.53377724e+00 -5.00610530e-01 1.02556276e+00
1.31012261e+00 -9.38580394e-01 -4.76629823e-01 1.23788444e-02
-8.21274459e-01 1.05458343e+00 3.01238805e-01 4.69478577e-01
5.29264569e-01 5.66191614e-01 4.89847153e-01 8.00989717e-02
-7.73656726e-01 -4.14809299e-04 5.36755741e-01 4.40330565e-01
5.75470924e-01 -3.75652388e-02 8.12435076e-02 8.19082499e-01
-9.39006209e-01 -7.90409863e-01 1.70653671e-01 7.16095626e-01
-4.22803313e-01 -8.37098002e-01 -8.27317238e-01 3.71692836e-01
-7.93885469e-01 -2.51708835e-01 -6.13563478e-01 4.86207604e-01
1.60637265e-03 1.37899661e+00 4.98952977e-02 -4.01284635e-01
7.58762598e-01 5.45709252e-01 2.57831186e-01 -7.20632255e-01
-1.36148751e+00 9.24267530e-01 5.32892466e-01 -5.61147109e-02
-3.87469947e-01 -1.26979613e+00 -8.98796201e-01 -2.74795499e-02
-6.98964477e-01 4.46882874e-01 1.26502621e+00 1.12441432e+00
-3.02810837e-02 7.18551159e-01 5.89377463e-01 -1.26165628e+00
-4.93754089e-01 -1.73371851e+00 -3.17367822e-01 1.03144385e-01
5.67428887e-01 1.10449204e-02 -6.13439023e-01 3.36298883e-01] | [14.502267837524414, 6.531845569610596] |
659f5db1-aa74-4c2e-8247-c8f723562a84 | 3d-axial-attention-for-lung-nodule | 2012.14117 | null | https://arxiv.org/abs/2012.14117v3 | https://arxiv.org/pdf/2012.14117v3.pdf | 3D Axial-Attention for Lung Nodule Classification | Purpose: In recent years, Non-Local based methods have been successfully applied to lung nodule classification. However, these methods offer 2D attention or limited 3D attention to low-resolution feature maps. Moreover, they still depend on a convenient local filter such as convolution as full 3D attention is expensive to compute and requires a big dataset, which might not be available. Methods: We propose to use 3D Axial-Attention, which requires a fraction of the computing power of a regular Non-Local network (i.e., self-attention). Unlike a regular Non-Local network, the 3D Axial-Attention network applies the attention operation to each axis separately. Additionally, we solve the invariant position problem of the Non-Local network by proposing to add 3D positional encoding to shared embeddings. Results: We validated the proposed method on 442 benign nodules and 406 malignant nodules, extracted from the public LIDC-IDRI dataset by following a rigorous experimental setup using only nodules annotated by at least three radiologists. Our results show that the 3D Axial-Attention model achieves state-of-the-art performance on all evaluation metrics, including AUC and Accuracy. Conclusions: The proposed model provides full 3D attention, whereby every element (i.e., pixel) in the 3D volume space attends to every other element in the nodule effectively. Thus, the 3D Axial-Attention network can be used in all layers without the need for local filters. The experimental results show the importance of full 3D attention for classifying lung nodules. | ['Maxine Tan', 'Kelvin Shak', 'Mundher Al-Shabi'] | 2020-12-28 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [ 6.76867738e-02 4.50314343e-01 -2.08964035e-01 -1.31772861e-01
-7.15055346e-01 -2.06866845e-01 4.80912566e-01 -6.91126287e-02
-4.59310830e-01 1.57903448e-01 1.16941303e-01 -4.15832132e-01
-2.75849879e-01 -6.83226943e-01 -7.85375595e-01 -8.45539570e-01
1.47638425e-01 2.49783233e-01 5.79202294e-01 1.28703818e-01
-1.46165341e-01 7.15564132e-01 -1.20611036e+00 6.58646077e-02
7.48374701e-01 1.15401912e+00 4.27014440e-01 5.66562355e-01
-1.42033741e-01 5.24815917e-01 -2.10419133e-01 -1.01273142e-01
3.57355326e-01 -4.09626305e-01 -8.66873622e-01 1.90217152e-01
5.19336879e-01 -4.08994019e-01 -2.99469382e-01 1.11378467e+00
5.94915092e-01 -2.25215420e-01 6.82851374e-01 -6.23700023e-01
-9.57581222e-01 3.83572400e-01 -7.26343036e-01 4.29915458e-01
-2.49717370e-01 1.52101308e-01 9.53552067e-01 -9.67805386e-01
2.59629995e-01 8.00164878e-01 7.82136440e-01 4.88328278e-01
-7.46546566e-01 -3.96321476e-01 7.84073621e-02 3.45155522e-02
-1.42853689e+00 2.09027722e-01 5.95284641e-01 -4.38551456e-01
6.32336318e-01 4.32191640e-01 7.93152153e-01 7.98785150e-01
3.86857241e-01 5.08968294e-01 8.98040473e-01 -1.97613731e-01
-1.66715294e-01 3.48531336e-01 4.73120861e-04 1.06927538e+00
3.88135344e-01 -1.18409775e-01 2.51404017e-01 -8.59742463e-02
1.20585382e+00 3.21299821e-01 -3.59335721e-01 -4.95407820e-01
-1.35050809e+00 8.14373076e-01 1.21163058e+00 5.64398944e-01
-5.40729463e-01 2.36150637e-01 1.88976377e-01 -1.76692009e-01
4.32301074e-01 4.69809592e-01 -3.43355328e-01 3.74202996e-01
-6.60400867e-01 -2.96535999e-01 4.53902721e-01 6.80832088e-01
7.06583202e-01 -3.95382434e-01 -5.67123353e-01 6.06778026e-01
2.36777455e-01 3.10966432e-01 9.61701632e-01 -2.71917254e-01
2.13890925e-01 7.50946999e-01 -2.13818356e-01 -1.03051364e+00
-7.15783775e-01 -8.19377720e-01 -1.12550044e+00 -1.50726577e-02
2.05657914e-01 9.68044251e-02 -1.01098609e+00 1.41394031e+00
4.38559204e-01 3.60498279e-01 -2.76598543e-01 1.09802926e+00
1.02532947e+00 1.43881783e-01 -9.99529213e-02 3.42245102e-02
1.69710445e+00 -1.37946391e+00 -5.12546301e-01 -2.77719554e-02
7.54473329e-01 -6.51475191e-01 1.18605888e+00 -3.37348223e-01
-1.03222501e+00 -6.88649714e-01 -8.00089777e-01 -1.34520844e-01
-3.02005887e-01 4.70700473e-01 4.92915362e-01 5.46213567e-01
-1.19512415e+00 2.57000387e-01 -9.92212296e-01 -6.14939094e-01
3.83817703e-01 4.97872978e-01 -4.35059905e-01 -1.40937582e-01
-1.10642493e+00 8.20329189e-01 9.73643437e-02 3.77483636e-01
-7.24439085e-01 -8.80842149e-01 -6.14708722e-01 2.75326669e-01
4.49393421e-01 -8.83618176e-01 1.14784634e+00 -5.70805550e-01
-1.23153615e+00 7.79911220e-01 3.10405064e-02 -3.48076701e-01
4.00851816e-01 -2.95779365e-03 -8.13871901e-03 3.08173299e-01
9.50816125e-02 6.54170156e-01 6.82307959e-01 -7.09331989e-01
-3.29765528e-01 -3.33064765e-01 1.78973362e-01 3.50457817e-01
-6.44496262e-01 -3.76099557e-01 -8.52662861e-01 -5.69749117e-01
3.71989548e-01 -1.07676744e+00 -4.33549464e-01 4.09003764e-01
-3.99697065e-01 -3.04777384e-01 1.06017363e+00 -5.83026469e-01
1.22835064e+00 -2.16460609e+00 -9.41307396e-02 1.24524750e-01
4.90466207e-01 3.60911340e-01 5.85851111e-02 -3.29016775e-01
-2.07310677e-01 4.92902428e-01 -2.13473558e-01 -7.21450076e-02
-2.78282672e-01 5.03815189e-02 6.31323159e-01 6.90157473e-01
4.88384902e-01 1.34629273e+00 -7.34756947e-01 -8.30899298e-01
4.70010340e-01 7.13937640e-01 -7.52505422e-01 9.46640149e-02
1.72886997e-01 2.93549895e-01 -7.37547398e-01 5.22679329e-01
6.27049923e-01 -7.83076763e-01 -1.40604734e-01 -5.16528249e-01
-1.41096383e-01 1.30692646e-01 -6.72811747e-01 1.65019178e+00
-7.04355180e-01 3.41706038e-01 -2.36415099e-02 -7.34696269e-01
7.27043808e-01 4.89867717e-01 6.11372411e-01 -3.88851047e-01
2.73114294e-01 3.76251549e-01 2.71324873e-01 -5.98890126e-01
-1.61539301e-01 1.01424240e-01 7.77986497e-02 3.35465133e-01
-2.47588567e-02 -4.40690629e-02 -2.02195793e-01 -7.45628253e-02
1.37087452e+00 -2.76681840e-01 5.63827217e-01 -4.91171122e-01
8.54188979e-01 -2.86757022e-01 2.36071050e-01 8.63572836e-01
-3.35621178e-01 7.85770237e-01 5.27035356e-01 -3.57003361e-01
-8.16719770e-01 -6.97182655e-01 -3.32635432e-01 6.55128300e-01
6.51764870e-02 -4.66401316e-03 -6.73755944e-01 -1.22262621e+00
-5.55375218e-02 -3.48323025e-02 -1.15755033e+00 -1.07836209e-01
-6.23418331e-01 -6.28615022e-01 4.19075817e-01 7.50553608e-01
6.44591868e-01 -9.06241834e-01 -7.50176251e-01 -6.87952340e-02
-5.72362095e-02 -8.52603376e-01 -8.86851847e-01 4.11516279e-01
-9.01076376e-01 -1.09618914e+00 -1.06286383e+00 -8.13154519e-01
1.07380354e+00 4.80314851e-01 7.36055195e-01 1.37141541e-01
-5.94816685e-01 4.06509250e-01 -1.99435502e-01 -1.34955078e-01
1.09991588e-01 4.96302783e-01 -4.38753933e-01 1.19931474e-02
3.29125196e-01 -1.43951148e-01 -8.14459145e-01 4.71144795e-01
-7.90040374e-01 -4.74687815e-02 1.29736829e+00 1.03442979e+00
8.01451981e-01 -1.07130431e-01 4.43535030e-01 -9.70290899e-01
3.31605881e-01 -4.99904871e-01 -3.35583895e-01 7.78614134e-02
-9.20605808e-02 8.23515058e-02 6.29027724e-01 -3.29811811e-01
-7.78181493e-01 1.81104481e-01 -1.30237937e-01 -6.11809909e-01
-4.74399840e-03 4.53554481e-01 -2.14541648e-02 -3.49001974e-01
6.41013205e-01 -4.32379656e-02 1.67552039e-01 -3.78967851e-01
6.84511140e-02 6.95846438e-01 1.36382833e-01 4.61191162e-02
9.14534688e-01 5.14092863e-01 1.20832525e-01 -7.92848647e-01
-1.01413476e+00 -7.04749644e-01 -7.28611112e-01 -2.49575786e-02
1.16413617e+00 -7.18178451e-01 -5.37640870e-01 8.14781934e-02
-8.05089653e-01 -2.05559283e-02 -5.36309540e-01 8.09786916e-01
-2.92368859e-01 2.03666657e-01 -4.79090720e-01 -2.95971453e-01
-4.43268687e-01 -1.44129491e+00 1.17122066e+00 2.68435180e-01
-1.06483400e-02 -9.65691566e-01 -2.20694885e-01 1.68288052e-01
8.18873644e-01 7.63261914e-02 8.51111829e-01 -6.93986595e-01
-6.47081077e-01 -4.64697838e-01 -6.34515405e-01 2.54864246e-01
5.11515856e-01 -2.87287951e-01 -9.64187860e-01 -2.93339461e-01
1.67583406e-01 -2.01112419e-01 9.13346350e-01 5.89367986e-01
1.72280037e+00 -6.00667447e-02 -4.26458836e-01 7.62493193e-01
1.22381401e+00 -1.51733965e-01 3.21098655e-01 1.52098276e-02
8.54022384e-01 2.98802078e-01 3.72947633e-01 1.79163754e-01
8.50140452e-02 4.73151594e-01 8.75274777e-01 -5.55600405e-01
-4.62526351e-01 2.49005947e-02 -1.55504361e-01 8.12774837e-01
-1.52776048e-01 -9.45456699e-02 -7.99082875e-01 6.42766058e-01
-1.55186284e+00 -4.49427128e-01 -1.32678524e-01 2.16074634e+00
5.33146203e-01 1.61363427e-02 -3.37810248e-01 -9.85854045e-02
7.15211689e-01 2.45029077e-01 -6.32737875e-01 -1.78299531e-01
2.79482365e-01 2.33638853e-01 6.61263287e-01 2.98429400e-01
-1.43175638e+00 4.57532018e-01 5.27648783e+00 7.97347188e-01
-1.37080562e+00 3.38569969e-01 5.99597871e-01 -1.09653071e-01
-2.36420214e-01 -4.86658961e-01 -7.11793065e-01 2.29184344e-01
6.28500879e-01 5.00221737e-02 -8.87180418e-02 9.69921708e-01
1.41596021e-02 1.19779088e-01 -9.36922014e-01 7.61660933e-01
2.13757426e-01 -1.14659727e+00 -5.86495958e-02 3.24386537e-01
7.13108480e-01 1.74828932e-01 2.06112579e-01 2.97110558e-01
-3.23211551e-01 -1.17364717e+00 -2.52471585e-02 4.79117781e-01
1.05314684e+00 -6.30185544e-01 1.25817192e+00 3.45529854e-01
-1.42166102e+00 -9.61135626e-02 -5.06791532e-01 6.14347398e-01
-3.11117321e-01 6.96869791e-01 -1.05153871e+00 7.05570579e-01
8.30719888e-01 6.10519707e-01 -9.23533201e-01 1.09279537e+00
-3.21281999e-02 3.41734231e-01 -4.51052129e-01 -1.50754794e-01
6.21638179e-01 2.18964309e-01 2.78673679e-01 1.14789891e+00
5.83062530e-01 -6.70725852e-02 5.05181141e-02 7.39047706e-01
-1.71449736e-01 2.47842625e-01 -5.33147991e-01 4.84533384e-02
1.32668149e-02 1.62890160e+00 -7.14427829e-01 -1.11532457e-01
-7.28561759e-01 9.69845057e-01 1.69310600e-01 2.68954616e-02
-1.14163744e+00 -4.18542534e-01 3.48266751e-01 4.81649637e-01
6.23988152e-01 2.76404589e-01 1.06471758e-02 -8.19511890e-01
-3.05499882e-03 -5.60612261e-01 2.82748580e-01 -5.72611392e-01
-1.33641326e+00 8.13245296e-01 -3.65428388e-01 -1.25421357e+00
3.21221538e-02 -7.66333997e-01 -5.27909279e-01 9.63150859e-01
-1.64347410e+00 -1.21229982e+00 -8.82263005e-01 6.07091248e-01
2.74386555e-01 1.66299656e-01 7.51150966e-01 3.91101390e-01
-3.88911217e-01 7.50623703e-01 -2.16069177e-01 1.30875334e-01
7.79955447e-01 -1.18847930e+00 3.48677710e-02 4.84420538e-01
-2.53271908e-01 5.39663136e-01 -1.24741003e-01 -4.37491089e-01
-1.32709444e+00 -1.47824347e+00 8.30904901e-01 -3.02545995e-01
6.14193201e-01 -1.57671645e-01 -9.06004667e-01 6.51796460e-01
4.80492227e-02 8.95269215e-01 5.50567627e-01 -2.38976613e-01
5.54295629e-02 -7.89183006e-02 -1.21152902e+00 3.42363775e-01
9.31398630e-01 -4.77831155e-01 -3.12190503e-01 3.86458278e-01
9.68229771e-01 -4.01036382e-01 -8.83241892e-01 6.51200712e-01
3.66921037e-01 -8.69163513e-01 9.38807189e-01 -1.63319543e-01
4.15357858e-01 -4.79552448e-01 1.05679445e-01 -8.96238506e-01
-8.20665359e-01 -2.92201638e-02 -1.52892685e-02 5.72366655e-01
4.93281007e-01 -7.36077189e-01 8.00249100e-01 1.32808790e-01
-4.85606253e-01 -1.29102850e+00 -8.95194471e-01 -4.49240357e-01
2.13130172e-02 -1.67651493e-02 4.23733264e-01 7.73492694e-01
-5.03726125e-01 -2.09145136e-02 9.71385315e-02 3.76129121e-01
2.72082776e-01 -1.00477912e-01 2.89821744e-01 -1.16239619e+00
-3.89949024e-01 -5.77447891e-01 -5.36374569e-01 -1.24800670e+00
-1.93970770e-01 -1.13558638e+00 -5.89479469e-02 -1.69842935e+00
4.90337759e-01 -3.96135628e-01 -5.39302111e-01 4.85485941e-01
-2.91291744e-01 4.69452590e-01 -1.21348716e-01 2.18809962e-01
-3.94421905e-01 4.09969956e-01 1.87046862e+00 -1.22399785e-01
6.97168335e-03 9.20431316e-02 -8.33606422e-01 7.67989635e-01
7.11116314e-01 -3.26154619e-01 -3.70029271e-01 -2.83816576e-01
-1.37406856e-01 -1.76213607e-01 5.59174061e-01 -1.12156773e+00
3.55295539e-01 1.49029523e-01 6.22207940e-01 -8.59728336e-01
2.19357252e-01 -1.07058072e+00 -1.96584180e-01 6.35632694e-01
-8.94509628e-02 -3.29617649e-01 1.05349995e-01 6.25738442e-01
-2.58769721e-01 -3.08889776e-01 9.58244205e-01 -3.25376898e-01
-2.46377811e-01 6.02033019e-01 -6.17202744e-02 -3.27894181e-01
1.36047637e+00 -1.63950741e-01 -1.64744154e-01 6.27302155e-02
-7.34732687e-01 3.51473503e-02 2.32361570e-01 1.32397026e-01
4.12883461e-01 -1.45096636e+00 -5.79446793e-01 4.72599298e-01
6.86224103e-02 4.09430653e-01 4.87654150e-01 1.47627068e+00
-5.71052313e-01 8.98454189e-01 5.96696399e-02 -9.13060546e-01
-1.11309040e+00 6.98831022e-01 8.04852962e-01 -6.31212294e-01
-7.51357436e-01 1.02951038e+00 7.49588668e-01 -6.48070335e-01
1.44420445e-01 -7.96192348e-01 -2.79770344e-01 -1.68426819e-02
3.14586401e-01 -4.26107906e-02 3.57977390e-01 -3.68352681e-01
-6.15502715e-01 8.34116161e-01 -2.24920467e-01 4.20853764e-01
9.71336246e-01 2.41895244e-02 -6.05720468e-02 2.22086400e-01
1.38908124e+00 -1.23781256e-01 -1.02739775e+00 -3.05406451e-01
-3.14768374e-01 -3.38715702e-01 3.93276632e-01 -4.55542386e-01
-1.37707269e+00 9.96780455e-01 7.57663667e-01 2.37069219e-01
1.18256164e+00 2.80547649e-01 6.27607167e-01 2.83665210e-01
-1.36811594e-02 -4.87368912e-01 2.69238859e-01 4.63460952e-01
8.89649808e-01 -1.35570490e+00 -6.08137324e-02 -4.03841496e-01
-3.81942213e-01 9.50427413e-01 9.42917585e-01 -2.32949585e-01
9.83146727e-01 1.72112793e-01 -1.18354782e-01 -1.85339212e-01
-5.80932856e-01 -3.33295971e-01 5.06620467e-01 2.07467064e-01
6.86664045e-01 -5.66805564e-02 -2.36948460e-01 6.51086509e-01
1.99589372e-01 -1.71626061e-01 3.58252257e-01 8.14118385e-01
-3.37226093e-01 -5.95293224e-01 -3.24322015e-01 7.40548074e-01
-4.58303630e-01 -3.45160700e-02 -1.65585384e-01 1.07723403e+00
3.13055366e-01 2.96363562e-01 2.37327650e-01 -1.74735710e-01
3.88148129e-01 -8.49297941e-02 2.61268169e-01 -7.47994959e-01
-7.44714618e-01 3.30719143e-01 -4.77906048e-01 -5.81728399e-01
-2.93482959e-01 -4.52283114e-01 -1.18728483e+00 6.34757727e-02
-6.28947318e-01 -5.47902435e-02 5.67141652e-01 4.97917771e-01
3.85201126e-01 1.15809035e+00 6.70254886e-01 -7.47125804e-01
-6.84735954e-01 -1.12598062e+00 -4.54710037e-01 1.40177729e-02
3.87889862e-01 -6.05287254e-01 -7.00062037e-01 -3.76796901e-01] | [15.331622123718262, -2.182124614715576] |
b8ef4ca8-dacd-4016-8ba8-2d1c312f1519 | a-benchmark-for-systematic-generalization-in | 2003.05161 | null | https://arxiv.org/abs/2003.05161v2 | https://arxiv.org/pdf/2003.05161v2.pdf | A Benchmark for Systematic Generalization in Grounded Language Understanding | Humans easily interpret expressions that describe unfamiliar situations composed from familiar parts ("greet the pink brontosaurus by the ferris wheel"). Modern neural networks, by contrast, struggle to interpret novel compositions. In this paper, we introduce a new benchmark, gSCAN, for evaluating compositional generalization in situated language understanding. Going beyond a related benchmark that focused on syntactic aspects of generalization, gSCAN defines a language grounded in the states of a grid world, facilitating novel evaluations of acquiring linguistically motivated rules. For example, agents must understand how adjectives such as 'small' are interpreted relative to the current world state or how adverbs such as 'cautiously' combine with new verbs. We test a strong multi-modal baseline model and a state-of-the-art compositional method finding that, in most cases, they fail dramatically when generalization requires systematic compositional rules. | ['Laura Ruis', 'Diane Bouchacourt', 'Marco Baroni', 'Jacob Andreas', 'Brenden M. Lake'] | 2020-03-11 | null | http://proceedings.neurips.cc/paper/2020/hash/e5a90182cc81e12ab5e72d66e0b46fe3-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/e5a90182cc81e12ab5e72d66e0b46fe3-Paper.pdf | neurips-2020-12 | ['systematic-generalization'] | ['reasoning'] | [ 3.74875426e-01 2.28619024e-01 6.80018142e-02 -6.30028903e-01
-4.78265136e-01 -9.58196461e-01 1.20325172e+00 2.05435559e-01
-5.25377035e-01 7.13063061e-01 8.77597332e-01 -3.77841383e-01
-2.94540934e-02 -8.20523977e-01 -8.00726354e-01 -5.06743610e-01
8.60536471e-02 7.25567698e-01 4.63495627e-02 -8.40523124e-01
1.37892693e-01 4.40511853e-01 -1.44837523e+00 9.36745822e-01
5.43833494e-01 6.36489213e-01 1.62982997e-02 2.46563405e-01
-2.16954514e-01 1.20678902e+00 -5.42087913e-01 -5.00680447e-01
1.40835643e-01 -8.13055813e-01 -1.16620910e+00 -4.10903126e-01
6.98218942e-01 -1.78144917e-01 -6.77097365e-02 1.14146376e+00
1.60903424e-01 5.80262661e-01 7.60664284e-01 -1.00033045e+00
-1.16802609e+00 1.60311306e+00 3.81123871e-02 2.19647363e-01
8.10263157e-01 4.88538653e-01 1.44222867e+00 -5.41195989e-01
8.84687781e-01 1.74910557e+00 8.17579746e-01 1.09284997e+00
-1.57774723e+00 -4.23989654e-01 6.95543170e-01 3.29530150e-01
-1.09215474e+00 -4.62917656e-01 6.15377605e-01 -1.22244470e-01
1.40405941e+00 4.65689331e-01 6.04020774e-01 1.58625782e+00
-8.22230149e-03 9.08595860e-01 1.43157089e+00 -3.76981199e-01
4.67961997e-01 -2.50164598e-01 1.42244726e-01 5.10717452e-01
1.86254576e-01 4.36804324e-01 -7.44410157e-01 -1.04670286e-01
3.08753550e-01 -3.07491928e-01 -1.21898346e-01 -1.35724261e-01
-1.59510875e+00 6.63939118e-01 7.86982179e-01 5.64711988e-01
-2.37353742e-01 6.30434692e-01 5.37383199e-01 2.40516365e-01
3.97503413e-02 9.80441034e-01 -4.71265227e-01 1.10542681e-02
-3.46743375e-01 6.55939281e-01 7.86887765e-01 8.91567290e-01
5.93164086e-01 1.56084642e-01 2.22519085e-01 5.01910985e-01
2.62944579e-01 3.84401470e-01 6.36495233e-01 -1.24454916e+00
2.38847122e-01 5.67712128e-01 -8.18400234e-02 -5.05980611e-01
-7.90947258e-01 -1.89560324e-01 -5.68367839e-01 9.65621769e-02
5.50593793e-01 -3.18432935e-02 -7.00976312e-01 2.24022150e+00
8.39148387e-02 -1.33811146e-01 6.12588108e-01 9.48512316e-01
8.36411178e-01 6.75934196e-01 5.64713180e-01 -7.94403031e-02
1.45907843e+00 -7.59332478e-01 -4.57519948e-01 -7.08106279e-01
8.32455099e-01 -1.06221259e-01 1.74358976e+00 1.29551098e-01
-1.15809345e+00 -2.96067655e-01 -9.60504651e-01 -3.94619852e-01
-7.09878445e-01 -5.97097218e-01 9.46150005e-01 2.31686249e-01
-9.89826739e-01 6.06678188e-01 -8.04324329e-01 -7.85953104e-01
3.79745811e-01 2.95034554e-02 -4.55167294e-01 2.03396887e-01
-1.08037603e+00 1.47405362e+00 6.97009146e-01 3.68833616e-02
-1.21104622e+00 -8.35996687e-01 -1.05460787e+00 1.43720666e-02
3.88191044e-01 -1.02179205e+00 1.74748087e+00 -1.42521691e+00
-1.60893977e+00 1.17687345e+00 -2.31552631e-01 -6.55776918e-01
3.33443910e-01 -1.87401757e-01 -4.56419617e-01 -1.32456660e-01
3.00383240e-01 7.93491840e-01 4.00351167e-01 -1.51345909e+00
-7.62188613e-01 -5.46705842e-01 6.67246759e-01 4.07605946e-01
2.74345696e-01 1.25533924e-01 6.67406917e-01 -6.18477404e-01
3.24368238e-01 -8.94035220e-01 -2.26499632e-01 -2.22512573e-01
-2.84211129e-01 -4.48254108e-01 3.83620203e-01 -2.10459277e-01
8.16973448e-01 -2.16255426e+00 5.63385427e-01 7.82094430e-03
6.36643246e-02 -4.49428439e-01 -1.46428391e-01 4.62058544e-01
-3.57178152e-01 3.21994394e-01 -3.45167071e-01 -4.72420966e-03
7.79458523e-01 5.12690306e-01 -6.99628353e-01 2.08225355e-01
7.23996758e-02 1.19373918e+00 -1.19345796e+00 -4.17617187e-02
-7.79525787e-02 -4.26293686e-02 -6.44609392e-01 -2.72451550e-01
-6.51112199e-01 2.73798704e-01 3.86610106e-02 4.57183838e-01
7.20764697e-02 6.03644550e-03 4.35009420e-01 -3.26910913e-01
9.00460128e-03 8.62016737e-01 -8.87807429e-01 1.93861926e+00
-5.05660176e-01 4.21362251e-01 -3.23564857e-01 -6.68458402e-01
3.47586900e-01 1.27045169e-01 -4.08367813e-01 -7.72857308e-01
1.20028861e-01 3.59371215e-01 4.52018619e-01 -4.25331742e-01
2.50397593e-01 -9.33557093e-01 -4.94954884e-01 5.83780289e-01
-6.92743286e-02 -7.00168133e-01 3.81062418e-01 2.51243562e-01
1.01298296e+00 6.17209852e-01 6.95807576e-01 -7.29588151e-01
5.00217617e-01 1.71998858e-01 4.86004949e-01 6.81979358e-01
-1.12560973e-01 7.97409639e-02 4.76654321e-01 -1.01088023e+00
-9.37522352e-01 -1.62068975e+00 2.08290488e-01 1.59864712e+00
5.47982976e-02 -4.58143532e-01 -6.87339604e-01 -6.35217488e-01
-2.35504806e-01 1.69466329e+00 -7.86610961e-01 -4.24928665e-01
-8.83864999e-01 -4.57327992e-01 9.41016197e-01 8.20197105e-01
4.84058440e-01 -1.39336205e+00 -1.07939231e+00 1.14471518e-01
-4.88041304e-02 -9.16913092e-01 -1.37365580e-01 4.91996646e-01
-7.20994234e-01 -6.78443074e-01 7.13453591e-02 -8.01490843e-01
4.90427017e-01 -3.59587610e-01 1.43367183e+00 -7.85977915e-02
1.85231790e-01 3.62095594e-01 -3.69475603e-01 -6.17708504e-01
-6.96299136e-01 9.21204872e-03 3.11192572e-01 -3.71621937e-01
5.58316290e-01 -7.40045667e-01 -2.73165833e-02 -5.77652939e-02
-1.09924090e+00 1.05432220e-01 2.38794401e-01 6.82054639e-01
1.76501244e-01 -3.38127315e-01 4.96899188e-01 -8.89573097e-01
8.75854254e-01 -3.08007658e-01 -2.56160021e-01 2.78995335e-01
-6.92216679e-02 5.15609205e-01 8.81656766e-01 -5.46742380e-01
-1.30153668e+00 -2.65041739e-01 1.94285944e-01 2.51675963e-01
-3.70225757e-01 4.94054556e-01 -2.22543448e-01 4.15042639e-01
1.46496463e+00 1.04513049e-01 -1.93089068e-01 -1.41573042e-01
9.42864001e-01 -4.27911952e-02 7.00606167e-01 -1.39900136e+00
7.97917783e-01 6.97513759e-01 1.45921975e-01 -5.76053917e-01
-1.08068812e+00 1.68657899e-01 -6.37153864e-01 2.11191237e-01
9.11734045e-01 -7.12522447e-01 -9.27209795e-01 2.62962192e-01
-1.25652099e+00 -7.82778502e-01 -8.76323342e-01 3.14310461e-01
-9.12270427e-01 1.40495107e-01 -7.48957455e-01 -5.04303217e-01
7.37991408e-02 -8.87131512e-01 9.38948810e-01 1.14657238e-01
-1.09545946e+00 -1.14354742e+00 -1.29750057e-03 -1.16055869e-02
4.42943335e-01 3.14482868e-01 1.59325063e+00 -9.28009868e-01
-3.05218399e-01 5.30062675e-01 2.02210457e-03 2.76657075e-01
2.03486785e-01 -1.81448102e-01 -7.75682390e-01 9.75675508e-02
-5.25662713e-02 -5.29431760e-01 5.29886842e-01 -1.95772856e-01
7.07608759e-01 -6.45461559e-01 -1.61408260e-01 5.20137489e-01
1.23312509e+00 3.82824183e-01 5.56057155e-01 4.94936079e-01
4.17723507e-01 5.77515602e-01 7.15130791e-02 3.17238607e-02
7.34497249e-01 3.70084971e-01 2.13945508e-01 4.59563285e-01
-2.38078237e-01 -4.70033348e-01 7.04263866e-01 3.25518996e-01
-3.66909146e-01 3.08719184e-02 -1.07568240e+00 5.29999256e-01
-1.78965509e+00 -1.23823953e+00 1.31233618e-01 1.66250658e+00
1.26613820e+00 4.61502708e-02 -1.32234037e-01 -2.64488637e-01
1.75691530e-01 3.22529525e-01 -5.05563855e-01 -6.99091792e-01
-4.63869393e-01 5.37074625e-01 -1.51055932e-01 8.70636642e-01
-7.35208154e-01 1.58360147e+00 6.67781210e+00 3.96711379e-01
-8.37163150e-01 -1.51911927e-05 1.52044356e-01 -8.03776309e-02
-7.23905444e-01 2.64948994e-01 -5.16451240e-01 1.25256544e-02
8.31794798e-01 -1.47784352e-01 9.52382684e-01 5.30314505e-01
-1.08255088e-01 -2.14549810e-01 -2.05518603e+00 8.06966841e-01
3.76663476e-01 -1.27073002e+00 6.49278462e-01 -5.16918242e-01
6.38400555e-01 -1.86238721e-01 1.49452984e-01 4.20397103e-01
7.68732488e-01 -1.13349330e+00 1.31320429e+00 5.21076322e-01
2.39284366e-01 -3.24929297e-01 2.74401844e-01 2.11726993e-01
-8.93649638e-01 -2.81814456e-01 -9.44007486e-02 -5.92257500e-01
3.23891848e-01 -2.75934964e-01 -5.72896481e-01 1.25934109e-01
6.30281925e-01 3.90840411e-01 -7.04267085e-01 1.99233502e-01
-7.70450950e-01 2.23501876e-01 -4.36211348e-01 -2.64728367e-01
5.62780023e-01 -2.11133689e-01 6.26526833e-01 1.10317171e+00
1.00507393e-01 4.57015008e-01 -1.77021191e-01 1.29435968e+00
1.42332884e-02 -3.51207927e-02 -8.23859215e-01 -1.35958180e-01
2.11953446e-01 7.94683874e-01 -4.82619554e-01 -6.25710189e-01
-1.88850597e-01 7.77569830e-01 3.91702324e-01 5.46453238e-01
-8.45211625e-01 1.31743655e-01 7.81207561e-01 1.73412133e-02
-9.96489078e-02 -3.02432418e-01 -3.21410716e-01 -1.15949714e+00
-1.41041473e-01 -1.36855638e+00 3.83915871e-01 -1.25683939e+00
-1.50551391e+00 5.14949739e-01 5.54137468e-01 -6.24519408e-01
-4.44559574e-01 -8.88813198e-01 -4.97964352e-01 4.08622175e-01
-8.96435618e-01 -1.49948943e+00 -1.23579256e-01 5.37396550e-01
6.15238130e-01 -9.11540464e-02 1.11827540e+00 -3.21865112e-01
-1.11851655e-01 2.44347721e-01 -5.38214684e-01 -8.37104693e-02
4.83411878e-01 -1.31806052e+00 5.66091597e-01 8.26864243e-01
4.17887092e-01 1.16403282e+00 9.43713307e-01 -4.46365386e-01
-1.06899083e+00 -6.42576933e-01 8.89398694e-01 -9.45402801e-01
1.00569737e+00 -2.76675105e-01 -9.10553217e-01 1.52369523e+00
5.16030669e-01 -1.77675754e-01 6.15448356e-01 3.75090182e-01
-1.11585391e+00 -1.07751690e-01 -9.65923786e-01 1.40724766e+00
1.79690385e+00 -8.30135167e-01 -1.55785930e+00 3.89843136e-01
8.66518140e-01 -3.02701652e-01 -5.19283295e-01 4.16823834e-01
5.48988819e-01 -9.96361017e-01 7.67324626e-01 -1.58047736e+00
4.24908102e-01 -6.00210786e-01 -6.96167946e-01 -1.76549709e+00
-4.50372010e-01 -5.19328237e-01 2.68543869e-01 7.67313182e-01
6.94343090e-01 -8.35571349e-01 3.13798308e-01 7.22630143e-01
-3.67453545e-01 -1.70393527e-01 -1.00830650e+00 -7.50819862e-01
3.75635684e-01 -4.53498542e-01 8.18881929e-01 9.49088991e-01
5.47234774e-01 8.77151072e-01 5.21234274e-01 -2.33660406e-03
4.44813520e-01 1.69044104e-03 4.79536861e-01 -9.54272568e-01
-2.63312370e-01 -8.78321111e-01 -3.68322819e-01 -7.31193185e-01
8.49473059e-01 -1.30003822e+00 4.01200727e-02 -1.44461703e+00
1.19142607e-01 -6.98999465e-02 1.20962914e-02 7.53797591e-01
8.61901343e-02 -5.96224889e-02 4.92260218e-01 -2.49878570e-01
-5.98378599e-01 4.49441910e-01 9.97820437e-01 -4.70581084e-01
1.75971184e-02 -6.90097034e-01 -1.04603922e+00 1.26088488e+00
8.29595447e-01 -1.76344290e-01 -5.75896680e-01 -7.54320383e-01
8.83766413e-01 -5.86107373e-01 7.13760316e-01 -8.33176911e-01
1.55867308e-01 -5.54172456e-01 4.90170233e-02 7.70657361e-02
2.76263773e-01 -8.77191305e-01 3.56580585e-01 4.44530368e-01
-7.34790325e-01 4.87055868e-01 4.34502035e-01 2.05851406e-01
-1.89023800e-02 -1.56618848e-01 7.70705163e-01 -5.07310390e-01
-1.07652009e+00 -4.19343561e-01 -5.06621361e-01 3.91569436e-01
8.88701737e-01 -1.33612333e-02 -8.27730834e-01 -3.95818323e-01
-9.43655849e-01 -8.33066478e-02 5.18528759e-01 4.69422877e-01
4.67636198e-01 -1.45349360e+00 -6.27027690e-01 -1.74804926e-01
3.42618287e-01 7.29051530e-02 9.30195376e-02 6.54681504e-01
-8.21714580e-01 5.93259989e-04 -3.46316785e-01 -2.86956459e-01
-7.36681104e-01 8.28788042e-01 7.65474141e-01 2.35547200e-01
-4.55623180e-01 1.07488477e+00 7.50723243e-01 -7.99495935e-01
-2.06238523e-01 -1.22961390e+00 2.69300103e-01 1.43706603e-02
4.60643530e-01 6.52112886e-02 -3.03373516e-01 -7.77417243e-01
-4.99760985e-01 4.06752944e-01 9.29425806e-02 -4.22058612e-01
1.08699107e+00 -3.26648168e-02 -7.50178814e-01 1.13748729e+00
8.68399858e-01 -2.48301425e-03 -7.42157340e-01 -2.97986150e-01
2.23759204e-01 3.52033786e-02 -7.24593699e-01 -1.18168759e+00
-2.42508486e-01 5.20343363e-01 9.56263319e-02 1.21724792e-01
7.58238316e-01 4.33078408e-01 3.97849172e-01 1.01669621e+00
6.02770090e-01 -1.06468105e+00 3.91675979e-02 9.21054423e-01
1.35062194e+00 -9.49291050e-01 -7.27609513e-05 5.14653176e-02
-8.40995431e-01 8.86450589e-01 7.10283577e-01 -1.19621791e-01
7.71384686e-02 1.94734842e-01 2.12145925e-01 -2.52988219e-01
-9.87164557e-01 -1.83438897e-01 6.12608567e-02 7.15779901e-01
2.70546675e-01 4.67338890e-01 -1.15596421e-01 5.66664040e-01
-8.54361773e-01 -5.56683779e-01 4.64277387e-01 7.07568347e-01
-2.53774256e-01 -7.63332784e-01 -3.03445071e-01 5.08164465e-02
7.12408044e-04 -4.31354970e-01 -7.50867128e-01 1.05909860e+00
3.35961312e-01 6.62745237e-01 1.34909049e-01 2.74542421e-02
5.86695552e-01 5.23338795e-01 9.35387015e-01 -7.50053883e-01
-7.96805084e-01 -5.26309133e-01 4.89374340e-01 -7.49168336e-01
-7.15832949e-01 -9.35733497e-01 -1.82142746e+00 -2.89311022e-01
3.27487916e-01 -8.94528478e-02 2.03188702e-01 1.22260106e+00
-3.70284840e-02 3.93262714e-01 -2.32234254e-01 -7.58671403e-01
-7.88644791e-01 -1.02793801e+00 -3.01869273e-01 9.07555938e-01
2.95124203e-01 -4.25525516e-01 -4.30290103e-01 1.77147508e-01] | [9.653207778930664, 7.179845333099365] |
bfa90be2-83a0-4cfc-8ad9-b84ae59f4878 | a-copula-based-boosting-model-for-time-to | 2210.04869 | null | https://arxiv.org/abs/2210.04869v2 | https://arxiv.org/pdf/2210.04869v2.pdf | A copula-based boosting model for time-to-event prediction with dependent censoring | A characteristic feature of time-to-event data analysis is possible censoring of the event time. Most of the statistical learning methods for handling censored data are limited by the assumption of independent censoring, even if this can lead to biased predictions when the assumption does not hold. This paper introduces Clayton-boost, a boosting approach built upon the accelerated failure time model, which uses a Clayton copula to handle the dependency between the event and censoring distributions. By taking advantage of a copula, the independent censoring assumption is not needed any more. During comparisons with commonly used methods, Clayton-boost shows a strong ability to remove prediction bias at the presence of dependent censoring and outperforms the comparing methods either if the dependency strength or percentage censoring are considerable. The encouraging performance of Clayton-boost shows that there is indeed reasons to be critical about the independent censoring assumption, and that real-world data could highly benefit from modelling the potential dependency. | ['Arne Bang Huseby', 'Riccardo De Bin', 'Alise Danielle Midtfjord'] | 2022-10-10 | null | null | null | null | ['time-to-event-prediction'] | ['time-series'] | [-1.17743343e-01 -2.28049502e-01 -4.38388675e-01 -7.28377044e-01
-8.24632347e-01 -4.08420533e-01 6.17038310e-01 5.53942204e-01
-2.37444878e-01 1.14642167e+00 2.39225432e-01 -7.93890834e-01
-4.41209555e-01 -7.76584506e-01 -6.33630335e-01 -6.86993957e-01
-3.63187104e-01 6.40302658e-01 1.50570244e-01 3.09854536e-03
1.26028910e-01 4.01957095e-01 -1.40802991e+00 1.75142512e-01
6.87860966e-01 3.64632040e-01 -3.59593928e-01 6.72072530e-01
-1.24732234e-01 7.16219902e-01 -4.98469889e-01 -2.78225601e-01
-2.36270785e-01 -4.96256083e-01 -4.68853027e-01 -4.55520540e-01
-7.80516788e-02 -2.20347330e-01 3.14769655e-01 2.93950766e-01
5.16944408e-01 -2.30208501e-01 7.83798993e-01 -1.50345838e+00
-1.35746449e-02 3.45289052e-01 -4.26675171e-01 2.92930514e-01
4.02554035e-01 -4.19152647e-01 6.34753466e-01 -8.98277760e-01
2.61544615e-01 8.76743436e-01 1.21003354e+00 2.17148021e-01
-1.21019590e+00 -5.41242659e-01 -2.43941564e-02 1.45268783e-01
-1.11070430e+00 -1.53107002e-01 5.99717617e-01 -3.15688580e-01
7.21987665e-01 3.99653316e-01 5.53094566e-01 1.17433536e+00
9.66435313e-01 3.39276969e-01 1.36233211e+00 -7.36320674e-01
5.97270429e-01 3.13911354e-04 5.26649833e-01 2.66061742e-02
5.04567087e-01 7.52124846e-01 -4.37285453e-01 -8.95829737e-01
3.37607056e-01 5.70341945e-01 1.03425793e-01 -4.43162084e-01
-9.30009961e-01 1.06818509e+00 1.33733243e-01 1.06159441e-01
-3.63865972e-01 -1.87978178e-01 6.50554419e-01 3.78586888e-01
5.21652877e-01 -3.52682889e-01 -5.70686400e-01 -1.77490637e-01
-1.38623166e+00 1.82221889e-01 1.02701020e+00 8.10945153e-01
1.91364288e-01 -1.84763327e-01 -1.74789369e-01 4.16440904e-01
2.05525205e-01 1.18043825e-01 2.55838275e-01 -4.93447244e-01
-3.71220037e-02 3.91675830e-01 3.94232959e-01 3.16928327e-02
-6.35295093e-01 -6.13772213e-01 -7.01774240e-01 7.34389246e-01
7.72707999e-01 -3.33712578e-01 -9.28046465e-01 1.50068760e+00
2.66805649e-01 -7.18924925e-02 -9.77623612e-02 4.51633215e-01
3.17891628e-01 2.33627602e-01 4.23552841e-01 -4.42644924e-01
1.24777865e+00 -3.46710742e-01 -8.15308571e-01 -2.10654199e-01
5.49331248e-01 -6.32174432e-01 7.16626883e-01 5.81213772e-01
-9.30606306e-01 -1.61724016e-01 -1.04367113e+00 2.23031759e-01
-2.37445310e-01 -3.93730283e-01 8.28218877e-01 9.28269386e-01
-5.48463941e-01 5.92070639e-01 -1.00447905e+00 -5.05254209e-01
4.16505598e-02 2.14401975e-01 -1.92669868e-01 -3.96314234e-01
-1.11624980e+00 1.03686500e+00 3.43728587e-02 -1.97629914e-01
-8.35348189e-01 -5.58806241e-01 -4.39147204e-01 2.61279643e-01
-2.27545321e-01 -8.68041754e-01 9.42837477e-01 -1.03306341e+00
-6.25587285e-01 5.29605210e-01 -2.74255395e-01 -4.59696800e-01
9.95300293e-01 -8.45846608e-02 -4.91740435e-01 -4.06189352e-01
-1.77617092e-02 -4.82033379e-02 8.11628342e-01 -9.53576982e-01
-7.07746327e-01 -4.77663755e-01 -4.29769218e-01 -1.95513904e-01
1.49208277e-01 3.03962409e-01 2.27755278e-01 -5.79120457e-01
-7.73079172e-02 -8.07982087e-01 -1.39815137e-01 -1.92023918e-01
-3.42752726e-04 -2.88369417e-01 4.79809105e-01 -6.05134368e-01
1.13652790e+00 -2.07416081e+00 -6.27391875e-01 5.43087572e-02
-1.26869813e-01 -3.65195811e-01 4.04881656e-01 1.10057974e+00
-5.68503737e-01 -1.89119250e-01 -1.50585189e-01 -3.13388675e-01
-2.85188317e-01 3.84440511e-01 -3.15836310e-01 7.72207975e-01
-7.11807609e-02 5.19774318e-01 -5.46437860e-01 -5.11187553e-01
2.11375356e-01 6.47408783e-01 -1.17286876e-01 1.01236440e-01
3.95744801e-01 3.12636435e-01 -8.89123604e-02 5.80514252e-01
7.31643438e-01 -2.16723382e-01 1.06126115e-01 7.76567817e-01
-2.40186572e-01 1.36061653e-01 -7.78412163e-01 9.35673654e-01
-2.60163605e-01 4.10678566e-01 -7.56108686e-02 -7.81877995e-01
1.11257792e+00 6.04992449e-01 1.74230695e-01 -1.98639885e-01
9.79676619e-02 3.92194152e-01 -1.40249565e-01 -4.96435501e-02
-5.78154661e-02 -9.14043427e-01 -7.40965903e-02 1.82052612e-01
-2.66543806e-01 1.99219212e-01 -3.10696095e-01 -1.16149612e-01
1.25403595e+00 1.60168454e-01 5.28705060e-01 -3.64094704e-01
2.17516229e-01 1.34540990e-01 7.77433455e-01 1.06594110e+00
-1.91785529e-01 9.69292641e-01 5.41704595e-01 -6.66716456e-01
-1.02082384e+00 -1.17951632e+00 -7.72175074e-01 1.16899121e+00
-5.43769419e-01 1.02959290e-01 -2.28069782e-01 -8.08289051e-01
2.17289671e-01 1.22012377e+00 -9.59321201e-01 -1.65408045e-01
-1.53003365e-01 -1.36540663e+00 2.41875604e-01 7.59520173e-01
-2.53890902e-01 -8.41731012e-01 -4.56709415e-01 5.12076735e-01
7.28238225e-02 -2.23538622e-01 -7.13369576e-03 9.12102103e-01
-1.37852001e+00 -9.99412954e-01 -1.01823044e+00 -7.38194048e-01
5.10780811e-01 -4.26913127e-02 1.23344314e+00 1.86042920e-01
-7.34250201e-03 1.98300511e-01 -5.15576124e-01 -8.82317543e-01
-5.09545326e-01 -1.73765004e-01 -3.44748110e-01 -3.82809520e-01
7.96510100e-01 -6.61540210e-01 -7.47614443e-01 4.41006809e-01
-8.69735122e-01 -2.80939519e-01 4.29911166e-01 1.08003604e+00
1.31441653e-01 3.07674217e-03 1.35859430e+00 -9.22115684e-01
2.51004130e-01 -8.57340634e-01 -3.55515689e-01 1.86000034e-01
-1.07325530e+00 -3.09108108e-01 4.83068466e-01 -4.62627977e-01
-8.42831850e-01 4.59972508e-02 -4.95542809e-02 2.70278309e-04
-3.16445202e-01 5.75883567e-01 -9.08195153e-02 2.17220783e-01
6.09876633e-01 8.11673775e-02 -9.10766248e-04 -5.61338305e-01
-3.45936835e-01 6.70893610e-01 1.66601419e-01 -1.81431249e-01
1.95599422e-01 7.54505992e-01 1.81956530e-01 -3.08915049e-01
-5.70892632e-01 -4.99430299e-01 -3.91620576e-01 -1.72673762e-02
4.24150020e-01 -8.54102910e-01 -5.71349144e-01 3.14644873e-01
-8.39240193e-01 1.50355007e-02 -4.39223945e-01 8.05308759e-01
-7.24205732e-01 3.16612840e-01 -3.17847461e-01 -1.16835821e+00
-2.57783443e-01 -3.72617155e-01 4.96134967e-01 -7.08473288e-03
-3.38440865e-01 -1.43107367e+00 3.26203853e-01 2.36891195e-01
4.62969869e-01 7.22237468e-01 1.23138964e+00 -1.29634428e+00
2.26083204e-01 -1.18735814e+00 1.40068486e-01 2.14976162e-01
7.97063932e-02 -2.07450822e-01 -7.38178849e-01 -7.20670283e-01
2.15579063e-01 1.48540616e-01 5.73146820e-01 7.37127185e-01
5.50702453e-01 1.24483541e-01 -3.55217069e-01 9.56411436e-02
1.41374695e+00 2.13161334e-01 5.02909184e-01 5.44808269e-01
-5.06188795e-02 6.36819422e-01 5.38419485e-01 7.49716997e-01
4.68944192e-01 3.14831972e-01 6.84901655e-01 -4.41670269e-01
8.39910656e-02 -3.24430913e-01 2.56055266e-01 2.08196610e-01
1.39880821e-01 -1.03657134e-01 -1.17082107e+00 8.95375192e-01
-1.74594212e+00 -9.02278364e-01 -8.79278600e-01 2.95661235e+00
6.25322700e-01 5.99151433e-01 3.31630468e-01 4.54873264e-01
9.41847742e-01 -3.72623652e-01 -4.97783065e-01 -7.41705120e-01
-1.47528753e-01 -3.38613503e-02 5.62051058e-01 4.81652349e-01
-7.20441699e-01 -9.33625475e-02 7.43029785e+00 4.63229448e-01
-6.84325576e-01 1.46486744e-01 4.76124495e-01 -4.48742881e-02
2.31175069e-02 6.95488334e-01 -7.53307700e-01 7.04584718e-01
1.16046822e+00 -1.37917683e-01 -1.72809422e-01 7.47731328e-01
3.80337566e-01 -4.18979943e-01 -1.17385364e+00 1.98199570e-01
-1.42020062e-01 -5.80536962e-01 -3.82335216e-01 1.05277993e-01
3.98529321e-01 -3.48378532e-02 -4.73249227e-01 4.94883418e-01
4.73320365e-01 -1.07077193e+00 4.94238853e-01 8.37510407e-01
5.03366470e-01 -9.95035291e-01 1.26377106e+00 7.94227719e-01
-4.60334748e-01 -1.15352072e-01 -2.52378196e-01 -4.08970088e-01
4.09312755e-01 1.14504635e+00 -8.70094597e-01 5.89194655e-01
7.44720280e-01 1.56636536e-01 -5.72559774e-01 1.66482806e+00
-1.44737944e-01 1.26047671e+00 -4.33550686e-01 1.57566607e-01
-1.28908977e-01 2.57047653e-01 4.80227470e-01 1.12805605e+00
4.16975737e-01 -3.61955404e-01 -2.37721398e-01 1.78119674e-01
6.99209511e-01 -4.35291044e-02 -3.44662577e-01 7.08324015e-01
3.13462228e-01 7.94895411e-01 -7.28496432e-01 -2.81727463e-01
-8.86811078e-01 6.94919467e-01 -1.37330666e-01 2.41968572e-01
-7.00539589e-01 -3.74058396e-01 -2.87639797e-01 5.93568444e-01
4.26989824e-01 -1.33055136e-01 -8.93588781e-01 -8.21627140e-01
-6.81950301e-02 -5.28179049e-01 9.73823488e-01 -7.58186281e-01
-1.66483426e+00 2.21774384e-01 1.64413512e-01 -1.06069934e+00
-2.60094315e-01 -2.35598356e-01 -1.20100534e+00 1.10310781e+00
-1.57911134e+00 -1.35443890e+00 -7.13484883e-02 7.05904424e-01
3.23520869e-01 2.72849888e-01 9.55343485e-01 2.57195793e-02
-5.77118658e-02 5.33168256e-01 5.41707933e-01 -2.56215304e-01
1.14630413e+00 -1.38345873e+00 1.58616975e-01 3.76348376e-01
-4.84950960e-01 5.65967202e-01 1.21026969e+00 -1.02436197e+00
-8.73911440e-01 -6.95504606e-01 1.42246330e+00 -7.70874381e-01
4.72330302e-01 -3.52762341e-01 -1.21785140e+00 6.23293698e-01
7.77857006e-02 -4.35178243e-02 1.16888440e+00 7.41227508e-01
-1.74648926e-01 -7.48144416e-03 -1.19292796e+00 -1.13682374e-02
3.17764312e-01 3.04650478e-02 -1.12075758e+00 4.31514978e-01
-1.76114552e-02 1.29603162e-01 -7.49806523e-01 4.18812335e-01
8.08704138e-01 -1.14615166e+00 8.23199689e-01 -8.12822521e-01
1.55299813e-01 9.22796279e-02 1.01014093e-01 -8.68905008e-01
-4.48158234e-01 -3.88411909e-01 6.87647164e-02 1.49612582e+00
4.98876423e-01 -7.77632833e-01 8.72984290e-01 8.26942027e-01
2.36235365e-01 -4.39069062e-01 -1.66499662e+00 -1.14928496e+00
6.55682325e-01 -3.88365567e-01 7.17982173e-01 9.23731327e-01
2.17002049e-01 7.24368915e-02 -4.77822274e-01 3.61107476e-02
8.75110090e-01 7.18314201e-02 3.38940859e-01 -1.70609343e+00
-1.70950815e-01 2.30118036e-01 -1.61072448e-01 -1.31403198e-02
-2.10571304e-01 -5.46667397e-01 -2.66727686e-01 -1.50292587e+00
4.31435138e-01 -3.92827570e-01 -6.15689099e-01 6.50469959e-01
-3.09490740e-01 -1.21176898e-01 -2.67529875e-01 3.75890523e-01
-9.86885875e-02 3.64088804e-01 5.67202866e-01 4.89482135e-01
-2.07678512e-01 7.63126612e-01 -5.41119039e-01 7.04837859e-01
7.73152173e-01 -1.01646304e+00 -1.38135524e-02 1.60349712e-01
3.79487991e-01 7.16602147e-01 7.12466002e-01 -7.34373212e-01
2.76527315e-01 -6.43899441e-02 7.13076353e-01 -7.90695131e-01
5.83462417e-02 -1.24630547e+00 5.65911174e-01 5.90178370e-01
-4.76394594e-03 2.54609585e-01 1.53445750e-01 8.16299379e-01
-6.28679022e-02 -3.12659860e-01 6.55474901e-01 2.24007934e-01
1.86135113e-01 -2.40370601e-01 -6.88609660e-01 -2.46327102e-01
1.03037512e+00 -4.24816996e-01 -7.47048259e-02 -5.76554060e-01
-9.87862825e-01 2.21890092e-01 6.66862369e-01 3.92184407e-01
3.23033452e-01 -1.23160172e+00 -1.05554891e+00 2.83046454e-01
1.55120507e-01 -5.41656435e-01 1.33296177e-01 1.16749191e+00
-8.45954269e-02 1.68438405e-01 -1.50872976e-01 -2.87506253e-01
-1.24937820e+00 9.27430212e-01 3.65406349e-02 -3.67318124e-01
-7.76356161e-01 2.16432452e-01 -1.85535759e-01 -4.44802269e-02
1.30832627e-01 1.64257720e-01 -6.01381212e-02 2.17055351e-01
4.38777596e-01 7.59873927e-01 3.86075020e-01 -1.27087295e-01
-6.79236710e-01 8.62668231e-02 -2.70966202e-01 -1.77219838e-01
1.41998434e+00 -2.40856454e-01 1.48103079e-02 7.48247325e-01
6.21340930e-01 9.03958678e-02 -1.27189052e+00 1.99132636e-01
3.18609148e-01 -3.38954836e-01 8.78309086e-02 -1.35432553e+00
-2.99962282e-01 9.49362516e-01 6.91682160e-01 4.18467522e-01
1.13739359e+00 -1.40068352e-01 1.90808788e-01 -5.32015860e-01
4.99187320e-01 -7.11599290e-01 -8.59080195e-01 2.79263437e-01
8.00011158e-01 -1.19963324e+00 2.72463292e-01 -4.22266275e-02
-3.92587185e-01 1.12786758e+00 -1.70252398e-01 -3.78317356e-01
8.68055940e-01 3.15163553e-01 7.91300759e-02 1.02253973e-01
-8.56665671e-01 2.95926869e-01 1.29891364e-02 6.84112608e-01
5.90861499e-01 9.17421132e-02 -1.06817663e+00 1.06732810e+00
-5.00886887e-02 5.42501450e-01 6.48954749e-01 1.27230418e+00
-2.73095757e-01 -1.31205356e+00 -9.19987142e-01 5.54522574e-01
-9.64665651e-01 -3.08350846e-02 -2.70897388e-01 9.74082351e-01
-1.61210865e-01 1.22419667e+00 -1.52561873e-01 3.88510197e-01
2.30113193e-01 5.49373746e-01 -5.63480929e-02 -2.71669596e-01
-8.47492635e-01 1.59928828e-01 6.95350468e-02 3.37770805e-02
-5.44110266e-03 -1.13318169e+00 -9.98013258e-01 -4.11212355e-01
-8.40523124e-01 4.69040275e-01 9.36601818e-01 7.19218075e-01
1.47761717e-01 3.40588331e-01 6.35063231e-01 -4.71527636e-01
-7.32805431e-01 -7.78737485e-01 -8.88130069e-01 2.82419994e-02
6.60407662e-01 -6.28066540e-01 -9.87444043e-01 -1.43678710e-01] | [7.816097259521484, 5.432304382324219] |
13463f80-faaf-4120-8326-b8bcac894e60 | traffic-signs-detection-and-recognition | 2003.03256 | null | https://arxiv.org/abs/2003.03256v1 | https://arxiv.org/pdf/2003.03256v1.pdf | Traffic Signs Detection and Recognition System using Deep Learning | With the rapid development of technology, automobiles have become an essential asset in our day-to-day lives. One of the more important researches is Traffic Signs Recognition (TSR) systems. This paper describes an approach for efficiently detecting and recognizing traffic signs in real-time, taking into account the various weather, illumination and visibility challenges through the means of transfer learning. We tackle the traffic sign detection problem using the state-of-the-art of multi-object detection systems such as Faster Recurrent Convolutional Neural Networks (F-RCNN) and Single Shot Multi- Box Detector (SSD) combined with various feature extractors such as MobileNet v1 and Inception v2, and also Tiny-YOLOv2. However, the focus of this paper is going to be F-RCNN Inception v2 and Tiny YOLO v2 as they achieved the best results. The aforementioned models were fine-tuned on the German Traffic Signs Detection Benchmark (GTSDB) dataset. These models were tested on the host PC as well as Raspberry Pi 3 Model B+ and the TASS PreScan simulation. We will discuss the results of all the models in the conclusion section. | ['Magdy El-Moursy', 'Marco Magdy William', 'Keroles Khalil', 'Kerolos Gamal Alexsan', 'Pavly Salah Zaki', 'Bolis Karam Soliman'] | 2020-03-06 | null | null | null | null | ['traffic-sign-detection'] | ['computer-vision'] | [-1.23781584e-01 -7.90023685e-01 1.01329178e-01 -1.36927679e-01
-5.25986910e-01 -4.86141723e-03 5.30894041e-01 -7.48697579e-01
-5.50983310e-01 3.90844405e-01 -6.93757176e-01 -8.07195187e-01
-1.67198777e-01 -5.58513761e-01 -4.39141482e-01 -7.56089211e-01
6.81083202e-02 2.04085007e-01 9.18664515e-01 -5.82584560e-01
1.16861098e-01 9.67412829e-01 -2.09622908e+00 2.58351505e-01
6.05898380e-01 1.16245389e+00 1.54238231e-02 1.18873048e+00
1.37285456e-01 1.11293125e+00 -6.80236220e-01 -2.60027796e-01
2.79901028e-01 -4.46749255e-02 1.23342305e-01 -3.04762363e-01
7.20526218e-01 -3.02771986e-01 -6.18036151e-01 6.11574113e-01
6.68255746e-01 1.61481366e-01 4.47588295e-01 -1.51917660e+00
-2.10627347e-01 -2.10344374e-01 -3.95413399e-01 7.15608954e-01
-4.21023130e-01 6.04158878e-01 5.04209399e-01 -8.56450796e-01
4.37487721e-01 1.17188919e+00 7.21420348e-01 5.07445753e-01
-3.31188917e-01 -7.62703180e-01 -2.28213012e-01 1.12447023e+00
-1.19482291e+00 -3.37947935e-01 6.77317977e-01 -3.54077429e-01
1.22870386e+00 3.22214037e-01 3.78004283e-01 8.61892581e-01
2.04057664e-01 1.10459840e+00 1.25561321e+00 -3.44652534e-01
-4.07130877e-03 2.14298919e-01 7.65323877e-01 8.80138159e-01
3.37170959e-01 5.46935916e-01 -1.77615672e-01 2.65232891e-01
5.96224129e-01 -1.69487775e-03 3.19755137e-01 3.36531475e-02
-9.81537998e-01 6.88677788e-01 3.82983953e-01 4.62071717e-01
-5.49856603e-01 6.17364228e-01 5.91841102e-01 3.48107070e-01
2.77232639e-02 -3.79736692e-01 -4.05553401e-01 -2.62797028e-01
-7.10246384e-01 1.19856559e-01 4.41194952e-01 6.31655872e-01
5.08659720e-01 5.99509120e-01 -3.42082888e-01 7.17833698e-01
4.49256331e-01 1.21696675e+00 5.23994982e-01 -4.02337283e-01
4.09972936e-01 3.57201338e-01 -6.44354746e-02 -8.23529840e-01
-4.18693721e-01 -2.98486024e-01 -5.86767077e-01 7.52530336e-01
4.43193316e-01 -3.83681715e-01 -1.41611159e+00 8.76238704e-01
9.42232162e-02 5.01962543e-01 1.19854786e-01 8.91720116e-01
1.16797662e+00 8.49217176e-01 9.18636471e-02 4.42182153e-01
1.48426485e+00 -1.08761013e+00 -6.43410683e-01 -1.85328677e-01
8.00105989e-01 -6.02820992e-01 4.91506219e-01 2.83060938e-01
-4.55821097e-01 -9.92466152e-01 -1.13791502e+00 1.10906232e-02
-9.74636376e-01 8.74952197e-01 4.67437804e-01 1.11711252e+00
-1.05507159e+00 2.21067399e-01 -7.73648560e-01 -6.61782146e-01
2.24312186e-01 2.30223805e-01 -7.28697926e-02 1.49588421e-01
-1.16446638e+00 1.37253606e+00 -8.50094408e-02 6.45314097e-01
-6.79717004e-01 -2.19352156e-01 -3.85132432e-01 -5.48377782e-02
4.29453880e-01 -2.12594360e-01 1.26850557e+00 -6.64432704e-01
-1.52559412e+00 8.38354111e-01 -9.64634866e-02 -8.85636449e-01
7.86350429e-01 -7.00799972e-02 -1.03794873e+00 -3.37529294e-02
-4.77439463e-01 3.84329915e-01 8.41843486e-01 -7.23405123e-01
-1.05806470e+00 -2.02509537e-02 -3.60137254e-01 -3.50401729e-01
2.80466318e-01 5.98343551e-01 -1.90823078e-01 -7.06020668e-02
-4.95732129e-01 -9.99164045e-01 -6.13234527e-02 -1.83783192e-02
-2.01771483e-01 -4.81722653e-01 1.68731046e+00 -6.18551970e-01
1.05820549e+00 -2.06505489e+00 -6.66303575e-01 3.13851297e-01
-1.62899569e-01 1.44739711e+00 -2.39418134e-01 3.43284346e-02
-8.19500014e-02 -4.62797493e-01 4.14803505e-01 1.70095600e-02
1.32868588e-01 2.16911733e-01 -3.36511344e-01 5.91831744e-01
1.10904649e-01 1.21429992e+00 -3.86004686e-01 -4.49580669e-01
1.05719972e+00 7.54971266e-01 2.28463024e-01 -1.71226442e-01
1.58693716e-01 -1.66824833e-01 -4.61083859e-01 8.45669270e-01
5.40563583e-01 1.44596040e-01 -6.98134840e-01 -2.10915670e-01
-6.93118215e-01 -1.49635881e-01 -1.25365531e+00 5.47962189e-01
-4.86113280e-01 1.38723886e+00 -1.82215735e-01 -9.23622966e-01
1.33447659e+00 3.20275843e-01 1.97514653e-01 -1.25186670e+00
3.84785056e-01 5.72108686e-01 1.63922101e-01 -8.67741525e-01
5.80145776e-01 2.47840956e-01 2.46294752e-01 1.09878242e-01
-1.98877051e-01 3.37915301e-01 5.28051734e-01 -1.08709067e-01
9.96102870e-01 -1.64609745e-01 -5.20549305e-02 4.22211766e-01
8.49254191e-01 -8.55276361e-03 2.80576080e-01 7.34164536e-01
-6.70634568e-01 3.04328829e-01 2.01665744e-01 -6.08568072e-01
-9.29305971e-01 -7.77201653e-01 6.91128597e-02 9.53810692e-01
-1.96346954e-01 2.97216177e-01 -3.15367341e-01 -4.60332245e-01
1.38272613e-01 8.00695479e-01 -5.24700940e-01 -6.10975549e-03
-8.59704852e-01 -7.72392809e-01 1.02045798e+00 7.76233613e-01
9.14484203e-01 -1.34543669e+00 -1.25867295e+00 2.19483852e-01
3.85490328e-01 -1.17280030e+00 1.55801892e-01 7.30592832e-02
-4.47881281e-01 -1.26659250e+00 -1.04308498e+00 -6.51048541e-01
1.16669312e-01 6.04546666e-01 4.17311013e-01 1.15829051e-01
-7.02785730e-01 3.14231902e-01 -3.33720744e-01 -7.21022427e-01
-4.09566194e-01 -2.76698679e-01 -2.32044235e-01 4.61356521e-01
6.89206719e-01 8.42211172e-02 -4.13984954e-01 7.44154334e-01
-4.45415586e-01 -2.51605958e-01 8.64443421e-01 4.87957835e-01
6.23009801e-02 -3.99283975e-01 4.68036085e-01 -4.97041196e-01
3.59823912e-01 7.77866244e-02 -1.19748378e+00 5.64828038e-01
-3.14029992e-01 -2.33932436e-01 2.30034411e-01 -3.34406734e-01
-1.01155329e+00 1.14049196e-01 -2.58763224e-01 -6.28599048e-01
-3.87623489e-01 4.21602279e-02 2.12101564e-01 -2.97837853e-01
5.62284231e-01 2.19357669e-01 -1.92693755e-01 -2.69362390e-01
3.26227605e-01 1.13356829e+00 6.83813870e-01 1.88287765e-01
7.43097126e-01 4.09888804e-01 1.96095437e-01 -1.35138559e+00
-1.94863051e-01 -8.59320998e-01 -5.23043513e-01 -8.48409534e-01
8.38797212e-01 -6.95441544e-01 -9.47064519e-01 1.20008349e+00
-1.22061908e+00 -3.63856852e-01 -7.52789006e-02 5.64999461e-01
-2.55502969e-01 1.52815580e-01 -4.63267297e-01 -1.14180934e+00
-4.92125839e-01 -1.08485401e+00 9.37460542e-01 4.33635503e-01
2.75523841e-01 -6.26954675e-01 1.92152634e-01 4.53904539e-01
9.96384323e-01 1.41348019e-01 3.81374210e-01 -5.80544114e-01
-8.12523663e-01 -7.79775083e-01 -5.97137451e-01 6.63339794e-01
-2.34823480e-01 5.99406004e-01 -1.19986045e+00 2.74498284e-01
-4.76424724e-01 2.96713784e-02 1.24512899e+00 6.64402843e-01
4.83615935e-01 3.20940048e-01 -3.36524874e-01 3.77149403e-01
1.25028479e+00 5.48950374e-01 9.82694805e-01 5.47578275e-01
7.03046679e-01 2.78149128e-01 6.27885401e-01 6.99957833e-02
2.47465044e-01 7.52503932e-01 1.21256940e-01 -2.99843997e-01
-6.33850157e-01 1.15392052e-01 6.51716888e-01 4.63360071e-01
-3.38678032e-01 -3.25325906e-01 -1.01985455e+00 4.08805817e-01
-1.91673434e+00 -1.32665372e+00 -7.81919539e-01 1.93205237e+00
-2.27184877e-01 2.54629046e-01 3.70374382e-01 4.56807733e-01
9.13104832e-01 1.83394983e-01 -4.01614249e-01 -7.68369377e-01
-2.03647390e-01 4.24645171e-02 9.29650605e-01 4.64731753e-02
-1.32775092e+00 9.78603899e-01 5.48528194e+00 8.81832421e-01
-1.63181472e+00 1.33478329e-01 3.40632051e-01 1.40852109e-01
6.76481605e-01 -4.43405658e-01 -1.20480013e+00 4.75685388e-01
1.49677706e+00 2.19091848e-01 1.17591061e-01 1.02999091e+00
4.54174310e-01 -1.83197156e-01 -5.35604656e-01 1.01879656e+00
1.63655922e-01 -1.13831413e+00 -4.52572823e-01 -1.90063044e-01
3.82811517e-01 7.34416366e-01 2.60899127e-01 7.48406649e-01
3.85402031e-02 -6.64324939e-01 5.36915421e-01 5.80620825e-01
7.39472568e-01 -5.34042776e-01 8.78817439e-01 5.58088422e-02
-1.38286173e+00 -2.86179721e-01 -3.40457708e-01 3.36215675e-01
5.49542308e-01 3.49633783e-01 -6.71906650e-01 2.88413018e-01
6.22486055e-01 6.90676928e-01 -8.55784297e-01 1.61972642e+00
-3.45398515e-01 7.72341490e-01 -5.30533433e-01 -6.83066666e-01
5.78923523e-01 -3.63882817e-02 4.63016123e-01 1.54889393e+00
4.46731746e-01 -3.56974274e-01 -3.78469586e-01 5.96755147e-01
3.52248400e-01 -2.37849608e-01 -6.12249553e-01 1.47945106e-01
-1.48577228e-01 1.41350126e+00 -8.15663099e-01 -7.80952632e-01
-5.29117882e-01 4.16441619e-01 -4.32596892e-01 3.88662040e-01
-1.34181106e+00 -7.78677702e-01 5.07207811e-01 -1.95495524e-02
8.23040485e-01 -1.17394820e-01 -1.96670499e-02 -8.09368134e-01
-1.13647178e-01 -5.40334523e-01 2.57397324e-01 -1.04468405e+00
-7.68838704e-01 4.45510805e-01 -1.14011832e-01 -1.43556356e+00
-6.09265156e-02 -1.24073017e+00 -9.60060239e-01 5.51931143e-01
-1.92286801e+00 -1.18866265e+00 -4.00007367e-01 4.53160077e-01
5.17160296e-01 -4.31996495e-01 2.08426788e-01 8.77917469e-01
-1.00071943e+00 6.48869216e-01 3.27059567e-01 3.25993121e-01
5.00374734e-01 -7.33943522e-01 7.64784813e-01 1.09476876e+00
8.14240351e-02 -1.37886003e-01 4.18141395e-01 -3.99246663e-01
-1.26068187e+00 -1.20813894e+00 1.03420174e+00 -1.83390141e-01
7.78368831e-01 -9.96258706e-02 -6.86185122e-01 6.62005842e-01
-1.76336601e-01 3.52152735e-01 -9.32627022e-02 -4.56606209e-01
-4.25924867e-01 -5.56043088e-01 -9.15996313e-01 4.15598929e-01
4.73600060e-01 -3.66734385e-01 -4.97823477e-01 2.06685647e-01
1.52665284e-02 -4.01989549e-01 1.10162599e-02 3.43070984e-01
7.26009607e-01 -8.73667181e-01 1.00769424e+00 -2.87389725e-01
-2.79158115e-01 -5.95670462e-01 7.64822885e-02 -8.89941514e-01
1.06375620e-01 -4.16511208e-01 -1.77562520e-01 8.58541310e-01
3.49907368e-01 -9.50612962e-01 6.70715213e-01 5.12234390e-01
-3.34102333e-01 -3.16261292e-01 -1.19811606e+00 -1.07181764e+00
-5.43969214e-01 -8.06664944e-01 2.30964661e-01 7.50579014e-02
-5.82704484e-01 9.80244204e-02 -5.22100925e-01 -4.47024330e-02
4.87926900e-01 -1.56864412e-02 9.80509400e-01 -1.09931493e+00
1.21221222e-01 -6.17078125e-01 -1.05806589e+00 -7.99964547e-01
-3.43554199e-01 -4.72704083e-01 1.41430542e-01 -1.40020120e+00
-3.79139304e-01 -1.72290340e-01 -5.69406331e-01 5.29363394e-01
2.21187487e-01 2.55765647e-01 4.18333530e-01 -2.27682769e-01
-9.48204398e-01 3.15043747e-01 8.53981256e-01 -2.01831743e-01
-7.07295313e-02 4.71427947e-01 2.06473857e-01 5.35937309e-01
6.96459234e-01 -2.87830532e-01 1.23148181e-01 -7.75402635e-02
-1.13203913e-01 -2.94279635e-01 8.95040452e-01 -1.42259312e+00
6.49083555e-01 3.55507612e-01 1.20553523e-01 -1.32160056e+00
4.97683138e-01 -7.64646113e-01 -9.17134285e-02 8.70616317e-01
4.79832888e-02 1.20802380e-01 3.09609234e-01 2.81125247e-01
-1.97443277e-01 -4.48084362e-02 8.56820822e-01 2.59394675e-01
-1.50691414e+00 -7.59218959e-03 -9.52982664e-01 -5.53392351e-01
1.28538251e+00 -5.71439922e-01 -7.48554945e-01 -1.14979163e-01
-1.26303107e-01 2.77156293e-01 -3.44874531e-01 6.80232108e-01
7.84413755e-01 -1.15238643e+00 -8.13525498e-01 3.53860974e-01
1.31377220e-01 -6.15003765e-01 5.20974696e-01 1.10794806e+00
-7.35800505e-01 1.22932899e+00 -4.37532574e-01 -6.50195599e-01
-1.84535587e+00 3.37702096e-01 6.44917190e-01 -1.59206659e-01
-5.86173058e-01 4.29635972e-01 -4.91712362e-01 -1.99638903e-01
4.56559837e-01 -7.20032573e-01 -4.30355936e-01 1.96287632e-01
6.88479722e-01 1.13397264e+00 3.44579428e-01 -9.46509480e-01
-5.42961180e-01 1.03107023e+00 1.81086846e-02 1.25229031e-01
1.17867863e+00 2.32580945e-01 5.06422162e-01 3.07674915e-01
1.18140113e+00 -7.68757343e-01 -9.48263884e-01 -3.08318585e-02
1.32823616e-01 -5.90989664e-02 2.20562264e-01 -8.93192112e-01
-1.33015335e+00 1.17972744e+00 1.25778306e+00 -1.24396861e-01
6.95710421e-01 -2.79229671e-01 1.09815776e+00 9.67406154e-01
1.76178113e-01 -1.26737392e+00 -1.69360906e-01 7.84921885e-01
4.80110198e-01 -1.39825618e+00 -3.60141248e-01 1.22992083e-01
-8.35032821e-01 1.27005875e+00 4.43497956e-01 -2.87608355e-01
5.66857934e-01 2.36580521e-01 5.22582948e-01 -3.07439178e-01
-7.34449923e-01 -8.93471241e-01 2.98309028e-01 5.38364112e-01
6.08028471e-02 1.25525162e-01 -2.63378341e-02 -7.07386881e-02
4.67022598e-01 4.22145993e-01 3.88843268e-01 9.01799023e-01
-6.48968518e-01 -6.21487141e-01 -5.75104892e-01 4.85593557e-01
-1.22876458e-01 2.78229505e-01 -2.23995253e-01 1.06019855e+00
1.05687752e-01 1.05882311e+00 -4.26878780e-02 -4.97100800e-01
7.33348250e-01 1.46461770e-01 8.09628069e-02 8.07713345e-02
-7.38454521e-01 -2.16597393e-01 2.26833045e-01 -5.01177788e-01
-3.79086912e-01 -8.42937708e-01 -1.21013176e+00 -2.69067943e-01
-5.51760674e-01 -3.50539923e-01 1.08258641e+00 9.26539600e-01
1.78175509e-01 7.11719215e-01 5.41608095e-01 -7.00856328e-01
-4.48825270e-01 -9.50032771e-01 -5.39312840e-01 -5.97548410e-02
4.49698806e-01 -5.72572589e-01 -2.32749507e-01 -2.76820183e-01] | [8.06602954864502, -0.7966309189796448] |
823b811a-8c71-42ea-a04b-b181fe6d308d | flar-a-unified-prototype-framework-for-few | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Fan_FLAR_A_Unified_Prototype_Framework_for_Few-Sample_Lifelong_Active_Recognition_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Fan_FLAR_A_Unified_Prototype_Framework_for_Few-Sample_Lifelong_Active_Recognition_ICCV_2021_paper.pdf | FLAR: A Unified Prototype Framework for Few-Sample Lifelong Active Recognition | Intelligent agents with visual sensors are allowed to actively explore their observations for better recognition performance. This task is referred to as Active Recognition (AR). Currently, most methods toward AR are implemented under a fixed-category setting, which constrains their applicability in realistic scenarios that need to incrementally learn new classes without retraining from scratch. Further, collecting massive data for novel categories is expensive. To address this demand, in this paper, we propose a unified framework towards Few-sample Lifelong Active Recognition (FLAR), which aims at performing active recognition on progressively arising novel categories that only have few training samples. Three difficulties emerge with FLAR: the lifelong recognition policy learning, the knowledge preservation of old categories, and the lack of training samples. To this end, our approach integrates prototypes, a robust representation for limited training samples, into a reinforcement learning solution, which motivates the agent to move towards views resulting in more discriminative features. Catastrophic forgetting during lifelong learning is then alleviated with knowledge distillation. Extensive experiments across two datasets, respectively for object and scene recognition, demonstrate that even without large training samples, the proposed approach could learn to actively recognize novel categories in a class-incremental behavior. | ['Ying Wu', 'Wei Wei', 'Peixi Xiong', 'Lei Fan'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['scene-recognition'] | ['computer-vision'] | [ 5.24518073e-01 2.75961459e-01 -3.51546377e-01 -3.91072124e-01
-6.50750279e-01 -5.54586828e-01 7.86198437e-01 6.25222325e-02
-6.25122070e-01 8.95962715e-01 -4.08708155e-02 9.03884768e-02
-3.32433805e-02 -6.46665514e-01 -8.33924949e-01 -1.04307008e+00
7.07705542e-02 4.22208011e-01 1.91050068e-01 1.47748113e-01
3.30475032e-01 6.36977017e-01 -1.86255562e+00 -9.34132934e-02
1.24124801e+00 9.32437599e-01 3.67206872e-01 4.46980447e-01
-1.80062398e-01 9.85222638e-01 -4.60145533e-01 -1.44793004e-01
2.89667428e-01 -1.88158900e-01 -3.75082016e-01 6.32834136e-01
3.57477725e-01 -4.95687127e-01 -1.15275756e-02 8.33602250e-01
2.18161777e-01 6.81360960e-01 3.73252213e-01 -1.13068950e+00
-8.42644811e-01 3.41885984e-01 -3.37301761e-01 1.85385033e-01
3.48486423e-01 7.49851942e-01 6.53023839e-01 -1.29942238e+00
4.53262627e-01 1.19833720e+00 4.00752842e-01 9.84223545e-01
-1.03630984e+00 -4.50594425e-01 8.90093625e-01 4.53876376e-01
-1.20226896e+00 -8.41726542e-01 1.05392075e+00 -2.98277557e-01
9.45046246e-01 1.64383605e-01 1.06480908e+00 1.38781893e+00
-2.35818267e-01 1.21787226e+00 9.72806573e-01 -4.14807737e-01
8.87913465e-01 3.76255631e-01 3.26428920e-01 5.65723181e-01
5.46086729e-01 4.11253780e-01 -6.47862554e-01 -1.84320062e-02
4.50968236e-01 4.84497130e-01 -2.03266367e-01 -1.04689145e+00
-9.44932938e-01 8.01397800e-01 3.64460409e-01 -6.36172146e-02
-5.79752624e-01 -1.82626992e-01 1.82614639e-01 4.43916291e-01
4.07149523e-01 4.48139548e-01 -3.78522426e-01 -1.43435031e-01
-4.07755733e-01 -3.25735472e-02 5.67368031e-01 8.29124510e-01
1.00182521e+00 5.09337604e-01 5.55049405e-02 9.02193487e-01
4.33411121e-01 5.57799876e-01 7.50955164e-01 -7.53085792e-01
2.63058841e-01 1.06296015e+00 1.74953565e-01 -7.12787747e-01
-2.00299043e-02 -6.80676639e-01 -5.88342011e-01 6.30866528e-01
8.38764235e-02 -6.98738620e-02 -1.24646711e+00 1.60248160e+00
5.18027067e-01 2.93065488e-01 5.58377743e-01 7.98735678e-01
4.12682056e-01 5.04227161e-01 9.34449360e-02 -6.45168602e-01
5.82020283e-01 -1.27962255e+00 -4.83293861e-01 -5.64440250e-01
3.20871055e-01 -1.50402620e-01 1.12920094e+00 5.23550451e-01
-8.89661431e-01 -7.87998140e-01 -1.11277854e+00 4.01569158e-01
-5.31055450e-01 -2.08407745e-01 7.15368271e-01 5.09089112e-01
-8.41293991e-01 2.24554002e-01 -9.38600481e-01 -3.46906066e-01
7.12041199e-01 3.27065200e-01 -2.01885000e-01 -1.73353136e-01
-5.80080509e-01 7.83010781e-01 5.17456293e-01 3.61843884e-01
-1.52671230e+00 -4.88155663e-01 -8.62203658e-01 -2.32608631e-01
8.61797571e-01 -5.54999292e-01 1.24190509e+00 -1.46532774e+00
-1.82370806e+00 3.84017885e-01 1.96916331e-02 -7.36135423e-01
4.89852279e-01 -4.95196998e-01 -4.22306389e-01 -1.59691691e-01
-2.77687848e-01 6.46385074e-01 1.40237546e+00 -1.58981144e+00
-6.75171077e-01 -5.63633204e-01 2.95376897e-01 6.88288093e-01
-4.62817997e-01 -6.47706091e-01 -8.64523873e-02 -4.96192217e-01
1.38416916e-01 -8.35804462e-01 -4.55291629e-01 6.44667670e-02
1.90326944e-01 -3.88404995e-01 9.49614108e-01 -1.13512717e-01
6.15599453e-01 -2.20444512e+00 3.54664981e-01 -1.06898779e-02
1.70008138e-01 3.87565821e-01 -3.06639820e-01 2.00376976e-02
2.35266939e-01 -4.40964133e-01 -2.82043606e-01 -5.00711679e-01
-2.18704253e-01 6.16940141e-01 -8.15519392e-01 2.38811016e-01
3.93393338e-01 1.11681807e+00 -9.92998362e-01 -1.55330390e-01
2.62850255e-01 2.45949328e-01 -5.04178464e-01 5.19818783e-01
-3.82005394e-01 7.58197248e-01 -3.31103444e-01 8.54719400e-01
5.64653397e-01 -1.59651741e-01 -3.57250683e-03 2.77108014e-01
-1.77760571e-01 -2.52730370e-01 -9.93548632e-01 1.82111526e+00
-4.15115982e-01 1.32808015e-01 -6.99644387e-02 -1.37067103e+00
1.23745453e+00 -1.30791634e-01 1.50888249e-01 -9.45753276e-01
-2.46335849e-01 1.61290661e-01 -9.57444236e-02 -4.07839447e-01
3.69652122e-01 1.93396613e-01 1.16518915e-01 2.38273934e-01
-4.71551158e-02 1.27599211e-02 1.17473222e-01 -7.89253488e-02
8.54966760e-01 2.65698224e-01 3.40644926e-01 2.73250490e-01
4.89384592e-01 1.05454005e-01 8.32399964e-01 1.17031491e+00
-3.92506301e-01 1.61921039e-01 -2.92028368e-01 -6.74058199e-01
-6.71661854e-01 -1.19735968e+00 2.05879331e-01 1.14509845e+00
3.40375453e-01 1.72774285e-01 -2.42741942e-01 -1.09837353e+00
-7.41590261e-02 7.20575690e-01 -6.08664751e-01 -6.12779081e-01
-8.10016215e-01 -5.43214500e-01 -1.25987828e-01 5.18328071e-01
5.90878189e-01 -1.35831749e+00 -1.11133754e+00 2.59696752e-01
4.73583758e-01 -5.42451203e-01 1.25941206e-02 3.44026864e-01
-1.20896673e+00 -9.65280831e-01 -9.63721097e-01 -6.90134227e-01
1.02581298e+00 4.83659208e-01 7.44363964e-01 -1.11835271e-01
-2.87031889e-01 8.86736274e-01 -6.85765028e-01 -3.16164881e-01
-1.76192164e-01 -1.05106823e-01 2.32372791e-01 2.87797838e-01
2.99581975e-01 -6.25876725e-01 -5.29959023e-01 8.94747600e-02
-7.36815393e-01 3.72640751e-02 1.11004245e+00 1.14456785e+00
5.63609421e-01 -1.84356660e-01 7.23351717e-01 -8.81431878e-01
3.69185418e-01 -4.52790827e-01 -6.83560729e-01 4.80527550e-01
-7.51005352e-01 3.98110179e-03 8.72079253e-01 -9.97528970e-01
-1.32877612e+00 2.42988348e-01 2.44395718e-01 -6.68841898e-01
-2.75660247e-01 3.42984349e-01 -2.16987133e-01 -9.75632444e-02
6.56413555e-01 7.36626804e-01 1.46296434e-02 -3.68323386e-01
5.31098425e-01 4.92521703e-01 3.80890399e-01 -4.18389618e-01
9.26491737e-01 4.28885818e-01 -5.30828476e-01 -6.21099830e-01
-9.22464311e-01 -3.60881180e-01 -5.91500640e-01 -3.69334877e-01
2.23864585e-01 -9.34131145e-01 -4.81396556e-01 5.71890593e-01
-7.46842623e-01 -5.64314783e-01 -1.03771257e+00 5.63627839e-01
-6.63908064e-01 3.13544691e-01 1.17928507e-02 -1.27433085e+00
-3.01332027e-01 -9.58697617e-01 7.07130611e-01 5.98456264e-01
6.35083541e-02 -7.83481061e-01 3.11021090e-01 2.93380201e-01
5.04225016e-01 -3.22707705e-02 5.67204475e-01 -7.02854693e-01
-9.82179344e-01 3.14391148e-03 1.83828667e-01 3.19431573e-01
2.22874075e-01 -4.08905298e-01 -9.93425190e-01 -8.41583133e-01
1.53139099e-01 -8.01892757e-01 1.13252926e+00 -1.39684200e-01
1.09628630e+00 -6.71117365e-01 -2.22934380e-01 4.86852467e-01
1.04928052e+00 8.73073518e-01 4.08133626e-01 2.24988267e-01
5.80746830e-01 4.20946270e-01 1.04219365e+00 6.45679474e-01
1.58352509e-01 5.04823685e-01 6.07970059e-01 2.06755735e-02
-1.07527770e-01 -2.56895661e-01 5.93215048e-01 9.39486861e-01
9.15127024e-02 -5.29948846e-02 -8.47072423e-01 5.07022500e-01
-2.02523494e+00 -8.88364911e-01 8.52513909e-01 2.25704026e+00
6.99479580e-01 2.56601632e-01 -8.19424819e-03 1.50514454e-01
5.84916234e-01 1.45202830e-01 -1.52091861e+00 -5.53240366e-02
-5.92105165e-02 -8.02110061e-02 -3.69470902e-02 4.17400032e-01
-9.09935057e-01 8.43282819e-01 5.50218439e+00 5.37013710e-01
-1.32481039e+00 8.57985094e-02 4.23479527e-01 1.45971198e-02
-1.52089402e-01 2.53361940e-01 -7.69374311e-01 3.14718455e-01
5.34056246e-01 -9.84564945e-02 5.61999261e-01 1.31217575e+00
-1.24621205e-01 -1.16729461e-01 -1.08196151e+00 1.15620494e+00
3.75734776e-01 -9.47774112e-01 3.54671240e-01 -1.96532756e-01
6.83802485e-01 -1.96221441e-01 3.51769626e-01 6.97449625e-01
3.90295327e-01 -6.19416535e-01 7.85020351e-01 9.77142453e-01
5.02936721e-01 -5.81264675e-01 2.04819098e-01 6.00042522e-01
-8.76616478e-01 -7.50759721e-01 -5.63305914e-01 -2.40074471e-02
-1.24088287e-01 3.60355645e-01 -9.66397524e-01 3.06795031e-01
5.28860986e-01 1.04963374e+00 -7.91586101e-01 1.17240810e+00
-9.30918008e-02 4.32873964e-01 -1.51827976e-01 -5.12853675e-02
-3.88320535e-02 -8.91343355e-02 7.46950507e-01 5.94287634e-01
2.51013070e-01 1.00657791e-01 4.97416645e-01 6.85795963e-01
2.02531561e-01 -2.45330527e-01 -6.80873871e-01 -2.35246301e-01
5.78276336e-01 9.52124298e-01 -6.13758326e-01 -4.82856452e-01
-2.79640228e-01 1.08492780e+00 6.25379026e-01 5.36144555e-01
-4.82252806e-01 -7.02554137e-02 1.08815387e-01 -2.96347052e-01
2.78678656e-01 -3.54482800e-01 7.86973014e-02 -1.46179700e+00
1.01406910e-01 -9.19788420e-01 5.56956470e-01 -5.55894434e-01
-1.29554796e+00 5.64513564e-01 -1.11675471e-01 -1.34069121e+00
-5.52016914e-01 -3.03519875e-01 -4.54116583e-01 2.35160634e-01
-1.58225441e+00 -1.02185857e+00 -5.74409664e-01 7.44000733e-01
1.28748620e+00 -6.34678781e-01 7.57941067e-01 -4.37982529e-02
-6.37161076e-01 6.63010836e-01 2.00317591e-01 -2.62060493e-01
3.35223556e-01 -1.01511872e+00 1.19819291e-01 9.11646247e-01
3.65653276e-01 6.60195827e-01 5.12414932e-01 -7.15291679e-01
-1.75359511e+00 -1.07858872e+00 7.09517524e-02 -2.22664222e-01
2.63059527e-01 -4.66802359e-01 -1.02679873e+00 5.52013516e-01
-1.47068858e-01 1.07622497e-01 4.72591549e-01 9.69714671e-02
-3.72963727e-01 -4.59237427e-01 -1.12287748e+00 5.97137928e-01
1.24603724e+00 -3.13708246e-01 -8.09486866e-01 7.10758641e-02
6.64176106e-01 3.35677713e-02 -3.48984599e-01 6.66062057e-01
4.16345835e-01 -7.08771050e-01 8.85118783e-01 -5.17570555e-01
-3.35798234e-01 -3.38838130e-01 -5.49241751e-02 -1.35170150e+00
-3.31682652e-01 -5.22623837e-01 -8.24362159e-01 1.02817798e+00
1.14184119e-01 -9.24369633e-01 9.02017534e-01 3.91240120e-01
-2.50347674e-01 -5.90258062e-01 -9.95589793e-01 -8.92745137e-01
-2.73152381e-01 -8.21786225e-02 4.81849760e-01 8.37878048e-01
-3.06370944e-01 3.03020895e-01 -3.51866364e-01 6.57315254e-02
7.87737489e-01 3.81088018e-01 8.99979591e-01 -1.32037318e+00
-3.78176063e-01 5.32450946e-03 -3.71794462e-01 -1.37744045e+00
9.48496312e-02 -4.64481771e-01 2.39361286e-01 -1.12335026e+00
3.08357328e-01 -6.88309073e-01 -4.34791684e-01 7.26815462e-01
-1.35758474e-01 -1.56497695e-02 2.38957226e-01 6.20325744e-01
-1.13424230e+00 1.17515302e+00 1.08857524e+00 -4.51605916e-01
-5.49792707e-01 2.98643578e-02 -5.62617302e-01 6.47205889e-01
6.79187238e-01 -2.12903887e-01 -9.34182107e-01 -3.61040711e-01
-4.79788445e-02 -1.20391749e-01 3.10833693e-01 -1.19557357e+00
4.95714039e-01 -5.17716110e-01 5.57637572e-01 -6.56774282e-01
6.71461344e-01 -9.07956123e-01 7.97834024e-02 5.21899343e-01
-4.71205413e-01 -1.74892679e-01 -4.58632708e-02 9.67827141e-01
-5.25648519e-02 -3.32078904e-01 8.19604516e-01 -3.51258814e-01
-1.24817395e+00 2.90293485e-01 -3.81394506e-01 -2.09971637e-01
1.29544282e+00 -7.60672033e-01 -3.93640250e-01 -1.27872797e-02
-1.08978248e+00 2.23160446e-01 5.17646015e-01 7.19894290e-01
1.07857490e+00 -1.14236188e+00 -3.22584450e-01 5.67954004e-01
3.52237791e-01 3.25844586e-01 3.05946350e-01 5.50317168e-01
1.06019683e-01 2.76726097e-01 -2.10802794e-01 -7.68631756e-01
-9.62162316e-01 1.01539159e+00 2.05194086e-01 -3.56068090e-02
-6.70057774e-01 6.57061577e-01 3.08740973e-01 -3.95839244e-01
5.88320315e-01 2.66369935e-02 -4.65582639e-01 2.32507735e-01
5.56159854e-01 3.30928355e-01 -6.66441321e-02 -3.58992696e-01
-2.17876643e-01 5.32260180e-01 -6.54465735e-01 6.76755756e-02
1.42448306e+00 -4.44664448e-01 3.97597820e-01 8.68674517e-01
7.57448256e-01 -2.90183544e-01 -1.89993334e+00 -3.79835755e-01
4.21656854e-02 -5.99265933e-01 -1.17338464e-01 -9.80201960e-01
-9.43807125e-01 6.44519508e-01 1.03440619e+00 -2.30850298e-02
1.04910553e+00 -1.59690738e-01 3.86183232e-01 1.11725068e+00
7.71669924e-01 -1.18423808e+00 7.17851222e-01 4.79808360e-01
9.68290865e-01 -1.24966764e+00 -1.37188837e-01 -6.94675669e-02
-6.05001271e-01 9.28957582e-01 1.00866544e+00 -6.31244481e-02
2.46511146e-01 -2.40975156e-01 1.16645351e-01 6.05189949e-02
-9.78897095e-01 -3.76622856e-01 -2.22754385e-02 7.81863213e-01
-5.11740506e-01 -1.89809144e-01 1.68778405e-01 1.08739026e-01
2.41597950e-01 -2.00449582e-02 3.69398028e-01 1.28351533e+00
-7.80343413e-01 -9.14123178e-01 -2.48653799e-01 2.98828602e-01
3.43983918e-01 3.46009761e-01 -3.13159108e-01 4.81154263e-01
2.01585323e-01 8.94454420e-01 1.37062907e-01 -6.55521601e-02
2.65116781e-01 2.87560448e-02 4.99472767e-01 -7.74606705e-01
-1.66633755e-01 -1.76115915e-01 -3.26217830e-01 -3.27799320e-01
-5.23049235e-01 -6.25808299e-01 -9.74284291e-01 2.68513560e-01
-3.79686892e-01 2.59447366e-01 3.44369352e-01 7.88127482e-01
4.45103168e-01 2.74981886e-01 8.63921642e-01 -6.64059162e-01
-8.50390851e-01 -7.37163544e-01 -1.97288796e-01 3.65010530e-01
5.69501996e-01 -1.04871547e+00 -5.47095239e-01 -1.79071892e-02] | [9.821172714233398, 3.1864731311798096] |
e75be8b7-0cd8-4e0f-8226-24e5598c5cfb | joint-voxel-and-coordinate-regression-for | 1801.09242 | null | http://arxiv.org/abs/1801.09242v1 | http://arxiv.org/pdf/1801.09242v1.pdf | Joint Voxel and Coordinate Regression for Accurate 3D Facial Landmark Localization | 3D face shape is more expressive and viewpoint-consistent than its 2D
counterpart. However, 3D facial landmark localization in a single image is
challenging due to the ambiguous nature of landmarks under 3D perspective.
Existing approaches typically adopt a suboptimal two-step strategy, performing
2D landmark localization followed by depth estimation. In this paper, we
propose the Joint Voxel and Coordinate Regression (JVCR) method for 3D facial
landmark localization, addressing it more effectively in an end-to-end fashion.
First, a compact volumetric representation is proposed to encode the per-voxel
likelihood of positions being the 3D landmarks. The dimensionality of such a
representation is fixed regardless of the number of target landmarks, so that
the curse of dimensionality could be avoided. Then, a stacked hourglass network
is adopted to estimate the volumetric representation from coarse to fine,
followed by a 3D convolution network that takes the estimated volume as input
and regresses 3D coordinates of the face shape. In this way, the 3D structural
constraints between landmarks could be learned by the neural network in a more
efficient manner. Moreover, the proposed pipeline enables end-to-end training
and improves the robustness and accuracy of 3D facial landmark localization.
The effectiveness of our approach is validated on the 3DFAW and AFLW2000-3D
datasets. Experimental results show that the proposed method achieves
state-of-the-art performance in comparison with existing methods. | ['Zhenan Sun', 'Qi Li', 'Hongwen Zhang'] | 2018-01-28 | null | null | null | null | ['3d-facial-landmark-localization'] | ['computer-vision'] | [-3.47870797e-01 5.43803163e-02 -9.94155556e-02 -5.31422555e-01
-8.60674858e-01 -3.03382933e-01 6.03472710e-01 -7.03822225e-02
-3.57500881e-01 1.39351472e-01 -4.08092216e-02 -6.22203154e-03
7.14300945e-02 -5.57040691e-01 -5.11392057e-01 -7.25064158e-01
-2.13206969e-02 4.86325532e-01 -5.31920567e-02 3.10841352e-01
3.98900986e-01 1.12867916e+00 -1.55058420e+00 -3.12342435e-01
3.92582029e-01 1.46126378e+00 -2.99730394e-02 1.80482030e-01
-3.13241720e-01 -3.25161293e-02 -2.15425730e-01 -2.46272579e-01
4.75491107e-01 -5.89168631e-02 -3.64903569e-01 3.23952377e-01
7.87789762e-01 -4.70375150e-01 -4.06280980e-02 8.42366755e-01
6.82575285e-01 1.19850822e-01 7.31985211e-01 -1.08222890e+00
-3.02501202e-01 -2.44120613e-01 -1.13185084e+00 -1.48664266e-01
3.29517454e-01 -1.79108560e-01 7.84542680e-01 -1.55275011e+00
5.15082121e-01 1.44995415e+00 7.25608647e-01 4.64905411e-01
-1.07913435e+00 -8.21746528e-01 3.33733290e-01 -2.04455703e-01
-1.99759579e+00 -6.83999300e-01 1.13294017e+00 -3.95959198e-01
7.17622757e-01 -8.26519281e-02 6.57886326e-01 5.02413452e-01
-3.52000771e-03 4.38581437e-01 7.99313784e-01 -4.01240408e-01
1.08709373e-01 -1.45367786e-01 -4.31785196e-01 1.31231999e+00
-9.65447538e-03 -3.58673558e-02 -6.05484784e-01 -1.77031443e-01
1.06418347e+00 3.13986465e-02 2.13738903e-02 -7.49323845e-01
-7.29076326e-01 7.51136482e-01 6.00444198e-01 1.59921646e-01
-4.13368732e-01 3.07373255e-02 1.72729805e-01 -3.49859893e-01
9.51547503e-01 -2.17559382e-01 -3.05545181e-01 1.69637427e-01
-1.13011169e+00 1.03318632e-01 3.80288184e-01 8.09571922e-01
8.80699277e-01 -1.83455981e-02 4.60094810e-02 8.52545738e-01
8.30154479e-01 5.06013572e-01 1.34206951e-01 -8.56663525e-01
2.89642781e-01 7.84768164e-01 -5.13643473e-02 -1.16803229e+00
-6.62236929e-01 -2.26008266e-01 -6.29630864e-01 4.36929703e-01
4.46579039e-01 2.17488274e-01 -1.10873270e+00 1.51916063e+00
9.18674469e-01 4.34706420e-01 -4.02666569e-01 1.08646536e+00
8.29924822e-01 3.97580057e-01 -7.58343637e-02 -1.22284368e-01
1.26206040e+00 -6.80528104e-01 -4.67429727e-01 -1.28671944e-01
4.00709689e-01 -7.66044617e-01 7.47728527e-01 -7.83294886e-02
-1.12379742e+00 -3.86723191e-01 -7.03633547e-01 -3.15131426e-01
-2.50766635e-01 3.35486889e-01 4.74392951e-01 6.77512407e-01
-1.11763930e+00 1.38187140e-01 -9.94653523e-01 -2.36409038e-01
7.83401430e-01 5.42935669e-01 -5.68400919e-01 4.57150228e-02
-5.69566965e-01 7.30794609e-01 3.34065035e-02 4.73968297e-01
-6.79388642e-01 -8.66199017e-01 -1.18817437e+00 2.39321794e-02
1.82402819e-01 -4.98297155e-01 8.81743073e-01 -3.60859126e-01
-1.61274385e+00 1.24513972e+00 -5.37481070e-01 1.36849925e-01
4.55106646e-01 1.35216303e-02 1.24156252e-01 2.42664456e-01
1.79718047e-01 7.37375677e-01 9.64890838e-01 -1.31026256e+00
-5.42656481e-01 -9.30894494e-01 -8.82194713e-02 2.80615568e-01
-1.39214158e-01 -1.82809025e-01 -9.42335725e-01 -2.00010672e-01
6.43183827e-01 -6.49060011e-01 -7.62380809e-02 6.23330414e-01
-2.91547716e-01 -4.32413489e-01 6.15569949e-01 -5.42240381e-01
8.21699083e-01 -2.28049469e+00 8.99745300e-02 4.98082191e-01
1.71607912e-01 3.00142560e-02 -9.70494747e-02 -4.05828804e-02
1.05918013e-01 1.21633351e-01 1.47612125e-03 -1.00279605e+00
-4.40741293e-02 -1.63081542e-01 5.34593314e-02 9.74622905e-01
3.26416701e-01 7.78901398e-01 -6.82096601e-01 -7.79780626e-01
4.44987625e-01 1.06323051e+00 -5.15985191e-01 1.98294967e-01
5.62482066e-02 5.27843058e-01 -5.79983771e-01 1.15490484e+00
1.05192292e+00 -4.75642309e-02 -1.78677943e-02 -4.24751699e-01
-2.56685913e-01 8.19003023e-03 -1.09510887e+00 1.94419146e+00
-6.64912283e-01 1.93386674e-01 2.07605571e-01 -5.97051799e-01
1.25592828e+00 1.85750499e-01 6.83248103e-01 -7.40837872e-01
2.57622629e-01 2.55529106e-01 -6.09694898e-01 -1.98498189e-01
8.11142996e-02 -1.88036203e-01 1.39586195e-01 2.07347587e-01
1.00097256e-02 -2.08737999e-01 -3.03853184e-01 -4.13740754e-01
4.53508317e-01 4.56175923e-01 3.16548675e-01 2.49414891e-02
6.67566538e-01 -4.25885677e-01 5.95779777e-01 9.54356566e-02
-1.30706653e-01 6.89098239e-01 4.45684105e-01 -5.41478872e-01
-7.88849473e-01 -1.03719234e+00 -4.18553561e-01 6.29183531e-01
1.65589616e-01 -3.19542438e-01 -7.12242126e-01 -8.20493698e-01
2.25700721e-01 2.69225061e-01 -7.89069831e-01 1.40003741e-01
-5.86292028e-01 -3.18175793e-01 3.88333887e-01 2.89799392e-01
3.97418648e-01 -5.86209714e-01 -5.77193856e-01 -1.22609191e-01
2.37218365e-01 -1.13638425e+00 -5.92699349e-01 -1.17926031e-01
-9.38283503e-01 -1.03111064e+00 -7.70300865e-01 -7.99076557e-01
1.23023665e+00 3.56126845e-01 5.96645713e-01 2.64238536e-01
-2.86396801e-01 2.95800894e-01 -2.24406086e-02 -1.15239657e-01
2.19435319e-01 -1.84074610e-01 1.57631055e-01 2.95737326e-01
3.95257980e-01 -5.29725254e-01 -8.60528529e-01 3.45440239e-01
-5.02135992e-01 -2.04328626e-01 4.55575705e-01 6.72985613e-01
1.02778435e+00 -9.43998098e-02 2.12802738e-01 -4.90165830e-01
1.12320907e-01 -2.27255955e-01 -9.30447578e-01 6.21928163e-02
-4.86715615e-01 -2.54209876e-01 3.55236620e-01 -2.28385925e-01
-8.51824880e-01 6.03313804e-01 -2.48036206e-01 -8.23900700e-01
-1.84785321e-01 2.80673057e-01 -3.13726038e-01 -4.93494838e-01
1.69033721e-01 2.10426405e-01 2.52908200e-01 -5.67473650e-01
3.97908032e-01 4.66057181e-01 1.93683609e-01 -4.74811703e-01
9.26017761e-01 6.81585848e-01 4.60535288e-01 -8.75782609e-01
-7.80738473e-01 -3.63828778e-01 -1.09468448e+00 -3.21044713e-01
7.70985723e-01 -9.68277991e-01 -8.76275778e-01 2.30670750e-01
-1.14336002e+00 -6.12397306e-02 9.98914912e-02 3.56001556e-01
-5.05347729e-01 2.89789289e-01 -2.29264989e-01 -9.46691573e-01
-2.69871116e-01 -1.39804661e+00 1.63985837e+00 2.22692788e-01
-1.30837678e-03 -9.57025111e-01 -2.06300750e-01 2.13602379e-01
1.13191597e-01 4.05057609e-01 8.26876760e-01 -1.24342054e-01
-6.00968778e-01 -4.87992227e-01 -3.67160350e-01 -7.01131448e-02
1.35500759e-01 -7.38910511e-02 -1.10168362e+00 -1.78242177e-01
-8.92530903e-02 -7.31329620e-02 4.79186058e-01 5.40346801e-01
1.22923303e+00 -6.39432594e-02 -3.44025701e-01 1.00285721e+00
1.32613170e+00 3.98658849e-02 2.58630931e-01 1.67780176e-01
7.38418221e-01 6.21048152e-01 7.36524820e-01 5.31778276e-01
5.92502058e-01 8.69170368e-01 6.62238359e-01 -2.28655010e-01
-1.22174092e-01 -4.86350626e-01 -4.45949920e-02 4.69911993e-01
-1.16087310e-01 2.72833139e-01 -7.18291819e-01 2.96164006e-01
-1.45007515e+00 -4.73242730e-01 3.03877026e-01 2.28961229e+00
5.27854741e-01 -1.59814134e-01 8.37065354e-02 -3.89438048e-02
5.48088372e-01 2.39837751e-01 -5.50413907e-01 -1.59415543e-01
2.42378786e-01 9.81032178e-02 2.70237505e-01 6.67397201e-01
-9.94474411e-01 1.14262712e+00 5.56955147e+00 8.21753621e-01
-1.30345726e+00 6.73347637e-02 8.38311732e-01 -2.52595156e-01
-9.18509290e-02 -2.48980939e-01 -1.16528273e+00 2.97093004e-01
3.50656807e-01 2.66372055e-01 2.92296082e-01 8.50226283e-01
3.08412343e-01 -5.83924167e-02 -8.80289793e-01 1.30560994e+00
3.28903139e-01 -1.19520712e+00 -2.60718819e-02 2.78837413e-01
3.97173673e-01 -2.43987724e-01 1.94129705e-01 2.33831706e-05
-5.59536219e-01 -1.00796044e+00 8.15636277e-01 5.61776757e-01
1.12408459e+00 -1.07712996e+00 4.35450464e-01 2.76935250e-01
-1.39521420e+00 1.97137833e-01 -4.76504624e-01 2.02244282e-01
1.28924236e-01 4.91427094e-01 -1.04916751e+00 2.95012057e-01
6.15973175e-01 4.98044908e-01 -4.58682775e-01 1.02047038e+00
-2.03638226e-01 -1.20268865e-02 -5.55048108e-01 6.53733388e-02
1.85006693e-01 -2.86469042e-01 3.62552762e-01 9.54452038e-01
5.86440563e-01 1.03184924e-01 1.77883133e-01 7.94461191e-01
-1.49132967e-01 3.09624761e-01 -7.19999731e-01 3.69709224e-01
5.81547320e-01 1.52990282e+00 -9.91502285e-01 1.60855547e-01
-3.44123840e-01 8.36473644e-01 6.40678346e-01 2.89217710e-01
-7.15877295e-01 -8.35335031e-02 5.97383082e-01 3.05825323e-01
4.20516282e-01 -3.66751701e-01 -2.23420724e-01 -8.18612993e-01
1.39091820e-01 -2.55365759e-01 1.59041688e-01 -5.31621277e-01
-1.04513049e+00 7.02355206e-01 -1.08539857e-01 -1.12819743e+00
-1.75725773e-01 -5.56627452e-01 -3.81448865e-01 8.95985961e-01
-1.80109608e+00 -1.41028237e+00 -3.45185965e-01 5.38106501e-01
3.55497658e-01 5.45099797e-03 8.55831325e-01 3.41441274e-01
-5.79961300e-01 8.73423755e-01 -3.54542345e-01 2.21567959e-01
5.58328450e-01 -9.27485168e-01 2.22818613e-01 5.56487322e-01
1.27855897e-01 5.79820096e-01 7.49100074e-02 -5.47239542e-01
-1.58945870e+00 -1.11507940e+00 9.53447163e-01 -3.11440319e-01
2.52697885e-01 -5.83458364e-01 -6.14178181e-01 5.14527798e-01
-4.61359531e-01 6.08012676e-01 5.51781833e-01 6.99055865e-02
-4.43639040e-01 -3.18281502e-01 -1.43660998e+00 5.70685804e-01
1.06299400e+00 -5.62294960e-01 -9.20837969e-02 2.28809759e-01
4.42451239e-01 -7.46274412e-01 -9.18904364e-01 3.26633662e-01
7.49220192e-01 -8.94962907e-01 1.12654161e+00 -1.76643595e-01
6.37884066e-02 -4.84210283e-01 -2.30050266e-01 -9.78940666e-01
-1.61022201e-01 -3.40294600e-01 -2.09694773e-01 1.21959138e+00
1.40158743e-01 -3.86498570e-01 1.00906062e+00 5.49086392e-01
-7.93454945e-02 -1.28361034e+00 -1.48181379e+00 -4.00340557e-01
-2.84036964e-01 -3.41169268e-01 7.96399891e-01 6.78387046e-01
-4.27546591e-01 9.30075720e-02 -1.22487120e-01 4.96499091e-01
8.22559893e-01 2.48307914e-01 7.84815192e-01 -1.23510802e+00
4.50049996e-01 -6.09788120e-01 -7.09939957e-01 -1.24149156e+00
4.85433131e-01 -1.04718041e+00 -6.97118491e-02 -1.21304202e+00
-1.30127594e-01 -5.91247678e-01 -5.12316823e-02 5.62404990e-01
-3.74187226e-03 6.11264110e-01 1.54188080e-02 5.57240564e-03
-3.48173887e-01 7.01037765e-01 1.29246271e+00 5.69400378e-02
-3.70731980e-01 -5.84782921e-02 -4.37862068e-01 8.08937550e-01
4.99834955e-01 -2.60378361e-01 -2.27338329e-01 -5.67180991e-01
-2.70834595e-01 1.55314848e-01 4.31356043e-01 -5.51116705e-01
2.89561063e-01 1.90858275e-03 8.67729068e-01 -9.56437767e-01
8.34784925e-01 -1.05160177e+00 -2.53107160e-01 9.42823142e-02
4.22881432e-02 -3.21338885e-02 1.59577280e-01 4.03811723e-01
-3.83118689e-02 -7.25369230e-02 9.54456985e-01 -1.30689954e-02
-4.87433821e-01 9.08585489e-01 3.35975409e-01 -3.12705368e-01
1.11383736e+00 -4.50846225e-01 3.01943392e-01 -9.17863399e-02
-5.17204523e-01 1.31425977e-01 5.87027133e-01 3.16313654e-01
1.10030270e+00 -1.54862750e+00 -5.70066333e-01 6.77349448e-01
8.35774466e-02 3.87789607e-01 2.53827184e-01 9.57948864e-01
-5.03404617e-01 3.96098137e-01 1.52583038e-02 -1.00086534e+00
-1.19963729e+00 3.34210038e-01 5.65842152e-01 1.76589698e-01
-6.83619201e-01 9.04343784e-01 1.58638954e-01 -5.64039946e-01
5.39933920e-01 -4.35561724e-02 -2.67382652e-01 2.65920848e-01
6.48179531e-01 3.47082987e-02 1.11947477e-01 -1.07273853e+00
-6.56231284e-01 1.31024194e+00 -1.56122660e-02 6.47275150e-02
1.43909848e+00 -3.42255652e-01 -1.95680767e-01 2.57785141e-01
1.53449357e+00 4.05731983e-02 -1.47862673e+00 -1.81881130e-01
-2.23575294e-01 -8.53746355e-01 3.11827511e-01 -3.09508115e-01
-1.39703226e+00 9.30862606e-01 7.62875080e-01 -3.56037229e-01
1.03603661e+00 1.02979252e-02 5.74562550e-01 -6.75397692e-03
3.95255119e-01 -6.20353580e-01 -1.31054789e-01 3.30830604e-01
8.87559533e-01 -1.26594949e+00 6.82294667e-02 -5.34353077e-01
-1.62625074e-01 1.30059457e+00 6.48030341e-01 -1.28304008e-02
8.97219121e-01 -5.77883283e-03 1.30278721e-01 -4.11413431e-01
-1.75844118e-01 -1.44483268e-01 5.67400694e-01 5.07345676e-01
3.19471627e-01 -1.98523968e-01 6.23790659e-02 2.53520429e-01
1.35207437e-02 -2.58655787e-01 -1.64718762e-01 6.33777320e-01
-2.49661297e-01 -8.19458961e-01 -4.49319810e-01 3.77556235e-02
-2.79975533e-01 1.22180730e-01 -1.04992971e-01 9.44587052e-01
1.80785269e-01 6.33105397e-01 2.58456916e-01 -1.74557641e-01
3.77744913e-01 7.43444338e-02 6.69093430e-01 -5.33353150e-01
-2.27084961e-02 5.11032283e-01 -4.18464988e-01 -8.09350431e-01
-3.38610888e-01 -5.09809792e-01 -1.43396771e+00 -1.19080253e-01
-2.45798394e-01 -7.05505684e-02 1.03959286e+00 7.34172761e-01
5.02024651e-01 4.32139225e-02 9.94819820e-01 -1.42497659e+00
-3.06159675e-01 -6.59459651e-01 -5.64339936e-01 6.09868318e-02
4.83437389e-01 -1.15020561e+00 -3.59266907e-01 -3.01224023e-01] | [13.474387168884277, 0.3218021094799042] |
26ff491a-dc37-4c06-b3f7-c1ece6212891 | simulation-and-measurement-of-human | 2208.00837 | null | https://arxiv.org/abs/2208.00837v2 | https://arxiv.org/pdf/2208.00837v2.pdf | The feasibility of Q-band millimeter wave on hand-gesture recognition for indoor FTTR scenario | The generalization for different scenarios and dif-ferent users is an urgent problem for millimeter wave gesture recognition for indoor fiber-to-the-room (FTTR) scenario. In order to solve this problem and verify the feasibility of FTTR Q-band millimeter wave in gesture recognition, we build a real-time millimeter wave gesture recognition system. The moving hand-gestures are represented as a variety of time-variant spec-trum features, such as micro-Doppler feature, and then the feature learning and classification is realized by using a convo-lution neural network (CNN). The experimental results show that the millimeter wave gesture recognition system can achieve the generalized gesture recognition for 2 scenarios and 4 users. | ['Yanbo Zhao', 'Feng Xu', 'Zhaoyang Xia', 'Yuxuan Hu'] | 2022-07-22 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.13579103e-01 -5.69716752e-01 5.31149060e-02 -5.47212243e-01
-4.02258456e-01 -2.99325496e-01 2.25402489e-01 -1.27464247e+00
-6.82133436e-01 4.37625557e-01 -6.95161298e-02 -6.60080016e-01
-5.00359774e-01 -9.59093809e-01 3.10388803e-01 -9.38148737e-01
-4.38006908e-01 9.89870504e-02 -2.08037533e-02 -9.29547772e-02
1.20827422e-01 8.51604939e-01 -6.36596024e-01 7.48996390e-03
3.27604622e-01 1.28394461e+00 -5.68372011e-02 1.01573849e+00
-9.84623060e-02 1.42087787e-01 -9.52193081e-01 2.17860743e-01
9.17711139e-01 -2.45769024e-01 -4.68696468e-02 -3.07914972e-01
4.33772475e-01 -6.41214132e-01 -1.14592564e+00 7.35278070e-01
9.30084765e-01 2.09557980e-01 4.70301300e-01 -1.00801945e+00
-4.41161692e-01 1.71180382e-01 -4.65617031e-02 3.60411584e-01
2.67966866e-01 1.49863869e-01 3.00985575e-01 -4.75595921e-01
4.17475045e-01 1.09400320e+00 5.92920363e-01 6.25567377e-01
3.39407101e-02 -1.24258149e+00 3.00915428e-02 2.21216664e-01
-1.51501656e+00 -6.14679009e-02 4.45919842e-01 -1.37174487e-01
8.78946185e-01 5.36651611e-01 8.36476266e-01 8.91852677e-01
6.02345347e-01 3.49629164e-01 6.18737519e-01 -1.45978630e-01
1.02812409e-01 -8.30218315e-01 3.79382312e-01 1.15494800e+00
3.45580369e-01 5.95436275e-01 -3.17038536e-01 3.12711000e-01
1.33755159e+00 6.28269255e-01 -5.09750605e-01 3.64134550e-01
-1.37885404e+00 2.08121911e-01 9.81875300e-01 7.28326023e-01
-1.93434283e-01 7.30409801e-01 -2.76353031e-01 7.61092246e-01
-1.97755739e-01 4.04618323e-01 -5.31510532e-01 -3.64640027e-01
-7.46137917e-01 1.37214493e-02 7.09810436e-01 1.24418402e+00
4.87333834e-01 6.07236266e-01 -2.32648894e-01 2.18938068e-02
8.79520416e-01 1.32497168e+00 8.72901440e-01 -1.33935288e-01
8.68615866e-01 7.85539672e-02 1.60050482e-01 -7.60573387e-01
-9.94234204e-01 -6.56700850e-01 -1.05694771e+00 3.30262065e-01
3.54154557e-01 -8.15145791e-01 -1.73150170e+00 8.00219536e-01
-1.68673292e-01 4.80641961e-01 1.53113544e-01 1.22744250e+00
1.40913403e+00 4.30440485e-01 -1.92604698e-02 -1.15618788e-01
1.00359416e+00 -6.60000861e-01 -6.06931567e-01 5.48048876e-02
5.77849209e-01 -5.28946400e-01 5.37308395e-01 1.55150220e-01
-1.71513274e-01 -5.72176456e-01 -1.22902977e+00 4.78932619e-01
-2.82168955e-01 3.51379514e-01 1.08541453e+00 1.51631749e+00
-3.56455326e-01 2.13533908e-01 -1.23112977e+00 -3.64042133e-01
4.44481730e-01 7.46043682e-01 1.00098409e-01 -1.04104631e-01
-8.89622748e-01 2.77706265e-01 -4.00172807e-02 7.13447809e-01
-4.02884632e-02 -5.01349330e-01 -1.62263483e-01 3.69391888e-02
-4.04342681e-01 -5.03834128e-01 1.13029122e+00 -2.91326225e-01
-1.84765077e+00 3.53895687e-02 2.14518677e-03 1.02064185e-01
4.42948550e-01 -7.83866039e-04 -1.17273033e+00 -1.03746362e-01
-4.61748123e-01 -1.67510912e-01 5.81192255e-01 -6.49346828e-01
-7.90705025e-01 -5.84308624e-01 -3.95572521e-02 -4.96776514e-02
-4.27700318e-02 -1.30658016e-01 -3.37292627e-02 -3.93817604e-01
1.23017848e+00 -8.13740492e-01 -2.03679860e-01 -1.06881976e-01
-3.91546518e-01 6.94213212e-02 1.23438907e+00 -4.44775075e-01
9.54369247e-01 -2.05466175e+00 -4.75778103e-01 6.17019534e-01
8.34826306e-02 3.14227611e-01 -1.86394706e-01 3.53687942e-01
3.15723419e-01 -1.01308361e-01 3.57194781e-01 3.95549476e-01
-7.21575841e-02 6.73097149e-02 -2.63568640e-01 4.24982637e-01
-6.16373003e-01 1.21413553e+00 -6.45390153e-01 2.22374335e-01
2.03683808e-01 5.85099235e-02 -6.97059780e-02 -2.16181912e-02
4.18691486e-01 8.85302603e-01 -1.22427702e+00 1.03072679e+00
9.52850580e-01 1.88097775e-01 -1.56533092e-01 -3.04396003e-01
-3.99584055e-01 -9.95361581e-02 -1.23951113e+00 1.46653295e+00
-5.96786380e-01 9.41625953e-01 -2.84546524e-01 -6.31996274e-01
1.29876196e+00 2.12742031e-01 4.99285191e-01 -1.05818307e+00
4.37967479e-01 3.33220422e-01 3.23981375e-01 -1.11788309e+00
9.77120101e-02 -4.94349420e-01 -6.78140447e-02 4.42229688e-01
-6.84355274e-02 -1.44548073e-01 -5.29890180e-01 -6.12703621e-01
1.44458485e+00 -3.10153157e-01 -1.49304360e-01 1.28791630e-01
1.64126605e-01 -2.77358741e-01 3.70212883e-01 9.36551630e-01
-3.33030641e-01 3.92732829e-01 -7.12492764e-01 -8.90251756e-01
1.36432927e-02 -1.28361607e+00 1.15424320e-01 3.97126287e-01
6.80959165e-01 1.87444359e-01 1.14583433e-01 -8.62467051e-01
-3.38411168e-03 3.92108299e-02 2.74380922e-01 8.54179338e-02
-8.60685766e-01 -6.45731509e-01 7.99699008e-01 6.43733382e-01
1.52898669e+00 -7.24847436e-01 -1.13670099e+00 2.20923752e-01
2.26588696e-01 -1.02054584e+00 -5.52648343e-02 1.76573232e-01
-8.00457478e-01 -9.54367936e-01 -6.09827399e-01 -9.40726638e-01
3.48135591e-01 6.86191320e-01 3.42166841e-01 2.59173036e-01
-5.84713221e-01 5.71002245e-01 -6.62739694e-01 -5.58096349e-01
5.99150538e-01 8.14870670e-02 2.04369500e-01 -2.06978470e-01
7.08332658e-01 -7.73602843e-01 -7.80080259e-01 3.02572191e-01
-2.79871672e-01 -3.98204118e-01 9.03266668e-01 5.00577331e-01
-1.73163056e-01 3.06882784e-02 1.60124943e-01 -2.37301707e-01
4.45725590e-01 2.38088205e-01 -6.24068499e-01 4.22222227e-01
-2.48807654e-01 -2.20272899e-01 3.40167791e-01 -4.78431225e-01
-8.75995457e-01 1.68387685e-02 -3.19757640e-01 2.27447748e-02
-1.06996343e-01 6.84740365e-01 -3.45803946e-01 -1.02470958e+00
6.21807456e-01 3.29853982e-01 -6.12289011e-01 -1.75866202e-01
2.46755615e-01 1.12387311e+00 3.38449538e-01 -1.78077668e-01
1.57741499e+00 4.83854085e-01 1.71530858e-01 -1.34524500e+00
-8.43001828e-02 -5.89538515e-01 -6.46762490e-01 -3.83742630e-01
8.59188735e-01 -7.58363724e-01 -9.19815361e-01 5.30560255e-01
-1.08145523e+00 -6.74852192e-01 1.62416071e-01 1.40128517e+00
-1.59740984e-01 1.68328583e-01 -4.42381084e-01 -8.69621873e-01
-1.21774025e-01 -6.30697608e-01 9.23388660e-01 8.84431601e-01
2.54597008e-01 -6.11262679e-01 -2.49499202e-01 -1.33515343e-01
8.75761390e-01 -4.32332531e-02 5.27130425e-01 -2.97038108e-02
-1.40022075e+00 -6.08160913e-01 -4.09781009e-01 -6.33117139e-01
3.95736963e-01 -3.94990027e-01 -7.49084830e-01 -1.48568615e-01
-9.29794312e-02 3.64662200e-01 8.08463395e-01 4.52646077e-01
1.08386862e+00 -3.09174601e-02 -5.33715725e-01 1.46889591e+00
1.52319944e+00 9.77706075e-01 1.06243932e+00 7.06650689e-02
7.11151063e-01 -4.88843143e-01 5.82274139e-01 4.36695486e-01
-1.94241911e-01 5.94986498e-01 2.52807617e-01 -1.29179865e-01
-1.80513129e-01 2.10558385e-01 1.39197311e-03 5.33122838e-01
-8.62907171e-01 -3.65308940e-01 -7.57079005e-01 -3.61392051e-01
-1.49667931e+00 -1.32489288e+00 -3.83348674e-01 1.74796546e+00
-3.48251283e-01 -3.30862582e-01 -4.99881715e-01 -1.89858496e-01
1.92950889e-02 2.59240627e-01 -4.79054719e-01 3.15172166e-01
4.57819015e-01 9.53287065e-01 9.45219457e-01 5.22831440e-01
-1.24346089e+00 1.02151680e+00 6.47883797e+00 3.39652270e-01
-1.96806848e+00 -8.92730877e-02 -1.51551872e-01 -3.90030667e-02
-3.22279433e-04 -5.51003873e-01 -7.31092870e-01 3.77174951e-02
3.23670000e-01 7.03735054e-02 4.04270142e-01 6.43160701e-01
3.86238322e-02 3.80336314e-01 -5.60108066e-01 1.77252686e+00
-1.01597659e-01 -1.24298179e+00 -2.17067748e-01 7.68521503e-02
3.90968591e-01 2.67070293e-01 1.70002326e-01 8.26653183e-01
9.20176581e-02 -1.09869874e+00 2.44212791e-01 6.65634394e-01
1.16919804e+00 -2.33194306e-01 8.00123692e-01 2.59504080e-01
-1.76048553e+00 -3.98877650e-01 -6.12157285e-01 -4.49445516e-01
6.81945831e-02 4.30617571e-01 -7.09288180e-01 8.20013404e-01
4.44133162e-01 3.68454397e-01 2.56676832e-03 1.40278840e+00
-2.06611022e-01 5.21262884e-01 -4.17308360e-01 -7.69004166e-01
3.04383159e-01 -1.65227950e-01 1.95140153e-01 1.24246955e+00
1.11674035e+00 9.59835291e-01 2.92512000e-01 3.93846750e-01
2.11403266e-01 -3.96917045e-01 -2.81296134e-01 -2.22262263e-01
3.74136448e-01 1.17618024e+00 -8.73716235e-01 -1.69550404e-02
-2.49876067e-01 1.08407736e+00 -7.40105033e-01 1.12137723e+00
-5.79579353e-01 -1.13179576e+00 9.08428311e-01 -3.93095851e-01
2.62356967e-01 -1.33034360e+00 -4.86015260e-01 -1.17574227e+00
1.17455028e-01 5.51401451e-03 4.00786623e-02 -5.69503546e-01
-9.54959393e-01 4.76997018e-01 -7.12935567e-01 -1.66550648e+00
-8.46552849e-02 -1.27905250e+00 -8.23101044e-01 1.00547183e+00
-1.55311954e+00 -1.30375051e+00 -1.31937361e+00 5.90062618e-01
4.27474976e-01 -5.63788056e-01 1.14960659e+00 3.59120965e-01
-1.85348749e-01 9.57543075e-01 1.38652146e-01 7.29874194e-01
3.53887081e-01 -6.32616580e-01 2.87881315e-01 6.42502129e-01
2.98091769e-01 7.44040251e-01 5.61264120e-02 -7.45974183e-01
-1.98339069e+00 -1.15650117e+00 4.86914903e-01 -2.05759093e-01
2.55560160e-01 -1.44601315e-01 6.26717210e-02 8.60134423e-01
-5.02664387e-01 3.70958656e-01 6.67977989e-01 2.38978922e-01
-2.49575168e-01 -3.62102687e-01 -1.22154641e+00 5.72403729e-01
1.79759622e+00 -3.04175556e-01 -2.90315807e-01 3.92239958e-01
3.45079035e-01 -7.27620840e-01 -3.82460445e-01 2.52975136e-01
1.22197890e+00 -5.67355871e-01 1.17448711e+00 -5.98910093e-01
-2.09902197e-01 -3.01709205e-01 -5.94908535e-01 -1.18544865e+00
-5.58966160e-01 -5.56061029e-01 2.83840418e-01 2.35014454e-01
8.68927613e-02 -9.17126656e-01 1.45238054e+00 3.61842602e-01
-3.98871273e-01 -3.56937170e-01 -1.56466722e+00 -1.33534348e+00
-3.16595823e-01 -6.34011090e-01 9.66638803e-01 7.59953678e-01
1.07541710e-01 -8.68391097e-02 -1.50680304e-01 7.09111869e-01
3.33582163e-01 2.39334375e-01 9.78971481e-01 -1.39127815e+00
-2.15407774e-01 -3.94983381e-01 -1.08567023e+00 -2.00057912e+00
-2.76477039e-01 -8.85599673e-01 3.30242127e-01 -1.88114524e+00
-6.87931716e-01 -8.54276478e-01 -2.72299409e-01 2.38480777e-01
4.15932953e-01 2.11152285e-01 8.64892602e-02 1.48145556e-01
-2.70114779e-01 6.20979011e-01 1.50433934e+00 -5.72587788e-01
-5.55005968e-01 6.26895189e-01 -1.12075701e-01 3.36835861e-01
5.78010321e-01 2.43351739e-02 -1.99551836e-01 -7.30110347e-01
-2.57718116e-01 3.16024631e-01 3.85198176e-01 -1.77427185e+00
4.35470462e-01 -6.52303040e-01 9.14623797e-01 -4.58129257e-01
4.11997885e-01 -1.11157095e+00 -8.99436027e-02 1.10382354e+00
4.75566775e-01 -3.95605892e-01 -2.56550103e-01 4.02985454e-01
7.04581290e-02 3.03663433e-01 1.40905812e-01 -4.87022214e-02
-1.02806354e+00 1.04356778e+00 -5.37992358e-01 -6.92382276e-01
5.49386203e-01 -6.02043211e-01 -5.56903183e-01 -5.30151010e-01
-5.35925686e-01 -1.58261031e-01 -5.00897169e-01 1.58607766e-01
1.01168454e+00 -1.58460951e+00 -3.62242132e-01 7.66163468e-01
1.52926922e-01 -4.84774739e-01 2.48313397e-01 5.02918303e-01
-9.48160708e-01 6.89325571e-01 -4.15423840e-01 -3.04694921e-01
-1.34280419e+00 -3.84810120e-01 9.07867730e-01 1.23422623e-01
-9.67775524e-01 1.11569023e+00 -3.46727133e-01 -4.38473105e-01
1.56975716e-01 -9.91621375e-01 7.02228844e-02 -6.92439675e-01
5.95463395e-01 4.43526246e-02 1.00176550e-01 -1.17053397e-01
-5.22557080e-01 1.08838379e+00 6.04915023e-01 -5.15402198e-01
1.25567043e+00 2.45194480e-01 4.77511436e-01 7.74113869e-04
1.03720903e+00 8.93608108e-03 -9.37543750e-01 1.09992117e-01
-3.06112826e-01 -8.05358529e-01 -1.97688356e-01 -1.12389290e+00
-1.19909978e+00 8.26608479e-01 1.42964578e+00 -5.01902103e-02
9.02556956e-01 -4.81868982e-01 1.03387499e+00 1.42201638e+00
1.32222748e+00 -5.85264921e-01 1.31754354e-01 9.75465357e-01
8.22352946e-01 -8.50630999e-01 -6.25744164e-02 -6.42434180e-01
4.62755978e-01 1.89294600e+00 5.57424664e-01 -4.64459538e-01
1.29848361e+00 1.97106108e-01 8.68944466e-01 -3.50959629e-01
4.11207944e-01 -2.81953841e-01 1.86948508e-01 1.17575252e+00
3.36472720e-01 4.44777906e-01 -4.28946555e-01 8.43417108e-01
-7.68699527e-01 7.96410814e-02 1.66418463e-01 9.07395065e-01
-7.52666831e-01 -9.59670186e-01 -2.98110843e-01 6.47653520e-01
4.20095831e-01 2.94547677e-01 -2.52533164e-02 8.58891785e-01
2.57478297e-01 1.03999805e+00 -1.21491238e-01 -1.26644635e+00
3.17922711e-01 1.59423202e-01 8.70811224e-01 -5.09418845e-01
-3.37609291e-01 -2.09938839e-01 -1.80377230e-01 -7.47714221e-01
-1.55139029e-01 -5.61866611e-02 -1.86732221e+00 9.17063430e-02
-2.05023199e-01 -8.24429393e-02 1.06833816e+00 1.32588089e+00
4.69335504e-02 8.15616250e-01 6.19377375e-01 -8.58878851e-01
-4.86512184e-01 -1.05071986e+00 -1.07150245e+00 -1.72948286e-01
3.14175814e-01 -4.70821202e-01 -1.56238452e-01 -4.70932752e-01] | [6.738157272338867, 0.3492282032966614] |
160da7d1-e881-47a8-b819-4b2dcff1fd66 | image-harmonization-with-transformer | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Guo_Image_Harmonization_With_Transformer_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Guo_Image_Harmonization_With_Transformer_ICCV_2021_paper.pdf | Image Harmonization With Transformer | Image harmonization, aiming to make composite images look more realistic, is an important and challenging task. The composite, synthesized by combining foreground from one image with background from another image, inevitably suffers from the issue of inharmonious appearance caused by distinct imaging conditions, i.e., lights. Current solutions mainly adopt an encoder-decoder architecture with convolutional neural network (CNN) to capture the context of composite images, trying to understand what it looks like in the surrounding background near the foreground. In this work, we seek to solve image harmonization with Transformer, by leveraging its powerful ability of modeling long-range context dependencies, for adjusting foreground light to make it compatible with background light while keeping structure and semantics unchanged. We present the design of our harmonization Transformer frameworks without and with disentanglement, as well as comprehensive experiments and ablation study, demonstrating the power of Transformer and investigating the Transformer for vision. Our method achieves state-of-the-art performance on both image harmonization and image inpainting/enhancement, indicating its superiority. Our code and models are available at https://github.com/zhenglab/HarmonyTransformer. | ['Junyu Dong', 'Bing Zheng', 'Zhaorui Gu', 'Haiyong Zheng', 'Dongsheng Guo', 'Zonghui Guo'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['image-harmonization'] | ['computer-vision'] | [ 5.30899107e-01 -3.02757978e-01 1.91363215e-01 -1.43685535e-01
-4.46908802e-01 -3.46765578e-01 3.71617615e-01 -4.59563196e-01
-1.21524766e-01 4.83808905e-01 2.96104431e-01 -4.59409431e-02
2.78675228e-01 -5.97265661e-01 -9.52743948e-01 -9.11145389e-01
6.21768534e-01 -3.63761425e-01 1.50762200e-01 -3.66214901e-01
3.69187444e-02 2.25300521e-01 -1.59986985e+00 4.81105298e-01
1.01966894e+00 9.66052115e-01 5.83330989e-01 5.56314886e-01
1.86659079e-02 9.84835625e-01 -6.88450575e-01 -7.56812453e-01
4.11231637e-01 -5.79367816e-01 -3.30392182e-01 3.90322894e-01
8.86690795e-01 -4.90398079e-01 -5.63521743e-01 1.26807356e+00
6.17144406e-01 -2.81314343e-01 1.02726258e-01 -1.25166810e+00
-1.32602072e+00 4.38374162e-01 -8.20330262e-01 1.25769153e-01
7.27566034e-02 5.75745642e-01 7.52477825e-01 -8.09263885e-01
4.48226810e-01 1.12636936e+00 5.49383044e-01 5.79236150e-01
-1.27385557e+00 -8.46088350e-01 7.07265884e-02 4.06112731e-01
-1.23385608e+00 -5.62739313e-01 9.50318515e-01 -8.63527134e-02
4.97447252e-01 4.88295346e-01 8.23930025e-01 1.29023993e+00
5.29103816e-01 8.18817377e-01 1.18201971e+00 -2.77283221e-01
-2.26324171e-01 -1.27089061e-02 -4.40017492e-01 4.30891156e-01
1.88443661e-01 2.38186166e-01 -5.98466933e-01 3.56900394e-01
7.02441990e-01 8.32464173e-02 -6.86678648e-01 -1.41593277e-01
-1.12884343e+00 3.19664806e-01 6.10490620e-01 1.73452780e-01
-2.38988832e-01 2.10845336e-01 6.58806190e-02 6.70515224e-02
2.72798359e-01 4.02393222e-01 -2.37600580e-01 2.65007198e-01
-8.33623886e-01 1.99154422e-01 1.88777745e-01 1.04700923e+00
5.67299485e-01 2.97533691e-01 -5.01102149e-01 7.79858351e-01
1.11276604e-01 5.94600260e-01 2.68617690e-01 -1.03803480e+00
2.92327881e-01 4.66550589e-01 4.44423407e-02 -9.38416541e-01
-1.21335052e-01 -6.40695751e-01 -1.15470064e+00 3.45567912e-01
8.46581906e-02 -5.72485849e-02 -8.14712703e-01 1.87824333e+00
3.46856922e-01 4.67522860e-01 1.82817932e-02 1.06137061e+00
1.00456405e+00 6.58465922e-01 2.92476788e-02 -2.59110630e-01
1.42717063e+00 -1.21190894e+00 -9.91432369e-01 -3.17428589e-01
-1.80601075e-01 -1.16613603e+00 1.20878494e+00 4.15567666e-01
-1.39922917e+00 -9.37022746e-01 -9.94473994e-01 -5.06305873e-01
-1.32129155e-02 1.75790608e-01 4.56681043e-01 3.92637044e-01
-1.25430882e+00 3.75838578e-01 -4.71627265e-01 6.33311085e-03
4.34362113e-01 -8.44017137e-03 -2.59089023e-01 -8.78161341e-02
-1.24582326e+00 8.86216044e-01 2.41675556e-01 3.96578938e-01
-7.71353722e-01 -8.36933851e-01 -8.23087454e-01 5.71783222e-02
3.89099538e-01 -9.97962356e-01 1.15659523e+00 -1.39721704e+00
-1.59387791e+00 8.09002459e-01 -1.34856820e-01 -3.29513222e-01
5.44332564e-01 -3.38100970e-01 -5.73758543e-01 1.96613260e-02
2.63396539e-02 6.55627489e-01 1.17986059e+00 -1.50034654e+00
-5.66440105e-01 -2.11362705e-01 -2.51753293e-02 2.95372993e-01
-1.92487374e-01 5.87351397e-02 -9.38961923e-01 -1.00093043e+00
-1.85141340e-01 -8.19076598e-01 -2.80047003e-02 2.69954950e-01
-3.96858156e-01 3.66043031e-01 1.00300217e+00 -9.71744478e-01
1.28562176e+00 -2.15045524e+00 1.66313395e-01 -3.45210999e-01
4.34888095e-01 5.44562697e-01 -4.30584818e-01 1.99839324e-01
-2.36716539e-01 -1.51575357e-01 -3.54659587e-01 -3.11664522e-01
-1.36726961e-01 1.26739889e-01 -4.36163932e-01 3.93470347e-01
5.24683833e-01 1.12701619e+00 -7.66143739e-01 -4.76036400e-01
4.25243467e-01 7.58504152e-01 -4.74403471e-01 4.00691152e-01
-1.58244401e-01 6.81969166e-01 -4.29480895e-02 7.70960748e-01
1.06276786e+00 -1.76502466e-01 9.97552648e-02 -8.65426660e-01
-1.59479216e-01 -2.83386350e-01 -9.76003945e-01 1.65086567e+00
-5.67234814e-01 8.55536819e-01 2.48696581e-01 -5.76518953e-01
6.92787349e-01 6.23907819e-02 3.53648007e-01 -1.16456485e+00
1.73809424e-01 -1.09677138e-02 -5.72023764e-02 -5.23387134e-01
5.12187421e-01 7.17047825e-02 2.65730739e-01 1.02075726e-01
-2.64644146e-01 -2.87375748e-01 5.28166490e-03 -5.29071465e-02
6.12524331e-01 2.18734652e-01 1.40236944e-01 9.37910378e-02
4.48480010e-01 -4.76891756e-01 6.28853321e-01 4.68753576e-01
-2.23336205e-01 9.73133624e-01 3.19315106e-01 -3.58079195e-01
-1.04663754e+00 -1.29036784e+00 -4.38372642e-02 7.65469968e-01
6.77662611e-01 -3.06130767e-01 -8.12832654e-01 -4.36008871e-02
-3.12261999e-01 7.80181885e-01 -5.98842919e-01 -4.01498258e-01
-5.81549883e-01 -7.04880059e-01 4.01049465e-01 2.14209184e-01
1.08336401e+00 -8.61936152e-01 -6.85871542e-01 6.36088327e-02
-5.94101012e-01 -1.15625358e+00 -9.00984883e-01 -1.24714918e-01
-3.38789374e-01 -1.04973626e+00 -7.00367808e-01 -7.53181994e-01
4.51471627e-01 7.53523707e-01 1.32390940e+00 1.10587068e-01
-5.12518287e-01 1.34164333e-01 -2.22730488e-01 -3.50359797e-01
-3.36964190e-01 -4.80423689e-01 -3.69915307e-01 3.51540267e-01
-1.63775325e-01 -6.95017695e-01 -1.08402061e+00 3.06880772e-01
-1.51216185e+00 6.37419522e-01 8.29141438e-01 9.26875651e-01
5.66163361e-01 2.32810870e-01 -9.38936695e-03 -6.03207767e-01
5.09390116e-01 -1.95565820e-01 -5.08556724e-01 5.43600559e-01
-3.58679473e-01 -2.65192211e-01 5.08540273e-01 -4.80307043e-01
-1.42125225e+00 -1.74253866e-01 -8.90209228e-02 -7.50182450e-01
1.22858487e-01 -7.61335045e-02 -6.45889401e-01 -1.93323642e-01
4.43221748e-01 4.91279691e-01 -2.30049416e-02 -2.30768561e-01
6.45250142e-01 5.27872086e-01 8.92932355e-01 -3.36338192e-01
7.04463661e-01 6.76539779e-01 -1.19087316e-01 -5.77724457e-01
-9.43173051e-01 5.25905825e-02 -3.74974906e-01 -4.37231541e-01
9.94211733e-01 -1.09685683e+00 -6.04812920e-01 8.06407094e-01
-1.27438235e+00 -4.86562252e-01 -2.00971708e-01 4.80039120e-02
-3.02639097e-01 4.59543556e-01 -6.39722228e-01 -2.84837425e-01
-3.85675073e-01 -1.29036212e+00 1.16505587e+00 5.27001917e-01
1.38403922e-01 -6.59037769e-01 -1.89632684e-01 5.57214439e-01
7.37279415e-01 3.11886311e-01 6.70860887e-01 2.93437839e-01
-7.97218680e-01 2.65001953e-01 -4.95135009e-01 5.58679223e-01
3.90141159e-01 3.30994934e-01 -1.08372009e+00 -2.35261217e-01
2.48258129e-01 2.67396346e-02 8.81151378e-01 4.84636724e-01
1.38705480e+00 -4.41511303e-01 6.71966895e-02 9.80545223e-01
1.43490160e+00 1.46743491e-01 1.11714780e+00 2.42938891e-01
8.17570746e-01 4.82131124e-01 5.15929878e-01 4.24445152e-01
3.16966027e-01 9.01892126e-01 6.38061106e-01 -7.44988143e-01
-8.14746141e-01 -2.04324245e-01 3.03494811e-01 4.73194867e-01
1.80971503e-01 -4.67051148e-01 -5.57862222e-01 2.11191282e-01
-1.73743045e+00 -1.02582014e+00 -5.30908480e-02 1.86620486e+00
1.09503102e+00 -1.45562470e-01 -3.47395569e-01 -3.72887105e-01
8.53644311e-01 3.55866909e-01 -7.05895483e-01 -8.45066458e-02
-6.25865579e-01 2.11757749e-01 3.68809789e-01 4.27020341e-01
-1.10826898e+00 9.64080989e-01 5.56694174e+00 9.70449686e-01
-1.52378595e+00 1.38254970e-01 1.05129230e+00 -2.86254466e-01
-3.45077097e-01 -6.42691925e-02 -3.59902352e-01 7.76540577e-01
3.84838521e-01 -6.81089982e-02 5.69288015e-01 3.95754069e-01
4.02848512e-01 -6.87876269e-02 -7.32085407e-01 1.19007599e+00
2.67182976e-01 -1.41113687e+00 2.24482745e-01 -4.24252689e-01
8.98564875e-01 -1.46085605e-01 5.96761048e-01 -7.84068406e-02
5.50847761e-02 -9.27408755e-01 1.09222949e+00 5.30139923e-01
9.52027261e-01 -4.10767883e-01 4.27666336e-01 -1.61879048e-01
-1.18545508e+00 3.15428264e-02 -1.22401327e-01 1.82993442e-01
9.70494896e-02 5.55438042e-01 -1.97988361e-01 5.87978065e-01
9.57269728e-01 9.15547013e-01 -8.10185492e-01 7.91108251e-01
-4.20358926e-01 2.03237057e-01 1.44932672e-01 5.62146425e-01
-1.89920485e-01 -4.09284115e-01 4.69336092e-01 1.17404401e+00
2.72283643e-01 5.87023906e-02 -6.94362074e-02 1.28482115e+00
2.37415428e-03 -2.11773023e-01 -4.44701880e-01 2.57646173e-01
3.25685233e-01 1.29854620e+00 -3.84197474e-01 -2.55850911e-01
-5.06958067e-01 1.30047023e+00 3.95525396e-02 5.44326067e-01
-1.32309175e+00 -2.22513065e-01 8.51484358e-01 -1.28191281e-02
3.20236713e-01 -5.46368808e-02 -3.72076839e-01 -1.31263888e+00
3.11033726e-01 -1.10903919e+00 -2.29991227e-03 -1.38944983e+00
-1.31075478e+00 6.71916962e-01 -1.09715022e-01 -1.48828709e+00
3.47963303e-01 -4.57588673e-01 -7.40956247e-01 8.01185668e-01
-1.71291208e+00 -1.49169564e+00 -6.86994553e-01 8.07655036e-01
5.70001423e-01 2.74818897e-01 2.41621897e-01 5.97404063e-01
-7.99531817e-01 5.96249640e-01 1.88409742e-02 -1.10432170e-01
1.03889537e+00 -9.42705750e-01 3.07204306e-01 1.23804998e+00
-2.76255552e-02 4.70540375e-01 8.68324697e-01 -5.24284065e-01
-1.54557133e+00 -1.34521484e+00 4.40115720e-01 -2.05435306e-01
4.90529269e-01 -2.35861108e-01 -9.43198979e-01 4.02236074e-01
7.10521579e-01 4.89809141e-02 3.71286213e-01 -5.11811137e-01
-4.88243490e-01 -4.37974960e-01 -9.29591238e-01 1.18370330e+00
1.07351553e+00 -4.77688015e-01 -2.30651021e-01 2.13154346e-01
9.65144098e-01 -6.28969848e-01 -5.95207691e-01 4.10897940e-01
5.01099348e-01 -1.45907760e+00 1.15362847e+00 -4.69684154e-02
7.86692500e-01 -6.84883296e-01 -1.44087136e-01 -1.20303082e+00
-4.23424661e-01 -9.00659919e-01 7.40193352e-02 1.33542848e+00
4.46392931e-02 -5.81289351e-01 2.81372011e-01 6.09109938e-01
-1.83796749e-01 -6.54248595e-01 -6.04963362e-01 -4.19265360e-01
-2.44613469e-01 -1.64435282e-01 7.18511283e-01 8.51295114e-01
-7.54742622e-01 3.11878234e-01 -7.03767002e-01 3.31598967e-01
5.80208063e-01 3.97936285e-01 9.18584764e-01 -4.31303710e-01
-3.53197128e-01 -4.50629532e-01 -4.25243601e-02 -8.25961828e-01
7.63491914e-03 -4.98390853e-01 -6.73613185e-03 -1.39685845e+00
3.00222635e-01 -3.62328477e-02 -2.45079949e-01 4.70624387e-01
-4.67858762e-01 6.24284625e-01 5.68798304e-01 1.38286099e-01
-5.96536815e-01 8.49217176e-01 1.74861646e+00 -3.67580682e-01
3.58737744e-02 -3.62150937e-01 -1.00104785e+00 5.94934225e-01
8.22223604e-01 -7.68694207e-02 -4.82448816e-01 -8.93084109e-01
4.58483621e-02 -4.99311369e-03 7.31787860e-01 -8.78036201e-01
1.51776910e-01 -2.44425163e-01 6.51686311e-01 -3.31462443e-01
3.90302002e-01 -8.90140474e-01 7.12821424e-01 4.25266653e-01
-2.56317407e-01 2.18709648e-01 4.41628814e-01 4.79705781e-01
-5.38384199e-01 2.01310292e-01 1.07641625e+00 -1.66803636e-02
-8.33092034e-01 4.41764295e-01 9.31339040e-02 9.62859467e-02
1.01851976e+00 -2.51955420e-01 -5.99291563e-01 -4.26237255e-01
-3.58904243e-01 4.72964682e-02 7.74606645e-01 5.08308470e-01
8.55683029e-01 -1.27464902e+00 -8.34497154e-01 2.98690498e-01
2.03996561e-02 -1.75627857e-01 8.25905979e-01 8.91882181e-01
-6.71427190e-01 -9.83260721e-02 -4.96292442e-01 -4.75712836e-01
-1.36625028e+00 7.16706276e-01 5.32593191e-01 -6.59470260e-02
-6.66007817e-01 8.04499924e-01 8.70989621e-01 8.54374319e-02
1.22223638e-01 -3.85709494e-01 1.13629125e-01 -2.47494340e-01
5.86759567e-01 9.93577950e-03 -1.00442402e-01 -4.25639033e-01
-9.76908021e-03 6.52029634e-01 -4.54804637e-02 2.96495587e-01
1.06448889e+00 -4.86720651e-01 -3.22163165e-01 -2.93424483e-02
1.02440119e+00 6.92024231e-02 -1.69250464e+00 -3.38263541e-01
-5.40677607e-01 -7.89438307e-01 3.73670757e-02 -8.61613333e-01
-1.54647315e+00 7.99501956e-01 7.58324444e-01 -3.45938988e-02
1.82028699e+00 -2.14156523e-01 9.49915528e-01 -1.25329599e-01
-2.54284479e-02 -8.63509297e-01 3.31377417e-01 1.45190820e-01
1.23424983e+00 -1.30603862e+00 -8.60580616e-03 -4.50329095e-01
-8.13749492e-01 1.05238378e+00 7.87970185e-01 2.46392209e-02
4.27665293e-01 3.70799035e-01 3.05750042e-01 -2.01722234e-02
-6.97879195e-01 -1.66530848e-01 3.57482851e-01 6.05256736e-01
2.99999475e-01 1.65187139e-02 1.07634686e-01 3.58539730e-01
-1.24624610e-01 -1.92057341e-01 3.98014665e-01 5.30199468e-01
-1.36040822e-01 -9.47230279e-01 -4.70494360e-01 -2.13193684e-03
-3.85407209e-01 -4.12422597e-01 -3.10289830e-01 7.23960102e-01
5.65259457e-01 1.01584470e+00 7.32872039e-02 -3.11575681e-01
3.42660695e-01 -6.49074793e-01 6.28482282e-01 -1.98691592e-01
-4.85269308e-01 2.30979234e-01 -2.52663106e-01 -8.69539440e-01
-5.99053919e-01 -3.09024870e-01 -7.49994636e-01 -3.00921261e-01
-2.54566502e-02 -4.90510404e-01 3.60744387e-01 4.93896097e-01
5.25537074e-01 8.51130664e-01 6.75886750e-01 -9.73678291e-01
-2.15720549e-01 -6.41815186e-01 -5.96279502e-01 6.14765346e-01
4.92954731e-01 -4.40496683e-01 -1.60256103e-01 3.85520965e-01] | [11.223129272460938, -1.3598802089691162] |
6ad3ea78-33b1-4e30-9ff2-407aab92b8e1 | leveraging-audio-tagging-assisted-sound-event | 2304.12688 | null | https://arxiv.org/abs/2304.12688v1 | https://arxiv.org/pdf/2304.12688v1.pdf | Leveraging Audio-Tagging Assisted Sound Event Detection using Weakified Strong Labels and Frequency Dynamic Convolutions | Jointly learning from a small labeled set and a larger unlabeled set is an active research topic under semi-supervised learning (SSL). In this paper, we propose a novel SSL method based on a two-stage framework for leveraging a large unlabeled in-domain set. Stage-1 of our proposed framework focuses on audio-tagging (AT), which assists the sound event detection (SED) system in Stage-2. The AT system is trained utilizing a strongly labeled set converted into weak predictions referred to as weakified set, a weakly labeled set, and an unlabeled set. This AT system then infers on the unlabeled set to generate reliable pseudo-weak labels, which are used with the strongly and weakly labeled set to train a frequency dynamic convolutional recurrent neural network-based SED system at Stage-2 in a supervised manner. Our system outperforms the baseline by 45.5% in terms of polyphonic sound detection score on the DESED real validation set. | ['Eng Siong Chng', 'Andrew Koh', 'Rohan Kumar Das', 'Tanmay Khandelwal'] | 2023-04-25 | null | null | null | null | ['audio-tagging', 'sound-event-detection'] | ['audio', 'audio'] | [ 4.31909800e-01 2.09044009e-01 -2.03605592e-02 -4.33866858e-01
-1.51827967e+00 -5.72493494e-01 5.30069619e-02 -2.07969978e-01
-4.00835663e-01 5.19392252e-01 2.33115956e-01 -1.28217295e-01
6.58603072e-01 -5.11975288e-01 -5.60822964e-01 -5.76849520e-01
1.39019117e-01 1.73339516e-01 4.67358291e-01 2.98421681e-01
-2.95871079e-01 6.87340274e-03 -1.55861843e+00 4.94941980e-01
4.49833989e-01 1.06964827e+00 3.23668242e-01 7.91125596e-01
1.17020167e-01 9.08696175e-01 -4.49816555e-01 4.02349621e-01
8.98548663e-02 -8.26443434e-01 -6.19006634e-01 2.41181254e-02
1.97810069e-01 -2.97473997e-01 -9.03374627e-02 8.96574259e-01
8.21574688e-01 2.55780339e-01 3.22780520e-01 -8.35737288e-01
-1.16214946e-01 1.25377548e+00 -2.83827722e-01 2.65639454e-01
1.57140389e-01 -6.75838366e-02 9.91560578e-01 -1.36519504e+00
8.86335969e-02 1.13734269e+00 6.97593153e-01 6.26285434e-01
-8.77258658e-01 -1.05126250e+00 9.32827741e-02 -1.54446185e-01
-1.30552745e+00 -7.13281035e-01 1.22797298e+00 -3.06452602e-01
7.68189847e-01 -8.59585479e-02 2.94170946e-01 9.11406219e-01
-3.48011792e-01 7.34888792e-01 1.01142669e+00 -7.35017121e-01
5.09640932e-01 3.15945983e-01 4.31470066e-01 5.52371502e-01
-4.31636184e-01 3.11904490e-01 -7.49898255e-01 -1.56792030e-01
4.13245529e-01 -2.67196953e-01 -1.20484821e-01 4.76083934e-01
-8.00449550e-01 5.91262937e-01 1.74764261e-01 3.20379615e-01
-4.07048732e-01 -3.05982605e-02 4.71266627e-01 3.47355008e-01
9.15225685e-01 1.30908921e-01 -5.26571572e-01 -1.10495523e-01
-1.24413323e+00 -1.22675076e-01 7.30106652e-01 5.27186930e-01
6.60909116e-01 4.88058150e-01 -1.99233502e-01 1.20481384e+00
6.18801773e-01 6.18428409e-01 6.72804236e-01 -6.49547100e-01
5.05768418e-01 3.22790086e-01 -3.30360904e-02 -2.11473942e-01
-3.59027892e-01 -8.54897141e-01 -3.62811208e-01 -1.05356812e-01
-1.53311193e-01 -6.90877914e-01 -1.04568386e+00 2.00622535e+00
4.68222022e-01 8.20528030e-01 1.54054388e-01 8.26254189e-01
9.62353766e-01 1.01065981e+00 1.67898491e-01 -4.32499498e-01
9.46919322e-01 -1.30200362e+00 -6.67166114e-01 -3.41927499e-01
5.93456149e-01 -8.97411644e-01 8.23086441e-01 3.22778732e-01
-1.13841510e+00 -1.08534324e+00 -9.83489513e-01 1.83232665e-01
-3.90300788e-02 6.77644372e-01 -9.71701741e-02 6.03398323e-01
-1.07686448e+00 4.02127594e-01 -8.92830908e-01 -1.55525163e-01
3.02684277e-01 2.98116446e-01 9.59571302e-02 1.23933926e-01
-1.37761986e+00 3.46753389e-01 5.67419469e-01 2.15508312e-01
-1.46187019e+00 -5.48956573e-01 -9.27100062e-01 1.28396854e-01
2.57227898e-01 -2.00063409e-03 1.77383852e+00 -8.67834508e-01
-1.66052186e+00 7.55111158e-01 -1.90426394e-01 -7.80039012e-01
1.74889266e-01 -5.01677155e-01 -7.82717168e-01 1.27294242e-01
3.37368906e-01 6.17766619e-01 1.10619950e+00 -1.10025132e+00
-8.46615553e-01 -1.34724826e-01 -3.86670113e-01 2.63444006e-01
-2.84203678e-01 1.55127302e-01 -3.71105075e-01 -8.33231270e-01
1.63242146e-01 -1.12882125e+00 1.88583992e-02 -5.53498447e-01
-5.28928876e-01 -3.72974902e-01 1.08402181e+00 -6.70875311e-01
1.45154667e+00 -2.46901584e+00 -3.50787759e-01 -2.41679717e-02
-1.18869990e-01 7.01826751e-01 -2.43749216e-01 2.99968809e-01
-5.33012211e-01 -2.97750682e-01 -7.42781162e-02 -7.00594842e-01
-2.10806340e-01 -2.12071583e-01 -7.54435658e-01 -2.22268328e-02
5.96382856e-01 3.87356818e-01 -1.20084321e+00 -3.83405328e-01
1.46915287e-01 2.41161391e-01 -5.15213609e-01 7.21121371e-01
-1.79995313e-01 4.45406705e-01 -4.00423348e-01 6.39733315e-01
5.47137082e-01 6.11046217e-02 6.08108239e-04 -2.04834625e-01
-2.12890476e-01 5.96248806e-01 -1.18911409e+00 1.63281131e+00
-6.23299718e-01 4.38636273e-01 -7.89284110e-02 -8.64495158e-01
9.48239386e-01 8.86244476e-01 3.24527353e-01 -1.77301139e-01
8.58015567e-02 3.07825983e-01 -6.51435107e-02 -3.50765407e-01
1.39317408e-01 -2.87558556e-01 -1.68384284e-01 8.42127025e-01
6.38683140e-01 -1.10870287e-01 -8.69362876e-02 5.39124124e-02
1.16899538e+00 5.08247137e-01 -1.02651045e-02 1.43843353e-01
6.13230348e-01 -3.63612175e-01 7.99668312e-01 6.73557103e-01
-3.88273925e-01 6.98754668e-01 2.08320562e-02 -7.74039999e-02
-7.68314242e-01 -9.78396475e-01 -1.30175680e-01 1.19918156e+00
-1.60977095e-01 -4.53819573e-01 -7.92607605e-01 -9.92532074e-01
-2.84455508e-01 6.75424516e-01 -3.80073279e-01 -3.02495807e-01
-4.74802583e-01 -4.29380625e-01 8.19659591e-01 6.60491168e-01
5.67661345e-01 -1.37125361e+00 -3.82828236e-01 5.47746778e-01
-1.97222739e-01 -1.04979050e+00 -5.16640007e-01 7.07327485e-01
-5.40093243e-01 -5.61841369e-01 -6.38572991e-01 -9.86090481e-01
4.13722992e-01 2.18198866e-01 7.54371107e-01 -4.46567386e-01
1.65705830e-01 2.05324753e-03 -4.68482733e-01 -7.02982128e-01
-7.24298477e-01 5.87283708e-02 3.08450848e-01 4.52033490e-01
3.92031431e-01 -6.14537477e-01 -3.69367123e-01 4.19652492e-01
-5.74790120e-01 -4.33544144e-02 5.81090689e-01 8.60916317e-01
7.27421403e-01 1.15184724e-01 1.17607236e+00 -1.01622272e+00
2.27624342e-01 -5.40082335e-01 -3.90082479e-01 -8.67860392e-03
-3.11946362e-01 -1.90544967e-02 7.20457673e-01 -6.71054900e-01
-1.43135858e+00 5.05173743e-01 -4.16767091e-01 -6.83081448e-01
-2.26483464e-01 4.61868286e-01 -1.62850037e-01 3.11954856e-01
7.71656096e-01 7.02277105e-03 -4.88194168e-01 -6.79886043e-01
1.71270087e-01 1.40414858e+00 7.18265653e-01 -2.69207358e-01
8.23854864e-01 1.73842072e-01 -6.24057710e-01 -6.79597318e-01
-1.63327503e+00 -7.38329649e-01 -6.43172860e-01 -3.31484705e-01
6.19748652e-01 -1.52887797e+00 3.47447582e-02 6.37368619e-01
-9.29952621e-01 -5.56574285e-01 -5.66235602e-01 8.12660575e-01
-2.53825992e-01 -6.05452843e-02 -6.58765912e-01 -1.23673368e+00
-4.85816419e-01 -7.88145721e-01 1.39208567e+00 3.77356201e-01
-2.61857927e-01 -6.24551952e-01 5.17275155e-01 4.28955227e-01
1.93612382e-01 -7.06036687e-02 3.09735954e-01 -1.08390975e+00
4.42855526e-03 -2.65246898e-01 1.51744187e-01 8.46724391e-01
2.30502427e-01 -2.73958951e-01 -1.80676079e+00 -1.96331605e-01
2.09703565e-01 -7.45034635e-01 1.10942006e+00 2.98004389e-01
9.70319033e-01 1.29692540e-01 -1.84555471e-01 2.80011505e-01
9.59781528e-01 3.44621152e-01 -8.62280503e-02 -3.82511616e-01
5.74706852e-01 2.61909753e-01 8.56369495e-01 4.26802665e-01
9.30449814e-02 3.08499217e-01 -1.06155379e-02 -2.21776187e-01
-5.30519545e-01 -6.03318036e-01 7.91824162e-01 1.43072987e+00
3.32659334e-01 -1.59563616e-01 -7.06585109e-01 6.01138771e-01
-1.78986073e+00 -8.57317090e-01 2.66950130e-01 2.12934709e+00
1.20081830e+00 4.46772099e-01 1.50913507e-01 5.30383587e-01
1.08909416e+00 1.31416366e-01 -6.05379283e-01 4.64951247e-02
2.54376084e-01 4.32472765e-01 -1.22430930e-02 3.02201986e-01
-1.44355416e+00 9.85203505e-01 5.35179663e+00 8.51154029e-01
-1.32070768e+00 3.52385938e-01 4.81286168e-01 -6.06308170e-02
9.02846828e-02 -3.37089673e-02 -1.05551505e+00 5.75582743e-01
1.59366250e+00 1.50471658e-01 1.79239288e-01 9.39214528e-01
5.35733283e-01 -5.38322292e-02 -1.01876509e+00 8.57992470e-01
1.04837880e-01 -1.02025616e+00 -1.67993173e-01 -4.13437188e-01
7.51112342e-01 2.98970997e-01 -1.62516888e-02 8.13695252e-01
1.83017343e-01 -4.16102827e-01 8.84590864e-01 1.89840332e-01
1.03648758e+00 -7.22729862e-01 6.93508565e-01 4.93695706e-01
-1.55582285e+00 -1.04334198e-01 -1.62422750e-02 1.42376661e-01
2.32833251e-01 6.88409030e-01 -1.04812610e+00 3.10400814e-01
5.51418185e-01 9.10102665e-01 -1.73574030e-01 8.37629616e-01
-4.69299406e-01 1.71721828e+00 -4.10004884e-01 4.04020876e-01
1.57798424e-01 2.38541067e-01 3.49359483e-01 1.26795459e+00
7.27368072e-02 7.90657550e-02 5.44348419e-01 6.91468239e-01
-3.39615971e-01 -1.14187509e-01 -2.29185775e-01 -2.60198563e-01
8.02475512e-01 1.28786623e+00 -7.32822299e-01 -4.29122567e-01
-4.75300580e-01 8.38643610e-01 2.20158204e-01 2.90874541e-01
-8.89491558e-01 -4.89451557e-01 -1.55920118e-01 -3.46839815e-01
1.66131184e-01 2.87162066e-01 2.36103199e-02 -1.13925838e+00
-3.21973681e-01 -4.86564636e-01 4.85543460e-01 -9.43983078e-01
-1.06016922e+00 8.63526464e-01 -2.68878967e-01 -1.49125826e+00
-4.06567246e-01 2.65628826e-02 -9.07758176e-01 9.91091669e-01
-1.61456728e+00 -1.22564471e+00 -1.64519563e-01 4.85675067e-01
1.09697926e+00 -2.16673926e-01 9.06036079e-01 5.35834730e-01
-8.32858682e-01 4.58818108e-01 -1.99305892e-01 2.19355866e-01
9.48233008e-01 -1.24293149e+00 3.89185756e-01 1.03215277e+00
4.72699106e-01 7.52817020e-02 3.81503701e-01 -6.79336727e-01
-6.38634562e-01 -1.64540327e+00 1.03262579e+00 -1.07917458e-01
5.46901584e-01 -3.83613825e-01 -8.91649365e-01 5.95488131e-01
-1.99558973e-01 3.10115665e-01 8.72050285e-01 -1.69936478e-01
-2.25596800e-01 -1.29562274e-01 -7.58027434e-01 1.16639370e-02
6.95571721e-01 -9.84769821e-01 -7.96954691e-01 3.43337327e-01
1.01909721e+00 -4.18867350e-01 -5.33904254e-01 4.57980841e-01
2.92771101e-01 -5.19282222e-01 5.85201144e-01 -2.49608472e-01
1.83539540e-01 -3.41582775e-01 -1.05079956e-01 -1.35733986e+00
1.86883267e-02 -8.81729901e-01 -1.40824482e-01 1.46483970e+00
5.21044970e-01 -2.64493108e-01 6.14798367e-01 7.07981456e-03
-4.42648917e-01 -4.49938774e-01 -6.92200959e-01 -6.38267457e-01
-4.93547529e-01 -7.91927457e-01 2.47036919e-01 7.48228908e-01
6.56799525e-02 7.79580534e-01 -3.52262586e-01 3.97272080e-01
4.05885786e-01 2.62767047e-01 2.98607528e-01 -1.27219594e+00
-3.97202432e-01 2.97357917e-01 1.05656020e-01 -8.86261523e-01
2.87780851e-01 -8.05292070e-01 5.25500059e-01 -9.42936897e-01
1.38431964e-02 -5.48058569e-01 -8.04249167e-01 7.67355859e-01
-2.94359565e-01 5.22040069e-01 -1.19806685e-01 3.50770384e-01
-6.84488714e-01 6.01424158e-01 5.42733073e-01 1.84690490e-01
-5.46947300e-01 4.75491375e-01 -2.91249067e-01 9.58793581e-01
7.53696501e-01 -6.80679142e-01 -6.05340183e-01 -3.51395048e-02
-2.99884439e-01 3.59714776e-01 1.05795003e-01 -1.28374803e+00
-6.65553510e-02 3.37357640e-01 1.35746270e-01 -8.66101801e-01
3.47838521e-01 -4.29290414e-01 -2.77610004e-01 3.33278000e-01
-7.93521404e-01 -6.28794730e-01 1.95584178e-01 6.31074369e-01
-4.10368949e-01 -2.05098391e-01 8.79817784e-01 1.17765173e-01
-5.43668032e-01 1.02791391e-01 -2.18883857e-01 2.17450961e-01
7.51602948e-01 2.72348434e-01 1.41714931e-01 -3.50804418e-01
-1.11287713e+00 1.29309863e-01 -3.45180035e-01 4.01615024e-01
3.23368490e-01 -1.34544849e+00 -8.00890625e-01 3.55366945e-01
-8.20013788e-03 6.09407492e-04 4.00837302e-01 4.12991524e-01
1.37638152e-01 3.46857935e-01 2.78546154e-01 -5.83340108e-01
-1.13284159e+00 2.43514091e-01 2.31157884e-01 -2.63790905e-01
-3.13618541e-01 1.20131218e+00 7.03999475e-02 -3.22462022e-01
6.49494827e-01 -5.12822390e-01 -3.96737903e-01 6.17459137e-03
4.95953023e-01 1.38481125e-01 1.74819097e-01 -7.73856819e-01
-2.49182552e-01 2.00668216e-01 8.23958814e-02 -4.59968328e-01
1.33424056e+00 -6.29378259e-02 3.69382679e-01 8.67858648e-01
1.15516424e+00 1.14401467e-01 -1.43254161e+00 -7.13236153e-01
-6.62002265e-02 1.70820102e-01 4.02314067e-01 -8.84952128e-01
-8.91837358e-01 1.03600919e+00 8.29563975e-01 9.34849232e-02
1.15937662e+00 1.51393011e-01 1.01214433e+00 1.59157574e-01
1.86974183e-02 -1.26261258e+00 1.82955652e-01 4.81360734e-01
6.06358290e-01 -1.30188680e+00 -5.36729157e-01 -1.69438377e-01
-6.59123063e-01 9.19425607e-01 6.88701153e-01 -2.62591302e-01
9.57363009e-01 4.24283564e-01 3.41821462e-01 9.59686860e-02
-9.66349900e-01 -4.52371925e-01 4.18915868e-01 2.92587787e-01
5.16852796e-01 -5.60391694e-02 1.79139256e-01 9.47495580e-01
-1.07174188e-01 2.01990038e-01 1.52293608e-01 9.41075385e-01
-5.95262229e-01 -1.07456827e+00 -2.57108539e-01 2.57832736e-01
-4.85689729e-01 -1.46390304e-01 -2.28196353e-01 1.44188955e-01
3.39633435e-01 1.29644620e+00 -4.10902221e-03 -6.84191823e-01
1.59568414e-01 4.85336691e-01 -1.13453448e-01 -1.26014614e+00
-8.47799003e-01 7.80610800e-01 2.89563954e-01 -2.51371533e-01
-5.68236768e-01 -6.00498140e-01 -1.32708538e+00 9.84113455e-01
-7.31165409e-01 5.00754595e-01 3.60489368e-01 1.18063629e+00
1.93972737e-01 6.81970596e-01 1.26383853e+00 -8.31649601e-01
-6.27747536e-01 -1.19496846e+00 -5.25131404e-01 8.73882994e-02
6.03474796e-01 -3.72195780e-01 -5.52619874e-01 3.92677784e-01] | [15.191133499145508, 5.150817394256592] |
2d6b48e5-293c-4ebe-952d-8183cda7c0f4 | from-multi-label-learning-to-cross-domain | 2207.11742 | null | https://arxiv.org/abs/2207.11742v1 | https://arxiv.org/pdf/2207.11742v1.pdf | From Multi-label Learning to Cross-Domain Transfer: A Model-Agnostic Approach | In multi-label learning, a particular case of multi-task learning where a single data point is associated with multiple target labels, it was widely assumed in the literature that, to obtain best accuracy, the dependence among the labels should be explicitly modeled. This premise led to a proliferation of methods offering techniques to learn and predict labels together, for example where the prediction for one label influences predictions for other labels. Even though it is now acknowledged that in many contexts a model of dependence is not required for optimal performance, such models continue to outperform independent models in some of those very contexts, suggesting alternative explanations for their performance beyond label dependence, which the literature is only recently beginning to unravel. Leveraging and extending recent discoveries, we turn the original premise of multi-label learning on its head, and approach the problem of joint-modeling specifically under the absence of any measurable dependence among task labels; for example, when task labels come from separate problem domains. We shift insights from this study towards building an approach for transfer learning that challenges the long-held assumption that transferability of tasks comes from measurements of similarity between the source and target domains or models. This allows us to design and test a method for transfer learning, which is model driven rather than purely data driven, and furthermore it is black box and model-agnostic (any base model class can be considered). We show that essentially we can create task-dependence based on source-model capacity. The results we obtain have important implications and provide clear directions for future work, both in the areas of multi-label and transfer learning. | ['Jesse Read'] | 2022-07-24 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 5.51962733e-01 4.07904498e-02 -4.71866190e-01 -6.91806674e-01
-8.57003093e-01 -8.71863008e-01 7.94910908e-01 2.73639619e-01
-4.03778166e-01 7.16037273e-01 -6.59446279e-03 -3.51333201e-01
-5.48743904e-01 -4.56520021e-01 -6.96485877e-01 -6.18925095e-01
1.13784231e-01 7.21442997e-01 1.52525276e-01 6.64029196e-02
1.67201236e-01 2.41651252e-01 -1.52364540e+00 3.95239770e-01
5.26113868e-01 7.28037179e-01 4.51765582e-03 6.04705215e-01
1.14663467e-01 7.83808887e-01 -3.65269810e-01 -3.66603673e-01
2.21872240e-01 -4.43130672e-01 -1.11458087e+00 1.63837299e-01
5.24999797e-01 1.58656925e-01 2.65777260e-01 8.61542702e-01
2.45789617e-01 -5.69512621e-02 1.10015321e+00 -1.53758144e+00
-7.25572169e-01 2.50153810e-01 -5.67660332e-01 -4.21950445e-02
-1.47845913e-02 -1.57687768e-01 1.35857832e+00 -4.96772200e-01
3.15072596e-01 1.21510589e+00 9.13475811e-01 4.77986813e-01
-1.68941188e+00 -8.43777835e-01 3.35731417e-01 5.76605089e-02
-1.01921916e+00 -3.54921609e-01 4.97837722e-01 -8.33646953e-01
6.10942483e-01 1.45455241e-01 5.44336326e-02 1.28296852e+00
2.89799303e-01 6.55515313e-01 1.78201866e+00 -8.71846795e-01
2.96810772e-02 5.48281968e-01 3.80563825e-01 2.41069078e-01
2.22405583e-01 2.54299700e-01 -4.18321520e-01 -2.07272097e-01
5.61941206e-01 -3.52794454e-02 6.21322915e-02 -6.52968884e-01
-1.20090663e+00 1.01551616e+00 5.58363507e-03 5.61201692e-01
2.78668515e-02 -4.77545522e-02 4.74665731e-01 7.21456826e-01
7.24796534e-01 5.14471591e-01 -8.57509971e-01 1.54633686e-01
-8.14003944e-01 2.31576294e-01 8.21682751e-01 7.48761296e-01
1.04899001e+00 -4.85693842e-01 -4.31607738e-02 9.69786644e-01
3.09688985e-01 1.17506549e-01 6.59680903e-01 -9.09064651e-01
2.53037632e-01 4.61638063e-01 1.58395812e-01 -5.61666787e-01
-6.90328121e-01 -4.45154369e-01 -4.83566046e-01 2.60288537e-01
8.18935990e-01 -1.49740815e-01 -5.76424837e-01 2.14245057e+00
1.96457550e-01 2.21311808e-01 -2.85444204e-02 5.08925319e-01
2.95141995e-01 1.72763437e-01 4.43702519e-01 -1.67577758e-01
1.23170543e+00 -8.34488511e-01 -3.90698671e-01 -5.51972985e-01
1.22078633e+00 -7.43841052e-01 1.09752774e+00 4.07344997e-01
-5.83244920e-01 -7.03659177e-01 -8.71533096e-01 3.39967087e-02
-4.18383449e-01 -2.09741637e-01 7.51031280e-01 7.34432518e-01
-9.63775456e-01 6.41742110e-01 -4.13213819e-01 -5.43210328e-01
1.89883247e-01 3.89671743e-01 -5.01529098e-01 -1.48726389e-01
-1.12438428e+00 1.36531067e+00 4.44457114e-01 -3.05698007e-01
-8.02219033e-01 -6.64954960e-01 -4.66332376e-01 -1.27891138e-01
3.84896457e-01 -4.82597262e-01 1.45021152e+00 -1.43632197e+00
-1.09728467e+00 1.14371812e+00 -4.86107953e-02 -1.14166833e-01
4.18531805e-01 1.61271449e-02 -4.01790440e-01 -3.60245854e-01
2.27486536e-01 4.83619690e-01 7.89407134e-01 -1.53879094e+00
-8.71150970e-01 -4.67779070e-01 2.13467166e-01 1.59725383e-01
-4.58762735e-01 3.00908864e-01 3.03611457e-01 -3.38628471e-01
-2.02016309e-01 -1.27026081e+00 -9.73679405e-03 -2.90150851e-01
-1.58981979e-01 -4.87046421e-01 5.42019546e-01 -2.42069766e-01
9.36284244e-01 -2.08664918e+00 1.27093479e-01 -1.49639547e-02
3.10786724e-01 9.46935415e-02 -3.12757552e-01 6.22093439e-01
-5.63637972e-01 3.03555608e-01 -2.22794726e-01 -4.75391626e-01
9.48590562e-02 2.48006493e-01 -2.80254662e-01 6.31060779e-01
2.36866385e-01 9.61346447e-01 -8.31426501e-01 -5.23704410e-01
6.87202588e-02 1.90145344e-01 -2.59371489e-01 1.55496418e-01
-8.01289976e-02 6.88123882e-01 -2.75938392e-01 2.61813521e-01
3.88345659e-01 -6.10740662e-01 3.14596117e-01 1.17457077e-01
-2.90424633e-03 1.85733080e-01 -1.19980597e+00 1.42651105e+00
-5.88076890e-01 5.62721968e-01 -3.70771512e-02 -1.40577066e+00
7.33410716e-01 5.63049376e-01 6.68847382e-01 -4.45348442e-01
-1.75712332e-02 3.40144694e-01 2.44519800e-01 -4.25128639e-01
8.49596784e-02 -8.60403836e-01 -1.28166035e-01 7.95264244e-01
3.06973368e-01 -1.15340035e-02 -8.08065757e-02 -2.28482962e-01
9.79928374e-01 2.56483376e-01 4.63041782e-01 -2.61950582e-01
3.06796670e-01 9.57473926e-03 4.69630301e-01 7.66576648e-01
-2.40015417e-01 3.17050755e-01 5.06836057e-01 -2.22283363e-01
-9.80442226e-01 -8.73576760e-01 -6.40491664e-01 1.81840920e+00
-2.42648035e-01 -1.17974855e-01 -3.69289696e-01 -9.83001351e-01
2.10463077e-01 6.42269731e-01 -1.01948512e+00 -1.17050514e-01
-3.64956468e-01 -8.01499486e-01 5.56840956e-01 3.74094993e-01
-8.98660645e-02 -8.56690109e-01 -3.57189864e-01 1.63790375e-01
4.14522644e-03 -8.92780483e-01 -3.56779695e-02 8.66475105e-01
-9.79852080e-01 -1.22629201e+00 -6.94572866e-01 -7.08991408e-01
3.12790334e-01 2.64920235e-01 1.34866738e+00 -6.23493418e-02
2.90501356e-01 6.29245639e-01 -3.94457906e-01 -6.88537419e-01
-7.36418128e-01 1.68648064e-01 6.57084957e-03 9.17362273e-02
6.27862811e-01 -7.61719108e-01 -1.55949801e-01 5.22493362e-01
-1.03027987e+00 6.27912581e-02 7.55455196e-01 8.47869277e-01
2.15912849e-01 -1.57799840e-01 1.15122318e+00 -1.31918168e+00
6.03145182e-01 -8.43702018e-01 -2.62520313e-01 7.03493714e-01
-9.01802421e-01 1.12363454e-02 4.52454150e-01 -6.04220092e-01
-8.46103966e-01 -8.55598152e-02 2.81912297e-01 -2.37751365e-01
-6.05005920e-01 5.77258587e-01 6.10136252e-04 -6.78703636e-02
8.16558182e-01 -9.39554945e-02 1.61294147e-01 -5.63207626e-01
4.13129032e-01 7.57480741e-01 -1.65656842e-02 -7.97671616e-01
5.42762756e-01 2.57757127e-01 1.83893785e-01 -4.78457838e-01
-1.22572207e+00 -6.13121510e-01 -1.02787650e+00 -2.35641092e-01
7.72869468e-01 -8.79764199e-01 -5.43590367e-01 2.14104801e-01
-1.09450793e+00 -4.94800717e-01 -1.33766815e-01 5.30202746e-01
-8.09289753e-01 1.91113696e-01 -4.37040865e-01 -7.60953486e-01
3.98878455e-01 -1.10504317e+00 8.84103000e-01 -1.88681960e-01
-5.05517125e-01 -1.77497172e+00 2.98497498e-01 4.13988054e-01
2.80237675e-01 2.24544518e-02 1.31638551e+00 -1.41286969e+00
-5.40456884e-02 -1.57516986e-01 -2.14850575e-01 5.76062202e-01
3.98655653e-01 -3.50352228e-01 -1.28481174e+00 -4.41040128e-01
2.88928330e-01 -8.84713769e-01 6.91661537e-01 3.07651579e-01
6.70646906e-01 1.03002027e-01 -3.98159474e-01 -7.26146344e-03
1.52954745e+00 -6.71102256e-02 1.11514524e-01 3.85297984e-01
6.07426167e-01 1.20262003e+00 4.41921413e-01 -2.85460763e-02
5.31127632e-01 8.08098972e-01 1.91936970e-01 -8.34716633e-02
-4.50826213e-02 3.05128992e-02 1.97770894e-01 6.56728029e-01
1.16897441e-01 -7.52183348e-02 -1.15575099e+00 4.70798194e-01
-1.86475885e+00 -7.74261653e-01 -1.80071861e-01 2.41682482e+00
1.00928330e+00 1.30555272e-01 3.64132404e-01 5.94610274e-02
6.72808647e-01 -8.58662203e-02 -5.33244491e-01 -3.03827077e-01
4.66429815e-02 1.54115677e-01 4.82844651e-01 5.36997199e-01
-1.24693727e+00 6.02796853e-01 6.99878836e+00 7.21163213e-01
-9.74543095e-01 3.87982666e-01 6.10346556e-01 1.28777057e-01
-2.04812139e-01 1.55906141e-01 -9.14673626e-01 2.38506719e-01
1.23945534e+00 -2.52412260e-01 1.61517784e-01 7.76235700e-01
-7.44662210e-02 -2.11024955e-01 -1.85405993e+00 5.47520280e-01
-3.84062976e-02 -6.16934001e-01 -3.20575058e-01 3.32537055e-01
7.07803071e-01 8.67020339e-02 1.31506264e-01 5.70021093e-01
6.22905612e-01 -1.22010636e+00 6.21343791e-01 3.35467637e-01
6.40647948e-01 -3.75247538e-01 5.16596019e-01 7.62515247e-01
-7.94514000e-01 -2.65344054e-01 -1.74074352e-01 -3.35295498e-01
-1.05132028e-01 4.52445596e-01 -6.63144052e-01 6.25044405e-01
2.40809590e-01 7.83650279e-01 -7.89268434e-01 8.56999218e-01
7.06359670e-02 7.26264000e-01 1.32878363e-01 3.81928712e-01
1.83096841e-01 -6.76467195e-02 1.10014841e-01 1.31240559e+00
6.17026687e-02 -3.83982778e-01 3.64562988e-01 5.25599360e-01
2.31974751e-01 2.35499531e-01 -9.07953620e-01 1.38622567e-01
3.51310074e-01 1.28750575e+00 -7.36429393e-01 -1.61050051e-01
-1.00376797e+00 6.17999434e-01 6.06427848e-01 2.66425073e-01
-7.50367045e-01 2.54677087e-01 4.49122965e-01 -1.76098999e-02
1.50134591e-02 -7.55102485e-02 -5.30169606e-01 -9.95983243e-01
-1.18575975e-01 -1.02086747e+00 4.25802797e-01 -3.30420732e-01
-1.78741527e+00 1.00686356e-01 3.40636790e-01 -1.18713892e+00
-4.51023281e-01 -7.80833423e-01 -2.91238964e-01 1.10331023e+00
-1.69841766e+00 -1.34278512e+00 1.35221556e-01 4.49733108e-01
3.97240281e-01 -6.90573175e-03 1.14219224e+00 2.57998228e-01
-2.76886851e-01 4.34390247e-01 2.97142386e-01 -1.26063257e-01
1.36682355e+00 -1.33190322e+00 -2.01508384e-02 2.52425581e-01
5.39289355e-01 5.19766927e-01 6.92230582e-01 -5.52291870e-01
-8.32663655e-01 -9.22870815e-01 9.84244943e-01 -9.73907590e-01
9.20780659e-01 -4.28484142e-01 -1.09944057e+00 1.07503855e+00
9.81564373e-02 -1.02552116e-01 1.35984254e+00 7.48951316e-01
-6.25783443e-01 4.93779108e-02 -8.70140553e-01 1.01538017e-01
7.77808785e-01 -7.17585206e-01 -6.47020340e-01 5.79899728e-01
4.55905259e-01 9.89742801e-02 -1.02172399e+00 4.14774120e-01
6.17922604e-01 -1.11520708e+00 7.61706293e-01 -9.97627258e-01
4.25935507e-01 1.65186241e-01 -3.52970213e-01 -1.55373728e+00
-6.15934670e-01 -2.61067860e-02 1.69939488e-01 1.33567405e+00
6.50518894e-01 -7.08095074e-01 5.52531660e-01 9.57204938e-01
-7.89209232e-02 -6.32737041e-01 -8.70717049e-01 -1.03127038e+00
7.90560544e-01 -6.19219303e-01 2.28186637e-01 1.37626386e+00
-1.66988298e-01 7.06181824e-01 -3.48477840e-01 -8.55753794e-02
4.96448427e-01 -1.61763909e-03 6.65216088e-01 -1.86384165e+00
-4.68547255e-01 -4.29154694e-01 -1.32004693e-01 -7.84704983e-01
6.39205277e-01 -1.17291701e+00 -3.99358310e-02 -1.30548322e+00
3.88172865e-01 -9.26924348e-01 -5.87934136e-01 6.15560889e-01
-2.29773328e-01 1.28258675e-01 2.81236738e-01 5.47863781e-01
-5.80521941e-01 6.98132142e-02 1.15839708e+00 1.72101498e-01
3.25858220e-02 1.97156206e-01 -9.52310741e-01 7.39319324e-01
7.16821074e-01 -7.98779726e-01 -5.87533772e-01 -3.21055949e-01
3.78807366e-01 -1.00821972e-01 5.15859604e-01 -8.33290458e-01
8.38939250e-02 -3.51239383e-01 1.61601529e-01 1.28704399e-01
2.44929790e-01 -1.14500546e+00 2.02898771e-01 1.74145520e-01
-7.81953633e-01 -7.78139010e-02 1.32297531e-01 7.25284874e-01
-3.21450010e-02 -5.41835964e-01 7.88106978e-01 -1.81245118e-01
-6.22029245e-01 2.08316166e-02 -1.74756944e-01 2.66140312e-01
1.21348727e+00 -2.00231701e-01 -1.83350265e-01 -2.69416720e-01
-8.56794357e-01 6.01051562e-02 4.29778278e-01 5.40596008e-01
-1.01517975e-01 -1.34576774e+00 -8.84199142e-01 6.34985883e-03
3.89798284e-01 -4.53344643e-01 6.82210252e-02 9.76166070e-01
4.75517094e-01 6.14991188e-01 -1.50300384e-01 -6.55156493e-01
-1.13323712e+00 8.46862972e-01 2.52624393e-01 -5.03736198e-01
-7.23529756e-02 5.37266970e-01 5.87835729e-01 -6.31959677e-01
-1.33357821e-02 1.91635303e-02 -2.37402946e-01 4.75143343e-01
2.23437712e-01 2.36061811e-01 8.95457044e-02 -6.32119656e-01
-1.45424008e-01 5.49129307e-01 -1.97149843e-01 -9.53980088e-02
1.27950454e+00 -1.15954675e-01 -9.79597494e-02 1.27432883e+00
1.24081171e+00 -2.22237349e-01 -1.23006284e+00 -4.60610092e-01
4.90714729e-01 -2.50431478e-01 -2.18907073e-01 -9.92267609e-01
-4.78347957e-01 9.00552392e-01 4.55803633e-01 5.63668370e-01
8.55599761e-01 2.38849401e-01 -4.71327733e-03 2.20872283e-01
4.55881506e-01 -9.70903218e-01 2.02084154e-01 4.98528153e-01
7.09099412e-01 -1.60316372e+00 -1.17186643e-01 -2.10320801e-01
-5.70340395e-01 1.07357907e+00 4.96230125e-01 1.75469778e-02
7.93702781e-01 1.03242181e-01 5.97418658e-02 -8.77330527e-02
-8.52489114e-01 -1.00560285e-01 3.73979807e-01 7.54586458e-01
8.55477154e-01 9.71456766e-02 -1.62872210e-01 4.81136978e-01
1.68846890e-01 3.42096053e-02 1.99144095e-01 8.46117318e-01
-5.00505805e-01 -1.71350253e+00 -3.82070184e-01 6.50002062e-01
-4.86989349e-01 -9.22588445e-03 -3.16011816e-01 9.98585701e-01
2.96419948e-01 1.10928798e+00 -1.10691644e-01 -2.81424224e-01
1.81308597e-01 7.02793896e-01 5.75666368e-01 -1.10968876e+00
-7.08547771e-01 -1.87875062e-01 1.24058679e-01 -1.23107314e-01
-7.94339538e-01 -7.21815467e-01 -5.88759780e-01 -1.47630692e-01
-5.97064495e-01 1.62865072e-01 5.65468431e-01 1.25801802e+00
1.31433412e-01 2.55019158e-01 6.05638385e-01 -7.77353644e-01
-7.92623758e-01 -1.20488632e+00 -8.04404914e-01 5.20427287e-01
3.18935931e-01 -9.55863297e-01 -4.96344268e-01 1.47050887e-01] | [9.295431137084961, 4.365570545196533] |
68464361-57cc-4e10-9cc3-ca196df41f10 | optimized-power-normalized-cepstral | 2109.12058 | null | https://arxiv.org/abs/2109.12058v1 | https://arxiv.org/pdf/2109.12058v1.pdf | Optimized Power Normalized Cepstral Coefficients towards Robust Deep Speaker Verification | After their introduction to robust speech recognition, power normalized cepstral coefficient (PNCC) features were successfully adopted to other tasks, including speaker verification. However, as a feature extractor with long-term operations on the power spectrogram, its temporal processing and amplitude scaling steps dedicated on environmental compensation may be redundant. Further, they might suppress intrinsic speaker variations that are useful for speaker verification based on deep neural networks (DNN). Therefore, in this study, we revisit and optimize PNCCs by ablating its medium-time processor and by introducing channel energy normalization. Experimental results with a DNN-based speaker verification system indicate substantial improvement over baseline PNCCs on both in-domain and cross-domain scenarios, reflected by relatively 5.8% and 61.2% maximum lower equal error rate on VoxCeleb1 and VoxMovies, respectively. | ['Tomi Kinnunen', 'Md Sahidullah', 'Xuechen Liu'] | 2021-09-24 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 1.94068104e-01 -1.70492381e-01 8.20886940e-02 -3.09191585e-01
-7.85496294e-01 -4.41388756e-01 4.27992731e-01 1.99771821e-02
-4.88334537e-01 5.46636403e-01 2.64459819e-01 -3.59121203e-01
1.06312215e-01 -1.45399153e-01 -1.67454094e-01 -8.73144686e-01
-5.44796661e-02 -5.11035204e-01 -3.57944697e-01 -1.70280069e-01
-5.59487492e-02 6.65545404e-01 -1.70147812e+00 -2.06601936e-02
7.68331826e-01 1.05227876e+00 -6.53526708e-02 8.67266655e-01
1.93889946e-01 2.97349870e-01 -1.11406362e+00 -5.03664434e-01
-2.82143969e-02 -3.78603250e-01 -2.79093117e-01 -1.44086435e-01
1.99254483e-01 -3.08989942e-01 -3.85622650e-01 1.15775526e+00
1.14004838e+00 2.51828969e-01 4.33276117e-01 -1.06010067e+00
-4.26613331e-01 9.29504216e-01 -2.11609110e-01 3.37301373e-01
3.36945355e-01 1.10021234e-01 8.52626622e-01 -1.02078319e+00
1.11426217e-02 1.05976367e+00 9.13833380e-01 6.43100441e-01
-9.97366071e-01 -8.13374877e-01 -1.57498762e-01 4.65109348e-01
-1.62874794e+00 -1.26409638e+00 1.05830503e+00 4.11853679e-02
1.44765115e+00 4.39994603e-01 3.83380294e-01 1.20671034e+00
-2.12454230e-01 6.05162799e-01 6.35481715e-01 -5.66986501e-01
3.84178348e-02 8.84522796e-02 9.54162031e-02 1.13384366e-01
-1.46794036e-01 4.21313167e-01 -8.80918384e-01 -7.61079118e-02
3.61317575e-01 -5.62750161e-01 -5.67405105e-01 4.06568348e-01
-1.02559614e+00 6.33441269e-01 3.76135781e-02 5.75772166e-01
-4.36638892e-01 -2.59311330e-02 6.36008143e-01 4.09139752e-01
4.14092481e-01 5.54301664e-02 -5.83399057e-01 -4.34104532e-01
-1.10907698e+00 -2.02448517e-01 6.83843493e-01 6.59639299e-01
1.23672292e-01 8.99533331e-01 -1.05312295e-01 1.23172212e+00
3.56831282e-01 7.74777651e-01 9.37570453e-01 -4.41397190e-01
2.65137255e-01 -1.15686201e-01 -2.82244146e-01 -7.32926190e-01
-5.03509581e-01 -7.13499606e-01 -1.01692784e+00 -1.52768165e-01
2.38677800e-01 -3.71091038e-01 -9.33341861e-01 1.72636473e+00
2.13165626e-01 2.45489508e-01 4.29480642e-01 8.26841414e-01
1.02129984e+00 6.31774783e-01 3.62289548e-02 -5.11705041e-01
1.39359581e+00 -6.11297011e-01 -1.28502774e+00 1.10299475e-02
7.95970634e-02 -1.08792579e+00 8.27832580e-01 5.74091613e-01
-1.02199352e+00 -9.28122580e-01 -1.30905676e+00 -1.31905805e-02
-2.44944677e-01 3.75241011e-01 1.70263886e-01 1.40806723e+00
-1.14801896e+00 5.42249739e-01 -8.38245332e-01 -1.27771810e-01
8.63937736e-02 2.97118932e-01 -2.47382641e-01 4.85730022e-01
-1.51157510e+00 8.13359559e-01 1.98582888e-01 4.29077059e-01
-7.23460853e-01 -6.06051445e-01 -9.61632788e-01 3.06962222e-01
-2.46668294e-01 -1.56664580e-01 1.35584879e+00 -6.62098408e-01
-2.14597392e+00 4.45033789e-01 -5.27659535e-01 -7.21987844e-01
1.85551241e-01 -1.09409727e-01 -1.35659111e+00 1.17310204e-01
-5.34671903e-01 4.68106508e-01 1.28214169e+00 -5.59865177e-01
-4.85141516e-01 -1.24269776e-01 -5.97670972e-01 3.98534238e-02
-5.50026715e-01 2.77986437e-01 -9.99000296e-02 -8.33189607e-01
4.01000202e-01 -5.74248910e-01 2.62350678e-01 -5.36483824e-01
-4.92687464e-01 -3.46066386e-01 9.27683294e-01 -1.32228434e+00
1.35825992e+00 -2.43610907e+00 -4.32874769e-01 2.15059713e-01
-4.41691726e-01 7.25941002e-01 -2.04993397e-01 1.49072140e-01
-3.23002040e-01 2.43041180e-02 -1.71512797e-01 -4.88639623e-01
1.90189987e-01 -1.67091563e-01 -2.90984809e-01 7.08036065e-01
5.24888813e-01 5.52983165e-01 -2.10723087e-01 -1.70754567e-01
3.94231141e-01 1.22279608e+00 -4.09806818e-01 -1.65202633e-01
5.66896379e-01 1.98782936e-01 3.52698982e-01 7.90011227e-01
9.69197154e-01 6.30594850e-01 1.41230255e-01 -3.13412130e-01
-8.25084299e-02 8.38920534e-01 -1.18890190e+00 1.49875546e+00
-5.37928879e-01 1.07993317e+00 5.38605154e-01 -7.72439301e-01
9.57680762e-01 9.27530527e-01 1.96377963e-01 -6.22986257e-01
3.03694904e-01 1.96821898e-01 2.83023715e-01 -2.82136589e-01
5.83863199e-01 -2.92556316e-01 1.79381073e-01 -2.90029682e-02
3.63184303e-01 -2.98854392e-02 -2.09846079e-01 -2.39625663e-01
5.98492146e-01 -4.86091495e-01 2.33203426e-01 -2.44171560e-01
9.63419378e-01 -8.08043540e-01 5.65960884e-01 3.92594397e-01
-7.44610250e-01 7.24003553e-01 8.67670849e-02 1.72260895e-01
-7.12068021e-01 -9.38092411e-01 -5.89049697e-01 9.11048591e-01
-3.32234770e-01 -2.85176873e-01 -8.29344332e-01 -2.48315781e-01
-2.73595005e-01 7.46701300e-01 -6.89558163e-02 -1.15670681e-01
-6.74396694e-01 -7.73673296e-01 1.24775255e+00 5.60160697e-01
4.53488082e-01 -9.25075710e-01 -1.95748687e-01 4.67330158e-01
-2.18246326e-01 -1.08212101e+00 -6.99307621e-01 4.78748918e-01
-5.66337347e-01 -6.06447637e-01 -7.49898970e-01 -7.52095580e-01
1.43209696e-01 -3.01923174e-02 6.18305147e-01 -1.63135216e-01
-6.73031285e-02 9.37877744e-02 -3.28299642e-01 -5.36717594e-01
-5.13968170e-01 1.37162045e-01 6.42864048e-01 1.25164166e-01
6.04203641e-01 -5.81582844e-01 -4.31183398e-01 2.85686225e-01
-5.77715695e-01 -6.41938448e-01 3.31899643e-01 1.02847028e+00
8.20668265e-02 2.24616438e-01 1.05904996e+00 9.09810215e-02
8.96807075e-01 8.53819251e-02 -5.64630926e-01 7.60464072e-02
-4.28383082e-01 -2.75026232e-01 5.05212009e-01 -4.74527329e-01
-1.33005178e+00 -3.26501913e-02 -8.24552715e-01 -3.21590155e-01
-2.24933207e-01 4.38084573e-01 -5.61111689e-01 1.16932593e-01
6.06710553e-01 3.99455816e-01 1.57897286e-02 -5.13586700e-01
1.52484223e-01 1.25620270e+00 6.81502581e-01 2.12461669e-02
8.33728194e-01 4.02968451e-02 -5.30996859e-01 -1.43460751e+00
1.56668220e-02 -5.44613957e-01 -4.06175464e-01 3.74078006e-02
5.55088699e-01 -1.10036898e+00 -9.97462690e-01 8.72030914e-01
-1.30039251e+00 1.78743809e-01 -6.73391968e-02 9.15025294e-01
-1.03815701e-02 4.84763741e-01 -7.38027871e-01 -1.24503398e+00
-8.11293602e-01 -9.68034148e-01 8.10543835e-01 4.00480777e-01
-3.69347483e-01 -7.24934101e-01 -1.48436099e-01 2.29578033e-01
8.63081038e-01 -4.39810455e-01 3.91886503e-01 -8.61841261e-01
6.52566403e-02 -3.30937773e-01 2.03209162e-01 9.21172917e-01
2.96574056e-01 2.08251148e-01 -1.94760203e+00 -4.47907269e-01
4.39594597e-01 1.58690736e-01 6.43899560e-01 4.95424241e-01
9.19901669e-01 -3.05047184e-01 5.05806170e-02 7.59748518e-01
8.71324658e-01 5.59778035e-01 5.96871316e-01 -2.59507984e-01
2.75689662e-01 4.23903972e-01 1.02755606e-01 2.89731354e-01
-2.20144734e-01 6.28857911e-01 -4.72497344e-02 1.16964690e-01
-4.25993621e-01 -1.30254731e-01 7.80341268e-01 1.36977935e+00
7.56631717e-02 -2.71414667e-01 -6.95113003e-01 6.58717394e-01
-9.61847901e-01 -1.00909209e+00 -9.94289815e-02 2.12914419e+00
7.65349269e-01 -4.37220745e-02 -2.62998659e-02 8.15511048e-01
9.78576481e-01 2.79256344e-01 -5.95333338e-01 -5.41960955e-01
-5.59690654e-01 5.03440380e-01 2.50926614e-01 4.84597862e-01
-1.17705798e+00 4.56339031e-01 6.21526766e+00 1.06807053e+00
-1.52273679e+00 2.34848261e-01 4.06051904e-01 -2.75255859e-01
1.50921777e-01 -7.00028598e-01 -7.97366261e-01 2.86246777e-01
1.58215559e+00 -1.23065494e-01 4.44939882e-01 7.85933137e-01
3.18311602e-01 2.36785755e-01 -1.08586192e+00 1.33739996e+00
1.78245977e-01 -7.22070992e-01 -5.15645862e-01 -7.33485371e-02
3.86894971e-01 1.63041607e-01 2.43566945e-01 4.09622669e-01
-3.25372130e-01 -1.04868197e+00 6.82409763e-01 -1.55479208e-01
9.65205193e-01 -1.09143436e+00 9.81568277e-01 -3.21675092e-02
-1.26638043e+00 6.49615601e-02 -1.91866070e-01 1.06183894e-01
2.47264355e-01 6.29186571e-01 -1.20776737e+00 4.70760316e-01
5.19397497e-01 3.79207581e-01 -7.92667493e-02 9.04769480e-01
-3.57279271e-01 9.74231720e-01 -5.39595127e-01 8.11694562e-02
-2.93928713e-01 2.44215131e-01 6.53989196e-01 1.42803907e+00
6.24668956e-01 -6.04823790e-02 -6.51967943e-01 4.11375821e-01
-2.30514869e-01 -7.69650936e-02 -1.62379935e-01 -2.58009993e-02
8.18434715e-01 1.15497506e+00 -3.35196227e-01 -1.85132071e-01
-1.29758731e-01 8.67998660e-01 -2.91675508e-01 5.97236156e-01
-1.06544518e+00 -9.21883345e-01 1.24711549e+00 -3.35116655e-01
4.46123481e-01 -6.02851212e-02 -3.46104026e-01 -1.14056289e+00
1.64371118e-01 -8.60764921e-01 -6.89224526e-03 -3.14202696e-01
-1.12000465e+00 8.55751514e-01 -5.21153808e-01 -1.29580986e+00
-3.49157244e-01 -5.81125855e-01 -6.20700181e-01 1.16827023e+00
-1.74253559e+00 -8.11304152e-01 1.61644608e-01 6.80767059e-01
5.23086011e-01 -3.08566064e-01 9.41759884e-01 6.80050850e-01
-7.82054424e-01 1.23310089e+00 1.61990166e-01 2.80631870e-01
5.90399981e-01 -8.95786583e-01 6.51460886e-01 1.24880254e+00
9.69529971e-02 7.17150807e-01 6.83720589e-01 -1.61480606e-01
-1.25268030e+00 -9.60247993e-01 1.23202491e+00 1.00559695e-02
3.62070948e-01 -4.99295622e-01 -9.12467241e-01 2.00698808e-01
6.49269164e-01 -1.14988498e-01 7.44895339e-01 6.70455694e-02
-2.58255839e-01 -3.37817252e-01 -1.34688997e+00 3.35787088e-01
8.10057878e-01 -1.13681149e+00 -5.60827017e-01 -2.21632677e-03
6.76865220e-01 -3.00801337e-01 -8.54637206e-01 3.70992243e-01
6.08825445e-01 -9.29474056e-01 9.52710330e-01 -9.67647955e-02
-2.56786793e-01 -3.51707816e-01 -2.81593353e-01 -1.54579067e+00
-3.42060775e-02 -1.01827466e+00 2.74633784e-02 1.72641206e+00
4.83851016e-01 -8.44279289e-01 4.80024636e-01 4.37344372e-01
-3.76411408e-01 7.70573243e-02 -1.43230629e+00 -9.34845209e-01
-2.56204039e-01 -9.43371236e-01 6.86887026e-01 8.79787266e-01
3.74942273e-01 1.13710545e-01 -3.91249001e-01 5.28492689e-01
3.28895599e-01 -4.63748246e-01 3.68906856e-01 -9.79134202e-01
-1.04682341e-01 -5.78907073e-01 -3.77983361e-01 -9.52974319e-01
2.23208100e-01 -5.56302786e-01 3.53250772e-01 -7.61602759e-01
-4.72875506e-01 1.49692312e-01 -5.77123106e-01 1.46923348e-01
-1.34237781e-01 8.84297639e-02 1.66893646e-01 -3.76974881e-01
2.02394217e-01 6.99785590e-01 4.21524435e-01 -3.79328579e-01
-3.20759684e-01 1.24860205e-01 -4.12946343e-01 5.91327488e-01
8.88490081e-01 -2.92422146e-01 -2.59806484e-01 -1.30044922e-01
-6.54510081e-01 7.44594783e-02 6.23960942e-02 -1.33160329e+00
3.20452631e-01 4.89704907e-01 2.77476341e-01 -7.91609228e-01
7.81933427e-01 -6.99962795e-01 1.59434438e-01 5.41402042e-01
-9.14787874e-02 -2.24487960e-01 5.48800409e-01 1.69169396e-01
-8.10695469e-01 -1.53920949e-01 7.95471370e-01 4.50827718e-01
-3.10196847e-01 -2.01559141e-01 -7.79731154e-01 -4.92379844e-01
4.98063624e-01 -2.54223973e-01 -4.01715100e-01 -4.59048152e-01
-6.35263145e-01 -3.80354732e-01 -6.25015274e-02 3.27651232e-01
6.39370322e-01 -1.22568405e+00 -7.84538031e-01 8.13528955e-01
-2.90125459e-01 -3.80980223e-01 6.99762940e-01 8.61260831e-01
-1.85453966e-01 8.93899083e-01 -4.54368303e-03 -5.80933273e-01
-1.53263605e+00 2.39705816e-01 4.97542799e-01 2.13318005e-01
-1.79337174e-01 1.24760497e+00 -1.70436829e-01 -2.18225583e-01
6.91226900e-01 -6.86623693e-01 -1.14074513e-01 4.69040453e-01
7.29188204e-01 4.91910964e-01 8.39185596e-01 -9.08724248e-01
-7.60902107e-01 2.66467839e-01 1.41875714e-01 -2.36052528e-01
1.26949584e+00 -2.96168327e-01 1.50943562e-01 1.46115065e-01
1.37751722e+00 2.75019825e-01 -9.57166731e-01 -6.39212579e-02
-6.74122423e-02 -8.94598942e-03 3.83058786e-01 -8.05454433e-01
-1.24942970e+00 1.11021209e+00 9.47928488e-01 4.09889609e-01
1.42152929e+00 -5.46309948e-01 8.59349728e-01 3.08135599e-01
-5.75235859e-02 -1.17622519e+00 -4.98100549e-01 6.17409825e-01
9.36775446e-01 -1.09096766e+00 -3.50709826e-01 -1.41509265e-01
-4.36343372e-01 1.26705229e+00 1.46132991e-01 4.45358008e-01
8.57216358e-01 3.33271831e-01 4.62500930e-01 3.85107458e-01
-5.52560270e-01 -5.02411723e-02 2.34750628e-01 7.99423218e-01
6.54965937e-01 1.32366329e-01 -1.32002190e-01 7.69136846e-01
-5.54432809e-01 -3.39620471e-01 3.52896720e-01 5.48106670e-01
-2.69643039e-01 -8.90903652e-01 -8.17859828e-01 5.90643324e-02
-7.54264951e-01 -3.28473806e-01 -1.49123102e-01 4.46172059e-01
-1.15650341e-01 1.57790267e+00 1.51533514e-01 -5.10638237e-01
4.47625250e-01 5.13919652e-01 5.48002832e-02 6.37961030e-02
-9.93449032e-01 5.16123474e-01 3.05599600e-01 -2.29027629e-01
-3.97576481e-01 -9.35323298e-01 -1.07844758e+00 -3.36034685e-01
-6.49489522e-01 8.42852816e-02 1.30433488e+00 6.63363039e-01
3.28901470e-01 7.30815947e-01 6.60935581e-01 -7.65612245e-01
-6.48277283e-01 -1.55285108e+00 -5.98820210e-01 1.29064411e-01
9.03968513e-01 -1.80146545e-01 -6.23450816e-01 1.69864967e-01] | [14.443700790405273, 6.032254219055176] |
44a9c409-445e-494f-8c01-4674b064b805 | deep-quantigraphic-image-enhancement-via | 2304.02285 | null | https://arxiv.org/abs/2304.02285v1 | https://arxiv.org/pdf/2304.02285v1.pdf | Deep Quantigraphic Image Enhancement via Comparametric Equations | Most recent methods of deep image enhancement can be generally classified into two types: decompose-and-enhance and illumination estimation-centric. The former is usually less efficient, and the latter is constrained by a strong assumption regarding image reflectance as the desired enhancement result. To alleviate this constraint while retaining high efficiency, we propose a novel trainable module that diversifies the conversion from the low-light image and illumination map to the enhanced image. It formulates image enhancement as a comparametric equation parameterized by a camera response function and an exposure compensation ratio. By incorporating this module in an illumination estimation-centric DNN, our method improves the flexibility of deep image enhancement, limits the computational burden to illumination estimation, and allows for fully unsupervised learning adaptable to the diverse demands of different tasks. | ['Akisato Kimura', 'Yongqing Sun', 'Xiaomeng Wu'] | 2023-04-05 | null | null | null | null | ['image-enhancement'] | ['computer-vision'] | [ 4.96606171e-01 -2.88664043e-01 2.24874347e-01 -5.13006330e-01
-4.08520877e-01 -4.60241884e-01 4.58906502e-01 -4.63986576e-01
-7.15117335e-01 5.91263890e-01 5.82843795e-02 -1.66234329e-01
-4.78768200e-02 -9.32020009e-01 -6.72996879e-01 -1.07545471e+00
8.68848026e-01 -3.24461758e-01 1.69838406e-02 -2.20434517e-01
1.19196273e-01 5.40823042e-01 -1.54047179e+00 -1.52577013e-02
1.09278274e+00 1.06460822e+00 4.69847679e-01 5.80568612e-01
-2.20974628e-02 6.13962650e-01 -4.71163839e-01 -3.19093704e-01
5.38987398e-01 -3.64305556e-01 -1.93117544e-01 4.09372717e-01
3.21060270e-01 -6.94500804e-01 -4.56429631e-01 1.12447798e+00
7.55668163e-01 1.38909176e-01 4.28090096e-01 -1.07273531e+00
-6.17094755e-01 1.16434336e-01 -6.24130130e-01 -1.13314196e-01
-1.80539548e-01 2.48431489e-01 4.60325420e-01 -8.03857505e-01
3.67362112e-01 8.65097165e-01 5.92229187e-01 2.70084947e-01
-1.19218540e+00 -3.92239362e-01 1.29608035e-01 3.94984111e-02
-1.04402614e+00 -5.72672069e-01 9.91596699e-01 -2.40315199e-01
5.60336709e-01 4.01073694e-02 5.42556286e-01 8.38871181e-01
7.11601526e-02 5.83191991e-01 1.52532601e+00 -4.43628609e-01
3.39580685e-01 2.69213110e-01 -2.47096926e-01 4.75744188e-01
1.17800370e-01 3.17340553e-01 -3.04221869e-01 3.12029541e-01
9.74354625e-01 1.03011981e-01 -5.15655458e-01 -4.53263998e-01
-7.89129496e-01 5.27306199e-01 4.94286478e-01 4.02293615e-02
-3.68687809e-01 4.44330368e-03 7.66445547e-02 2.47208044e-01
5.95743895e-01 5.33665180e-01 -5.07451236e-01 1.30211800e-01
-9.86614585e-01 -3.44634242e-02 3.97739708e-01 7.34720051e-01
1.14738655e+00 3.95038456e-01 -5.24499491e-02 9.63874221e-01
1.70595795e-01 5.84884524e-01 2.16091841e-01 -1.02734089e+00
3.22086275e-01 6.35156512e-01 2.78449535e-01 -7.87844956e-01
-5.15644014e-01 -8.44876468e-01 -8.25113475e-01 5.83224773e-01
2.24721834e-01 -5.30001163e-01 -9.04579818e-01 1.78121793e+00
4.32909548e-01 -2.43263051e-01 5.45350537e-02 1.08822083e+00
5.88002682e-01 6.48108125e-01 -4.28957120e-02 -2.95273960e-01
1.13562751e+00 -1.05769038e+00 -7.80764937e-01 -1.64612561e-01
2.16877848e-01 -8.61665905e-01 9.88486290e-01 3.41053665e-01
-1.23289597e+00 -7.40609467e-01 -1.19436049e+00 -4.28541362e-01
-4.22676414e-01 5.45965016e-01 5.48403442e-01 6.53068185e-01
-1.28556848e+00 4.75069821e-01 -5.30558705e-01 -1.16527140e-01
4.29304615e-02 4.24157143e-01 -8.32561105e-02 -5.82740493e-02
-1.05761611e+00 9.73621190e-01 4.01297748e-01 2.54589230e-01
-6.26012206e-01 -7.70034134e-01 -7.91846931e-01 2.28140146e-01
2.87543118e-01 -8.38923335e-01 9.87554431e-01 -1.26552737e+00
-2.08643603e+00 5.95549107e-01 2.55344808e-03 -1.24735925e-02
6.22109771e-01 -9.45775136e-02 -2.83238024e-01 1.43747777e-01
-5.15434146e-01 6.11945152e-01 1.31615996e+00 -1.46666563e+00
-5.01356244e-01 -1.68946624e-01 1.81955576e-01 5.67160606e-01
-6.45013988e-01 4.23790701e-02 -5.84113121e-01 -6.06978297e-01
-4.75049131e-02 -6.86647117e-01 -1.27397612e-01 4.45911318e-01
-8.90933648e-02 4.20237154e-01 9.83300567e-01 -6.74653172e-01
8.90042603e-01 -2.24036169e+00 -1.07946398e-03 -1.11624852e-01
8.38926211e-02 3.79631579e-01 -5.62536538e-01 2.01515287e-01
-1.45155832e-01 -3.85480404e-01 -3.53949308e-01 -5.08802235e-01
4.35525738e-03 1.66958183e-01 -8.78912434e-02 5.12919247e-01
3.22406113e-01 7.14654505e-01 -9.04815376e-01 -3.29056323e-01
6.94056213e-01 1.06404555e+00 -5.05430222e-01 5.95397115e-01
3.05632614e-02 5.16795754e-01 -6.47843480e-02 3.78202856e-01
1.06422842e+00 4.44124863e-02 1.24084845e-01 -6.97661698e-01
-6.52286530e-01 -1.12725042e-01 -1.32253265e+00 1.40859294e+00
-8.09798837e-01 5.49141884e-01 4.56260979e-01 -7.59294033e-01
1.04516602e+00 1.65211961e-01 5.33702731e-01 -8.25364590e-01
3.98674369e-01 -1.26892226e-02 -1.98468253e-01 -4.58849728e-01
7.42258430e-01 -9.49752256e-02 4.45506275e-01 3.85190278e-01
4.52300198e-02 -2.66886503e-01 3.85634340e-02 -2.95775175e-01
5.97343266e-01 5.93625426e-01 2.05726191e-01 -2.21219137e-01
4.67042387e-01 -5.84165335e-01 7.45805860e-01 6.10990405e-01
-2.85318550e-02 6.70516431e-01 1.79283708e-01 -2.98287511e-01
-1.15446877e+00 -1.05101347e+00 -1.02077939e-01 1.16551149e+00
1.88628167e-01 1.13508720e-02 -9.12848949e-01 -4.68793958e-01
-3.88038307e-01 4.68587816e-01 -3.75760078e-01 -1.47650093e-01
-4.70481694e-01 -9.75974679e-01 3.35666947e-02 5.08691013e-01
1.02922726e+00 -7.57192731e-01 -6.22405529e-01 3.15498821e-02
-3.09732646e-01 -1.20413566e+00 -6.07960999e-01 3.82653415e-01
-7.91082978e-01 -7.94068336e-01 -8.07313740e-01 -6.70135856e-01
9.98383880e-01 4.21591461e-01 8.69097114e-01 -1.86125711e-01
-9.90612656e-02 2.18306810e-01 -1.54823735e-01 -5.03622532e-01
-1.86951563e-01 -1.29534811e-01 -1.45542711e-01 5.09199321e-01
-3.77888493e-02 -6.36993527e-01 -9.43733573e-01 2.91945904e-01
-1.19740045e+00 3.53818804e-01 8.20617497e-01 8.03398907e-01
5.55695176e-01 2.27606446e-01 2.52781093e-01 -5.91810465e-01
2.21811861e-01 3.21851112e-02 -8.71624589e-01 1.92298010e-01
-7.13906109e-01 1.31944835e-01 7.69962192e-01 -5.67431331e-01
-1.77340734e+00 4.92663860e-01 -2.28724688e-01 -2.32331067e-01
-1.54419571e-01 1.50829047e-01 -5.35634160e-01 -6.30412698e-01
4.53993499e-01 3.62832248e-01 -1.72533870e-01 -4.92029011e-01
4.12146091e-01 5.00306010e-01 7.62887776e-01 -3.94416779e-01
1.07652164e+00 5.52516758e-01 1.30338460e-01 -9.03474510e-01
-8.90583992e-01 -3.12623501e-01 -5.52566648e-01 -4.81961191e-01
8.44668388e-01 -9.23175335e-01 -6.44210935e-01 8.30734432e-01
-1.05218601e+00 -5.35969675e-01 -4.36799586e-01 4.23670769e-01
-3.62413377e-01 4.74067301e-01 -6.47421479e-01 -7.07564712e-01
-3.62195253e-01 -1.24172187e+00 1.05003345e+00 4.99525726e-01
3.78823936e-01 -9.81735587e-01 -5.74773662e-02 2.33855575e-01
8.58001351e-01 1.81172505e-01 8.43797505e-01 3.32390696e-01
-5.99034905e-01 2.16740184e-02 -7.12400734e-01 8.50856006e-01
1.73806503e-01 -3.61392424e-02 -1.37820721e+00 -3.15124959e-01
3.92580092e-01 -1.45248055e-01 8.99076581e-01 5.79960346e-01
1.29676414e+00 -1.37378544e-01 1.79392979e-01 1.14427686e+00
1.88991094e+00 9.81488079e-02 9.45131481e-01 4.27110106e-01
6.04100108e-01 5.67583680e-01 3.24963838e-01 4.87647861e-01
2.75778502e-01 7.05923438e-01 6.07020497e-01 -9.76772308e-01
-3.55647951e-01 4.04774304e-03 4.10003960e-01 7.15997338e-01
-2.86972851e-01 -2.58706003e-01 -2.66211182e-01 3.02666605e-01
-1.68654287e+00 -7.63696790e-01 1.30875021e-01 2.04911065e+00
9.06785369e-01 -2.74722338e-01 -9.65749323e-02 6.14047423e-02
6.81030393e-01 2.97299206e-01 -6.00636482e-01 -2.85615176e-01
-3.39983076e-01 1.49767339e-01 5.69658518e-01 5.59207916e-01
-1.05539727e+00 6.80598259e-01 6.62018490e+00 4.85670269e-01
-1.33322847e+00 4.24781032e-02 5.05807102e-01 4.39747050e-02
-3.37911814e-01 -1.11030579e-01 -6.86490297e-01 4.10914332e-01
3.90579641e-01 2.15609103e-01 5.38077295e-01 6.17473662e-01
4.15341377e-01 -9.89945307e-02 -9.21615958e-01 1.09646916e+00
1.74697250e-01 -8.09978664e-01 3.97807769e-02 3.15487720e-02
8.99685681e-01 -1.10745586e-01 3.44668716e-01 7.90968016e-02
-1.02112088e-02 -5.95371246e-01 5.77978730e-01 6.88329995e-01
7.36071765e-01 -6.47214413e-01 6.21546805e-01 4.15218696e-02
-9.63399291e-01 -1.97491884e-01 -4.29859161e-01 -4.33595553e-02
1.52195305e-01 8.27429771e-01 -2.07975626e-01 5.30724525e-01
4.89905447e-01 5.47572732e-01 -3.90665591e-01 9.91195977e-01
-4.81040955e-01 1.94908023e-01 -1.14103176e-01 3.46388191e-01
-1.11770453e-02 -6.58386111e-01 5.18439472e-01 1.30404377e+00
2.83362567e-01 -5.83399534e-02 7.35343993e-02 9.31754887e-01
-1.55551404e-01 -4.86583188e-02 -2.50372708e-01 4.03717995e-01
4.20053229e-02 1.77888751e+00 -4.92988646e-01 -5.63093014e-02
-4.67810452e-01 1.00494933e+00 1.09379075e-01 7.79430509e-01
-8.64633262e-01 -4.71575737e-01 4.91933256e-01 1.23111665e-01
3.21679771e-01 -1.36389941e-01 -2.67929375e-01 -1.28892422e+00
1.53511211e-01 -8.40668857e-01 3.04492898e-02 -1.22916913e+00
-9.43178177e-01 7.04139054e-01 -3.02674830e-01 -1.11880994e+00
2.74397194e-01 -9.10920322e-01 -5.30742347e-01 8.83038163e-01
-2.26035166e+00 -1.31300402e+00 -6.64493918e-01 7.53098190e-01
2.74063319e-01 1.99183926e-01 5.23095787e-01 5.95834076e-01
-6.59634292e-01 4.76752222e-01 2.62775600e-01 -1.34599909e-01
1.00886512e+00 -1.24516368e+00 -1.52579948e-01 1.18548274e+00
-4.80658054e-01 2.93440491e-01 5.79133689e-01 -1.50386140e-01
-1.26217806e+00 -1.27391469e+00 6.47587776e-01 -3.99199389e-02
3.34980488e-01 -3.17046940e-01 -7.35584140e-01 3.43410552e-01
2.66823530e-01 -9.65977535e-02 3.78501981e-01 -1.80380374e-01
-3.08844239e-01 -7.07661450e-01 -1.02959657e+00 6.84328437e-01
7.90137529e-01 -4.85880077e-01 -1.73518226e-01 2.60364830e-01
4.40194130e-01 -4.89598900e-01 -8.31836700e-01 3.61693799e-01
4.55344856e-01 -9.80494797e-01 9.38947499e-01 1.13125797e-02
5.75966656e-01 -5.28819680e-01 -8.00975114e-02 -1.45935512e+00
-3.68873209e-01 -6.93870068e-01 4.46547195e-02 1.27873981e+00
2.20260218e-01 -7.23188519e-01 5.90066314e-01 6.36135638e-01
-3.46361488e-01 -6.19331956e-01 -3.20252001e-01 -6.77144170e-01
-2.91450113e-01 -2.18347043e-01 6.48715794e-01 8.14965546e-01
-5.91959536e-01 9.79067292e-03 -6.86311305e-01 3.30925226e-01
9.17039275e-01 1.08701251e-01 6.83426499e-01 -9.60135579e-01
-5.38865924e-01 -3.72717530e-01 3.97442561e-03 -1.25610602e+00
-4.01772149e-02 -4.83811200e-01 3.43787551e-01 -1.42333412e+00
2.28693768e-01 -1.65929586e-01 -4.15891469e-01 5.14789701e-01
-1.02638938e-01 3.54247391e-01 9.69517380e-02 -7.08333999e-02
-3.97007108e-01 6.79807544e-01 1.39184618e+00 -2.73068435e-02
-3.91926467e-01 -2.47223470e-02 -8.87696147e-01 8.29246700e-01
6.53929055e-01 -1.43298671e-01 -4.27273750e-01 -8.92983079e-01
4.04078335e-01 -2.71216482e-01 3.58707130e-01 -8.98513615e-01
3.26418579e-01 -2.06549197e-01 5.85425973e-01 -3.78091246e-01
3.91493618e-01 -9.30596530e-01 4.01662514e-02 1.91923752e-01
-2.90605247e-01 -1.42664596e-01 6.98505044e-02 2.34543324e-01
-1.95718735e-01 -1.61576480e-01 1.32686162e+00 1.30172297e-01
-7.77594268e-01 3.98422390e-01 -2.30039909e-01 -2.51281679e-01
6.46787703e-01 -3.26288283e-01 -4.61661726e-01 -5.02191424e-01
-2.52345115e-01 -1.06713392e-01 5.27689934e-01 1.03494167e-01
5.46715856e-01 -1.05596280e+00 -6.51137650e-01 2.50108242e-01
-1.01254441e-01 -1.16016582e-01 4.82074648e-01 8.57970893e-01
-3.27650011e-01 -3.26395705e-02 -2.94876486e-01 -3.98621261e-01
-1.09137130e+00 5.61251044e-01 6.71488345e-01 -2.81405419e-01
-4.54638422e-01 4.36615169e-01 5.30436397e-01 -4.18564141e-01
1.40267462e-01 -6.76112622e-02 -1.47344559e-01 -2.84798682e-01
6.27920151e-01 4.94674981e-01 1.15528107e-01 -4.07659709e-01
6.71490207e-02 7.89855599e-01 1.25196606e-01 1.41679406e-01
1.53880930e+00 -5.18580258e-01 -1.30205676e-01 6.95789307e-02
1.11259913e+00 -2.32965518e-02 -1.89991605e+00 -3.16092312e-01
-6.09229445e-01 -4.38918978e-01 5.35235047e-01 -9.65914547e-01
-1.47227883e+00 8.58181775e-01 7.69937694e-01 -1.29719764e-01
1.88074267e+00 -6.74118817e-01 7.32746601e-01 3.84064674e-01
-2.16272529e-02 -1.43659461e+00 1.20362699e-01 5.08827269e-01
7.97179401e-01 -1.25582099e+00 2.34997958e-01 -4.02445704e-01
-3.21674794e-01 1.29504800e+00 6.92451894e-01 9.56541896e-02
5.24032772e-01 5.41594446e-01 3.51222992e-01 1.35474220e-01
-2.58476198e-01 -3.36954474e-01 3.81323814e-01 7.53468871e-01
2.72730947e-01 -3.78036499e-01 -1.56741053e-01 1.98697522e-01
2.16071039e-01 5.06580481e-03 4.42858070e-01 5.43482542e-01
-3.05663198e-01 -9.66480553e-01 -1.92550257e-01 3.60199511e-02
-2.67396837e-01 -2.23276213e-01 6.23568520e-02 6.91172600e-01
3.44022274e-01 9.43607152e-01 -2.31522694e-03 -1.32632658e-01
4.22454506e-01 -3.09529543e-01 5.13554633e-01 -2.71080673e-01
-5.17854452e-01 2.93632776e-01 -1.30339816e-01 -4.41642225e-01
-7.31010973e-01 -2.72000283e-01 -7.01682448e-01 -1.06034972e-01
-6.34613216e-01 -3.08374584e-01 9.78052318e-01 9.84564126e-01
2.82852978e-01 5.91716826e-01 9.12745833e-01 -1.03764665e+00
-5.19912899e-01 -5.67176521e-01 -6.94440901e-01 3.82273972e-01
3.71197850e-01 -4.57517058e-01 -4.32117939e-01 2.29172587e-01] | [10.720809936523438, -2.4353718757629395] |
a4a71909-e129-4b38-a4df-dc9902e6dd63 | jccs-pfgm-a-novel-circle-supervision-based | 2306.07824 | null | https://arxiv.org/abs/2306.07824v1 | https://arxiv.org/pdf/2306.07824v1.pdf | JCCS-PFGM: A Novel Circle-Supervision based Poisson Flow Generative Model for Multiphase CECT Progressive Low-Dose Reconstruction with Joint Condition | Multiphase contrast-enhanced computed tomography (CECT) scan is clinically significant to demonstrate the anatomy at different phases. In practice, such a multiphase CECT scan inherently takes longer time and deposits much more radiation dose into a patient body than a regular CT scan, and reduction of the radiation dose typically compromise the CECT image quality and its diagnostic value. With Joint Condition and Circle-Supervision, here we propose a novel Poisson Flow Generative Model (JCCS-PFGM) to promote the progressive low-dose reconstruction for multiphase CECT. JCCS-PFGM is characterized by the following three aspects: a progressive low-dose reconstruction scheme, a circle-supervision strategy, and a joint condition mechanism. Our extensive experiments are performed on a clinical dataset consisting of 11436 images. The results show that our JCCS-PFGM achieves promising PSNR up to 46.3dB, SSIM up to 98.5%, and MAE down to 9.67 HU averagely on phases I, II and III, in quantitative evaluations, as well as gains high-quality readable visualizations in qualitative assessments. All of these findings reveal our method a great potential to be adapted for clinical CECT scans at a much-reduced radiation dose. | ['Ge Wang', 'Daoqiang Zhang', 'Yang Chen', 'Cong Xia', 'Yuting He', 'Rongjun Ge'] | 2023-06-13 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 8.29100832e-02 -3.96028161e-01 4.98253629e-02 -4.82317545e-02
-6.41022563e-01 -2.21508935e-01 2.50726789e-01 1.52242929e-01
-2.78813630e-01 6.07492626e-01 3.66392940e-01 -4.52897996e-01
-3.10981452e-01 -7.86928475e-01 -3.75215948e-01 -9.77015376e-01
-3.79061371e-01 3.70869517e-01 6.17443919e-01 1.50975019e-01
-2.08985787e-02 2.89706767e-01 -7.24024713e-01 1.62615106e-01
1.06692219e+00 8.73798549e-01 4.60106939e-01 6.58447325e-01
2.55029023e-01 9.51102734e-01 -2.26304948e-01 -2.20850706e-01
1.05412588e-01 -5.85806251e-01 -7.39784241e-01 1.26800463e-01
-1.21030614e-01 -4.48337853e-01 -3.43360245e-01 8.45655024e-01
7.13228106e-01 -1.59667298e-01 6.98922276e-01 -7.49506652e-01
-3.36894363e-01 3.79072726e-01 -1.02476597e+00 6.87262058e-01
1.59952953e-01 4.78476077e-01 1.16553612e-01 -6.68353975e-01
5.62955976e-01 8.32453966e-01 5.96574545e-01 1.87668815e-01
-1.08482015e+00 -5.34631550e-01 -3.17223579e-01 1.56654611e-01
-1.12871802e+00 3.01249743e-01 3.52869034e-01 -4.24435526e-01
4.37693566e-01 5.78014433e-01 1.02595353e+00 5.38008451e-01
9.85735834e-01 5.04136145e-01 1.49384582e+00 -7.73100927e-02
-1.50805427e-04 -2.16750056e-01 -2.49042302e-01 7.95591950e-01
5.08045435e-01 1.60880893e-01 -1.52089715e-01 -9.19068083e-02
1.15023804e+00 1.10451922e-01 -8.03863406e-01 -3.59569609e-01
-1.44858408e+00 4.94879365e-01 6.19044721e-01 3.03092837e-01
-4.58105236e-01 1.48627996e-01 6.08293474e-01 -1.18057646e-01
1.80435061e-01 3.72198708e-02 2.66716570e-01 -7.03771189e-02
-6.15607321e-01 -1.53713375e-01 4.71652746e-01 8.70127797e-01
3.60704772e-02 -4.70942408e-02 -4.46970820e-01 5.42194188e-01
3.15303832e-01 8.48358095e-01 4.55726475e-01 -8.25761437e-01
2.62084544e-01 8.73852000e-02 -1.83660239e-01 -7.47183740e-01
-3.04831713e-01 -8.38800311e-01 -1.40358412e+00 9.67359394e-02
1.48901924e-01 1.37089700e-01 -8.35952997e-01 1.23208606e+00
5.02066195e-01 1.90033630e-01 -2.11932987e-01 1.32844698e+00
8.58813345e-01 8.06603432e-01 4.41823512e-01 -7.21880376e-01
1.64359260e+00 -8.86871517e-01 -9.18568552e-01 3.95124517e-02
4.42594737e-01 -9.99517500e-01 1.11519003e+00 3.89191449e-01
-1.36566043e+00 -1.98940158e-01 -1.15498567e+00 4.37815011e-01
7.74700582e-01 3.17439884e-02 4.54077393e-01 7.14120030e-01
-6.65111959e-01 5.40464818e-01 -1.03738058e+00 1.91240445e-01
5.75316429e-01 -1.34988213e-02 -1.49846405e-01 -5.10660470e-01
-9.09510911e-01 8.04651499e-01 4.35216501e-02 1.71971887e-01
-7.96095729e-01 -1.18941367e+00 -6.15789294e-01 -1.65083423e-01
3.22726429e-01 -1.01187611e+00 1.16396832e+00 -1.50984898e-01
-1.31236005e+00 5.72002232e-01 -2.57925689e-02 -1.95639148e-01
9.02794242e-01 7.64499903e-02 -3.47257495e-01 7.35528350e-01
2.79400319e-01 1.55928671e-01 3.91261995e-01 -1.25486422e+00
-3.13368917e-01 -2.43075177e-01 -4.94296879e-01 4.46340233e-01
2.41024897e-01 3.93751869e-03 -4.52425689e-01 -7.14280605e-01
4.88989621e-01 -9.15838182e-01 -5.59011579e-01 3.32365602e-01
-1.84981629e-01 4.93125081e-01 6.40903771e-01 -6.48126245e-01
1.39257824e+00 -1.98265743e+00 -3.05763036e-01 1.93049118e-01
4.90107417e-01 1.32170513e-01 5.10957062e-01 -8.86010006e-02
-1.05280809e-01 1.28162295e-01 -4.81949747e-01 2.09840074e-01
-6.19253516e-01 9.57044438e-02 1.16779916e-01 6.35879159e-01
-2.67288387e-01 1.06036937e+00 -1.11324978e+00 -8.72622669e-01
3.49192023e-01 5.00201106e-01 -5.08275449e-01 8.70316550e-02
4.94626313e-01 9.31594908e-01 -5.44918418e-01 5.33432961e-01
1.09979737e+00 -7.13474452e-01 2.68578857e-01 -4.13041383e-01
-2.45411351e-01 -8.17977488e-02 -9.29833293e-01 1.48456407e+00
-4.10680264e-01 1.17134839e-01 4.13401239e-02 -3.13944906e-01
4.48187619e-01 3.83090675e-01 7.52504408e-01 -1.19453669e+00
3.24968010e-01 4.24859285e-01 2.04550236e-01 -7.51628876e-01
6.28828183e-02 -6.28587008e-01 1.75358966e-01 6.54765069e-01
-4.19302493e-01 -2.08355337e-01 9.50817540e-02 3.40533227e-01
9.03508246e-01 -4.58927304e-01 2.18252853e-01 -6.05800331e-01
5.76536477e-01 -7.73897022e-02 5.22749782e-01 5.69067955e-01
-4.31903541e-01 8.65019977e-01 3.40777814e-01 -2.33238921e-01
-1.17969882e+00 -1.12562561e+00 -4.50101703e-01 1.13225274e-01
6.70401096e-01 -1.67451710e-01 -4.29884464e-01 -3.94258767e-01
-4.82533127e-01 3.21859807e-01 -2.25939497e-01 -1.56923935e-01
-9.98032331e-01 -1.09116805e+00 3.09650511e-01 5.13372362e-01
9.25794840e-01 -7.27255464e-01 -1.02476919e+00 2.55699396e-01
-6.03297174e-01 -1.08819056e+00 -6.10825896e-01 -8.83545950e-02
-1.29984355e+00 -1.05936348e+00 -1.26744819e+00 -8.19552660e-01
6.07707679e-01 4.06078011e-01 1.19249094e+00 1.34193659e-01
-5.04077911e-01 4.13082913e-02 -2.26175174e-01 -1.38912769e-02
-5.18715918e-01 -6.01800799e-01 -3.91232848e-01 -3.58063549e-01
-5.24454951e-01 -5.74657917e-01 -1.34872580e+00 5.60934961e-01
-8.84182394e-01 3.89082044e-01 8.49540591e-01 1.03427315e+00
9.55601990e-01 -4.64724749e-03 6.24016970e-02 -9.17820275e-01
4.05777574e-01 -3.84686112e-01 -3.34391654e-01 2.79448628e-01
-6.86904430e-01 -1.69850886e-01 5.02156973e-01 -3.59238893e-01
-1.33467758e+00 -3.70171905e-01 -2.89475083e-01 -3.44845444e-01
1.88015565e-01 3.43287379e-01 2.05167502e-01 -2.52021641e-01
6.34392917e-01 5.84654808e-01 5.54988645e-02 -1.10925615e-01
-8.86541754e-02 1.67496383e-01 5.72898746e-01 -4.88429904e-01
6.16953790e-01 6.25636399e-01 3.92176509e-01 -5.27564228e-01
-4.25114840e-01 -3.01679164e-01 -1.92415312e-01 -4.85709578e-01
7.95679986e-01 -8.85172307e-01 -7.91130066e-01 5.38786888e-01
-7.05641627e-01 -1.45559520e-01 -2.15198547e-01 9.54141319e-01
-3.80824775e-01 8.62274647e-01 -1.14463317e+00 -3.46357495e-01
-7.29918838e-01 -1.55000138e+00 5.94563305e-01 2.72350013e-01
1.96306363e-01 -9.39288080e-01 -1.09156966e-01 2.40517601e-01
7.45189130e-01 4.74287599e-01 1.16971600e+00 2.07712263e-01
-8.95070195e-01 2.46637344e-01 -4.10184890e-01 1.26672283e-01
-2.26624170e-03 -4.94062513e-01 -4.08738405e-01 -4.27311420e-01
5.57890654e-01 2.63066031e-02 4.73927885e-01 8.32353175e-01
1.44035172e+00 4.53137532e-02 -1.91858098e-01 9.11721826e-01
1.65437555e+00 6.03182137e-01 8.49900782e-01 2.02802762e-01
4.57167476e-01 -4.52369340e-02 7.63575077e-01 5.05594194e-01
1.55205041e-01 3.39580506e-01 4.73386407e-01 -4.33349341e-01
-3.63815248e-01 -1.91013455e-01 -2.78846502e-01 1.23094165e+00
-2.50238210e-01 -1.61483824e-01 -1.01623535e+00 3.40130955e-01
-1.26243031e+00 -6.20201588e-01 -6.32580936e-01 2.03966355e+00
7.40370154e-01 7.39752222e-03 -2.17397958e-01 1.18625730e-01
7.54162610e-01 -1.58580113e-02 -3.47910672e-01 6.12149015e-02
5.20458072e-02 3.44182998e-01 6.24066651e-01 3.25916886e-01
-7.64972270e-01 -1.11749150e-01 6.44488621e+00 9.20276463e-01
-1.32168388e+00 4.17490155e-01 1.01261294e+00 2.68847812e-02
-4.55290258e-01 -2.30657607e-02 -8.99381191e-02 7.48789608e-01
4.39894170e-01 -3.01056296e-01 -2.18778849e-01 4.33986753e-01
2.68357068e-01 -6.98176563e-01 -5.11809349e-01 1.15157497e+00
-2.33376071e-01 -1.39316607e+00 8.69527981e-02 2.41288319e-01
4.81412441e-01 -3.49505544e-01 1.48825631e-01 -1.01446755e-01
-1.97592080e-01 -9.43009973e-01 5.27083337e-01 2.79022694e-01
1.18072248e+00 -7.74321795e-01 7.47717977e-01 2.91850030e-01
-1.11642444e+00 3.10593873e-01 -1.10271603e-01 4.85652447e-01
6.16960287e-01 7.54495084e-01 -7.09588945e-01 1.02322698e+00
6.68390036e-01 6.14814043e-01 -3.80818516e-01 1.59310067e+00
-9.94991660e-02 5.56645334e-01 -1.76200673e-01 1.57544941e-01
2.08279327e-01 -2.36743689e-01 6.71837926e-01 1.14577210e+00
4.71348256e-01 6.66441917e-01 9.53720659e-02 5.84284425e-01
3.30834150e-01 2.51607984e-01 -1.35537669e-01 7.34467447e-01
3.29709589e-01 1.09862912e+00 -1.02380323e+00 -5.43773532e-01
-2.26808831e-01 7.42369294e-01 -4.40831065e-01 1.27314731e-01
-1.24532104e+00 5.17451167e-02 -1.92491904e-01 6.00292981e-01
-1.18853651e-01 -6.20218329e-02 -4.47403789e-01 -1.13153732e+00
1.37786925e-01 -5.96065223e-01 5.05947471e-01 -8.94726753e-01
-1.02362537e+00 7.70589709e-01 1.80228814e-01 -1.78799677e+00
2.56981730e-01 -2.65240848e-01 -6.36569142e-01 7.85938621e-01
-1.64776921e+00 -7.17684209e-01 -7.36357689e-01 5.51048160e-01
2.40077808e-01 4.04323757e-01 4.86456424e-01 7.30873585e-01
-3.02908093e-01 2.51286179e-01 -6.66478351e-02 -8.85372832e-02
5.69084048e-01 -1.15080178e+00 -9.50281993e-02 7.85520375e-01
-6.33984566e-01 4.99533206e-01 5.18413961e-01 -7.15527594e-01
-1.24393165e+00 -8.81595075e-01 2.52422810e-01 1.29781231e-01
2.47871488e-01 3.17906857e-01 -9.30267215e-01 2.41197765e-01
3.01956058e-01 4.39328074e-01 6.16502881e-01 -6.21065676e-01
2.03010619e-01 1.34959696e-02 -1.34729898e+00 4.33452129e-01
7.55126178e-01 5.35655469e-02 -1.89460143e-01 1.43769816e-01
5.34715772e-01 -1.11709499e+00 -1.22591627e+00 6.64661407e-01
4.14932311e-01 -1.10894704e+00 1.10511053e+00 1.96722940e-01
8.13776314e-01 -4.28847879e-01 2.44129851e-01 -1.11060250e+00
-4.77625519e-01 -3.39689195e-01 1.02110498e-01 4.70400900e-01
8.42788815e-02 -6.74146950e-01 6.45937026e-01 2.34385729e-01
-3.98030370e-01 -1.07477748e+00 -8.95456374e-01 -6.03713870e-01
1.69068485e-01 -1.62453115e-01 2.17241213e-01 8.28388453e-01
-1.89365283e-01 -2.26053279e-02 -2.78152883e-01 1.92308843e-01
1.05137908e+00 1.75337687e-01 1.67573258e-01 -6.85871899e-01
-3.77707779e-01 -2.41078198e-01 -2.08596632e-01 -1.23642087e+00
-8.31278920e-01 -7.94219315e-01 -2.54460782e-01 -1.63442028e+00
6.93565607e-01 -8.45959604e-01 -5.86598143e-02 -2.44275182e-02
-2.34448150e-01 3.07115674e-01 8.09236839e-02 3.88284862e-01
-3.86900842e-01 4.58174318e-01 2.38711405e+00 4.51030806e-02
2.08422348e-01 1.89881232e-02 -3.82628918e-01 5.71673930e-01
4.03590918e-01 -5.32710373e-01 -5.10654867e-01 -3.53176594e-01
-5.26534207e-03 8.65029216e-01 4.90626693e-01 -1.18430281e+00
3.28964561e-01 -1.33873075e-01 6.82136178e-01 -7.91908085e-01
6.73478991e-02 -7.00475395e-01 5.09676874e-01 1.23532414e+00
1.70653999e-01 7.58948699e-02 1.35724783e-01 5.89035034e-01
-4.36545402e-01 -7.66865835e-02 1.31209612e+00 -4.39716101e-01
-4.49859381e-01 3.78572017e-01 -3.96718115e-01 2.14045882e-01
1.14026546e+00 -2.17125878e-01 -2.08282873e-01 -2.50173271e-01
-8.18113983e-01 3.05781037e-01 2.25102305e-01 -1.35149121e-01
8.12481403e-01 -1.37005007e+00 -8.23684812e-01 3.56741965e-01
-1.25498116e-01 2.93579608e-01 1.00916755e+00 1.63418579e+00
-1.28064561e+00 2.66433984e-01 -2.90373057e-01 -1.24223661e+00
-1.13861287e+00 1.86720595e-01 5.19172847e-01 -7.42846549e-01
-1.08158195e+00 6.08895898e-01 4.48600173e-01 1.20892279e-01
-3.63520771e-01 -3.99672389e-01 1.07342847e-01 -3.85365129e-01
3.98586154e-01 2.87070304e-01 2.45396048e-01 -5.03177106e-01
-3.09608847e-01 6.69585109e-01 -9.84681770e-02 2.75197830e-02
1.13538158e+00 -3.39222491e-01 1.35477513e-01 -2.85059512e-02
9.11845565e-01 -2.05546841e-02 -1.19909990e+00 -1.69940531e-01
-4.96050447e-01 -9.86718416e-01 1.02946609e-01 -9.17346895e-01
-1.27941799e+00 7.96256423e-01 7.69638956e-01 -7.89007470e-02
1.34482849e+00 -1.54614255e-01 9.98672664e-01 -6.01468027e-01
6.50771976e-01 -3.39272588e-01 4.45941329e-01 2.05674455e-01
6.90354705e-01 -9.74497676e-01 3.90665323e-01 -9.58671212e-01
-9.83719707e-01 9.56407309e-01 4.99288529e-01 -1.76474378e-01
6.84136987e-01 6.04240119e-01 3.19321789e-02 -2.51259238e-01
-5.82695305e-01 2.91993678e-01 7.23338351e-02 2.67408580e-01
5.23890316e-01 6.42995164e-02 -6.68070078e-01 2.38394856e-01
1.11032166e-01 -9.87490416e-02 6.64417267e-01 1.11791158e+00
-3.20013553e-01 -6.71671331e-01 -4.00268346e-01 2.85541713e-01
-7.00905025e-01 -4.67730239e-02 6.76780403e-01 8.26185703e-01
-1.68816328e-01 5.12535572e-01 -2.24186361e-01 1.56927437e-01
4.64087784e-01 -6.20826602e-01 6.68793201e-01 -3.71865720e-01
-6.68922186e-01 6.43887460e-01 -3.49952579e-01 -5.57871282e-01
-4.71051097e-01 -3.79155517e-01 -1.43612063e+00 -4.50950503e-01
-1.46124884e-01 8.53329450e-02 5.17808139e-01 6.08319461e-01
2.50060037e-02 8.95524561e-01 7.37363875e-01 -2.69571662e-01
-2.43019238e-01 -7.15332031e-01 -6.20615840e-01 3.25039089e-01
1.70100152e-01 -4.70879138e-01 -2.14403644e-01 -1.42207786e-01] | [13.553671836853027, -2.520820140838623] |
67d7376a-9b8d-426b-94ff-608520fcde0e | multi-level-graph-convolutional-networks-for | 2006.01963 | null | https://arxiv.org/abs/2006.01963v1 | https://arxiv.org/pdf/2006.01963v1.pdf | Multi-level Graph Convolutional Networks for Cross-platform Anchor Link Prediction | Cross-platform account matching plays a significant role in social network analytics, and is beneficial for a wide range of applications. However, existing methods either heavily rely on high-quality user generated content (including user profiles) or suffer from data insufficiency problem if only focusing on network topology, which brings researchers into an insoluble dilemma of model selection. In this paper, to address this problem, we propose a novel framework that considers multi-level graph convolutions on both local network structure and hypergraph structure in a unified manner. The proposed method overcomes data insufficiency problem of existing work and does not necessarily rely on user demographic information. Moreover, to adapt the proposed method to be capable of handling large-scale social networks, we propose a two-phase space reconciliation mechanism to align the embedding spaces in both network partitioning based parallel training and account matching across different social networks. Extensive experiments have been conducted on two large-scale real-life social networks. The experimental results demonstrate that the proposed method outperforms the state-of-the-art models with a big margin. | ['Hongxu Chen', 'Hongzhi Yin', 'Tong Chen', 'Katarzyna Musial', 'Bogdan Gabrys', 'Xiangguo Sun'] | 2020-06-02 | null | null | null | null | ['anchor-link-prediction'] | ['graphs'] | [ 7.08070919e-02 -1.81335777e-01 -3.94264877e-01 -2.68391341e-01
6.50115311e-02 -3.48529071e-01 4.72373724e-01 3.93018037e-01
-3.55438828e-01 5.24874210e-01 1.87969014e-01 -3.01275313e-01
-5.82996070e-01 -1.10474288e+00 -1.01817302e-01 -3.80655140e-01
8.08645412e-03 5.10762334e-01 2.49542877e-01 -2.98137546e-01
1.89457834e-01 1.63573936e-01 -1.13524294e+00 -1.79496571e-01
1.10996282e+00 5.79744875e-01 -8.33055824e-02 2.12463960e-01
-3.71328831e-01 2.80098617e-01 -1.78675056e-01 -9.55949306e-01
2.58904487e-01 -3.76586914e-01 -6.89653754e-01 1.15749210e-01
1.67877585e-01 -4.30732034e-03 -4.79429573e-01 1.38310993e+00
6.32345140e-01 6.87430501e-02 4.97583777e-01 -1.54656708e+00
-7.55515218e-01 7.98515320e-01 -8.86832237e-01 2.75404483e-01
1.47585139e-01 2.54505798e-02 1.41286480e+00 -5.09550035e-01
4.45068210e-01 1.21438253e+00 8.40812922e-01 2.98966199e-01
-1.30103934e+00 -8.92260492e-01 3.70411098e-01 2.80780017e-01
-1.23849201e+00 -1.68255642e-01 1.14353335e+00 -5.22617757e-01
6.10804915e-01 2.12843016e-01 9.73604977e-01 8.90171587e-01
-2.42931828e-01 5.41707218e-01 8.33579659e-01 -2.08754674e-01
-7.89023936e-02 3.70376557e-01 3.30961645e-01 6.65838003e-01
5.89855969e-01 -4.18844551e-01 -2.79070646e-01 -5.39464056e-01
7.57568955e-01 2.97895163e-01 -3.75146896e-01 -6.30973399e-01
-1.30140460e+00 1.00248647e+00 7.31367469e-01 4.80265170e-01
-2.04283670e-01 -2.78845966e-01 5.92743814e-01 2.97950923e-01
6.12074435e-01 3.59285504e-01 -2.28757843e-01 1.46730497e-01
-6.94889843e-01 4.47950847e-02 8.31157625e-01 7.56424725e-01
7.23803103e-01 -2.66017079e-01 2.31024802e-01 1.05243385e+00
6.23011649e-01 7.53512904e-02 5.54047942e-01 -3.78366172e-01
7.93409288e-01 1.22358060e+00 -3.29531938e-01 -1.70053923e+00
-4.25645262e-01 -7.31548071e-01 -1.32721317e+00 -5.39466083e-01
4.07795489e-01 -1.12295754e-01 -3.16946149e-01 1.79411972e+00
5.85502386e-01 4.84229445e-01 -2.90895343e-01 8.37984681e-01
7.53026783e-01 4.37292725e-01 5.96640743e-02 -1.69795170e-01
1.33045149e+00 -1.04189825e+00 -6.86758995e-01 -1.94867760e-01
5.80669940e-01 -5.66201091e-01 1.04509962e+00 -9.81603414e-02
-8.65570366e-01 -4.86174315e-01 -1.12687540e+00 2.43275046e-01
-4.22661483e-01 -1.66228428e-01 9.24594104e-01 6.81966603e-01
-1.18851340e+00 6.98457658e-01 -2.78555185e-01 -8.03198814e-01
4.16970521e-01 4.77999657e-01 -4.36413914e-01 -3.61402407e-02
-1.62393498e+00 4.63796765e-01 3.64989817e-01 3.13529968e-01
1.44973606e-01 -5.60777247e-01 -7.46501029e-01 3.16521108e-01
2.01175585e-01 -8.87224793e-01 7.67491281e-01 -9.21026230e-01
-1.08390105e+00 7.35329151e-01 1.87389016e-01 -8.36088434e-02
5.74391365e-01 2.84820527e-01 -7.47454464e-01 -1.18868783e-01
3.05141732e-02 -7.96752125e-02 3.71677279e-01 -9.32700634e-01
-2.47730151e-01 -6.25785053e-01 2.93305755e-01 3.06351691e-01
-1.22906923e+00 3.13060246e-02 -7.03603506e-01 -5.74698865e-01
3.07860672e-01 -9.82359469e-01 -3.42536062e-01 -1.20816603e-01
-5.62869132e-01 -2.14312613e-01 4.14619416e-01 -3.63426000e-01
1.65001011e+00 -1.79597557e+00 2.38206849e-01 5.04549623e-01
6.73882842e-01 4.11123216e-01 -6.52595907e-02 5.83786666e-01
-1.74318194e-01 2.85883904e-01 -2.89944351e-01 -5.54465175e-01
6.87759817e-02 -2.32332483e-01 2.05329537e-01 4.51874644e-01
-2.49587819e-01 7.97902346e-01 -1.07895291e+00 -8.46130967e-01
-5.44183739e-02 4.49088722e-01 -6.14422977e-01 6.84273466e-02
2.95315921e-01 2.85849839e-01 -5.91391206e-01 3.39027971e-01
8.34586978e-01 -7.42291451e-01 7.37773597e-01 -2.27390394e-01
4.24432188e-01 8.53249282e-02 -1.25132036e+00 1.66007781e+00
-1.75401449e-01 3.34730417e-01 1.45699114e-01 -9.35325861e-01
9.12330925e-01 1.17697760e-01 7.37909377e-01 -1.88461587e-01
2.40300193e-01 2.45767117e-01 2.51141787e-01 -3.31012249e-01
4.26131874e-01 1.16369431e-03 7.45981187e-02 7.25518942e-01
-1.10196903e-01 5.08323312e-01 2.61667728e-01 3.84578764e-01
9.83292460e-01 -3.44669163e-01 3.59925389e-01 -3.60783160e-01
8.25201988e-01 -5.18085182e-01 6.69832051e-01 2.53792495e-01
-4.69319701e-01 3.96039218e-01 6.93172395e-01 -2.79755384e-01
-8.43660235e-01 -6.53796494e-01 -4.61642891e-02 8.25716257e-01
3.26566905e-01 -5.45967937e-01 -7.11273611e-01 -9.13517296e-01
2.07590312e-01 -6.15273640e-02 -5.92185855e-01 -8.48286673e-02
-5.62040806e-01 -1.32286060e+00 2.64773905e-01 2.82758832e-01
6.21498346e-01 -1.05955136e+00 5.57862401e-01 2.70921946e-01
-2.97535539e-01 -1.06853855e+00 -7.27342427e-01 -5.99456489e-01
-8.17661762e-01 -1.36297703e+00 -6.82069004e-01 -8.30545008e-01
7.83586025e-01 6.77964568e-01 1.14484107e+00 6.34866536e-01
1.39443815e-01 -1.95756368e-03 -2.54395783e-01 1.12337396e-01
-1.06337152e-01 7.90523767e-01 1.51327923e-01 5.43483317e-01
5.80730140e-01 -1.12099040e+00 -7.76964247e-01 4.81402189e-01
-8.08048546e-01 4.77314256e-02 5.90242326e-01 6.66972816e-01
-2.79988404e-02 1.36199281e-01 1.03660703e+00 -1.39810991e+00
1.09339881e+00 -9.31925952e-01 -3.14463824e-01 2.40838811e-01
-1.07064974e+00 -2.19641343e-01 5.94927728e-01 -4.49351728e-01
-6.67902231e-01 -3.49040598e-01 4.36443798e-02 -1.79411829e-01
3.35469812e-01 7.55876184e-01 -4.91207868e-01 2.43231189e-02
2.77609646e-01 9.00402442e-02 4.93072271e-02 -6.00400567e-01
1.86666727e-01 7.82081485e-01 -3.73422019e-02 -3.31508607e-01
9.58362758e-01 2.59851217e-01 -1.42628819e-01 -4.47450697e-01
-4.80534494e-01 -6.60712063e-01 -7.57034540e-01 -1.22845411e-01
6.30798817e-01 -7.72056103e-01 -7.99550593e-01 6.92319214e-01
-1.00176549e+00 1.54078677e-01 2.60990471e-01 4.60999876e-01
2.50676107e-02 7.19253421e-01 -7.71923840e-01 -5.81452489e-01
-4.58070427e-01 -9.69065011e-01 3.29525173e-01 2.72604644e-01
-8.08305442e-02 -1.42444551e+00 2.39044532e-01 5.80143034e-01
5.24187684e-01 2.23723173e-01 9.23658133e-01 -8.58523369e-01
-6.35037482e-01 -2.30611250e-01 -8.05384636e-01 7.19674826e-02
3.57395351e-01 -1.02792270e-01 -7.29771972e-01 -5.24399042e-01
-3.78057331e-01 7.22562820e-02 9.04199898e-01 -1.60916105e-01
1.04886317e+00 -2.81906933e-01 -4.63075429e-01 5.11345088e-01
1.44089317e+00 -4.42693204e-01 4.47519243e-01 2.71148175e-01
1.15713596e+00 8.16206157e-01 2.15109348e-01 3.91941607e-01
8.30472589e-01 6.86674178e-01 5.26071370e-01 -2.21006975e-01
1.38612404e-01 -4.41434562e-01 -1.11568168e-01 1.33758390e+00
2.56108213e-02 -2.82195807e-01 -7.50588417e-01 6.55340731e-01
-2.26343632e+00 -8.16637337e-01 -4.37537044e-01 2.07779264e+00
6.73447073e-01 2.67732531e-01 4.22508061e-01 2.19805673e-01
1.15652072e+00 5.40635288e-01 -5.38494349e-01 -9.11326930e-02
4.47008014e-02 -3.71060938e-01 4.05883253e-01 3.72284919e-01
-1.00545204e+00 5.26338160e-01 4.97234821e+00 6.71240389e-01
-7.57294834e-01 2.11013436e-01 3.44133496e-01 1.88171118e-01
-5.12175798e-01 1.36859240e-02 -4.68287021e-01 8.90410066e-01
6.78178072e-01 -2.90077001e-01 4.25651878e-01 5.86485624e-01
-2.29278170e-02 4.68708754e-01 -8.16289008e-01 1.03447950e+00
-1.26928361e-02 -1.20957828e+00 -2.18604039e-02 4.06045824e-01
7.13090420e-01 -9.61112827e-02 -1.20998695e-01 2.50903010e-01
1.21479765e-01 -7.99218774e-01 5.79547584e-02 3.85869682e-01
5.83337903e-01 -7.47672558e-01 9.13239658e-01 3.28800589e-01
-1.59458971e+00 -9.35297087e-02 -2.43055925e-01 -5.75292371e-02
7.11748973e-02 6.79164469e-01 -6.92484617e-01 7.80352592e-01
3.46162766e-01 1.02132332e+00 -6.52076244e-01 1.31190252e+00
-4.61132489e-02 4.20621425e-01 -1.31453916e-01 -1.67893782e-01
4.13951129e-02 -4.80509400e-01 3.19089651e-01 9.61207211e-01
1.15796380e-01 -1.36112481e-01 2.59572893e-01 5.53632736e-01
-4.63898689e-01 5.89562476e-01 -5.76248109e-01 -5.73327430e-02
6.23258412e-01 1.53204215e+00 -7.34051347e-01 -1.19220041e-01
-6.18845403e-01 8.30284774e-01 5.68822503e-01 2.72746205e-01
-7.81455159e-01 -5.05354702e-01 7.54675865e-01 6.07898057e-01
-1.80029988e-01 -5.15686199e-02 -2.53885817e-02 -1.43344474e+00
1.07608393e-01 -6.83590770e-01 5.93184054e-01 -3.40749055e-01
-1.78502738e+00 6.04685068e-01 -2.70159632e-01 -1.33713078e+00
1.22037992e-01 -3.30004543e-01 -8.95115137e-01 8.75424385e-01
-1.66990018e+00 -1.32690263e+00 -4.44195569e-01 5.89440346e-01
2.25359332e-02 -1.94137216e-01 7.94905603e-01 1.01427484e+00
-8.80542994e-01 9.44246650e-01 1.41949013e-01 4.11108255e-01
7.66415179e-01 -1.10410297e+00 5.62828183e-01 5.81758976e-01
-1.64984033e-01 8.17690194e-01 3.82505745e-01 -6.11490071e-01
-1.16380692e+00 -1.08743298e+00 1.09397101e+00 -1.64888844e-01
9.06650305e-01 -4.37758118e-01 -1.06481338e+00 5.75004518e-01
1.46063000e-01 4.06040438e-02 8.63802850e-01 6.44461095e-01
-3.68225157e-01 -3.85144502e-01 -1.25298786e+00 6.09972179e-01
1.37728655e+00 -5.06178141e-01 -1.53472081e-01 4.61439610e-01
6.32870257e-01 4.07193638e-02 -9.90168750e-01 2.44776681e-01
5.27318716e-01 -9.46898103e-01 9.44572210e-01 -7.46169925e-01
1.72004491e-01 -2.09020227e-01 1.98028028e-01 -1.33360970e+00
-5.60344875e-01 -7.95796752e-01 -1.40499234e-01 1.55651343e+00
5.58726370e-01 -1.12976408e+00 1.02831435e+00 5.81184566e-01
4.40536708e-01 -8.67568016e-01 -6.05360806e-01 -3.99882048e-01
-2.16170654e-01 9.95637774e-02 1.08102679e+00 1.51298571e+00
2.71313399e-01 7.48207808e-01 -4.66947794e-01 5.71193695e-02
6.71378613e-01 2.98212111e-01 8.62484455e-01 -1.75968850e+00
-2.59651929e-01 -6.56372190e-01 -5.52226722e-01 -7.77012169e-01
2.38719374e-01 -1.10834455e+00 -6.68416023e-01 -1.76369238e+00
6.07670248e-01 -8.17164898e-01 -6.72560871e-01 1.24646217e-01
-4.48183894e-01 2.88637370e-01 2.02343807e-01 4.81164664e-01
-6.87433004e-01 5.82046986e-01 1.21711671e+00 -1.17696173e-01
-1.31756127e-01 1.52631387e-01 -1.07874703e+00 6.38755143e-01
8.76984060e-01 -3.08696449e-01 -8.18645537e-01 -3.93326372e-01
5.89135349e-01 -6.17325678e-02 9.95679051e-02 -7.33886659e-01
3.64166051e-01 -7.71133453e-02 7.13405246e-03 -7.34104365e-02
2.55960971e-01 -9.17915821e-01 1.48487329e-01 2.39027396e-01
-1.38855681e-01 3.48380476e-01 -2.61191666e-01 9.51628923e-01
-1.05734110e-01 -1.25378445e-02 6.67936504e-01 -7.04727247e-02
-2.96951741e-01 7.56794930e-01 3.79807293e-01 6.16347753e-02
8.39012504e-01 -1.49218261e-01 -4.50002939e-01 -4.16643947e-01
-5.49771130e-01 5.14418960e-01 6.81591570e-01 5.94775736e-01
4.30114329e-01 -1.66550946e+00 -7.95222938e-01 3.68143708e-01
2.77611911e-01 -2.91659385e-01 4.51702505e-01 9.60307539e-01
-1.55900866e-01 7.38010854e-02 -4.78934534e-02 -3.56920868e-01
-1.12683856e+00 4.51667637e-01 1.38423666e-01 -7.43371248e-01
-3.37930292e-01 7.36709833e-01 -1.03066526e-01 -8.98501158e-01
2.53013633e-02 2.18893349e-01 -5.34314752e-01 4.09791619e-01
3.19908500e-01 2.77697593e-01 -1.25430385e-02 -8.84840608e-01
-3.23279947e-01 6.05930209e-01 -1.61554053e-01 2.82673389e-01
1.27231634e+00 -2.78872758e-01 -2.89564341e-01 2.57741928e-01
1.43065548e+00 1.89253673e-01 -7.09210694e-01 -7.26079166e-01
-1.79237828e-01 -7.30648458e-01 -2.20156446e-01 -2.23334193e-01
-1.40079546e+00 8.23643386e-01 2.41267920e-01 8.12370420e-01
8.59374404e-01 -3.32004070e-01 1.20047069e+00 2.73714960e-02
3.29066098e-01 -8.24850917e-01 2.57963147e-02 1.90713391e-01
4.24213797e-01 -1.37665248e+00 2.35892236e-01 -6.54174685e-01
-4.51921403e-01 7.75105476e-01 8.46693635e-01 -2.60682032e-02
1.15018892e+00 -4.85290229e-01 -2.50522971e-01 -2.07143873e-01
-3.46865296e-01 -1.78371862e-01 2.05781683e-01 3.65037262e-01
3.78020465e-01 1.17231965e-01 -4.50324714e-01 6.43746376e-01
-2.79464386e-02 -3.24043125e-01 3.94788861e-01 6.11311793e-01
-1.10824585e-01 -1.48249972e+00 1.50644749e-01 6.82310879e-01
-3.12905043e-01 -1.54052928e-01 -4.74602252e-01 7.36437201e-01
1.27471024e-02 7.84886301e-01 -2.04372823e-01 -7.00422883e-01
3.00832301e-01 -1.80387765e-01 1.27487347e-01 -5.76276183e-01
-6.83039784e-01 -2.81123430e-01 6.72647506e-02 -2.35271275e-01
-4.51868802e-01 -6.57636106e-01 -8.38396132e-01 -9.40536618e-01
-5.57542980e-01 2.48515651e-01 4.45523411e-01 6.15316033e-01
6.41593874e-01 3.86397392e-01 8.39269042e-01 -6.32816195e-01
-4.99865025e-01 -1.10040474e+00 -8.28154743e-01 5.93761027e-01
1.42766923e-01 -6.91050172e-01 -5.35060704e-01 -5.25822818e-01] | [7.375385284423828, 6.286040782928467] |
2a9f438b-5fb0-4747-a73f-7947a3bb964f | wikineural-combined-neural-and-knowledge | null | null | https://aclanthology.org/2021.findings-emnlp.215 | https://aclanthology.org/2021.findings-emnlp.215.pdf | WikiNEuRal: Combined Neural and Knowledge-based Silver Data Creation for Multilingual NER | Multilingual Named Entity Recognition (NER) is a key intermediate task which is needed in many areas of NLP. In this paper, we address the well-known issue of data scarcity in NER, especially relevant when moving to a multilingual scenario, and go beyond current approaches to the creation of multilingual silver data for the task. We exploit the texts of Wikipedia and introduce a new methodology based on the effective combination of knowledge-based approaches and neural models, together with a novel domain adaptation technique, to produce high-quality training corpora for NER. We evaluate our datasets extensively on standard benchmarks for NER, yielding substantial improvements up to 6 span-based F1-score points over previous state-of-the-art systems for data creation. | ['Roberto Navigli', 'Francesco Cecconi', 'Niccolò Campolungo', 'Valentino Maiorca', 'Simone Tedeschi'] | null | null | null | null | findings-emnlp-2021-11 | ['multilingual-named-entity-recognition', 'multilingual-nlp'] | ['natural-language-processing', 'natural-language-processing'] | [-2.29301482e-01 -1.67126057e-03 -2.09980514e-02 -2.38624603e-01
-1.44433069e+00 -8.91319454e-01 6.95703864e-01 3.54641020e-01
-1.12906027e+00 1.32417023e+00 6.89582765e-01 -3.01681906e-01
-3.90181318e-02 -6.24590456e-01 -7.72858083e-01 -1.35738507e-01
5.88608086e-02 7.72438288e-01 -9.15601198e-03 -6.66192055e-01
1.65513948e-01 6.02245390e-01 -1.18264627e+00 1.31943673e-01
1.22034800e+00 3.45758438e-01 3.12698446e-03 5.22065401e-01
-5.25945961e-01 6.82371974e-01 -8.65316212e-01 -7.68135607e-01
1.16452679e-01 -1.31412178e-01 -1.12304521e+00 -6.44616246e-01
4.73239422e-01 3.35000336e-01 -1.68096930e-01 8.30742538e-01
9.15920019e-01 3.41284305e-01 4.67771918e-01 -5.71240067e-01
-8.84692132e-01 1.07508898e+00 -1.70583092e-02 3.19645017e-01
3.96173686e-01 -2.12182462e-01 9.25123394e-01 -1.03704667e+00
1.19811141e+00 7.25496590e-01 1.01773214e+00 7.62255907e-01
-1.10721517e+00 -3.90029311e-01 -3.66463572e-01 1.56160325e-01
-1.27182913e+00 -7.87226796e-01 3.93043488e-01 -1.84220627e-01
1.42862606e+00 -1.47255778e-01 1.67631343e-01 1.11737025e+00
-2.08673552e-01 5.60078800e-01 1.10111201e+00 -9.71975684e-01
2.86425464e-02 1.32658482e-01 2.39979997e-02 1.47338688e-01
3.04618448e-01 9.57866237e-02 -4.34943289e-01 9.18077305e-03
5.81958771e-01 -7.05287337e-01 -1.04731761e-01 -2.14930221e-01
-1.52509677e+00 6.59943938e-01 2.65923411e-01 9.08569336e-01
-7.28199005e-01 -2.42370456e-01 5.74521303e-01 3.57082903e-01
4.51736510e-01 1.12003064e+00 -1.01846385e+00 -4.20608789e-01
-1.01814830e+00 1.85728326e-01 1.29453969e+00 9.12787557e-01
5.70675075e-01 9.05062407e-02 -2.93769747e-01 1.18205535e+00
-1.45379379e-01 4.18655604e-01 4.25359905e-01 -6.72643006e-01
9.66274858e-01 5.96618474e-01 3.88975531e-01 -5.14394522e-01
-5.36994100e-01 -2.40415767e-01 -6.08352840e-01 -2.54717112e-01
5.96880674e-01 -5.85076809e-01 -8.29063058e-01 1.68987548e+00
4.83672500e-01 -1.42018661e-01 6.51225686e-01 3.11233670e-01
8.91324878e-01 6.44704521e-01 4.79333699e-01 9.12374780e-02
1.24229228e+00 -9.70388234e-01 -7.01464951e-01 -1.86932217e-02
9.22875404e-01 -9.10883427e-01 7.80902088e-01 1.82211056e-01
-8.65417659e-01 -4.42625940e-01 -7.93505251e-01 -3.27832669e-01
-1.17426026e+00 3.66014838e-01 5.02017617e-01 6.55904233e-01
-1.09959400e+00 6.49297297e-01 -3.35578859e-01 -7.67044961e-01
2.18940862e-02 2.53817588e-01 -9.06304479e-01 -2.27607936e-01
-1.65589070e+00 1.36635840e+00 9.42213655e-01 1.67806312e-01
-4.98386711e-01 -8.39835286e-01 -9.99152958e-01 -2.09002078e-01
2.56098062e-01 -4.53612536e-01 1.28654027e+00 -3.34595829e-01
-1.28225100e+00 8.68102133e-01 4.98355739e-02 -6.07954025e-01
3.87726992e-01 -5.28267324e-01 -9.60949838e-01 -2.06263006e-01
3.38430583e-01 7.91039467e-01 5.69189899e-02 -1.05542028e+00
-7.92115867e-01 -2.47186702e-03 -3.75034027e-02 2.27057293e-01
-4.70925093e-01 4.95127141e-01 -2.40574211e-01 -6.36449099e-01
-7.53062248e-01 -6.77976131e-01 -5.60966790e-01 -1.01590800e+00
-4.23571646e-01 -3.79901767e-01 2.73426145e-01 -1.25263309e+00
1.42782104e+00 -1.64678061e+00 5.87957026e-03 -1.34054050e-01
-1.56717137e-01 6.05680168e-01 -4.11683053e-01 8.25766981e-01
-1.99072734e-01 3.71311098e-01 -3.34710270e-01 -6.21068068e-02
8.70761424e-02 1.01085618e-01 -9.54474881e-02 -1.37554213e-01
4.50474977e-01 9.50704932e-01 -1.08986127e+00 -4.35456067e-01
6.03515981e-03 6.16060078e-01 -1.92828223e-01 2.40077898e-01
-6.75213635e-02 6.36120498e-01 -1.86938033e-01 4.72879678e-01
4.58340168e-01 3.45019966e-01 1.49798587e-01 -1.85836896e-01
-6.45247400e-01 6.10220671e-01 -1.27249753e+00 1.97665143e+00
-7.59991646e-01 4.00395602e-01 -3.35118711e-01 -5.93538821e-01
1.03860247e+00 7.64113903e-01 2.50571847e-01 -6.97333694e-01
3.06479502e-02 6.10115051e-01 -3.43273699e-01 -3.57822090e-01
1.03587151e+00 -3.25317495e-02 -6.37102664e-01 2.58225590e-01
6.72983944e-01 2.25975841e-01 9.10433233e-01 -5.91759011e-02
1.20264387e+00 3.23043138e-01 6.76936448e-01 -3.45047325e-01
5.32891333e-01 5.01545966e-01 5.55432379e-01 6.43960893e-01
-1.33950725e-01 3.46439272e-01 9.35711339e-02 -4.84888285e-01
-1.47272015e+00 -7.04003513e-01 -2.01069430e-01 1.15503860e+00
-6.46599174e-01 -3.09256557e-02 -9.90618467e-01 -1.17702472e+00
-3.31322491e-01 1.04435194e+00 -4.86273646e-01 3.19273889e-01
-1.03044558e+00 -8.05826902e-01 1.25934422e+00 4.87189680e-01
3.96875650e-01 -1.57531512e+00 -6.20948337e-02 6.71871722e-01
-4.64202881e-01 -1.41743016e+00 -3.53190035e-01 3.63295406e-01
-7.08737195e-01 -8.67995322e-01 -1.01699615e+00 -9.36110318e-01
3.59548181e-01 -2.50688106e-01 1.52042842e+00 -3.73132795e-01
-1.20691285e-01 2.78172523e-01 -5.46037078e-01 -4.21691328e-01
-7.95044422e-01 8.64183545e-01 -1.10743955e-01 -4.92816389e-01
7.45662153e-01 -2.57658750e-01 -1.57393087e-02 1.18806511e-02
-1.00975072e+00 -3.70502234e-01 8.44084859e-01 8.80708218e-01
5.53226948e-01 -5.96129112e-02 1.05299914e+00 -1.10817337e+00
7.94644475e-01 -6.29135191e-01 -4.02692884e-01 6.15993202e-01
-3.83791953e-01 3.09426457e-01 7.98915923e-01 -2.17359304e-01
-1.41824985e+00 1.87154248e-01 -6.00548923e-01 2.18518108e-01
-6.11065328e-01 7.46083677e-01 -3.77619654e-01 -7.56425187e-02
1.03923416e+00 1.03780560e-01 -8.96700799e-01 -9.06711519e-01
8.81445587e-01 8.86747420e-01 7.29753315e-01 -7.58695304e-01
5.67547858e-01 -5.15625253e-02 -3.51933867e-01 -7.98373282e-01
-9.04293835e-01 -6.63741291e-01 -1.12280869e+00 1.65661544e-01
7.74047315e-01 -1.04563200e+00 1.25050336e-01 3.86508137e-01
-1.42505229e+00 -2.23522767e-01 -4.72549528e-01 4.81923997e-01
-3.09932441e-01 8.44529271e-02 -6.89261973e-01 -3.59948277e-01
-7.42745221e-01 -6.15145326e-01 8.13321471e-01 4.59026664e-01
-6.37287647e-02 -1.36857009e+00 8.27163517e-01 3.78000468e-01
5.16596675e-01 2.85161644e-01 7.75486529e-01 -1.39046133e+00
9.21986103e-02 -1.79832831e-01 -8.81060734e-02 4.14362401e-01
-9.53240842e-02 -2.67409831e-01 -8.30536783e-01 2.92116171e-03
-5.52554607e-01 -3.50701451e-01 5.35516441e-01 -1.90216750e-01
1.98951647e-01 -2.09522039e-01 -1.42597079e-01 2.21457884e-01
1.66138458e+00 1.68001503e-02 7.51505792e-01 8.49898100e-01
7.15579629e-01 6.42983913e-01 4.61617827e-01 2.14386091e-01
7.27987051e-01 5.38466513e-01 -2.14949653e-01 -3.32251757e-01
-3.39652896e-01 -4.49748963e-01 1.49104252e-01 1.12268329e+00
-2.24805370e-01 -2.73649246e-01 -1.35324728e+00 1.07742345e+00
-1.52011144e+00 -8.62498105e-01 -1.92377105e-01 2.17228127e+00
1.29439902e+00 -2.11810723e-01 -9.23973471e-02 -2.20372513e-01
9.74362493e-01 -1.67586207e-01 -2.40639016e-01 -5.11132240e-01
-3.60503912e-01 6.21713877e-01 6.82984769e-01 8.18366110e-02
-1.39109504e+00 1.43029249e+00 6.76608562e+00 7.80743957e-01
-8.73836160e-01 2.75166631e-01 1.75074980e-01 4.66493219e-01
-4.88019660e-02 -6.78531602e-02 -1.43103063e+00 1.18342467e-01
1.54095519e+00 -2.53696352e-01 4.28253919e-01 8.43541503e-01
-1.60393044e-01 2.58840233e-01 -9.93176401e-01 5.35350084e-01
8.08851719e-02 -1.21412671e+00 6.28779829e-02 -9.33371633e-02
1.14489615e+00 5.71823299e-01 -5.94385505e-01 6.58531010e-01
9.06571150e-01 -8.01563978e-01 4.77306396e-01 6.16598547e-01
8.32027078e-01 -8.86972785e-01 1.10162258e+00 2.24021435e-01
-1.16168249e+00 1.81437761e-01 -3.19049805e-01 4.43061918e-01
4.44118232e-01 7.30893552e-01 -9.05623794e-01 1.07213950e+00
7.02437341e-01 5.32000601e-01 -4.76219177e-01 1.31313443e+00
-5.25986314e-01 4.02565628e-01 -3.17690253e-01 1.29329458e-01
1.68936491e-01 2.75526792e-01 5.02186358e-01 1.77356970e+00
4.42071378e-01 -4.77357775e-01 -9.44085047e-02 4.57482904e-01
-7.16669321e-01 7.22252846e-01 -8.41502607e-01 -3.86582285e-01
7.54446685e-01 1.50453579e+00 -3.94173741e-01 -3.23719501e-01
-4.48046565e-01 9.80780423e-01 9.33178961e-01 2.26517856e-01
-4.25912619e-01 -9.85348105e-01 1.82224050e-01 -3.12816352e-01
5.03860176e-01 -4.29017037e-01 -1.03436545e-01 -1.37529361e+00
3.08678318e-02 -1.02029824e+00 7.54214823e-01 -3.98352206e-01
-1.55666780e+00 9.57652688e-01 -2.00268000e-01 -1.02838612e+00
-4.57034290e-01 -4.38997597e-01 -4.03707504e-01 9.56464171e-01
-1.94084859e+00 -1.42410827e+00 9.15748104e-02 4.00707334e-01
3.67419809e-01 -9.38982964e-02 1.35189903e+00 8.04950595e-01
-4.56726223e-01 6.23329818e-01 4.66906101e-01 7.35800743e-01
1.22062957e+00 -1.49446332e+00 8.96583617e-01 1.13342977e+00
5.17876685e-01 7.42666245e-01 3.94978404e-01 -7.27763176e-01
-1.03103137e+00 -1.13353443e+00 1.92719734e+00 -8.55512202e-01
9.18405414e-01 -3.20453525e-01 -7.61377931e-01 6.23474061e-01
5.16886830e-01 -1.36580795e-01 8.67753386e-01 4.17845875e-01
-3.93938988e-01 1.60715073e-01 -1.23396564e+00 4.60592002e-01
9.54638243e-01 -4.37601745e-01 -1.23620868e+00 1.84378639e-01
6.51480138e-01 -4.56447631e-01 -1.56457150e+00 2.82334954e-01
2.41798550e-01 -4.32407916e-01 8.87308359e-01 -9.43098068e-01
1.12785064e-01 -3.07342619e-01 -4.80805039e-02 -1.74999881e+00
-5.41557483e-02 -5.90250969e-01 2.13757753e-01 1.88799620e+00
9.14503098e-01 -4.65904027e-01 2.82400310e-01 3.76766264e-01
-2.30140984e-01 -3.25896665e-02 -1.16816139e+00 -9.25275505e-01
4.86686051e-01 -2.45301932e-01 6.93410337e-01 1.44908619e+00
7.35089034e-02 5.78405440e-01 -4.48836565e-01 -2.08552405e-02
4.34408218e-01 -5.82211316e-01 6.84488237e-01 -1.32555902e+00
1.71793118e-01 -3.11193079e-01 -4.98641878e-02 -5.70585966e-01
2.35155270e-01 -9.87142980e-01 3.68433654e-01 -1.83391309e+00
-5.65370098e-02 -5.53130567e-01 -4.71669316e-01 7.08728135e-01
-1.88347504e-01 1.18672162e-01 1.35047451e-01 1.30901381e-01
-8.37456107e-01 2.75851458e-01 7.49609888e-01 3.18136394e-01
-3.49261671e-01 -5.70875704e-01 -7.63389885e-01 3.89912188e-01
6.57232106e-01 -8.80516589e-01 2.70710766e-01 -6.97299778e-01
4.97699708e-01 -1.56493351e-01 -2.21157953e-01 -1.21847916e+00
3.47196519e-01 1.13665208e-01 3.32581371e-01 -3.61197829e-01
-2.18993485e-01 -6.24037206e-01 4.55194600e-02 9.59890112e-02
-3.10554713e-01 2.01537833e-01 3.58313948e-01 2.26574406e-01
-4.03098792e-01 -4.33455616e-01 5.09589911e-01 -3.67411941e-01
-1.17512298e+00 -2.51076613e-02 -1.10277362e-01 5.90716541e-01
7.53389776e-01 2.74428368e-01 -4.94363993e-01 1.96899995e-01
-6.12703741e-01 1.99376687e-01 2.88137704e-01 6.24922514e-01
-2.78164353e-02 -1.35898113e+00 -1.12115479e+00 -1.81377456e-01
4.73183393e-01 -3.13120157e-01 2.12372199e-01 5.48466384e-01
-5.26644826e-01 7.08535135e-01 -4.19242293e-01 2.24716648e-01
-6.72525644e-01 3.54674697e-01 2.04287976e-01 -1.06273723e+00
-3.61620128e-01 4.21749592e-01 -6.05529904e-01 -1.40357161e+00
-4.60455604e-02 1.20987393e-01 -7.79745936e-01 2.25556225e-01
6.78732097e-01 5.71781516e-01 4.56039995e-01 -8.97540987e-01
-2.17991352e-01 3.44885111e-01 -1.94765329e-02 -3.04057389e-01
1.58811271e+00 -1.54447462e-02 -4.47535403e-02 3.76065999e-01
8.45622063e-01 3.31220716e-01 -5.34766316e-01 -4.38430578e-01
8.40988100e-01 1.88859656e-01 -2.03258306e-01 -1.47566366e+00
-4.50650334e-01 6.46166086e-01 3.45794082e-01 1.22881338e-01
8.43494356e-01 -3.30043933e-03 1.02929461e+00 8.05114388e-01
5.25081635e-01 -1.50483286e+00 -6.24213517e-01 1.10372829e+00
6.33804321e-01 -1.29129791e+00 -3.40249628e-01 -1.07052505e-01
-6.06879950e-01 1.21629930e+00 3.20518047e-01 -2.47399526e-04
3.85488391e-01 2.25340381e-01 4.15035576e-01 1.46867216e-01
-4.76239145e-01 -4.80557919e-01 3.49029332e-01 8.26454878e-01
7.29034901e-01 -2.84660477e-02 -6.18967712e-01 7.28079319e-01
-9.92413610e-02 2.93634474e-01 3.23041081e-01 9.88055825e-01
-3.06712300e-01 -1.67918289e+00 -2.99994022e-01 8.12712014e-02
-8.79212737e-01 -6.52733147e-01 -3.54506910e-01 1.05956531e+00
1.40592754e-01 9.21938956e-01 -2.89312601e-01 -1.35751665e-01
6.94786727e-01 4.73948658e-01 9.55109894e-02 -6.75072849e-01
-1.08097279e+00 -3.38728070e-01 8.57211769e-01 -1.58045694e-01
-6.88722908e-01 -7.71617770e-01 -1.20796418e+00 -2.89695822e-02
-2.18883887e-01 5.23113012e-01 1.15653515e+00 9.76332307e-01
6.00796640e-01 1.76565558e-01 3.24701995e-01 -5.90097904e-01
-4.29689080e-01 -1.23621094e+00 -2.01159567e-01 4.88090515e-01
-3.05421293e-01 -2.47736976e-01 -1.66237950e-01 2.35162973e-01] | [9.866133689880371, 9.722396850585938] |
4441e8bc-3565-4c89-940d-bb90fe344b1d | distilling-transformers-for-neural-cross | 2108.03322 | null | https://arxiv.org/abs/2108.03322v1 | https://arxiv.org/pdf/2108.03322v1.pdf | Distilling Transformers for Neural Cross-Domain Search | Pre-trained transformers have recently clinched top spots in the gamut of natural language tasks and pioneered solutions to software engineering tasks. Even information retrieval has not been immune to the charm of the transformer, though their large size and cost is generally a barrier to deployment. While there has been much work in streamlining, caching, and modifying transformer architectures for production, here we explore a new direction: distilling a large pre-trained translation model into a lightweight bi-encoder which can be efficiently cached and queried. We argue from a probabilistic perspective that sequence-to-sequence models are a conceptually ideal---albeit highly impractical---retriever. We derive a new distillation objective, implementing it as a data augmentation scheme. Using natural language source code search as a case study for cross-domain search, we demonstrate the validity of this idea by significantly improving upon the current leader of the CodeSearchNet challenge, a recent natural language code search benchmark. | ['Neel Sundaresan', 'Dawn Drain', 'Chen Wu', 'Colin B. Clement'] | 2021-08-06 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [ 2.14625746e-01 4.99541275e-02 -5.02279222e-01 -2.67264575e-01
-1.37214291e+00 -7.68860638e-01 6.96348011e-01 1.17305852e-01
-3.10699493e-01 3.45837086e-01 5.37262321e-01 -9.92818475e-01
1.66728515e-02 -6.18974209e-01 -9.06965196e-01 -2.18193814e-01
-6.16538227e-02 7.40327120e-01 2.97377914e-01 -4.27977204e-01
3.57795179e-01 1.67974398e-01 -1.53394687e+00 3.56125116e-01
8.38984132e-01 7.20502496e-01 3.27825546e-01 5.62289834e-01
-2.35255867e-01 9.29447770e-01 -3.27549279e-01 -5.82509816e-01
1.69825837e-01 -1.45341307e-01 -1.32517207e+00 -4.54991877e-01
4.13937539e-01 -1.16161972e-01 -2.70513326e-01 8.67229402e-01
4.14082348e-01 -3.41190308e-01 2.38434538e-01 -1.16604638e+00
-8.62103522e-01 8.91542435e-01 -4.93783951e-01 1.23266503e-01
5.20794034e-01 1.85518876e-01 1.32796192e+00 -6.80153072e-01
7.99177885e-01 9.81654346e-01 7.64656544e-01 4.24813777e-01
-1.35165679e+00 -4.65079337e-01 -2.55578339e-01 -1.55286361e-02
-1.18307996e+00 -5.97593188e-01 3.17788601e-01 -5.50758421e-01
1.62157500e+00 1.03441715e-01 2.81012893e-01 1.16482639e+00
2.98699260e-01 1.09816122e+00 5.81692338e-01 -5.15368700e-01
-4.20893393e-02 1.41682267e-01 -1.82161629e-01 1.01172602e+00
3.99309471e-02 1.15109332e-01 -6.23947740e-01 -6.75401986e-01
2.99110532e-01 -1.19560868e-01 -1.80012509e-01 -6.99340105e-01
-1.26271498e+00 7.18131125e-01 3.14044356e-01 4.25296873e-01
8.23948358e-04 4.63418186e-01 9.57788527e-01 7.33652890e-01
2.89868325e-01 8.32806408e-01 -6.44353509e-01 -6.82083070e-01
-1.05930543e+00 2.77603060e-01 1.07554805e+00 1.23262405e+00
6.20927870e-01 -1.86017692e-01 -6.78014681e-02 7.21615195e-01
1.43641546e-01 3.96124125e-01 7.02440262e-01 -6.14845872e-01
5.04866421e-01 7.14928567e-01 2.57355925e-02 -6.46514535e-01
1.01891935e-01 -3.95271212e-01 -3.01043868e-01 -6.96948245e-02
2.01420903e-01 3.10042202e-01 -6.27217114e-01 1.68556273e+00
-1.27155796e-01 -1.61899373e-01 -1.91391587e-01 7.09893465e-01
-4.42924015e-02 6.09570861e-01 4.06104364e-02 2.39082113e-01
1.30587995e+00 -1.02140594e+00 -6.96314275e-02 -3.42851490e-01
1.12399685e+00 -9.87164915e-01 1.36326933e+00 4.57033932e-01
-1.31444335e+00 -1.35025457e-01 -1.06158495e+00 -3.71210098e-01
-1.98361889e-01 -3.12257931e-02 9.42332327e-01 5.08060694e-01
-1.44219875e+00 5.29186606e-01 -9.36769962e-01 -5.63575268e-01
1.28933698e-01 2.56586343e-01 -4.14575487e-01 -1.04528017e-01
-9.41574633e-01 1.22354603e+00 7.41310790e-02 -4.84847099e-01
-9.74339128e-01 -8.59542310e-01 -5.54171443e-01 3.27302784e-01
3.36039186e-01 -6.94539011e-01 1.68334198e+00 -9.13565397e-01
-1.20071840e+00 1.03701782e+00 -1.08149759e-01 -5.02269506e-01
1.76026478e-01 -3.14373255e-01 -3.68865788e-01 -3.54626149e-01
2.55918354e-01 4.14178759e-01 6.84110820e-01 -7.61033654e-01
-5.42141795e-01 -1.27343386e-01 1.10903949e-01 -1.95499346e-01
-4.37533230e-01 5.97580135e-01 -3.89194340e-01 -5.66196263e-01
-3.82194102e-01 -8.59584868e-01 -6.71721995e-02 5.26182950e-02
-1.53024733e-01 -4.90069896e-01 5.38326323e-01 -5.14983177e-01
1.39974546e+00 -2.19077992e+00 2.25547567e-01 5.10971211e-02
2.21701965e-01 3.29862356e-01 -2.93226242e-01 8.15643489e-01
-2.13805065e-01 2.39922523e-01 -2.66053587e-01 -2.94320490e-02
4.33495015e-01 -9.41460133e-02 -7.18670666e-01 3.13078374e-01
3.41485590e-01 1.02545762e+00 -1.20183563e+00 -4.15638745e-01
-3.05046737e-01 3.14235657e-01 -7.49540269e-01 4.15126830e-02
-5.95770121e-01 -3.75420809e-01 -4.81724232e-01 6.02053165e-01
1.63995743e-01 -5.87568283e-01 2.41007805e-01 2.36822411e-01
-2.14940637e-01 9.37913060e-01 -5.64893961e-01 2.35009027e+00
-8.87780905e-01 7.83907831e-01 1.08197339e-01 -9.67017531e-01
7.85128951e-01 3.44254375e-01 4.90443647e-01 -8.68950069e-01
-1.23495303e-01 6.46300435e-01 -1.16853446e-01 -4.76491064e-01
7.93182075e-01 -4.61240038e-02 -3.37531269e-01 8.47151816e-01
3.22123729e-02 -2.44088635e-01 1.31948451e-02 4.26573396e-01
1.76420987e+00 5.03580451e-01 6.19902909e-02 -4.78194088e-01
2.68787295e-01 5.12979746e-01 6.69230297e-02 6.72414184e-01
2.48553574e-01 3.51346284e-01 4.91759658e-01 -4.53313857e-01
-1.39272654e+00 -9.44059372e-01 3.12069803e-02 1.38492310e+00
-1.46126658e-01 -9.53071296e-01 -6.03469849e-01 -7.07952142e-01
6.06968179e-02 7.05435932e-01 -2.40412891e-01 -4.47494864e-01
-7.13961959e-01 -4.63587493e-01 9.84735250e-01 5.60539782e-01
1.89127047e-02 -7.09094822e-01 -8.37804973e-01 3.66642267e-01
-9.49708074e-02 -7.57165372e-01 -6.93976521e-01 5.27132630e-01
-6.35303974e-01 -7.95300364e-01 -4.48195815e-01 -8.91143620e-01
2.91466594e-01 1.65747225e-01 1.90251005e+00 3.58157068e-01
-4.75298017e-01 3.02874178e-01 -3.77147317e-01 -7.80007094e-02
-9.07725811e-01 5.29841483e-01 -2.45946005e-01 -7.55157411e-01
6.72867715e-01 -7.36308455e-01 -5.79648197e-01 2.08277687e-01
-9.44550395e-01 -6.11879267e-02 7.80231178e-01 9.60883498e-01
-3.87973525e-02 -3.94626111e-01 2.10064888e-01 -7.35784054e-01
8.27895522e-01 -7.43870854e-01 -8.55589449e-01 5.61180770e-01
-9.25884843e-01 6.17645979e-01 6.41597807e-01 -2.39629358e-01
-7.01400995e-01 -8.64249021e-02 -2.27108762e-01 -2.87045121e-01
3.12048376e-01 6.33723497e-01 2.76726782e-01 -5.83920702e-02
8.97693634e-01 3.63788217e-01 1.19018592e-01 -5.97033560e-01
5.53895950e-01 6.34142160e-01 4.80988085e-01 -1.23299646e+00
8.39634180e-01 1.67233661e-01 -2.88870811e-01 -2.69373208e-01
-2.63977617e-01 -5.62422156e-01 -1.36690676e-01 5.29747486e-01
4.47289050e-01 -9.54302669e-01 -6.01082325e-01 -9.12448913e-02
-1.23683584e+00 -3.40881437e-01 -2.91358441e-01 -5.31789027e-02
-5.70863962e-01 3.47231865e-01 -6.06043160e-01 -2.64069319e-01
-4.30546641e-01 -1.42582989e+00 1.42001283e+00 -4.05751646e-01
-3.31771761e-01 -7.27605641e-01 5.03041446e-01 1.15792342e-01
1.00530136e+00 -3.88180852e-01 1.30414462e+00 -7.08839417e-01
-8.85040998e-01 -3.14122587e-01 -3.52776259e-01 2.33145937e-01
-2.94237852e-01 -4.41597849e-02 -7.69950569e-01 -3.57647657e-01
-2.29412422e-01 -6.30204320e-01 6.62621915e-01 -3.94434690e-01
1.04814875e+00 -2.74095625e-01 -6.20674908e-01 4.78266120e-01
1.70626271e+00 -1.13841064e-01 5.46553016e-01 4.01537180e-01
2.74194121e-01 8.15176740e-02 2.16148928e-01 3.16329658e-01
4.48193789e-01 8.59006882e-01 2.39199072e-01 9.32124630e-02
-9.54864919e-02 -6.10598207e-01 4.60157007e-01 1.05234456e+00
3.54528427e-01 -1.74904168e-01 -1.38759458e+00 8.84371877e-01
-1.66932774e+00 -7.81106055e-01 3.64519268e-01 2.13413048e+00
1.15373158e+00 -3.63800395e-03 5.01559675e-02 -3.22335213e-01
5.80719672e-02 -3.62347700e-02 -3.76528174e-01 -5.61431110e-01
2.54765451e-01 6.74427569e-01 4.88768399e-01 2.78424978e-01
-6.60106421e-01 7.89852142e-01 6.58009148e+00 8.68731856e-01
-1.36498475e+00 2.52216071e-01 2.90340871e-01 5.37862666e-02
-6.74757838e-01 4.65905219e-01 -6.02347732e-01 4.78981256e-01
1.28956783e+00 -4.31263775e-01 7.35813439e-01 1.26486695e+00
-3.84979248e-01 -2.65126303e-02 -1.49404538e+00 8.23288500e-01
-5.38036339e-02 -1.46515560e+00 -6.61463514e-02 2.09937035e-03
4.50258791e-01 7.70731449e-01 4.33057845e-02 5.98013163e-01
8.40016484e-01 -1.12687695e+00 8.03265214e-01 6.44299388e-02
8.95776033e-01 -3.07980418e-01 2.54787564e-01 2.51164138e-01
-1.08338177e+00 -1.44751877e-01 -1.80531517e-01 1.97361410e-01
-1.28670841e-01 3.55402976e-01 -9.32719707e-01 3.85765940e-01
4.95910257e-01 4.43848193e-01 -6.37065411e-01 9.63236988e-01
2.08837554e-01 5.06124020e-01 -3.80779505e-01 -1.24642044e-01
4.94502455e-01 2.03822330e-01 3.13214839e-01 1.45669591e+00
5.59178114e-01 -5.24756312e-01 7.98804760e-02 9.72933829e-01
-2.03485265e-01 -4.94534671e-02 -8.32968116e-01 -4.72778559e-01
5.04205346e-01 1.01201320e+00 -3.51148903e-01 -3.62799793e-01
-8.56270015e-01 8.56615603e-01 3.82157326e-01 1.55031636e-01
-8.60994697e-01 -5.03628314e-01 6.82517767e-01 1.00526169e-01
1.88096091e-01 -1.81298375e-01 -6.30671531e-02 -1.40187728e+00
3.74162853e-01 -1.45962131e+00 9.43631902e-02 -7.05411673e-01
-1.05108786e+00 6.72379613e-01 2.82342662e-03 -9.96202826e-01
-6.23843968e-01 -4.52295661e-01 -1.88461766e-01 9.43198442e-01
-1.38548100e+00 -1.01094818e+00 1.67193413e-01 3.73505473e-01
5.33613026e-01 -2.43706435e-01 9.21993315e-01 6.31059885e-01
9.06469375e-02 6.40457332e-01 3.22895408e-01 -6.09909520e-02
6.76016986e-01 -1.17271686e+00 9.32387412e-01 7.60653079e-01
3.28073412e-01 1.05276418e+00 5.96086860e-01 -2.82260895e-01
-2.09229159e+00 -5.52964687e-01 1.22754407e+00 -7.20027328e-01
1.29191792e+00 -5.82188249e-01 -8.01765919e-01 7.24346757e-01
4.79545683e-01 -2.18887895e-01 3.10372412e-01 2.90685892e-01
-9.08588052e-01 8.77965391e-02 -6.83166802e-01 5.22111416e-01
1.09517252e+00 -1.16650999e+00 -4.44501728e-01 5.64421296e-01
8.19630742e-01 -3.23249161e-01 -9.02737856e-01 1.14098616e-01
6.75410330e-01 -8.39354992e-01 9.77087557e-01 -5.99729896e-01
7.01800764e-01 -3.05527076e-02 -3.01796764e-01 -1.23927855e+00
-8.45517218e-02 -1.04838717e+00 1.47205994e-01 1.03887677e+00
5.58339477e-01 -3.21056604e-01 9.30489480e-01 6.84913993e-01
-2.05833927e-01 -8.50492716e-01 -7.88966954e-01 -8.23623180e-01
3.68419409e-01 -4.98882681e-01 8.12175691e-01 7.96557784e-01
6.32990241e-01 5.14796317e-01 -1.31271914e-01 -3.57256711e-01
2.39157453e-01 2.13732049e-01 8.29749346e-01 -9.72661376e-01
-7.53034115e-01 -6.67478621e-01 -2.08256990e-01 -1.37397397e+00
1.71363980e-01 -1.32697237e+00 1.22384280e-01 -1.22009170e+00
4.00933772e-01 -6.53298616e-01 -1.49945751e-01 6.04030073e-01
2.14290261e-01 -2.66041219e-01 2.21797754e-03 2.76129901e-01
-4.91559535e-01 2.55454451e-01 6.66015029e-01 -2.73560822e-01
4.03853983e-01 -2.90080637e-01 -8.33287656e-01 6.55382797e-02
3.75542879e-01 -5.13241291e-01 -7.01961040e-01 -1.01267457e+00
8.65378737e-01 3.40076834e-01 2.75776684e-01 -8.55399311e-01
4.08193707e-01 1.22259356e-01 -4.78307158e-01 -8.20022076e-02
1.19381145e-01 -9.59657848e-01 1.64334223e-01 3.54111552e-01
-6.77781880e-01 5.63015997e-01 3.14524204e-01 3.22723895e-01
-2.64107734e-01 -3.26572061e-01 5.04643619e-01 -3.33896279e-01
-6.26575112e-01 1.95265219e-01 -2.42920756e-01 4.35913622e-01
5.79363465e-01 8.62253383e-02 -5.30087233e-01 -1.20094664e-01
-7.39521859e-03 4.11748588e-02 9.10981119e-01 7.23478854e-01
2.83080310e-01 -8.71893167e-01 -4.78515118e-01 2.93335319e-01
3.46411526e-01 -3.96913230e-01 -4.79020000e-01 7.28886724e-01
-5.22682667e-01 7.15895057e-01 4.91958112e-02 -4.32409972e-01
-9.66956019e-01 7.69156337e-01 1.87513202e-01 -5.63790560e-01
-5.56887448e-01 9.25986886e-01 -5.36694713e-02 -4.08616453e-01
2.12054521e-01 -4.88581002e-01 7.08784103e-01 -5.16218424e-01
2.74649322e-01 -1.37276947e-01 4.55268055e-01 -1.23569727e-01
-4.06120300e-01 1.71211988e-01 -2.55943328e-01 -5.19852079e-02
1.46864116e+00 2.31989950e-01 -4.78376001e-01 7.64440820e-02
1.45479143e+00 9.19587910e-02 -6.01627290e-01 -4.33465511e-01
6.96091294e-01 -4.24317330e-01 -4.54594381e-03 -1.00173116e+00
-7.28805721e-01 9.05319750e-01 2.95558691e-01 2.17146918e-01
1.00084352e+00 7.32144639e-02 9.91528749e-01 7.39385247e-01
8.74011397e-01 -7.61285603e-01 -3.68024521e-02 5.35407245e-01
5.23440838e-01 -9.80723739e-01 -1.18617736e-01 8.08188170e-02
-2.84795761e-01 1.00413907e+00 3.11394185e-01 -7.54916668e-02
4.74632949e-01 7.72160590e-01 -3.08273226e-01 -3.19448650e-01
-1.31909359e+00 1.76078662e-01 -6.48139641e-02 3.69793206e-01
8.24300528e-01 -3.23140532e-01 -2.09072465e-03 -4.59916368e-02
-1.97008606e-02 3.68430108e-01 1.45953640e-01 1.15766585e+00
-3.10274214e-01 -1.63054812e+00 1.12828486e-01 4.06230748e-01
-5.78807533e-01 -5.93824506e-01 -2.22058222e-01 6.87783062e-01
-2.87718058e-01 5.47571003e-01 -1.83329925e-01 -5.38069487e-01
1.27548441e-01 3.98704380e-01 4.56307143e-01 -6.97729290e-01
-9.04039204e-01 -3.74055982e-01 1.81617662e-01 -8.14323723e-01
-7.02026486e-03 -4.82979655e-01 -7.35825479e-01 -5.40648818e-01
-3.51702720e-01 5.21081924e-01 8.87171626e-01 6.46097481e-01
8.30105662e-01 1.66674882e-01 2.68430471e-01 -3.57202113e-01
-1.12115204e+00 -6.69843316e-01 -4.08447832e-02 1.90324351e-01
2.80084729e-01 -3.62908602e-01 -1.67101249e-01 1.64755099e-02] | [7.566506385803223, 8.04464340209961] |
e32b9a70-1ff6-488c-9d5e-68c0d8010d98 | marc-memory-by-association-and-reinforcement | 1312.2844 | null | http://arxiv.org/abs/1312.2844v1 | http://arxiv.org/pdf/1312.2844v1.pdf | mARC: Memory by Association and Reinforcement of Contexts | This paper introduces the memory by Association and Reinforcement of Contexts
(mARC). mARC is a novel data modeling technology rooted in the second
quantization formulation of quantum mechanics. It is an all-purpose incremental
and unsupervised data storage and retrieval system which can be applied to all
types of signal or data, structured or unstructured, textual or not. mARC can
be applied to a wide range of information clas-sification and retrieval
problems like e-Discovery or contextual navigation. It can also for-mulated in
the artificial life framework a.k.a Conway "Game Of Life" Theory. In contrast
to Conway approach, the objects evolve in a massively multidimensional space.
In order to start evaluating the potential of mARC we have built a mARC-based
Internet search en-gine demonstrator with contextual functionality. We compare
the behavior of the mARC demonstrator with Google search both in terms of
performance and relevance. In the study we find that the mARC search engine
demonstrator outperforms Google search by an order of magnitude in response
time while providing more relevant results for some classes of queries. | ['Patrice Descourt', 'Norbert Rimoux'] | 2013-12-10 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-1.78775296e-01 -1.92380071e-01 4.20095623e-02 -6.13608882e-02
-7.15876937e-01 -6.58402741e-01 1.01499486e+00 4.59878683e-01
-9.55859125e-01 5.79362094e-01 1.38166100e-01 -3.93925369e-01
-8.11898828e-01 -9.86682951e-01 -2.93469757e-01 -5.52826107e-01
-2.45576248e-01 7.00360239e-01 7.20899403e-01 -7.95375049e-01
7.66444206e-01 4.14029777e-01 -2.17370677e+00 1.56632721e-01
4.78580892e-01 1.09772527e+00 5.51040471e-01 9.21766460e-01
-2.37836204e-02 7.19302595e-01 -4.41033244e-01 -5.72139435e-02
1.03943661e-01 -4.00450975e-02 -9.82834339e-01 -7.71439075e-01
3.99888633e-03 1.14470258e-01 -7.43441939e-01 1.06942225e+00
8.64444137e-01 3.73510450e-01 4.63018596e-01 -7.74831891e-01
-3.71012598e-01 5.70914857e-02 2.55652398e-01 8.19193661e-01
7.92535424e-01 -3.20014745e-01 9.33378875e-01 -1.00206089e+00
9.90595162e-01 1.11111128e+00 3.58084381e-01 2.26556957e-01
-1.12344432e+00 -2.36869678e-01 -6.76332951e-01 5.27294040e-01
-1.48219526e+00 -2.55104303e-01 4.68129843e-01 -3.23164761e-01
1.32277560e+00 5.78912616e-01 5.19523561e-01 6.58743680e-01
6.61048591e-01 3.81923050e-01 8.99166524e-01 -7.99166918e-01
4.79158789e-01 3.03383082e-01 3.37595820e-01 3.31052393e-01
-4.09404933e-02 4.90689814e-01 -1.13590670e+00 -3.24715227e-01
3.83216560e-01 -4.69774842e-01 2.10365713e-01 -1.35084808e-01
-1.16850102e+00 7.57183433e-01 3.54019403e-01 6.66326582e-01
-4.30756330e-01 3.78037184e-01 3.41638774e-01 6.99593067e-01
1.04178287e-01 8.19143414e-01 -1.25696182e-01 -5.96420825e-01
-7.07831740e-01 5.06772280e-01 8.19793761e-01 1.06723619e+00
6.44931853e-01 -2.86047310e-01 5.73266372e-02 5.88976979e-01
3.39238733e-01 3.93883884e-01 1.01266980e+00 -8.20663452e-01
9.02688950e-02 2.82066077e-01 3.28135133e-01 -9.38779771e-01
-4.87383872e-01 -2.83772469e-01 -2.72611111e-01 -9.83993709e-02
-1.76663503e-01 2.26093367e-01 -5.39530337e-01 1.35520494e+00
2.68032968e-01 -2.26950780e-01 2.14478567e-01 7.71381319e-01
8.35152507e-01 5.89389980e-01 -7.89916664e-02 -1.09727129e-01
1.39177144e+00 -3.22217345e-01 -8.66518319e-01 2.65048921e-01
8.12973738e-01 -9.31622565e-01 8.51248562e-01 7.27174640e-01
-1.03485787e+00 -4.79338616e-01 -1.17672479e+00 -1.76758841e-01
-1.07455873e+00 -5.86681783e-01 6.50495589e-01 7.48349905e-01
-1.31071556e+00 7.07402289e-01 -6.25419736e-01 -7.18780220e-01
-2.37631142e-01 4.19870108e-01 -1.80741534e-01 2.98623323e-01
-1.58105838e+00 1.09057474e+00 7.04978764e-01 -4.71293569e-01
-6.27197683e-01 -2.40659535e-01 -2.23646507e-01 -6.97767958e-02
3.11536193e-01 -3.47381592e-01 1.16669178e+00 -1.37594402e-01
-1.19284618e+00 8.10059488e-01 6.53517395e-02 -5.73700726e-01
3.20419297e-02 -1.77701026e-01 -1.09154570e+00 1.29439801e-01
2.16519147e-01 1.93230033e-01 1.59964487e-01 -5.96693993e-01
-7.74711132e-01 -5.29503644e-01 -4.61215992e-03 4.43258733e-01
-1.68080345e-01 4.98713069e-02 -4.55439389e-01 -4.63723838e-01
3.32036465e-01 -8.78001809e-01 -8.43608528e-02 -4.73294914e-01
1.12633072e-01 -3.27045649e-01 9.16313648e-01 -2.78684705e-01
1.74453306e+00 -2.07862782e+00 1.53070554e-01 4.38343406e-01
-2.64036357e-02 1.37329295e-01 2.69292384e-01 1.14359927e+00
1.95339531e-01 1.51599467e-01 3.98105770e-01 2.07519814e-01
3.42012227e-01 3.59174013e-01 -1.67123929e-01 1.62454158e-01
-4.55393016e-01 9.05794919e-01 -9.93111789e-01 -5.50695300e-01
-7.84512088e-02 2.24533863e-02 -6.06064320e-01 -2.08982900e-02
-3.80651236e-01 -1.74591411e-02 -4.07777011e-01 6.26898170e-01
2.12935880e-02 -3.47917587e-01 1.46034900e-02 2.47410923e-01
-4.64563102e-01 4.71013844e-01 -1.31702065e+00 2.08632708e+00
-3.12425643e-01 9.42512512e-01 -1.49050713e-01 -9.39280152e-01
8.30732822e-01 3.71377707e-01 3.13689828e-01 -1.41349876e+00
-2.44634449e-01 5.34216285e-01 -7.33181462e-02 -8.95251274e-01
1.39484334e+00 3.89063992e-02 -2.84036160e-01 4.29329157e-01
1.81341380e-01 -8.80214944e-02 2.41691023e-01 4.21224117e-01
1.37341118e+00 -1.47357555e-02 1.38110891e-01 -7.31165469e-01
3.59762400e-01 1.24805510e-01 -1.14055842e-01 9.74690676e-01
-5.48441596e-02 7.39531294e-02 -3.32555436e-02 -2.91921914e-01
-9.75130141e-01 -1.11239970e+00 -4.32491362e-01 1.43832731e+00
6.02899790e-01 -8.94090533e-01 -4.92985606e-01 1.74642429e-01
-2.40526587e-01 8.02751601e-01 -2.63468027e-01 -1.47653788e-01
-2.97428388e-02 -5.65045118e-01 5.22945046e-01 -6.74555227e-02
5.41737378e-01 -1.29324114e+00 -1.09010136e+00 5.10204256e-01
-5.20446673e-02 -7.33797610e-01 4.01943065e-02 5.70104659e-01
-7.07168043e-01 -7.78877378e-01 3.17112282e-02 -6.39895856e-01
-3.44942272e-01 7.31618609e-03 1.23656607e+00 8.49102214e-02
-5.64533114e-01 9.49796200e-01 -6.82729006e-01 -5.50641477e-01
-3.73779446e-01 1.80470124e-01 3.18644196e-01 -5.64681888e-01
8.32910717e-01 -7.33139753e-01 -8.85501683e-01 2.82988459e-01
-1.20631313e+00 -5.56379914e-01 5.11922121e-01 7.19259739e-01
4.32674885e-01 3.16253066e-01 5.58537006e-01 -5.71521878e-01
1.18235123e+00 -8.39713752e-01 -5.89423716e-01 2.48132735e-01
-1.08574903e+00 3.77148598e-01 -4.66443934e-02 -1.85785517e-01
-5.18307030e-01 -2.85833865e-01 2.63969251e-03 1.50200695e-01
1.28968107e-02 8.99591506e-01 1.71316326e-01 -1.35594577e-01
1.18224764e+00 5.37813783e-01 -1.55347392e-01 -4.80541706e-01
4.53162044e-01 1.09242916e+00 5.41936338e-01 -4.55164522e-01
2.31089130e-01 3.30401391e-01 2.12665841e-01 -1.30373037e+00
5.82844168e-02 -9.46682751e-01 -4.47083831e-01 -2.83825636e-01
8.56127143e-01 -4.14672166e-01 -6.86677814e-01 5.39609492e-02
-1.01815319e+00 6.37216493e-02 -2.98875898e-01 3.71553510e-01
-6.97575390e-01 2.13667125e-01 -2.84792513e-01 -9.40647006e-01
-2.05874577e-01 -8.36632490e-01 9.45380032e-01 2.18281701e-01
-1.10437356e-01 -9.05174255e-01 3.94309580e-01 5.97791225e-02
5.64749718e-01 -1.13539912e-01 8.26091230e-01 -8.59195650e-01
-7.48064816e-01 -3.95142376e-01 1.93516329e-01 -1.79512545e-01
-3.84313673e-01 -5.30301690e-01 -7.56775081e-01 -8.06702301e-02
8.90878215e-02 -2.32089385e-01 3.95241827e-01 -4.20849830e-01
6.76857948e-01 -4.09531966e-02 -3.72522533e-01 1.00430707e-02
1.74315822e+00 8.12133431e-01 1.01725733e+00 8.61467957e-01
-2.63646722e-01 4.42985177e-01 9.19932604e-01 4.87856179e-01
1.37296066e-01 1.05085409e+00 4.93907303e-01 7.97012866e-01
1.38951033e-01 -1.39275327e-01 2.88104173e-02 9.62065399e-01
-2.54428033e-02 -1.99611321e-01 -1.17151654e+00 5.88462532e-01
-1.95329821e+00 -1.32209527e+00 -9.16750282e-02 2.30122471e+00
6.71775758e-01 2.70792633e-01 -4.43626456e-02 -5.35559691e-02
3.85686636e-01 -3.44659865e-01 -3.75616968e-01 -5.26436269e-01
-4.02444303e-02 3.97177815e-01 4.77902979e-01 5.14212191e-01
-6.99078321e-01 7.33444691e-01 6.62853670e+00 1.07113409e+00
-9.50691521e-01 3.02464157e-01 -1.86305553e-01 -1.42187895e-02
-1.69669136e-01 1.46569252e-01 -5.20105541e-01 7.74088800e-01
1.42054749e+00 -5.49841762e-01 6.47296548e-01 7.66111195e-01
3.43637690e-02 -6.13929331e-01 -1.04248118e+00 1.35083508e+00
-1.45950973e-01 -1.53467822e+00 -1.13023646e-01 3.26716334e-01
3.03279191e-01 3.77454907e-01 7.43993595e-02 2.65142143e-01
-9.86705571e-02 -1.00849760e+00 7.22032189e-01 1.06811774e+00
6.43520951e-01 -5.61417997e-01 6.55621350e-01 6.59459472e-01
-9.14278626e-01 -1.47267118e-01 -3.41617823e-01 -3.42320949e-02
2.36791782e-02 8.08308348e-02 -5.72469771e-01 8.33711684e-01
9.25707340e-01 -6.21620528e-02 -7.14083254e-01 1.25120461e+00
6.49823248e-01 1.86877221e-01 -6.73227072e-01 -6.36209548e-01
2.65381873e-01 -2.00573280e-01 6.82378232e-01 1.14063466e+00
5.41445315e-01 3.91525179e-01 -2.24823445e-01 4.56719011e-01
3.32573235e-01 2.30188116e-01 -8.64290059e-01 -1.67658240e-01
8.03446114e-01 7.92882383e-01 -5.84305286e-01 -4.44938749e-01
-1.10889055e-01 8.77720654e-01 -1.23944320e-01 1.17982954e-01
-3.75188649e-01 -8.21093559e-01 9.12948400e-02 4.27249432e-01
1.83778450e-01 -4.59266067e-01 1.96295097e-01 -8.25257063e-01
-1.36529997e-01 -6.84743404e-01 2.85170376e-01 -1.06271672e+00
-9.97275770e-01 7.00410604e-01 2.72566497e-01 -1.12717211e+00
-6.28231525e-01 -6.16880774e-01 -1.37897626e-01 8.11599910e-01
-8.18733633e-01 -5.48228204e-01 -2.24161148e-01 7.75930643e-01
1.82530761e-01 -6.02172315e-01 1.09895754e+00 4.75385368e-01
1.29482940e-01 9.75134000e-02 7.53778756e-01 -4.50256079e-01
4.13426012e-01 -1.32350242e+00 5.99404871e-02 4.03791934e-01
4.75373924e-01 1.00970876e+00 1.03516650e+00 -5.51887631e-01
-1.69572175e+00 -3.57399613e-01 9.19656813e-01 -6.67396247e-01
9.00977492e-01 -4.37483698e-01 -6.16650164e-01 1.94956213e-01
2.10656703e-01 -3.60255629e-01 6.71317458e-01 5.14917672e-02
-3.78114916e-02 -3.95944752e-02 -9.63029504e-01 5.34841359e-01
1.06533015e+00 -1.06975377e+00 -7.84526944e-01 6.02274537e-01
7.03208566e-01 -2.64137477e-01 -1.01956117e+00 4.23965380e-02
5.92651069e-01 -1.45783663e+00 9.10974801e-01 -4.33142483e-01
-1.75895587e-01 -1.75339356e-01 -7.19385445e-01 -7.77679265e-01
-1.30923301e-01 -9.99582529e-01 -1.44680589e-01 7.71011055e-01
3.54653925e-01 -6.35088742e-01 7.63511121e-01 5.44515312e-01
5.06963208e-02 -5.63326120e-01 -1.49560857e+00 -9.49573100e-01
-4.36479300e-02 -9.05868828e-01 4.35740173e-01 5.20724773e-01
5.30612886e-01 5.51290810e-01 -2.51965672e-02 -1.42954394e-01
4.74297166e-01 -1.89827383e-01 4.73667771e-01 -1.37327492e+00
-4.75560218e-01 -3.14189076e-01 -9.05488431e-01 -1.04226172e+00
-5.60637712e-01 -9.01292026e-01 -2.16353118e-01 -1.22230411e+00
-2.65822887e-01 -5.36596477e-01 -3.63458425e-01 -3.31676692e-01
5.14887869e-01 -5.22483028e-02 -3.27929156e-03 6.78456485e-01
-1.02314508e+00 3.20123613e-01 6.36366129e-01 2.10508987e-01
-1.81816921e-01 -1.54298857e-01 -3.05249721e-01 1.32672638e-01
5.70636511e-01 -3.24208707e-01 -5.69631219e-01 4.10196744e-02
8.99424732e-01 2.69325197e-01 3.45274359e-01 -1.12000179e+00
8.56175423e-01 2.03825071e-01 -1.47318974e-01 -8.10946167e-01
3.91163111e-01 -8.80747855e-01 3.63327652e-01 3.61536294e-01
-2.17059955e-01 4.58622813e-01 4.97734733e-02 8.78264487e-01
-3.64359170e-01 -6.11700773e-01 2.70912766e-01 -2.88551688e-01
-1.26351845e+00 -1.16102159e-01 -5.32750666e-01 -1.49201632e-01
9.30191517e-01 -3.51971924e-01 -3.34693283e-01 -6.18059278e-01
-9.01067972e-01 5.96880876e-02 3.47702384e-01 4.42185551e-01
4.51688915e-01 -1.42154348e+00 4.54299785e-02 1.36635855e-01
3.10206711e-01 -4.09723371e-01 -4.65935096e-02 5.32804668e-01
-7.99714327e-01 1.14425266e+00 -5.93709461e-02 -4.23584819e-01
-9.69369233e-01 8.83553684e-01 1.97288468e-01 -1.82417840e-01
-1.61691412e-01 4.98121113e-01 -3.78377855e-01 -2.12691501e-01
2.26584971e-01 1.20892614e-01 -2.24844083e-01 9.17437747e-02
4.40676421e-01 5.62746108e-01 3.74497771e-01 -6.90742195e-01
-4.13740546e-01 2.58208930e-01 1.50833055e-01 -7.25299537e-01
1.09728813e+00 -2.92620420e-01 -3.05543035e-01 7.88160145e-01
1.13048792e+00 -1.61959648e-01 -1.62004471e-01 -1.58320963e-01
7.17225373e-01 -2.25930303e-01 1.26442343e-01 -9.10896599e-01
2.69947853e-02 7.02403605e-01 1.23220384e+00 9.61621642e-01
9.40502226e-01 1.57224193e-01 2.43627951e-01 8.41576099e-01
1.04587710e+00 -1.58192241e+00 3.45411479e-01 5.25513291e-01
8.24524343e-01 -9.06811237e-01 -9.35757086e-02 2.12295189e-01
-2.33881965e-01 1.06702387e+00 1.32406831e-01 -6.55685365e-02
1.03904700e+00 3.57903019e-02 -2.82555431e-01 -8.59150410e-01
-9.87670958e-01 -3.11900705e-01 2.10569561e-01 6.57699585e-01
3.67529720e-01 -1.50693819e-01 -8.08290601e-01 3.06713998e-01
-4.76756901e-01 1.27174938e-02 4.53677505e-01 1.27930188e+00
-7.37807631e-01 -1.37577820e+00 -3.92653346e-01 4.56281692e-01
-5.20680010e-01 -3.57044429e-01 -2.79435456e-01 8.60612512e-01
9.45505500e-02 1.13604462e+00 4.03666263e-03 -7.34577715e-01
3.43295366e-01 3.22213948e-01 2.11607650e-01 -4.99279916e-01
-5.02110481e-01 -3.51040028e-02 2.07327738e-01 -7.73829460e-01
-5.70585608e-01 -5.89221179e-01 -1.12450993e+00 -2.07562983e-01
-5.80850720e-01 7.15873599e-01 1.10968709e+00 6.79313540e-01
7.07856178e-01 2.26481020e-01 1.03371985e-01 -4.67262745e-01
-4.93545383e-01 -9.99196112e-01 -8.28169823e-01 2.15252161e-01
1.26240805e-01 -8.50888729e-01 -1.86899170e-01 -2.35681295e-01] | [11.309934616088867, 7.533426761627197] |
e7b290c7-9cad-4456-b4f8-0f3560ec2549 | improving-robustness-using-joint-attention | 2005.08094 | null | https://arxiv.org/abs/2005.08094v2 | https://arxiv.org/pdf/2005.08094v2.pdf | Improving Robustness using Joint Attention Network For Detecting Retinal Degeneration From Optical Coherence Tomography Images | Noisy data and the similarity in the ocular appearances caused by different ophthalmic pathologies pose significant challenges for an automated expert system to accurately detect retinal diseases. In addition, the lack of knowledge transferability and the need for unreasonably large datasets limit clinical application of current machine learning systems. To increase robustness, a better understanding of how the retinal subspace deformations lead to various levels of disease severity needs to be utilized for prioritizing disease-specific model details. In this paper we propose the use of disease-specific feature representation as a novel architecture comprised of two joint networks -- one for supervised encoding of disease model and the other for producing attention maps in an unsupervised manner to retain disease specific spatial information. Our experimental results on publicly available datasets show the proposed joint-network significantly improves the accuracy and robustness of state-of-the-art retinal disease classification networks on unseen datasets. | ['Stewart Lee Zuckerbrod', 'Sharif Amit Kamran', 'Alireza Tavakkoli'] | 2020-05-16 | null | null | null | null | ['retinal-oct-disease-classification'] | ['computer-vision'] | [ 3.64966571e-01 -1.77967444e-01 1.29019201e-01 -5.95356226e-01
-5.66505849e-01 -3.30241770e-01 2.97620803e-01 -1.40470698e-01
-2.26302758e-01 6.68625474e-01 2.86907285e-01 -1.58782125e-01
-6.41821325e-01 -3.01250964e-01 -4.10795987e-01 -9.60136771e-01
9.35833901e-02 1.83342978e-01 1.84024632e-01 6.31676167e-02
4.61791396e-01 7.96160281e-01 -1.84330249e+00 5.44819117e-01
1.09751201e+00 9.62688625e-01 2.18958035e-01 6.47452891e-01
1.54429525e-01 5.94749510e-01 -5.37204146e-01 -1.13896102e-01
2.82487571e-01 -4.80622202e-01 -6.23224080e-01 1.38069808e-01
1.00352824e+00 -2.87707835e-01 -2.07084030e-01 1.11339188e+00
8.76121700e-01 -1.96548894e-01 7.49997556e-01 -5.88644147e-01
-6.89331830e-01 -2.57890195e-01 -4.29971486e-01 6.17689908e-01
-1.46640211e-01 5.43349802e-01 6.51059389e-01 -4.54219997e-01
8.09039056e-01 9.89765704e-01 4.58285540e-01 6.00240707e-01
-1.21176755e+00 -2.04771608e-01 1.01970108e-02 3.13972831e-01
-1.23036885e+00 -5.32836318e-01 6.01132572e-01 -7.84759879e-01
9.72587347e-01 3.38149428e-01 7.46608377e-01 7.60107458e-01
3.03067237e-01 6.27262816e-02 1.44644439e+00 -4.33653057e-01
-1.71309605e-01 6.90971166e-02 1.87452063e-01 7.12165415e-01
5.85698962e-01 2.64956087e-01 -3.58703613e-01 -1.43154442e-01
8.73633325e-01 -2.51207441e-01 -4.64914620e-01 -2.73626536e-01
-8.68164480e-01 4.71819311e-01 6.00023389e-01 1.20247692e-01
-2.96205610e-01 1.03381276e-01 1.72885835e-01 1.65920720e-01
5.41127980e-01 6.61207497e-01 -3.37863356e-01 2.05444679e-01
-8.14427972e-01 1.32971630e-01 1.54634207e-01 3.98106396e-01
2.70461529e-01 -2.75145352e-01 -3.13366055e-01 8.86651754e-01
4.07764614e-01 7.38624409e-02 6.18091762e-01 -7.17802942e-01
1.63352072e-01 9.18532133e-01 1.90532595e-01 -8.65760028e-01
-5.97997606e-01 -8.12974751e-01 -6.57311261e-01 4.59422588e-01
5.47025204e-01 -4.07621413e-02 -1.36148822e+00 1.33155632e+00
2.49012798e-01 3.50903988e-01 -9.83881727e-02 1.10034418e+00
8.16041231e-01 1.56665929e-02 -2.21410673e-02 1.87701534e-03
1.42733228e+00 -6.54702246e-01 -5.27473450e-01 -1.88558534e-01
7.22476721e-01 -8.28792214e-01 7.54062712e-01 1.96200714e-01
-9.60604072e-01 -3.85190964e-01 -7.49127328e-01 -3.03435326e-01
-3.85214120e-01 7.37481177e-01 5.03258169e-01 3.73991758e-01
-1.14708376e+00 3.19736302e-01 -8.51991832e-01 -7.17543662e-01
9.46730554e-01 5.51152229e-01 -4.58392322e-01 -2.49795616e-01
-4.97522801e-01 1.20567143e+00 -1.77536090e-03 5.82999587e-01
-4.71387595e-01 -7.20179558e-01 -5.17291427e-01 -2.61856556e-01
-2.09829718e-01 -1.05450928e+00 7.75104702e-01 -7.55822957e-01
-1.17203462e+00 1.19551730e+00 -4.76496100e-01 -3.55526567e-01
4.99030203e-01 -7.21358582e-02 -4.67181504e-01 3.23132992e-01
-1.63902432e-01 6.09533966e-01 7.77802706e-01 -9.32437956e-01
-6.42116368e-01 -6.81260109e-01 -3.58534604e-01 6.54509589e-02
-7.40231797e-02 2.86542773e-01 -4.09207076e-01 -4.53622431e-01
2.14889809e-01 -9.63441074e-01 -2.88731545e-01 4.57119375e-01
-4.78098601e-01 -1.07357867e-01 6.37383699e-01 -6.03217363e-01
1.08703792e+00 -2.14551783e+00 1.21630795e-01 1.91890284e-01
5.35454273e-01 6.07339323e-01 -3.21318984e-01 1.88649893e-02
-2.63823956e-01 4.29316163e-02 1.98904231e-01 1.64017994e-02
-6.90950930e-01 -1.52224656e-02 -7.44768083e-02 7.38852262e-01
6.47514105e-01 8.44152629e-01 -6.83284044e-01 -1.95063815e-01
2.23930031e-01 6.59477651e-01 -4.35568035e-01 1.93393454e-01
-7.38573596e-02 7.30202734e-01 -5.85034966e-01 6.93092525e-01
5.44927180e-01 -6.10192001e-01 -1.97217148e-02 -5.35127461e-01
1.35333454e-02 3.12859654e-01 -7.81642914e-01 1.28907561e+00
-3.13147046e-02 8.97791624e-01 -3.60403240e-01 -6.25749588e-01
6.57956541e-01 1.18709520e-01 2.89687276e-01 -4.36268359e-01
2.21336663e-01 3.21204633e-01 5.83942235e-01 -9.25386429e-01
2.37876009e-02 5.77505231e-02 6.46796405e-01 4.50727977e-02
-3.53127681e-02 3.63154233e-01 1.79048598e-01 -4.40082431e-01
9.08482611e-01 -1.62326638e-02 1.66690528e-01 -1.55055508e-01
3.24019343e-01 2.75577530e-02 5.93912303e-01 5.35916388e-01
-4.63486493e-01 9.82718050e-01 7.02170491e-01 -7.00042903e-01
-8.12813699e-01 -7.08531678e-01 -6.47592902e-01 3.58996987e-01
2.08347384e-02 6.38934970e-02 -5.35599768e-01 -6.50323629e-01
1.66292533e-01 1.85545966e-01 -8.82742465e-01 -3.17911357e-01
-3.62289459e-01 -1.09199691e+00 4.28836644e-01 5.49474895e-01
3.25336009e-01 -6.44629896e-01 -6.26272380e-01 -7.84745887e-02
1.41765192e-01 -8.48975003e-01 -1.30837679e-01 -1.78465739e-01
-8.95930231e-01 -1.43817949e+00 -7.75366545e-01 -9.84855592e-01
1.21364021e+00 2.32349485e-01 7.09754348e-01 2.05141798e-01
-1.02666676e+00 1.34840474e-01 -1.08182952e-02 -4.96652693e-01
-2.30330423e-01 -1.82890654e-01 -4.60734591e-02 4.57578957e-01
6.59619689e-01 -3.34762454e-01 -1.13016939e+00 2.04757929e-01
-6.29580438e-01 -2.29341164e-01 8.46518576e-01 1.00089407e+00
7.24692643e-01 4.46231738e-02 3.59800071e-01 -7.93595910e-01
6.25164926e-01 -1.86765522e-01 -7.21615016e-01 4.07670200e-01
-5.91839969e-01 -1.42641976e-01 1.52394667e-01 -3.57953906e-01
-1.02022147e+00 2.93846726e-02 2.74737507e-01 -2.57274777e-01
-4.67100352e-01 3.61352652e-01 1.93011418e-01 -5.49736917e-01
9.10861433e-01 5.50358854e-02 3.45679879e-01 -6.14203990e-01
1.19885974e-01 8.58047307e-01 4.15143192e-01 6.83916965e-03
4.04155642e-01 5.46330988e-01 3.22203487e-01 -8.30697238e-01
-9.51182306e-01 -5.87443590e-01 -6.13722801e-01 -2.60206405e-02
9.66781080e-01 -8.49093497e-01 -3.16933542e-01 7.92714715e-01
-1.13539422e+00 -3.34912650e-02 -3.70496139e-02 4.97652918e-01
-2.48032957e-01 6.46896437e-02 -3.79018307e-01 -3.94480109e-01
3.25941332e-02 -1.33966291e+00 9.08554971e-01 3.87280643e-01
-1.12791628e-01 -8.95822525e-01 1.41476020e-01 6.71633005e-01
3.93510371e-01 9.43534672e-02 1.22800720e+00 -4.50955212e-01
-8.74383748e-01 -2.11259112e-01 -6.19815290e-01 4.80650306e-01
4.02057648e-01 3.46079737e-01 -1.08061504e+00 -3.24173480e-01
-4.42811370e-01 -3.22849244e-01 1.15107393e+00 7.99469888e-01
1.11034632e+00 -1.16940789e-01 -4.69137400e-01 7.07587183e-01
1.44350922e+00 3.93853448e-02 8.69517922e-01 1.53956473e-01
6.02585375e-01 8.19699705e-01 3.76954466e-01 1.11295894e-01
1.42651886e-01 5.96627295e-01 3.78150821e-01 -4.24111098e-01
-7.20938861e-01 2.45636448e-01 -3.16030860e-01 2.55411983e-01
-4.01476115e-01 -1.70875371e-01 -9.73394394e-01 8.36306989e-01
-1.71305275e+00 -7.39608586e-01 -1.81443363e-01 2.33474994e+00
9.00935411e-01 -2.77896106e-01 -5.95285445e-02 -4.73746657e-01
6.46072388e-01 -1.04661487e-01 -6.13145232e-01 -1.16892569e-02
-1.99524105e-01 2.16817204e-02 3.72864455e-01 1.39223278e-01
-1.09047532e+00 8.14142585e-01 6.69249535e+00 3.76623422e-01
-1.48775339e+00 -2.19167888e-01 6.86752319e-01 -5.03340304e-01
1.99649528e-01 -1.91569716e-01 -8.83814216e-01 4.74331737e-01
7.43918002e-01 3.59057456e-01 1.24555849e-01 1.30630344e-01
5.15064836e-01 -7.30718374e-02 -9.93651986e-01 9.09801245e-01
7.47980252e-02 -1.51438665e+00 2.56648064e-01 1.60243422e-01
7.56220400e-01 3.61288816e-01 5.02293348e-01 -5.54066598e-01
-1.32520780e-01 -1.30539751e+00 -1.06578737e-01 1.25204241e+00
9.55075204e-01 -2.57764250e-01 9.13317442e-01 -3.39706391e-01
-5.68216145e-01 -1.73602775e-01 -5.23718834e-01 2.02664614e-01
-2.37467945e-01 4.51110333e-01 -9.62990224e-01 9.06414911e-02
7.33681738e-01 7.80619264e-01 -9.92391646e-01 1.84296453e+00
-9.49136540e-02 5.60402513e-01 -6.34787530e-02 2.17999205e-01
3.69180590e-02 -1.00752421e-01 7.57490516e-01 8.72433484e-01
3.68405163e-01 -7.32095465e-02 -2.91339606e-01 8.04173112e-01
1.49407446e-01 3.19834910e-02 -5.99636376e-01 -1.93719015e-01
2.33240455e-01 1.14282596e+00 -4.04239208e-01 1.79587379e-01
-5.16464174e-01 6.62252963e-01 3.94542992e-01 6.79810345e-01
-3.01188827e-01 -4.08152282e-01 9.41379726e-01 4.44171041e-01
3.49165440e-01 8.68562683e-02 -3.40119243e-01 -8.64812076e-01
2.21314996e-01 -7.00924158e-01 2.24580854e-01 -1.02812254e+00
-1.48735118e+00 6.91265047e-01 -5.19339561e-01 -1.42757928e+00
-4.25558053e-02 -8.92285883e-01 -5.74097693e-01 1.27684784e+00
-1.99498188e+00 -1.39701581e+00 -5.09983122e-01 4.28333461e-01
-1.24434913e-02 -5.70634305e-01 7.98383176e-01 1.21687800e-01
-8.19238842e-01 5.11181593e-01 1.47897825e-01 5.47344200e-02
1.13276541e+00 -1.17222643e+00 -3.93213443e-02 7.96009660e-01
-1.10022455e-01 7.65060246e-01 5.36547303e-01 -6.10688031e-01
-9.06235456e-01 -1.38694370e+00 1.01371729e+00 -5.29163003e-01
5.87662756e-01 2.67414391e-01 -9.47660446e-01 5.27224720e-01
-1.34697566e-02 6.55271292e-01 8.76298904e-01 1.49436608e-01
-4.07706559e-01 -4.78772372e-01 -1.13862562e+00 6.96574390e-01
8.90209377e-01 -5.43390870e-01 -4.71904755e-01 5.03798783e-01
2.21369609e-01 -3.70657235e-01 -8.14912856e-01 5.28600752e-01
4.73766983e-01 -9.51090038e-01 7.43437409e-01 -1.22684085e+00
3.90097231e-01 -5.54209352e-01 1.80501044e-01 -1.24698186e+00
-3.83542240e-01 -5.17837524e-01 1.37099981e-01 7.76972711e-01
4.12934899e-01 -1.04299855e+00 6.08708143e-01 5.84703863e-01
-1.75677743e-02 -1.17222559e+00 -9.39342499e-01 -4.09853935e-01
-8.79406855e-02 1.62971109e-01 3.14995706e-01 7.98525214e-01
-5.53475797e-01 -3.29928696e-02 -6.02289774e-02 6.82493329e-01
4.14389521e-01 1.17902108e-01 4.04422849e-01 -1.42425108e+00
-6.58151582e-02 -5.44833362e-01 -1.16962624e+00 -5.90470731e-01
-1.18208796e-01 -8.27410281e-01 -3.04972857e-01 -1.57795572e+00
2.28737116e-01 -4.68059719e-01 -5.88922501e-01 4.85933989e-01
-9.36357304e-02 5.36891222e-01 -3.34756404e-01 2.77991086e-01
-3.26191843e-01 2.07395777e-01 1.51105392e+00 7.59847322e-03
-3.11199546e-01 6.54660091e-02 -9.37230825e-01 7.02406824e-01
7.08442748e-01 -3.49879682e-01 -5.96147954e-01 -6.00408077e-01
1.02233119e-01 -3.30174863e-01 8.62825215e-01 -9.90183234e-01
3.38623732e-01 5.34221642e-02 5.31387150e-01 -3.67849529e-01
2.32934713e-01 -5.37364781e-01 -2.91720241e-01 2.31653869e-01
-2.82506645e-01 -3.71789426e-01 2.61270761e-01 7.83612490e-01
-4.43061054e-01 1.08666547e-01 1.04937899e+00 3.49386185e-02
-4.16464716e-01 3.01849335e-01 -9.92256105e-02 -1.85487885e-02
1.01923215e+00 -3.70595485e-01 -8.09452415e-01 8.78680497e-02
-8.89568865e-01 -7.22301677e-02 4.79704410e-01 3.25419456e-01
6.83981359e-01 -9.41327810e-01 -8.23832452e-01 3.98993045e-01
5.03630698e-01 -2.34190375e-01 4.78366822e-01 1.20415020e+00
-5.92119217e-01 6.19008243e-01 -3.85993838e-01 -7.69007742e-01
-1.54992497e+00 -6.19270690e-02 9.64484990e-01 -6.16250895e-02
-4.94255155e-01 1.01316440e+00 1.24548912e-01 2.25940570e-02
1.02579996e-01 -4.41662520e-01 -3.75639737e-01 2.80353706e-02
5.23499966e-01 2.35680625e-01 4.08914536e-01 -5.22891104e-01
-2.15877920e-01 8.73562098e-01 -6.03010178e-01 5.27940631e-01
1.37924576e+00 -3.73947881e-02 -4.16097373e-01 -2.73365658e-02
9.11469519e-01 -1.82933941e-01 -1.34906936e+00 -1.63332522e-01
-1.85756549e-01 -6.72542632e-01 4.47246552e-01 -1.44342160e+00
-1.08476770e+00 8.83892596e-01 1.16557014e+00 -9.99748260e-02
1.35499716e+00 -3.02650277e-02 3.35623384e-01 3.31751168e-01
1.43818974e-01 -7.63311565e-01 -3.22296023e-01 7.89557919e-02
9.44085896e-01 -1.34278393e+00 -1.29225303e-03 -4.97139543e-01
-4.08944726e-01 1.12232363e+00 6.86129034e-01 -2.07095072e-02
5.84830523e-01 -2.27181211e-01 5.66451013e-01 -4.20623660e-01
-7.81670749e-01 -4.06280071e-01 1.14436162e+00 9.17064250e-01
4.70441997e-01 -2.01657772e-01 -2.11342961e-01 4.03376400e-01
1.59435794e-01 8.89121816e-02 5.63046038e-01 5.74945211e-01
-2.67917067e-01 -1.18277049e+00 1.33518837e-02 1.00943649e+00
-5.06840169e-01 -1.30808935e-01 -5.87986708e-01 6.66085958e-01
3.27608466e-01 6.90099835e-01 2.90920287e-01 -4.35545556e-02
7.10054114e-02 7.92574584e-02 7.58919477e-01 -8.85257363e-01
-4.09723371e-01 2.19523489e-01 3.06902200e-01 -7.09105730e-01
-3.91367644e-01 -5.91335237e-01 -8.73810709e-01 3.56461346e-01
-2.86791861e-01 -6.23888433e-01 3.93296540e-01 8.24463248e-01
1.14517605e+00 5.89305222e-01 3.07940602e-01 -4.38370019e-01
-4.74219501e-01 -9.72428501e-01 -6.48176551e-01 2.67874599e-01
7.08499730e-01 -8.83703768e-01 -2.16569200e-01 1.22388341e-01] | [15.792003631591797, -3.9550223350524902] |
9d202ad0-718d-4fdd-8a02-6879d7d8709c | blind-direction-of-arrival-estimation-in | 2005.08318 | null | https://arxiv.org/abs/2005.08318v2 | https://arxiv.org/pdf/2005.08318v2.pdf | Blind Direction-of-Arrival Estimation in Acoustic Vector-Sensor Arrays via Tensor Decomposition and Kullback-Leibler Divergence Covariance Fitting | A blind Direction-of-Arrivals (DOAs) estimate of narrowband signals for Acoustic Vector-Sensor (AVS) arrays is proposed. Building upon the special structure of the signal measured by an AVS, we show that the covariance matrix of all the received signals from the array admits a natural low-rank 4-way tensor representation. Thus, rather than estimating the DOAs directly from the raw data, our estimate arises from the unique parametric Canonical Polyadic Decomposition (CPD) of the observations' Second-Order Statistics (SOSs) tensor. By exploiting results from fundamental statistics and the recently re-emerging tensor theory, we derive a consistent blind CPD-based DOAs estimate without prior assumptions on the array configuration. We show that this estimate is a solution to an equivalent approximate joint diagonalization problem, and propose an ad-hoc iterative solution. Additionally, we derive the Cram\'er-Rao lower bound for Gaussian signals, and use it to derive the iterative Fisher scoring algorithm for the computation of the Maximum Likelihood Estimate (MLE) in this particular signal model. We then show that the MLE for the Gaussian model can in fact be used to obtain improved DOAs estimates for non-Gaussian signals as well (under mild conditions), which are optimal under the Kullback-Leibler divergence covariance fitting criterion, harnessing additional information encapsulated in the SOSs. Our analytical results are corroborated by simulation experiments in various scenarios, which also demonstrate the considerable improved accuracy w.r.t. a competing state-of-the-art blind estimate for AVS arrays, reducing the resulting root mean squared error by up to more than an order of magnitude. | ['Amir Weiss'] | 2020-05-17 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 1.13332316e-01 -1.40557304e-01 5.44558167e-01 -1.33271396e-01
-1.08235562e+00 -7.56267011e-01 2.19920471e-01 -1.88682899e-01
-3.11702549e-01 2.31840238e-01 5.35639584e-01 -1.97223037e-01
-5.94422281e-01 -2.24300548e-01 -5.99269092e-01 -1.22971344e+00
-3.32985938e-01 1.23015232e-01 -1.59481108e-01 -5.17012551e-02
-2.69421581e-02 6.37160838e-01 -1.36746025e+00 -3.37606817e-01
4.23739642e-01 1.33772016e+00 -9.77700055e-02 1.07428181e+00
3.18296373e-01 3.37647557e-01 -4.36458856e-01 -2.81999111e-01
3.89455646e-01 -3.65897745e-01 2.39795055e-02 1.83208764e-01
3.25985670e-01 -2.57235408e-01 -3.90551776e-01 1.11839914e+00
5.66525996e-01 1.38194654e-02 7.39756167e-01 -8.66336226e-01
-1.22552633e-01 3.09805661e-01 -3.96565944e-01 4.31485008e-03
4.96146321e-01 -4.08019930e-01 1.12447965e+00 -1.37003708e+00
7.56140426e-02 9.33571756e-01 9.76514161e-01 -7.08910823e-02
-1.00926435e+00 -4.23222452e-01 -4.92994070e-01 -8.55302066e-02
-1.68778813e+00 -9.40909028e-01 8.12735140e-01 -4.97203767e-01
4.90714729e-01 5.51539183e-01 4.71005768e-01 7.86205471e-01
-7.94712529e-02 7.04111576e-01 9.70284879e-01 -6.02286577e-01
4.05892015e-01 -2.60861278e-01 1.24378957e-01 6.95497930e-01
6.62157476e-01 2.07505256e-01 -8.17490280e-01 -4.71417576e-01
4.02791351e-01 -3.33013564e-01 -6.21233165e-01 -5.49328923e-01
-1.33400548e+00 6.02078438e-01 -2.52834428e-02 4.60012525e-01
-6.89430416e-01 2.14731127e-01 -3.69292521e-03 2.75752455e-01
3.63714337e-01 1.58693001e-01 -1.04051247e-01 -5.16312681e-02
-9.30323422e-01 1.21260315e-01 1.23865020e+00 7.80392289e-01
7.10145593e-01 6.00351453e-01 1.42524108e-01 7.17154562e-01
8.66863430e-01 1.45852256e+00 -7.50047266e-02 -8.26495051e-01
4.41133469e-01 -4.51526016e-01 4.82589960e-01 -1.27716100e+00
-5.48188806e-01 -1.07797480e+00 -9.20857489e-01 -2.20573768e-01
6.10640764e-01 -4.08216596e-01 -4.91776109e-01 1.65737057e+00
3.79823565e-01 3.54092032e-01 1.59426287e-01 9.74824786e-01
1.78582922e-01 5.04778206e-01 -7.62219906e-01 -6.41778946e-01
1.14871681e+00 -4.16030809e-02 -9.00391996e-01 -2.08975598e-01
4.05763119e-01 -9.64159429e-01 2.47642681e-01 8.45022917e-01
-9.30120528e-01 -1.30755886e-01 -1.31596148e+00 5.80196083e-01
3.61803681e-01 1.60988882e-01 2.83826679e-01 1.13306952e+00
-1.05255628e+00 1.95100158e-01 -8.95538211e-01 1.95132252e-02
-3.29841286e-01 -4.41666059e-02 -3.79117846e-01 -1.98188037e-01
-6.50617063e-01 4.22618419e-01 -3.15777838e-01 6.06315255e-01
-8.33464503e-01 -5.67364156e-01 -7.17025340e-01 -7.21145868e-02
2.02787235e-01 -4.80743319e-01 1.17471969e+00 -3.39109063e-01
-1.46956789e+00 1.40232995e-01 -5.90305388e-01 -2.23142102e-01
9.87412930e-02 -3.63999218e-01 -5.13521194e-01 4.77889955e-01
1.00786246e-01 -5.50303757e-01 1.33468556e+00 -1.39043164e+00
-2.18676224e-01 -4.38604295e-01 -5.12946546e-01 -1.92113947e-02
-3.23591083e-01 -9.89449322e-02 -1.84562147e-01 -8.22972119e-01
9.75085855e-01 -9.94048774e-01 -3.06242406e-01 -2.48070240e-01
-5.00747442e-01 4.06823993e-01 6.63551390e-02 -8.76069844e-01
1.24532712e+00 -2.47019148e+00 2.14787692e-01 7.82855093e-01
2.56537914e-01 -1.61478028e-01 -1.20828487e-01 8.38135421e-01
-1.90229252e-01 -3.75290602e-01 -3.86843443e-01 -5.80201626e-01
4.11723107e-02 7.15462565e-02 -3.65552634e-01 1.06791615e+00
-2.97251225e-01 2.63150334e-01 -8.39665532e-01 7.42100030e-02
1.40584596e-02 4.34573233e-01 -5.87179601e-01 4.14537996e-01
4.51775104e-01 5.42438626e-01 -5.23066223e-01 3.83460313e-01
9.28857565e-01 1.78858310e-01 8.16894025e-02 -6.01367354e-01
-2.62757868e-01 -1.37968287e-02 -1.79427767e+00 1.63112056e+00
-5.40989578e-01 5.96473992e-01 8.81415188e-01 -1.21077490e+00
7.77270377e-01 5.24466336e-01 6.04330361e-01 -2.81925619e-01
1.18211448e-01 7.11357176e-01 3.78872454e-02 -5.07483900e-01
1.65373787e-01 -3.73274088e-01 -2.36822572e-02 4.24760789e-01
7.27449954e-02 -1.79284632e-01 -2.12704167e-02 2.13021919e-01
1.21507752e+00 -3.69508237e-01 3.42252225e-01 -4.87768531e-01
5.96238196e-01 -5.91175079e-01 2.82937050e-01 1.00922155e+00
2.34377831e-01 5.55269480e-01 5.02565093e-02 9.70663354e-02
-7.37094820e-01 -1.13619089e+00 -2.22877726e-01 6.55049026e-01
-3.25671345e-01 -2.29118094e-01 -7.03516841e-01 4.17009108e-02
-7.93524232e-05 6.33633137e-01 -2.46234789e-01 2.26686731e-01
-4.93167430e-01 -8.64376843e-01 5.72119713e-01 1.86375044e-02
1.03798382e-01 1.99493974e-01 -2.97500640e-01 3.55851263e-01
-1.46283552e-01 -1.40853643e+00 -3.39230239e-01 2.10128516e-01
-6.04300320e-01 -7.27253258e-01 -9.23431218e-01 -2.21868619e-01
5.77798426e-01 6.78410828e-01 4.38622564e-01 -4.09998089e-01
2.32685521e-01 1.22756004e+00 -4.21361506e-01 -2.92333364e-01
-3.70120674e-01 -7.74050534e-01 7.30885923e-01 8.86533499e-01
-8.76522884e-02 -8.73840511e-01 -4.71341163e-01 4.41327035e-01
-6.68679178e-01 -4.03540134e-01 5.71363330e-01 6.34680390e-01
2.92689085e-01 7.93287978e-02 3.24429631e-01 -8.05426165e-02
5.45141339e-01 -4.79832023e-01 -7.79435217e-01 -6.11843914e-02
-4.48312193e-01 2.81764150e-01 4.74190772e-01 -1.99087068e-01
-9.00179982e-01 1.02417506e-01 -2.31161162e-01 -3.56797248e-01
1.41967818e-01 7.71169841e-01 -1.04245335e-01 -5.02104342e-01
4.67419595e-01 4.02941823e-01 -1.48089215e-01 -7.96930075e-01
4.31142390e-01 8.17990839e-01 6.57272041e-01 -4.96291518e-01
1.38337636e+00 6.52414799e-01 5.12184620e-01 -1.48764348e+00
-9.05530751e-01 -8.61543298e-01 -4.78349447e-01 -2.13001192e-01
6.18011594e-01 -1.04390681e+00 -8.36083174e-01 5.19465148e-01
-1.10142791e+00 2.75507331e-01 -3.29112262e-02 1.18583977e+00
-3.36893886e-01 7.85594046e-01 -4.17592585e-01 -1.46865380e+00
-1.76217221e-02 -7.88297832e-01 9.41097021e-01 -4.20137197e-01
8.20471048e-02 -8.52384388e-01 2.64804065e-01 2.27437407e-01
4.05240953e-01 3.27606909e-02 4.49495673e-01 -5.59619367e-01
-4.76716518e-01 -5.80332100e-01 -2.80306097e-02 4.98969346e-01
-1.92379296e-01 -4.95455682e-01 -1.02776515e+00 -4.30285662e-01
7.72610843e-01 4.22583371e-01 4.95288044e-01 5.98606825e-01
2.64542639e-01 -5.33324659e-01 -3.85981910e-02 6.57665610e-01
1.49671423e+00 -1.35870546e-01 1.50242135e-01 -2.37685576e-01
8.13759387e-01 4.82019067e-01 4.14813548e-01 8.23155165e-01
1.92139477e-01 7.26477444e-01 3.89239252e-01 3.61923486e-01
-1.15776949e-01 1.23728747e-02 4.29118872e-01 1.70154452e+00
-6.73171207e-02 -1.41426757e-01 -7.60123551e-01 6.17720187e-01
-1.46559227e+00 -7.12965906e-01 -7.13614941e-01 2.76148844e+00
4.42891717e-01 -2.90441006e-01 -1.25269711e-01 5.84707022e-01
3.58169049e-01 1.94907799e-01 -1.77473113e-01 -2.56396877e-03
-1.93896741e-01 2.59778738e-01 9.07981873e-01 8.74162078e-01
-7.70483792e-01 1.05823778e-01 6.32117939e+00 6.31418049e-01
-7.20387578e-01 2.60340512e-01 -2.84232140e-01 3.19855154e-01
-5.14559031e-01 4.05624360e-02 -4.61349547e-01 1.05382182e-01
1.03158426e+00 -7.48300105e-02 5.97888529e-01 2.97721386e-01
2.82141984e-01 -1.32264107e-01 -8.95880342e-01 1.27099884e+00
2.33064637e-01 -7.63759553e-01 -4.09768879e-01 3.27950269e-01
5.42404652e-01 1.68840870e-01 -5.33216633e-03 -3.72188926e-01
-9.90914330e-02 -4.58174348e-01 9.84722316e-01 6.45874918e-01
5.81604362e-01 -3.61595720e-01 5.43295503e-01 3.38938475e-01
-1.25494802e+00 -2.62930214e-01 -1.80710644e-01 -1.34608701e-01
4.33090746e-01 1.41238487e+00 -7.92914569e-01 7.87749588e-01
3.89324456e-01 4.62027848e-01 5.67518100e-02 1.28000140e+00
-2.28335038e-01 1.00019419e+00 -8.52750957e-01 -6.88166246e-02
2.52250820e-01 -4.51238483e-01 1.32908702e+00 1.27304542e+00
9.45729196e-01 2.71127015e-01 -1.44135624e-01 2.54429013e-01
2.83109725e-01 1.72672406e-01 -4.38760489e-01 -1.52362371e-02
4.46395725e-01 1.04600358e+00 -2.90975660e-01 -7.58125354e-03
-3.52569103e-01 6.32513463e-01 -4.93988365e-01 9.10143018e-01
-2.31579497e-01 -4.07144159e-01 6.94888115e-01 -4.62262072e-02
7.25086212e-01 -7.78075635e-01 -2.60957539e-01 -1.15889752e+00
2.98875928e-01 -8.29421699e-01 -6.58478737e-02 -5.63777447e-01
-1.16429389e+00 2.93070436e-01 3.35480012e-02 -1.59390235e+00
-4.65364516e-01 -6.64025486e-01 -3.67299348e-01 8.31439555e-01
-1.26536596e+00 -7.35771239e-01 1.10501267e-01 4.76124167e-01
-1.57488152e-01 -5.05217612e-02 8.69498014e-01 4.72548395e-01
-2.92035937e-01 4.54118878e-01 7.82169580e-01 6.22676648e-02
2.39782438e-01 -1.05828476e+00 1.05409548e-02 1.21137285e+00
3.14424366e-01 7.61393249e-01 1.37303090e+00 -3.98776829e-01
-2.28446770e+00 -4.56155419e-01 6.18225217e-01 -2.70133197e-01
1.21675849e+00 -4.12161350e-01 -5.45480609e-01 4.65903670e-01
-7.25484118e-02 1.34000361e-01 9.10444081e-01 3.22444700e-02
-5.79259396e-01 -2.97574967e-01 -6.84087992e-01 1.95072547e-01
9.47961330e-01 -6.15423381e-01 -4.88026232e-01 3.51686358e-01
3.55540127e-01 -2.68478215e-01 -9.49189305e-01 3.76586467e-01
7.46425331e-01 -9.66855824e-01 1.08044946e+00 -2.22151399e-01
-2.57813811e-01 -6.69743776e-01 -1.08141577e+00 -1.28238320e+00
-2.80240119e-01 -1.15818071e+00 -1.41784236e-01 1.05324483e+00
2.94914097e-01 -7.50941217e-01 4.09065366e-01 8.45506340e-02
-3.44972491e-01 -2.84691155e-01 -1.39074969e+00 -8.98640871e-01
-4.68941212e-01 -1.18771672e+00 1.65058479e-01 7.03582227e-01
-6.69641197e-02 2.67923862e-01 -6.38900399e-01 1.06856012e+00
1.32671130e+00 -1.08715385e-01 7.63324201e-01 -1.09270406e+00
-8.80394518e-01 -8.00048932e-02 -5.39313376e-01 -1.52796638e+00
-1.08423218e-01 -7.14117348e-01 3.82690251e-01 -1.26469123e+00
-3.74186993e-01 -3.83103907e-01 -1.63565353e-01 -9.12180170e-02
2.33677998e-01 2.56181091e-01 2.20139250e-01 1.00948349e-01
-2.05026448e-01 5.13625741e-01 7.61080682e-01 1.46655962e-01
1.38280883e-01 4.62464720e-01 -4.45577174e-01 8.37934077e-01
1.34056449e-01 -6.28678143e-01 -2.06603751e-01 -5.52385926e-01
7.08248913e-01 4.33427602e-01 2.16232210e-01 -1.16518176e+00
4.23624575e-01 2.38153875e-01 -9.17732809e-03 -3.84998739e-01
6.70261502e-01 -8.64487171e-01 3.11778516e-01 2.98311025e-01
8.49763229e-02 -5.04679203e-01 -2.46015135e-02 8.67669046e-01
-2.00730965e-01 -5.10655701e-01 3.88987750e-01 3.70363533e-01
-1.22048460e-01 -7.19272941e-02 -6.69827700e-01 -1.85079962e-01
2.65219033e-01 7.66171096e-03 7.05316961e-02 -9.60914969e-01
-6.61936700e-01 -3.67336720e-01 -1.30534634e-01 -2.46193320e-01
5.95876038e-01 -1.20602500e+00 -1.04526997e+00 3.72260958e-01
4.26044054e-02 -5.24367273e-01 3.63634080e-01 1.23446488e+00
-3.05332422e-01 5.09416580e-01 5.12092948e-01 -7.47105777e-01
-9.47280228e-01 1.77895993e-01 2.31212318e-01 1.55449539e-01
-2.69633055e-01 9.74725246e-01 2.44464204e-01 -2.11247265e-01
8.80704895e-02 -2.24023238e-01 5.16683199e-02 4.97639515e-02
7.18130708e-01 5.02093494e-01 2.96530634e-01 -1.14811540e+00
-4.31988597e-01 1.08593285e+00 5.73065042e-01 -7.61655927e-01
1.15867233e+00 -4.77382958e-01 -2.65235990e-01 4.46088672e-01
1.64682972e+00 1.05784988e+00 -8.23248863e-01 -3.38025451e-01
-3.49340141e-02 -6.04849637e-01 3.05170774e-01 -3.18715006e-01
-8.30173194e-01 8.60258400e-01 6.01112902e-01 4.42667544e-01
1.11824107e+00 -3.35759707e-02 4.58318919e-01 6.43077850e-01
6.53468013e-01 -3.80092919e-01 -1.71734408e-01 3.89769703e-01
1.08882368e+00 -6.09486520e-01 1.09638087e-01 -4.73985523e-01
-1.13229677e-01 1.06395578e+00 -6.41835988e-01 -9.39030796e-02
9.58895087e-01 2.25427583e-01 1.50589600e-01 -1.13777608e-01
-1.52696744e-01 -1.25932470e-01 4.62078810e-01 6.75602138e-01
1.45443708e-01 2.70649940e-01 -1.12282857e-01 4.74725544e-01
-4.01640862e-01 -5.65839231e-01 7.59353876e-01 4.69280779e-01
-6.07222140e-01 -8.33814621e-01 -9.99138296e-01 2.72209764e-01
-5.10710597e-01 -1.56654969e-01 1.59386039e-01 1.59456715e-01
-4.99516189e-01 1.41965294e+00 -3.89173537e-01 -4.29602981e-01
4.17371511e-01 -6.86858669e-02 3.62235963e-01 -2.85269737e-01
1.97682276e-01 6.23667657e-01 4.26488996e-01 -4.85420525e-01
-3.93689096e-01 -9.51858759e-01 -8.57840836e-01 -1.61763141e-03
-4.23604459e-01 5.08538485e-01 1.19051540e+00 8.82597327e-01
7.23116919e-02 1.26639202e-01 1.04194069e+00 -9.39136982e-01
-8.89742970e-01 -7.93506444e-01 -1.03110850e+00 -1.71602387e-02
7.62734354e-01 -5.06321609e-01 -1.00569189e+00 -1.41081885e-01] | [6.551137924194336, 1.3878005743026733] |
dd35f4ef-2bbb-47dc-a069-a87e2a342e26 | a-ga-based-approach-for-selection-of-local | 1501.05495 | null | http://arxiv.org/abs/1501.05495v1 | http://arxiv.org/pdf/1501.05495v1.pdf | A GA Based approach for selection of local features for recognition of handwritten Bangla numerals | Soft computing approaches are mainly designed to address the real world
ill-defined, imprecisely formulated problems, combining different kind of novel
models of computation, such as neural networks, genetic algorithms (GAs.
Handwritten digit recognition is a typical example of one such problem. In the
current work we have developed a two-pass approach where the first pass
classifier performs a coarse classification, based on some global features of
the input pattern by restricting the possibility of classification decisions
within a group of classes, smaller than the number of classes considered
initially. In the second pass, the group specific classifiers concentrate on
the features extracted from the selected local regions, and refine the earlier
decision by combining the local and the global features for selecting the true
class of the input pattern from the group of candidate classes selected in the
first pass. To optimize the selection of local regions a GA based approach has
been developed here. The maximum recognition performance on Bangla digit
samples as achieved on the test set, during the first pass of the two pass
approach is 93.35%. After combining the results of the two stage classifiers,
an overall success rate of 95.25% is achieved. | ['Mahantapas Kundu', 'Ram Sarkar', 'Nibaran Das', 'Subhadip Basu', 'Mita Nasipuri', 'Punam Kumar Saha'] | 2015-01-22 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.82751673e-01 -3.46098654e-02 5.01405634e-02 -3.95438075e-01
-2.71883458e-01 -3.33009213e-01 5.81349909e-01 4.59269047e-01
-4.37196016e-01 8.55302870e-01 -8.01613033e-02 -4.75608036e-02
-5.62640667e-01 -1.09624755e+00 -1.24455974e-01 -9.64129567e-01
1.01622820e-01 6.80340350e-01 3.30199122e-01 -7.21251145e-02
8.45858634e-01 8.46909523e-01 -1.80876541e+00 5.46574593e-01
8.80439818e-01 9.90346730e-01 1.34046867e-01 5.32187998e-01
-2.70168275e-01 6.56944454e-01 -6.90186858e-01 1.36305898e-01
1.89348683e-01 -6.14580810e-01 -7.31992483e-01 3.88109624e-01
-3.55750173e-01 1.25735343e-01 2.86589414e-01 9.90276456e-01
3.77878666e-01 3.08119208e-01 8.01128507e-01 -6.96808934e-01
-5.90808541e-02 4.83161867e-01 -1.79033726e-01 1.56420693e-02
2.59250432e-01 -1.68299869e-01 5.15533090e-01 -7.91525960e-01
3.96113187e-01 1.01178062e+00 2.91561812e-01 1.74882561e-01
-1.01454377e+00 -2.51524091e-01 -5.97166047e-02 2.84784198e-01
-1.61593604e+00 -5.37462309e-02 7.25356698e-01 -4.60417151e-01
8.71935248e-01 4.09905881e-01 5.65779209e-01 -7.75863975e-03
2.63556480e-01 3.40378374e-01 1.26532590e+00 -8.88829291e-01
6.61243916e-01 5.27793169e-01 5.09246647e-01 2.69898713e-01
2.59828091e-01 5.73179610e-02 -8.45265687e-02 -1.28260106e-01
2.60223448e-01 -4.82652485e-02 -2.53283791e-02 -1.08784005e-01
-7.13818073e-01 8.20441484e-01 3.74385238e-01 1.15616906e+00
-6.84435666e-01 -6.30905628e-01 1.81479678e-01 1.31460959e-02
1.54527113e-01 5.82698166e-01 -2.97508210e-01 2.60030711e-03
-1.13177955e+00 6.49934188e-02 9.03885961e-01 4.23821032e-01
7.69105196e-01 -2.17764258e-01 -1.39279768e-01 6.81328177e-01
2.75736839e-01 3.44402157e-02 6.65819645e-01 1.45862550e-01
2.70321608e-01 1.18524909e+00 -1.09583400e-01 -9.52250302e-01
-2.26954132e-01 -4.30285513e-01 -6.35226667e-01 5.82291186e-01
5.09573340e-01 -3.14329803e-01 -1.27742505e+00 1.11602557e+00
3.75166148e-01 -4.34964508e-01 3.83951724e-01 7.56612718e-01
5.35779178e-01 9.82205927e-01 6.67915940e-02 -3.04130346e-01
1.14134288e+00 -6.03391349e-01 -4.39447612e-01 1.26489401e-01
2.23186553e-01 -7.52943754e-01 3.15056503e-01 8.45308483e-01
-6.43747211e-01 -7.76212335e-01 -1.26206505e+00 5.78225434e-01
-6.78616881e-01 7.71116555e-01 3.40970248e-01 7.27897525e-01
-6.48698866e-01 4.87015456e-01 -5.08268714e-01 -3.13535899e-01
1.50756136e-01 7.06153989e-01 -1.87504217e-01 -2.10484475e-01
-7.35703707e-01 9.88387525e-01 7.60430813e-01 4.60646749e-01
-4.53645915e-01 1.25554390e-03 -2.25040376e-01 1.56163082e-01
2.19895169e-01 1.84852242e-01 4.31522667e-01 -1.34613645e+00
-1.48253334e+00 7.44541347e-01 6.16192482e-02 -3.27713549e-01
4.63491410e-01 3.37093055e-01 -3.30849290e-01 2.98320986e-02
-3.13043952e-01 1.85624927e-01 5.16493738e-01 -1.02542627e+00
-9.43126380e-01 -5.41268408e-01 -3.01528215e-01 6.75528497e-02
7.58996084e-02 1.49511829e-01 -6.57687187e-02 -4.11078304e-01
4.11563456e-01 -6.88676536e-01 -2.95317262e-01 -5.48866928e-01
-1.03148185e-01 -3.17226142e-01 8.46424162e-01 -6.92378700e-01
1.15150130e+00 -2.25161719e+00 2.82197267e-01 1.00725663e+00
-4.15453076e-01 5.77174723e-01 3.40529680e-01 1.34349912e-01
-1.50091186e-01 -1.43166557e-01 -4.17336673e-01 2.54033834e-01
-3.41381431e-01 4.94888164e-02 1.25291705e-01 2.14545205e-02
4.11484271e-01 -1.26124874e-01 -4.68768656e-01 -2.86768913e-01
3.05735707e-01 2.05604091e-01 -1.98072776e-01 1.92123145e-01
-1.87781245e-01 1.46752775e-01 -6.46514297e-01 5.94911516e-01
8.71736944e-01 3.28837514e-01 4.09177572e-01 -2.32121542e-01
-2.49101102e-01 -9.19276699e-02 -1.76894557e+00 7.74589062e-01
-9.72325802e-02 2.97832161e-01 -2.35573813e-01 -1.47764456e+00
1.65857911e+00 3.75217795e-01 4.17982131e-01 -4.16857421e-01
3.66614312e-01 4.80868608e-01 3.56843919e-01 -5.18746078e-01
3.18838954e-01 3.15825231e-02 1.36686340e-01 2.78476804e-01
8.79028961e-02 -8.92279670e-02 4.12083805e-01 -4.65389609e-01
4.61937010e-01 2.08065569e-01 6.43011212e-01 -1.34208634e-01
9.83999908e-01 2.71301359e-01 4.19426024e-01 4.85966295e-01
2.64303479e-03 5.80912292e-01 3.79741251e-01 -6.74132645e-01
-9.73123550e-01 -2.53420442e-01 -3.26925367e-01 5.54545343e-01
-1.19808376e-01 3.31833154e-01 -7.76013434e-01 -6.31247044e-01
-1.59625128e-01 6.14917874e-01 -3.85204554e-01 -1.12089001e-01
-5.07408559e-01 -1.23243546e+00 2.99984097e-01 -7.71802440e-02
7.68582702e-01 -1.19742012e+00 -8.00215662e-01 5.22246003e-01
4.70010102e-01 -4.14040715e-01 5.59405565e-01 5.95021546e-01
-9.01982725e-01 -8.34074855e-01 -5.67450464e-01 -8.75870824e-01
8.61692905e-01 -4.64648098e-01 4.66064185e-01 3.12978178e-01
-3.76217186e-01 -3.73880267e-01 -7.48300076e-01 -3.16499084e-01
-4.65208262e-01 6.25156313e-02 -3.53102058e-01 4.69399363e-01
5.60399473e-01 -2.13142395e-01 6.19375668e-02 2.12268695e-01
-7.08854198e-01 -2.66422302e-01 8.51428866e-01 8.33366513e-01
5.06015301e-01 5.43483317e-01 7.38709629e-01 -5.07220089e-01
5.15311658e-01 -4.40383703e-01 -6.50018156e-01 6.01021647e-01
-4.04698491e-01 2.40321174e-01 6.96900606e-01 -5.16268551e-01
-1.08914793e+00 3.43824327e-01 -1.06427453e-01 2.93350458e-01
-4.91662860e-01 7.86464572e-01 -4.01129603e-01 -2.94755369e-01
8.25339139e-01 4.25656140e-01 1.83707952e-01 -3.51891756e-01
-4.74119455e-01 1.01487327e+00 2.43185073e-01 -4.92377162e-01
2.94331104e-01 -2.27813408e-01 -6.88155219e-02 -7.45729804e-01
-3.64772789e-02 -2.26360410e-01 -7.06416667e-01 -4.30353433e-01
8.38587761e-01 -2.30149344e-01 -2.19957307e-01 7.45233059e-01
-8.28547716e-01 1.44411534e-01 -1.83020517e-01 8.76580417e-01
-1.72624499e-01 1.19770780e-01 4.77153435e-02 -1.21697783e+00
-2.48774186e-01 -1.34319222e+00 4.62776989e-01 6.64354384e-01
-2.60690451e-01 -6.17061079e-01 -2.71846145e-01 2.75227632e-02
1.99189499e-01 2.99132735e-01 1.04104114e+00 -1.29252303e+00
-5.43847203e-01 -6.49826705e-01 1.36085078e-01 5.87812066e-01
4.34730142e-01 3.61274689e-01 -6.03030682e-01 -3.18203457e-02
5.85501716e-02 -1.11586615e-01 6.22073710e-01 3.39812964e-01
7.71262109e-01 8.50875974e-02 -3.88708681e-01 1.29169732e-01
1.62707651e+00 1.15581667e+00 7.74858952e-01 4.32191283e-01
3.61913927e-02 3.32759470e-01 1.10289669e+00 6.17008924e-01
-2.01507539e-01 5.73813438e-01 7.89462030e-02 -7.30132237e-02
1.58695951e-01 3.52316082e-01 -1.85557120e-02 1.45389408e-01
-2.92305797e-01 -2.26839423e-01 -9.90406096e-01 4.01089221e-01
-1.63240302e+00 -7.39519179e-01 1.76012963e-01 2.27994609e+00
6.17231190e-01 4.20532942e-01 2.22727701e-01 1.02966738e+00
1.11432171e+00 -3.91146570e-01 -1.52040720e-01 -8.35165858e-01
-2.08282799e-01 4.46297497e-01 2.74876446e-01 5.96511662e-01
-9.78421330e-01 4.99646187e-01 4.95672560e+00 7.26561129e-01
-1.38882732e+00 -6.07364297e-01 8.03550541e-01 1.10450879e-01
1.37712657e-01 8.35288540e-02 -8.51283371e-01 5.67139268e-01
5.75256228e-01 -1.06559023e-02 3.86443019e-01 5.81183016e-01
7.85987526e-02 -7.69394517e-01 -9.37296748e-01 5.94777286e-01
1.48307681e-01 -9.71822441e-01 1.53325394e-01 -2.30652802e-02
7.51155734e-01 -7.12956607e-01 -2.57828414e-01 -5.12108691e-02
-1.35198638e-01 -1.05312693e+00 6.91467524e-01 9.00421679e-01
1.47656664e-01 -1.12198722e+00 1.28252375e+00 6.07057452e-01
-9.13047075e-01 -3.80663186e-01 -2.61689574e-01 -7.84505606e-02
-2.92780548e-01 6.22122645e-01 -1.00399411e+00 7.59825706e-01
4.42308366e-01 1.07764043e-01 -4.67512190e-01 1.40998328e+00
1.12026311e-01 3.85966361e-01 -5.66282749e-01 -7.32731104e-01
4.69811678e-01 -2.73134172e-01 4.42604274e-01 1.17928576e+00
4.08674419e-01 3.90591860e-01 -7.57752955e-02 7.29261637e-01
7.06227720e-01 4.36062634e-01 -2.82828391e-01 9.60946381e-02
4.53071654e-01 9.31584895e-01 -1.08291221e+00 -5.22282720e-01
1.20447114e-01 4.93755102e-01 -1.37351369e-02 5.80437370e-02
-3.73118997e-01 -9.19079006e-01 -2.94847578e-01 -2.66558081e-01
4.21474099e-01 -5.76768368e-02 -6.30995512e-01 -6.99672937e-01
1.74476981e-01 -1.12113118e+00 5.14481068e-01 -3.59780073e-01
-7.04190254e-01 6.92308664e-01 -1.02551319e-02 -9.87043321e-01
-4.83171880e-01 -8.05246890e-01 -7.16685951e-01 1.40332925e+00
-8.27292144e-01 -8.04643810e-01 -2.30667502e-01 4.35399920e-01
4.60732102e-01 -5.12300253e-01 7.64113843e-01 -2.61159278e-02
-5.17521143e-01 3.81038010e-01 2.91883081e-01 2.93145664e-02
2.14108571e-01 -9.29631531e-01 -5.54573953e-01 1.09038031e+00
-2.39250347e-01 2.65379548e-01 5.85249424e-01 -7.44790673e-01
-7.22629726e-01 -5.23829401e-01 1.01911473e+00 4.61064428e-01
-8.19814280e-02 6.56425208e-02 -9.13801551e-01 7.79302325e-03
-2.17655674e-01 -3.65197688e-01 3.48159999e-01 -3.10223401e-01
4.17589247e-01 -2.54326105e-01 -1.70029545e+00 1.94535568e-01
4.74158861e-02 -4.54731248e-02 -6.35175645e-01 5.94426170e-02
-4.08892065e-01 -1.69004604e-01 -9.52482700e-01 3.74923170e-01
5.56673825e-01 -8.14737856e-01 5.64207375e-01 -2.71162331e-01
2.48220205e-01 -6.65566087e-01 -2.42983624e-01 -1.19130969e+00
-2.70948797e-01 1.37808630e-02 5.64358413e-01 1.33792162e+00
7.93031633e-01 -6.28087997e-01 7.93986022e-01 5.69192231e-01
2.05728054e-01 -7.99494088e-01 -7.96945751e-01 -2.44793802e-01
-3.37766767e-01 3.26586574e-01 6.48788869e-01 7.42242277e-01
3.12714353e-02 -2.43341327e-01 7.19166547e-02 1.71462551e-01
3.89275938e-01 2.57147908e-01 3.74079674e-01 -1.25143087e+00
-3.69029224e-01 -3.10697705e-01 -8.44549000e-01 -1.93327472e-01
-3.51719558e-01 -5.20457268e-01 7.47068301e-02 -1.34748387e+00
-5.36487699e-02 -7.87051201e-01 -1.97427794e-01 4.87554461e-01
-1.70223504e-01 1.33960182e-02 2.94473588e-01 4.80209626e-02
2.33052894e-01 -9.77116525e-02 9.01117802e-01 -3.87699809e-03
-5.49445629e-01 4.05314505e-01 -2.18204662e-01 4.66117501e-01
8.19346130e-01 -2.41606891e-01 -2.53207386e-01 -3.49709168e-02
-3.23361129e-01 1.25610560e-01 -1.66226670e-01 -1.29493165e+00
3.47170144e-01 -4.82668966e-01 7.95631051e-01 -7.50035822e-01
6.93172589e-02 -9.58553493e-01 4.90133077e-01 7.51599371e-01
-3.21931273e-01 -2.41895989e-01 1.27834141e-01 1.96613315e-02
-5.12721360e-01 -8.76248419e-01 9.23335075e-01 -1.44534200e-01
-7.86665440e-01 -3.10923487e-01 -5.29689968e-01 -6.38180614e-01
1.40476334e+00 -9.93787169e-01 2.26822987e-01 5.84988184e-02
-1.18152487e+00 -2.92458504e-01 8.84211564e-04 -1.10892709e-02
3.13902080e-01 -9.41419005e-01 -7.17054904e-01 2.93452352e-01
-1.73316091e-01 -2.05570459e-01 -1.66318007e-03 3.69388610e-01
-7.54984736e-01 4.18012083e-01 -8.08078885e-01 -3.77629787e-01
-1.53993738e+00 4.32631582e-01 5.28907657e-01 -3.53167027e-01
-1.36029124e-01 6.84641302e-01 -6.36715293e-01 -6.05633371e-02
5.71392179e-02 -3.83786112e-01 -8.73253584e-01 2.69632548e-01
3.09240431e-01 4.51612890e-01 5.75860977e-01 -6.20444655e-01
-4.58343178e-01 7.87060499e-01 7.42456689e-02 -2.24363148e-01
1.48236537e+00 3.55685830e-01 -2.59278893e-01 1.09366953e-01
9.18509305e-01 5.42740598e-02 -6.84849858e-01 -6.88763708e-02
1.07200377e-01 -4.13026869e-01 -1.26253942e-03 -1.29737413e+00
-8.70749414e-01 4.73582923e-01 7.91292131e-01 3.42157632e-01
1.65006769e+00 -4.39501226e-01 -1.65999740e-01 3.83822858e-01
3.32774460e-01 -1.23289847e+00 -5.32737792e-01 3.90648276e-01
8.73422801e-01 -7.69258201e-01 1.06336378e-01 -3.69176060e-01
-5.09206772e-01 1.71432543e+00 5.07345676e-01 -3.55933875e-01
5.12233317e-01 2.07209274e-01 -3.67574066e-01 1.98837951e-01
-5.06137013e-01 -8.97296518e-03 2.56551534e-01 4.08807933e-01
4.59819168e-01 1.56568110e-01 -1.11376274e+00 6.51265442e-01
-7.39106312e-02 3.53638232e-01 3.98177892e-01 1.25393891e+00
-7.96360552e-01 -1.15126777e+00 -9.17168617e-01 4.36096460e-01
-4.21486765e-01 1.77399859e-01 -5.57951033e-01 7.08448350e-01
6.85574651e-01 9.02044415e-01 3.17749023e-01 -4.39372599e-01
2.60771960e-01 6.30679801e-02 5.42019963e-01 -3.67867798e-01
-8.39199185e-01 -1.64591044e-01 1.91921994e-01 2.20445454e-01
-3.29565883e-01 -7.98855484e-01 -1.14666665e+00 -7.45638832e-02
-4.53988224e-01 5.91189682e-01 1.10348904e+00 1.10702491e+00
1.64103359e-02 2.63528019e-01 8.25907707e-01 -8.58808517e-01
-6.19864702e-01 -8.96168470e-01 -4.37821776e-01 9.16989371e-02
9.85357463e-02 -5.70914030e-01 -2.45276690e-01 -2.13669278e-02] | [7.934452056884766, 3.6769635677337646] |
4c9d9be7-e933-422e-9de3-587a8b21ea27 | ran4iqa-restorative-adversarial-nets-for-no | 1712.05444 | null | http://arxiv.org/abs/1712.05444v1 | http://arxiv.org/pdf/1712.05444v1.pdf | RAN4IQA: Restorative Adversarial Nets for No-Reference Image Quality Assessment | Inspired by the free-energy brain theory, which implies that human visual
system (HVS) tends to reduce uncertainty and restore perceptual details upon
seeing a distorted image, we propose restorative adversarial net (RAN), a
GAN-based model for no-reference image quality assessment (NR-IQA). RAN, which
mimics the process of HVS, consists of three components: a restorator, a
discriminator and an evaluator. The restorator restores and reconstructs input
distorted image patches, while the discriminator distinguishes the
reconstructed patches from the pristine distortion-free patches. After
restoration, we observe that the perceptual distance between the restored and
the distorted patches is monotonic with respect to the distortion level. We
further define Gain of Restoration (GoR) based on this phenomenon. The
evaluator predicts perceptual score by extracting feature representations from
the distorted and restored patches to measure GoR. Eventually, the quality
score of an input image is estimated by weighted sum of the patch scores.
Experimental results on Waterloo Exploration, LIVE and TID2013 show the
effectiveness and generalization ability of RAN compared to the
state-of-the-art NR-IQA models. | ['Hongyu Ren', 'Diqi Chen', 'Yizhou Wang'] | 2017-12-14 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 3.78014117e-01 8.72894451e-02 4.42855716e-01 -1.29945084e-01
-6.84704542e-01 -3.34024668e-01 4.79752809e-01 -3.19057226e-01
-5.30170090e-02 6.18441463e-01 5.31300783e-01 1.44936770e-01
9.67657566e-03 -9.65103269e-01 -7.93904245e-01 -1.06740212e+00
1.01187430e-01 -2.12245256e-01 1.00678764e-01 -1.52697548e-01
2.35263512e-01 2.07705900e-01 -1.45328867e+00 3.59305114e-01
1.27665806e+00 1.35624719e+00 1.40633792e-01 6.14036322e-01
3.91523272e-01 8.02763879e-01 -8.55278552e-01 -3.45552474e-01
5.16618431e-01 -9.13309693e-01 -5.39111257e-01 -1.94319338e-02
3.36959243e-01 -4.30507869e-01 -6.41129196e-01 1.51295912e+00
6.77734256e-01 1.20138131e-01 8.13204229e-01 -1.19669604e+00
-1.36590695e+00 6.53385594e-02 -3.83981109e-01 2.97865003e-01
3.85068268e-01 3.65603298e-01 5.66302717e-01 -8.85453880e-01
4.54411060e-01 1.32578063e+00 4.25426006e-01 4.80137587e-01
-1.34476328e+00 -4.53985512e-01 -2.24399894e-01 3.83517474e-01
-1.38096607e+00 -4.02770877e-01 9.00662303e-01 -4.00174856e-01
6.04636610e-01 3.92542958e-01 6.74943686e-01 7.73285747e-01
6.76474035e-01 5.16655207e-01 1.41965890e+00 -3.36180210e-01
4.83628929e-01 -9.24011022e-02 -4.04291064e-01 3.04501086e-01
-2.66778886e-01 6.91220343e-01 -3.05133611e-01 1.30064413e-01
8.25994194e-01 -6.00457601e-02 -8.58296156e-01 -3.59006256e-01
-9.09574628e-01 5.01253247e-01 1.16492736e+00 7.72354677e-02
-6.33540928e-01 -7.75343850e-02 4.92725298e-02 4.77465302e-01
3.69728118e-01 2.10183322e-01 3.05058420e-01 3.52759480e-01
-8.99244487e-01 3.01896762e-02 3.11626166e-01 5.28154075e-01
5.94530702e-01 3.86682242e-01 -4.74803865e-01 8.50088000e-01
2.65559256e-01 5.83807468e-01 5.46173751e-01 -1.09094334e+00
-2.95146275e-02 5.81762552e-01 7.54184797e-02 -1.14912772e+00
2.57186562e-01 -5.02308011e-01 -1.41964161e+00 8.87885869e-01
1.07326452e-03 3.33072215e-01 -1.19296503e+00 1.70528746e+00
9.01002064e-02 2.00518802e-01 1.82714388e-01 1.20084751e+00
8.43678951e-01 8.40482354e-01 -3.21023948e-02 -3.12763810e-01
8.75170052e-01 -9.94038224e-01 -7.95330882e-01 2.93817241e-02
-3.90071332e-01 -6.51543379e-01 1.10778213e+00 4.33562338e-01
-1.40756357e+00 -1.08325720e+00 -1.44321477e+00 7.04847835e-03
-8.98761079e-02 -3.12251031e-01 -9.95691270e-02 5.49856365e-01
-1.41223717e+00 7.52416492e-01 -5.32299638e-01 1.51877135e-01
6.15523636e-01 -1.01901554e-01 -2.80247957e-01 -3.28247011e-01
-1.24288189e+00 8.64097714e-01 1.07759088e-01 1.79435626e-01
-1.51663697e+00 -8.43324900e-01 -7.77380228e-01 7.66527727e-02
-1.17621437e-01 -8.65904927e-01 6.93763196e-01 -1.37430739e+00
-1.46798658e+00 7.27500260e-01 8.10757056e-02 -3.63682061e-01
6.31875038e-01 1.48080885e-01 -7.14935362e-01 1.89667895e-01
5.26074395e-02 6.91703618e-01 1.19209254e+00 -1.65108275e+00
-2.78916121e-01 -4.08316344e-01 -3.17444392e-02 3.17709297e-01
1.50218353e-01 -4.61926341e-01 -2.07403421e-01 -8.65357816e-01
9.78360549e-02 -4.50032383e-01 7.34280795e-02 1.20650239e-01
-4.77782011e-01 3.00346106e-01 6.34378791e-01 -1.01494646e+00
8.94254386e-01 -2.45803928e+00 2.82741338e-01 2.67139286e-01
2.71222800e-01 2.80795664e-01 -5.09886026e-01 2.45566055e-01
-2.61632890e-01 4.37640734e-02 -5.42591274e-01 1.21750208e-02
-9.44075882e-02 1.23081788e-01 -4.13144529e-01 4.71249819e-01
7.56968483e-02 9.63554263e-01 -8.37861478e-01 -1.62001818e-01
3.18777829e-01 8.75639200e-01 -2.41921499e-01 5.39443314e-01
2.10290194e-01 6.43963575e-01 3.27599347e-02 6.14162743e-01
1.05883396e+00 1.62632152e-01 -2.75868118e-01 -3.32172394e-01
1.94948524e-01 6.71603084e-02 -1.03609931e+00 1.56052029e+00
-2.98476636e-01 5.29283583e-01 -1.61545277e-02 -7.56139040e-01
1.00288582e+00 2.48015970e-01 -1.02558017e-01 -1.35151315e+00
6.02853149e-02 8.04133192e-02 1.74625404e-02 -3.80414605e-01
3.72085124e-01 -2.66616404e-01 2.61819392e-01 1.95294201e-01
-1.08829737e-02 -1.48101658e-01 -2.94692546e-01 9.33098570e-02
1.04324436e+00 7.40788653e-02 2.94577569e-01 -1.85717911e-01
5.71404278e-01 -5.79046071e-01 6.02951288e-01 4.48538721e-01
-4.43203956e-01 8.86649430e-01 2.32163191e-01 -3.21351856e-01
-1.12889457e+00 -1.64509344e+00 -1.25333235e-01 6.40696645e-01
6.62020326e-01 3.11239600e-01 -1.02052701e+00 -3.80541176e-01
-1.67865738e-01 8.76792967e-01 -7.38488138e-01 -6.09022975e-01
-2.89738208e-01 -3.53546530e-01 3.09934914e-01 4.42818284e-01
1.04627955e+00 -1.52007353e+00 -4.41161901e-01 7.80380517e-02
-3.57442051e-01 -5.10599315e-01 -6.59823895e-01 -2.85528183e-01
-5.41475058e-01 -9.73623514e-01 -8.12309742e-01 -7.86336601e-01
8.52690756e-01 1.97543368e-01 1.20395696e+00 2.56840050e-01
-2.00739428e-01 1.51006162e-01 -1.34259239e-01 -1.29229948e-02
-7.53197134e-01 -8.91309261e-01 -1.90839618e-01 2.72521824e-01
-2.95030743e-01 -9.31243420e-01 -1.21370518e+00 5.23068309e-01
-1.16014409e+00 -7.94206411e-02 7.23860979e-01 9.08255041e-01
1.14245450e+00 4.51196164e-01 5.96678674e-01 -4.57619697e-01
6.42310560e-01 -2.44128212e-01 -3.00672382e-01 3.13360006e-01
-7.57033527e-01 -1.38030410e-01 6.25043988e-01 -3.85903388e-01
-1.10248673e+00 -3.74990553e-01 -1.20210461e-01 -5.01164436e-01
-1.45015717e-01 1.99474007e-01 -7.47979522e-01 -1.69721901e-01
7.75220752e-01 7.01174200e-01 -2.33939543e-01 -1.45916775e-01
6.30055010e-01 5.53825140e-01 1.29102612e+00 -1.34747162e-01
8.56368899e-01 4.31243122e-01 -1.69748440e-01 -3.71940225e-01
-4.01246011e-01 1.55408885e-02 -3.05419654e-01 -5.53768396e-01
8.03296089e-01 -8.42950344e-01 -4.47370797e-01 6.72559321e-01
-1.02137554e+00 -2.58712471e-01 -8.13276768e-01 1.48889571e-01
-7.01659620e-01 2.40438253e-01 -2.83883393e-01 -6.25302315e-01
-5.79169512e-01 -1.03417301e+00 7.04079986e-01 5.82669020e-01
2.61714160e-01 -8.13542306e-01 1.43741861e-01 4.05932754e-01
5.96488833e-01 7.56869853e-01 1.20182562e+00 -2.89336424e-02
-7.08899438e-01 -2.06008509e-01 -1.66091383e-01 1.01058114e+00
1.54362440e-01 -4.79157150e-01 -1.09479713e+00 -5.75799048e-01
4.88777488e-01 -2.12180629e-01 9.01897490e-01 5.85511684e-01
9.23286319e-01 -4.88642037e-01 2.25661203e-01 7.38283277e-01
1.65861821e+00 4.89077240e-01 1.61729729e+00 2.01804757e-01
4.45384443e-01 1.80991307e-01 2.58104503e-01 -8.18818435e-02
1.45695880e-01 4.33305711e-01 7.17035770e-01 -3.75589132e-01
-7.39582241e-01 -5.67581594e-01 4.66817051e-01 7.67582893e-01
-2.34584317e-01 -4.70777482e-01 -4.14783835e-01 5.62204123e-01
-1.40104282e+00 -1.10841715e+00 1.77500099e-01 2.34810662e+00
7.66626239e-01 1.23665325e-01 -2.36310184e-01 4.83890802e-01
7.67604887e-01 2.43167192e-01 -1.01591933e+00 -4.27581936e-01
-3.32370400e-01 1.62950650e-01 1.24785408e-01 4.65141416e-01
-6.89332485e-01 4.43179965e-01 6.05947113e+00 1.02752709e+00
-9.93730903e-01 7.57300332e-02 8.29195321e-01 4.70071472e-02
-4.22581255e-01 -1.43446997e-01 2.78207034e-01 6.05165541e-01
8.09989691e-01 -4.31935728e-01 1.01125503e+00 7.04223633e-01
1.57611698e-01 -1.76261261e-01 -1.08773887e+00 8.74774635e-01
2.94898152e-01 -1.00911415e+00 3.27072918e-01 3.23529467e-02
9.94103074e-01 -2.87613899e-01 4.71802086e-01 2.05251828e-01
1.32918566e-01 -1.35508311e+00 8.67212057e-01 1.01964426e+00
9.63993311e-01 -8.55188191e-01 8.24241638e-01 3.54669601e-01
-9.99381185e-01 1.77980870e-01 -6.36649787e-01 1.45479739e-01
-3.59975696e-02 6.47403717e-01 -2.56933421e-01 7.36915827e-01
8.75090003e-01 4.97717619e-01 -5.08058012e-01 1.08642590e+00
-5.85350931e-01 3.82096231e-01 4.51594323e-01 8.18806350e-01
-2.08070919e-01 -3.68310511e-01 7.58146107e-01 5.84504485e-01
2.40551248e-01 3.06579679e-01 -1.19886912e-01 1.21689546e+00
-3.67575228e-01 -1.49367183e-01 -3.58494461e-01 3.63974601e-01
4.36052799e-01 9.66189086e-01 -3.53780985e-01 -2.48979792e-01
-3.24333925e-03 1.34511042e+00 -9.89608616e-02 5.39967835e-01
-6.92564428e-01 -4.62858111e-01 4.66921151e-01 4.93111089e-02
3.03810179e-01 3.68391842e-01 -2.45218769e-01 -6.95241868e-01
2.52316117e-01 -1.01786304e+00 -5.15937991e-03 -1.34064543e+00
-1.40561175e+00 1.08403599e+00 -4.49079990e-01 -1.47032201e+00
7.44345877e-03 1.36808842e-01 -8.39347959e-01 1.27731574e+00
-1.65500188e+00 -9.93514955e-01 -7.56488621e-01 8.03135216e-01
4.54485059e-01 -6.14865944e-02 7.95941114e-01 1.22732483e-01
-2.21927077e-01 6.38744652e-01 1.69522077e-01 9.32982489e-02
4.15746987e-01 -1.27278626e+00 3.73024642e-01 1.03033066e+00
-8.35841820e-02 1.68106809e-01 6.77462459e-01 -5.20638227e-01
-9.90952969e-01 -1.21673608e+00 3.58043402e-01 -1.09709680e-01
1.23417661e-01 1.33737922e-01 -1.26893663e+00 1.49586707e-01
6.69822991e-01 2.17760801e-01 3.95256579e-01 -7.79149115e-01
-5.32526731e-01 -2.97132641e-01 -1.69760466e+00 2.99164653e-01
1.02311790e+00 -7.04325855e-01 -7.46973574e-01 -1.26190335e-01
6.97847724e-01 -3.39620739e-01 -9.13085818e-01 4.35671479e-01
4.57076639e-01 -1.43124747e+00 1.07199180e+00 2.37870216e-02
7.34606087e-01 -5.58308303e-01 -3.21781904e-01 -1.90982938e+00
-6.13103926e-01 -2.46829078e-01 1.07139759e-01 1.23487890e+00
-2.45974474e-02 -4.74442065e-01 1.76005021e-01 2.85439432e-01
-3.09549809e-01 -5.85198283e-01 -1.10687959e+00 -1.00756919e+00
1.48955941e-01 -1.41457375e-03 9.22044277e-01 6.34737074e-01
-4.86710578e-01 7.23038763e-02 -4.17658389e-01 3.62953484e-01
9.13198292e-01 3.00088730e-02 3.64178538e-01 -8.32044721e-01
-2.36377299e-01 -3.40566546e-01 -6.78325593e-01 -7.48004615e-01
-1.79214135e-01 -8.69672060e-01 2.58881748e-01 -1.70839572e+00
3.09368640e-01 -1.03813060e-01 -7.52123177e-01 2.24668279e-01
-2.53223419e-01 5.82950413e-01 1.84415609e-01 5.51930726e-01
-3.03746104e-01 9.46562052e-01 1.69092214e+00 -4.99473512e-01
-2.06935748e-01 -2.53633708e-01 -7.83129871e-01 5.24974287e-01
5.84481657e-01 -4.89946514e-01 -5.45929849e-01 -3.98601592e-01
-1.22573227e-01 2.46297210e-01 7.07243323e-01 -1.18918824e+00
6.11280352e-02 -7.53496680e-03 6.78965032e-01 -3.02406251e-01
4.08463590e-02 -6.87495887e-01 5.12782216e-01 4.94845629e-01
-4.16657001e-01 -1.02636442e-01 -5.34016341e-02 8.18338990e-01
-4.52720970e-01 2.14758798e-01 1.28924549e+00 5.70325367e-02
-5.99292040e-01 1.98075905e-01 1.86153129e-02 9.06291325e-03
9.81459558e-01 -4.27803338e-01 -5.81951201e-01 -4.62021649e-01
-6.78130627e-01 -1.51317239e-01 6.31766856e-01 3.47031653e-01
1.18440092e+00 -1.67271662e+00 -9.63004947e-01 4.79541719e-01
-2.90177185e-02 -1.81284890e-01 7.24434912e-01 4.95041817e-01
-3.81844848e-01 -1.59463808e-01 -6.33487642e-01 -4.15475935e-01
-9.77575421e-01 7.85513401e-01 5.76356292e-01 -1.09239705e-01
-7.51908660e-01 7.04670668e-01 5.18533885e-01 3.84685248e-02
2.14520797e-01 -7.94580206e-02 -2.47905657e-01 -4.24021840e-01
8.55288506e-01 4.91206348e-01 1.54432029e-01 -7.44844496e-01
-1.93632543e-01 4.24046785e-01 2.48396873e-01 -6.85566738e-02
1.25284791e+00 -3.00584853e-01 -1.96354076e-01 1.51049405e-01
1.02415967e+00 -2.15658218e-01 -1.49212801e+00 -1.50930926e-01
-7.25260913e-01 -7.79312313e-01 4.42407459e-01 -1.43892252e+00
-1.49095881e+00 7.59610474e-01 1.33100212e+00 2.74589211e-01
1.87336004e+00 -2.24888563e-01 7.64688253e-01 -4.45245832e-01
3.29153031e-01 -6.82325482e-01 2.69370288e-01 -2.34458134e-01
1.64176881e+00 -9.50295866e-01 -1.86555922e-01 -1.94943380e-02
-6.67008936e-01 5.80335736e-01 3.58545840e-01 -3.82380068e-01
5.42174578e-01 -1.15389310e-01 1.51138529e-01 5.02824597e-03
-4.77597058e-01 2.01376408e-01 6.19177401e-01 1.10191762e+00
-2.32800990e-01 1.31352872e-01 -2.08714455e-02 5.27311444e-01
-3.95278722e-01 -2.61070561e-02 5.20915568e-01 4.78696346e-01
-3.05910200e-01 -5.85123599e-01 -4.69897896e-01 1.43184960e-01
-6.41823858e-02 -2.72367448e-01 -1.86499193e-01 4.39285398e-01
4.48871613e-01 1.21674955e+00 5.63728623e-02 -7.18922734e-01
6.67030990e-01 -2.75190026e-01 3.84379745e-01 -1.36207953e-01
-4.80497211e-01 -2.24355212e-03 -5.62199295e-01 -7.77966321e-01
-4.06274438e-01 -2.14565024e-01 -1.00630724e+00 -4.46403950e-01
-6.28110617e-02 -5.75102009e-02 6.04063392e-01 6.52374268e-01
3.43094319e-01 9.65067208e-01 1.04749715e+00 -8.24749172e-01
-4.73950326e-01 -9.37838793e-01 -8.00494671e-01 5.42753339e-01
4.63764191e-01 -4.40936714e-01 -6.66376770e-01 2.94481963e-01] | [11.617920875549316, -1.7699822187423706] |
4d513e8b-c3b4-40e8-9e76-f8397107eaf2 | mine-your-own-anatomy-revisiting-medical | 2209.13476 | null | https://arxiv.org/abs/2209.13476v5 | https://arxiv.org/pdf/2209.13476v5.pdf | Mine yOur owN Anatomy: Revisiting Medical Image Segmentation with Extremely Limited Labels | Recent studies on contrastive learning have achieved remarkable performance solely by leveraging few labels in the context of medical image segmentation. Existing methods mainly focus on instance discrimination and invariant mapping. However, they face three common pitfalls: (1) tailness: medical image data usually follows an implicit long-tail class distribution. Blindly leveraging all pixels in training hence can lead to the data imbalance issues, and cause deteriorated performance; (2) consistency: it remains unclear whether a segmentation model has learned meaningful and yet consistent anatomical features due to the intra-class variations between different anatomical features; and (3) diversity: the intra-slice correlations within the entire dataset have received significantly less attention. This motivates us to seek a principled approach for strategically making use of the dataset itself to discover similar yet distinct samples from different anatomical views. In this paper, we introduce a novel semi-supervised 2D medical image segmentation framework termed Mine yOur owN Anatomy (MONA), and make three contributions. First, prior work argues that every pixel equally matters to the model training; we observe empirically that this alone is unlikely to define meaningful anatomical features, mainly due to lacking the supervision signal. We show two simple solutions towards learning invariances - through the use of stronger data augmentations and nearest neighbors. Second, we construct a set of objectives that encourage the model to be capable of decomposing medical images into a collection of anatomical features in an unsupervised manner. Lastly, we both empirically and theoretically, demonstrate the efficacy of our MONA on three benchmark datasets with different labeled settings, achieving new state-of-the-art under different labeled semi-supervised settings | ['James S. Duncan', 'Lawrence Staib', 'David A. Clifton', 'Xiaoxiao Li', 'Xiaoran Zhang', 'Haoran Su', 'Yifei Min', 'Fenglin Liu', 'Weicheng Dai', 'Chenyu You'] | 2022-09-27 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 6.47160530e-01 4.51484323e-01 -5.50923765e-01 -6.11080408e-01
-9.57180619e-01 -5.69130003e-01 4.05890763e-01 5.21715656e-02
-4.74485099e-01 6.23657107e-01 1.31170049e-01 -3.08150500e-01
-4.97125953e-01 -3.20089072e-01 -5.64117968e-01 -7.78405190e-01
-2.33376786e-01 4.33964640e-01 2.01473117e-01 -1.61741264e-02
7.63130561e-02 4.77004498e-01 -1.23034453e+00 1.63729459e-01
8.19509864e-01 9.20501053e-01 1.73509296e-04 3.53475809e-01
1.17943384e-01 7.61995316e-01 -2.06672281e-01 -1.59027323e-01
6.64446712e-01 -6.77394450e-01 -1.09701753e+00 3.00669074e-01
6.42100453e-01 -1.56779736e-01 -1.81594733e-02 1.18137300e+00
3.99750501e-01 -2.11960211e-01 7.95966685e-01 -1.06283188e+00
-6.04100943e-01 5.27621806e-01 -8.21903825e-01 4.06169385e-01
-2.48767108e-01 1.60357818e-01 1.00260985e+00 -4.06829655e-01
5.93194604e-01 6.39862895e-01 8.64505649e-01 7.05453038e-01
-1.64711201e+00 -4.06740159e-01 7.47568309e-02 -2.33684689e-01
-1.21019375e+00 -3.56245697e-01 7.92219162e-01 -5.75510085e-01
3.17435682e-01 6.93559408e-01 4.66685265e-01 9.88656461e-01
1.37090385e-01 7.73805797e-01 1.77621067e+00 -4.10476625e-01
1.76488370e-01 1.84982985e-01 2.86465496e-01 9.02468562e-01
1.35572076e-01 1.11267373e-01 -2.77152717e-01 -9.33089778e-02
7.76740611e-01 -9.21673104e-02 -4.23762083e-01 -8.05196106e-01
-1.38176847e+00 7.64355838e-01 6.36129141e-01 5.05210638e-01
-1.85822234e-01 -1.92653716e-01 4.21990335e-01 3.40966046e-01
3.00238073e-01 7.59125471e-01 -4.76196647e-01 2.43604064e-01
-1.19910121e+00 -1.80980071e-01 4.83065575e-01 5.18636703e-01
7.44805157e-01 -2.89572686e-01 -1.13225117e-01 7.63380468e-01
5.92517667e-02 1.40015319e-01 8.76848102e-01 -8.66928339e-01
1.16540469e-01 5.26105464e-01 -3.89067322e-01 -7.73695409e-01
-6.82901740e-01 -7.12765872e-01 -1.00627637e+00 3.70637953e-01
7.61505604e-01 1.89419314e-02 -1.23611915e+00 2.02037811e+00
2.31919259e-01 -4.32810411e-02 -2.30473980e-01 9.24058914e-01
6.63156569e-01 -2.24887788e-01 5.42174578e-02 -2.17621282e-01
1.16688180e+00 -9.71258879e-01 -4.58756924e-01 -4.00504351e-01
7.18587101e-01 -4.70099181e-01 1.32783413e+00 2.92176932e-01
-8.99867892e-01 -4.01324302e-01 -1.10319650e+00 1.10051796e-01
-2.46867985e-01 3.87689583e-02 8.38875592e-01 8.78433347e-01
-9.94409859e-01 6.17872477e-01 -9.09353793e-01 -2.48497337e-01
8.11054587e-01 5.55051029e-01 -5.91486633e-01 -3.01172864e-02
-7.16197729e-01 8.30246747e-01 1.97243586e-01 -8.73563718e-03
-6.68821275e-01 -9.44782972e-01 -6.86104238e-01 -2.61055648e-01
4.80893970e-01 -7.39100158e-01 8.90624821e-01 -1.39565110e+00
-1.10231256e+00 1.46040702e+00 3.87269035e-02 -3.89807224e-01
7.53585398e-01 1.25271946e-01 -1.45371377e-01 2.37487882e-01
3.48951221e-01 7.54841626e-01 7.91863799e-01 -1.43543768e+00
-4.03048664e-01 -7.84621775e-01 -5.01312949e-02 1.50868371e-01
-3.45862448e-01 -2.61683017e-01 -3.02940726e-01 -9.50514793e-01
4.27400708e-01 -1.10303473e+00 -5.05811393e-01 1.80021167e-01
-6.91535354e-01 3.07971627e-01 3.69679034e-01 -4.28298265e-01
9.99892294e-01 -2.20489073e+00 4.27818522e-02 3.69518399e-01
5.78066826e-01 3.03771421e-02 4.60271277e-02 -2.80990422e-01
-3.98489267e-01 2.59928852e-01 -6.32200956e-01 -1.32339314e-01
-2.45322242e-01 2.71697909e-01 -1.46404371e-01 7.29210496e-01
1.25300676e-01 1.07328522e+00 -8.69040370e-01 -7.30675042e-01
9.28545967e-02 1.47812754e-01 -4.94336277e-01 -3.86369675e-02
3.72551948e-01 8.23368132e-01 -4.49029952e-01 6.38247192e-01
5.54434121e-01 -4.88429755e-01 1.95490450e-01 -3.87161106e-01
1.85723752e-01 3.77470851e-02 -9.50697541e-01 1.82417858e+00
-1.58433989e-01 3.03828210e-01 1.00580007e-01 -1.46310210e+00
5.00650346e-01 1.49625093e-01 9.24617171e-01 -7.80287921e-01
4.89175543e-02 4.64064419e-01 2.32012048e-01 -4.78276193e-01
6.25234609e-03 -4.97693896e-01 1.03927329e-01 4.91904408e-01
2.05135539e-01 9.92096122e-03 -8.85372907e-02 3.04206405e-02
1.12070751e+00 1.24621475e-02 2.84593701e-01 -5.94671428e-01
3.20766151e-01 2.64951400e-02 5.81674397e-01 9.99010086e-01
-6.36271775e-01 9.08615410e-01 4.70761597e-01 -3.80725890e-01
-8.04634631e-01 -1.25384152e+00 -6.44596696e-01 8.66805732e-01
9.39598307e-02 -3.13721634e-02 -7.69700229e-01 -1.24795628e+00
-1.04416996e-01 1.97415918e-01 -1.23438215e+00 -1.87676892e-01
-4.90875065e-01 -1.20309055e+00 6.36484027e-01 6.16412580e-01
2.46862411e-01 -7.90529013e-01 -8.88493240e-01 -1.27460539e-01
-5.60935698e-02 -9.47902918e-01 -3.51747036e-01 7.55835891e-01
-1.07996857e+00 -1.12754250e+00 -8.86996329e-01 -7.18848169e-01
1.02083433e+00 2.47066900e-01 1.08431184e+00 6.36667237e-02
-5.01375198e-01 3.70647341e-01 -1.67503625e-01 -1.43409595e-01
-3.26936871e-01 2.45719612e-01 -1.29602864e-01 6.04299940e-02
6.02645017e-02 -6.60799623e-01 -6.79859042e-01 3.85499150e-01
-9.26391482e-01 4.32033800e-02 9.06311691e-01 1.05469489e+00
8.27074230e-01 -1.79083273e-01 6.01941526e-01 -1.44404149e+00
2.94563681e-01 -3.20915192e-01 -1.44136578e-01 3.78965467e-01
-8.07776272e-01 1.50965378e-01 5.06423295e-01 -2.27497682e-01
-7.20203221e-01 2.49368623e-01 4.12793234e-02 -1.94474801e-01
-3.41724336e-01 4.26806808e-01 3.67605612e-02 -3.99585903e-01
9.15997207e-01 1.90743133e-01 3.96874636e-01 -2.72365808e-01
2.70585984e-01 4.01703119e-01 6.70202434e-01 -6.22943997e-01
7.96817720e-01 9.32003736e-01 2.23863602e-01 -5.70005417e-01
-1.17735279e+00 -5.47979116e-01 -1.04449487e+00 -7.80785605e-02
9.22930181e-01 -6.30115509e-01 -1.94571733e-01 1.62038967e-01
-2.97547609e-01 -4.85002041e-01 -7.01765001e-01 4.30413872e-01
-7.11420536e-01 5.76259017e-01 -4.31930453e-01 -3.35002601e-01
-2.21473902e-01 -1.39466023e+00 9.03406739e-01 6.97478130e-02
-3.47379982e-01 -1.09380472e+00 5.51093966e-02 4.87173080e-01
3.80493760e-01 3.96879315e-01 9.81427968e-01 -8.34811807e-01
-2.05071464e-01 -6.27395809e-02 -2.57121980e-01 5.08976400e-01
5.53646088e-01 -3.13198179e-01 -1.10267544e+00 -3.81812274e-01
2.08602145e-01 -5.04673839e-01 9.12897527e-01 5.77703595e-01
1.41581297e+00 -5.01865856e-02 -2.84485400e-01 9.75443900e-01
1.34318388e+00 -2.09533826e-01 3.37291360e-01 4.14203167e-01
7.62404203e-01 7.75412798e-01 4.42767352e-01 6.02970757e-02
2.18420655e-01 5.21392167e-01 3.97005022e-01 -6.50618196e-01
-3.34492624e-01 -5.88867767e-03 -6.10779710e-02 6.68684423e-01
-1.07620105e-01 3.75070989e-01 -8.90920818e-01 6.11529291e-01
-1.65916097e+00 -7.12074459e-01 -5.43632805e-02 2.33619595e+00
1.01996827e+00 1.86005652e-01 1.70507208e-01 1.31611347e-01
3.38528246e-01 4.01618816e-02 -6.36245966e-01 4.57728878e-02
-1.95484608e-01 3.89992923e-01 8.27821672e-01 2.12459981e-01
-1.42316866e+00 4.77782547e-01 6.41858768e+00 7.54585564e-01
-1.35609531e+00 1.72431260e-01 1.11127007e+00 -9.70229320e-03
-2.15990037e-01 -1.17330983e-01 -2.78968662e-01 4.04651165e-01
5.06093919e-01 7.26213753e-02 1.85828254e-01 6.42742991e-01
-3.69890966e-02 -2.82698423e-02 -1.27767885e+00 9.02693391e-01
1.82308316e-01 -1.19591475e+00 -5.06784655e-02 2.50842005e-01
8.13985288e-01 7.66136050e-02 3.66216451e-01 -3.12867239e-02
5.69430292e-02 -1.27457178e+00 6.25172496e-01 2.96894014e-01
8.19678128e-01 -4.29177493e-01 6.04821026e-01 2.84604698e-01
-6.97980285e-01 -4.95029939e-03 -3.48549597e-02 2.85517156e-01
-1.67744502e-01 5.10170758e-01 -6.71862960e-01 6.15215838e-01
7.36424148e-01 5.36892772e-01 -9.04499412e-01 1.02165079e+00
-8.86823088e-02 6.77000463e-01 -2.00356051e-01 5.82759142e-01
3.58588219e-01 -5.48746483e-03 2.87626624e-01 1.02906883e+00
-1.29844055e-01 -1.53784886e-01 2.78600872e-01 7.55474091e-01
3.29669341e-02 2.61887699e-01 -3.74513656e-01 3.22880358e-01
1.93794146e-02 1.34717011e+00 -1.17335165e+00 -1.25351667e-01
-4.41907078e-01 8.56290817e-01 2.94756413e-01 1.37621537e-01
-7.27581620e-01 3.94858643e-02 2.05018505e-01 2.46227786e-01
1.65824652e-01 1.33218437e-01 -7.85777211e-01 -1.17593193e+00
5.07904589e-02 -9.79478240e-01 7.23903239e-01 -2.30481669e-01
-1.55030763e+00 4.75611061e-01 -1.16519049e-01 -1.18579781e+00
2.24914365e-02 -6.39053524e-01 -2.78629959e-01 6.65449083e-01
-1.74145246e+00 -1.24005532e+00 -2.81488627e-01 6.53689981e-01
2.19106078e-01 -6.46370603e-03 8.48320603e-01 4.45899576e-01
-4.81843144e-01 8.69465709e-01 1.32199926e-02 2.55729109e-01
1.01477051e+00 -1.57811403e+00 -1.77443743e-01 7.02520072e-01
3.58009368e-01 7.01874673e-01 4.28858757e-01 -3.34349453e-01
-9.50142920e-01 -8.93316984e-01 3.87071252e-01 -6.28391922e-01
5.44452548e-01 -1.38491884e-01 -9.72283423e-01 8.09101760e-01
-2.29159147e-01 6.13471568e-01 9.49905872e-01 1.41022846e-01
-3.00562590e-01 -7.19868094e-02 -1.35640597e+00 5.52510977e-01
9.25240099e-01 -5.20596862e-01 -5.19472241e-01 3.72309238e-01
2.25004002e-01 -3.33130598e-01 -9.03515697e-01 7.11154282e-01
4.91793007e-01 -1.11497951e+00 9.75956619e-01 -7.34815896e-01
4.30307597e-01 -1.64133713e-01 -7.23905563e-02 -1.09194040e+00
-3.32947522e-01 -4.03542966e-01 1.96512222e-01 9.76984262e-01
4.82745200e-01 -6.61561608e-01 9.94011104e-01 6.94435894e-01
-1.59594491e-01 -1.01999617e+00 -1.08311427e+00 -7.38823950e-01
4.45236266e-01 -3.05842966e-01 2.55231529e-01 1.40418851e+00
-5.90789318e-02 2.96374947e-01 -2.43376195e-01 7.16449246e-02
7.20757067e-01 2.11422920e-01 5.55173576e-01 -1.18412161e+00
-4.74716246e-01 -6.01988673e-01 -5.22966146e-01 -8.93350959e-01
1.72611121e-02 -1.19638658e+00 1.13352537e-01 -1.06005263e+00
6.18237674e-01 -8.42908800e-01 -6.15416110e-01 7.08388627e-01
-2.98142016e-01 6.51568055e-01 -3.04665212e-02 4.91596907e-01
-4.92819011e-01 7.32797161e-02 1.47794211e+00 -1.42305911e-01
-3.12878452e-02 1.10492840e-01 -1.04157388e+00 8.46300483e-01
6.82123065e-01 -4.80454087e-01 -4.86582160e-01 -2.97054112e-01
-4.81656753e-02 -2.86063194e-01 6.71528697e-01 -9.13648427e-01
4.30356674e-02 4.54845615e-02 4.03917551e-01 -5.13356812e-02
-9.22997370e-02 -8.17403316e-01 -2.55031973e-01 5.27225554e-01
-5.40143192e-01 -1.56030640e-01 9.41310003e-02 4.60827440e-01
-1.51305825e-01 -2.78212547e-01 1.02615905e+00 -2.90316045e-01
-5.93124092e-01 3.10465783e-01 5.69983572e-02 4.83912379e-01
9.78996396e-01 -3.92961800e-01 -1.01177923e-01 -1.72719806e-02
-9.03192461e-01 -5.01483195e-02 5.99375427e-01 1.92979813e-01
1.94061875e-01 -9.93002594e-01 -5.38745880e-01 2.02832237e-01
1.77937940e-01 -7.40172267e-02 3.01992208e-01 1.37184048e+00
-3.00033510e-01 2.68951267e-01 -3.40280533e-01 -8.73728633e-01
-1.08627176e+00 5.79434633e-01 5.64127147e-01 -4.52796906e-01
-7.35695779e-01 8.12308073e-01 5.69846451e-01 -6.66278958e-01
4.50888872e-02 -3.47770244e-01 -4.11842763e-03 2.08093598e-02
2.06813917e-01 -4.52864543e-02 2.18123585e-01 -7.28108108e-01
-4.95726705e-01 6.45940065e-01 -1.94482207e-01 7.45493770e-02
1.34191477e+00 -1.37582552e-02 4.89335209e-02 3.77523988e-01
1.25906944e+00 1.18391432e-01 -1.21432078e+00 -2.65398830e-01
1.30441308e-01 -4.62714106e-01 1.93003953e-01 -1.03163731e+00
-1.27066302e+00 7.97178626e-01 9.14980590e-01 1.83493719e-01
1.12898707e+00 1.86246529e-01 5.15090466e-01 -3.17805959e-03
1.45967290e-01 -9.94463742e-01 -5.60097210e-02 -8.25680036e-05
4.09733772e-01 -1.52708435e+00 1.87371016e-01 -4.11662072e-01
-7.28246808e-01 9.48097587e-01 4.11460429e-01 -6.15472766e-03
6.10548019e-01 3.75858784e-01 4.05566365e-01 -4.69401866e-01
-2.17141464e-01 -3.36828053e-01 5.30578315e-01 6.65185809e-01
4.72226024e-01 5.71450070e-02 -2.08761036e-01 5.58208942e-01
-2.45366707e-01 -4.09370139e-02 2.56407231e-01 9.07079816e-01
-1.26915410e-01 -1.04301536e+00 -1.56390876e-01 6.37393415e-01
-7.46091306e-01 1.26717426e-02 -3.34602118e-01 9.88744795e-01
2.07176164e-01 4.23610419e-01 -1.41634479e-01 -7.93592781e-02
1.69718519e-01 -1.13209702e-01 6.43550634e-01 -7.21430898e-01
-5.38989604e-01 2.40368117e-02 -3.72971445e-01 -6.27010107e-01
-6.35688782e-01 -7.53739536e-01 -1.20599544e+00 2.82588094e-01
-2.49753386e-01 -1.15325265e-01 3.28880966e-01 1.03759909e+00
1.73457950e-01 6.44868135e-01 6.03137672e-01 -5.51021457e-01
-8.66183877e-01 -5.54033875e-01 -6.13557994e-01 7.03001738e-01
4.55572009e-01 -6.06589377e-01 -4.06931728e-01 8.83211344e-02] | [14.724120140075684, -2.1872053146362305] |
8248d91e-2e32-4c68-ae30-5befeb5ea203 | corefud-1-0-coreference-meets-universal | null | null | https://aclanthology.org/2022.lrec-1.520 | https://aclanthology.org/2022.lrec-1.520.pdf | CorefUD 1.0: Coreference Meets Universal Dependencies | Recent advances in standardization for annotated language resources have led to successful large scale efforts, such as the Universal Dependencies (UD) project for multilingual syntactically annotated data. By comparison, the important task of coreference resolution, which clusters multiple mentions of entities in a text, has yet to be standardized in terms of data formats or annotation guidelines. In this paper we present CorefUD, a multilingual collection of corpora and a standardized format for coreference resolution, compatible with morphosyntactic annotations in the UD framework and including facilities for related tasks such as named entity recognition, which forms a first step in the direction of convergence for coreference resolution across languages. | ['Daniel Zeman', 'Amir Zeldes', 'Zdeněk Žabokrtský', 'Martin Popel', 'Michal Novák', 'Anna Nedoluzhko'] | null | null | null | null | lrec-2022-6 | ['coreference-resolution'] | ['natural-language-processing'] | [-2.33835608e-01 4.22120601e-01 -5.29643536e-01 -6.13449752e-01
-1.06809533e+00 -9.99629736e-01 5.85548401e-01 6.22730553e-01
-6.62731588e-01 1.25927281e+00 8.82207215e-01 -2.31359288e-01
-2.17099383e-01 -2.88756967e-01 -2.45859325e-01 -1.43674016e-01
-1.37604978e-02 1.12708044e+00 4.45999324e-01 -5.75047851e-01
1.59349367e-01 5.97220659e-01 -1.19004476e+00 4.46288913e-01
7.93138921e-01 1.92726493e-01 1.77802786e-01 1.82816789e-01
-5.80760360e-01 7.87212133e-01 -4.51125652e-01 -8.09517205e-01
-3.65363002e-01 -4.09335285e-01 -1.84683633e+00 -7.59360313e-01
4.23278689e-01 5.47532797e-01 -6.11565895e-02 1.22485602e+00
4.84652489e-01 1.81159750e-01 1.62074462e-01 -1.02065623e+00
-3.76921982e-01 1.34177792e+00 -1.48979560e-01 2.94820845e-01
7.60640800e-01 -6.55182362e-01 1.55563235e+00 -5.23636460e-01
1.71148813e+00 1.38514948e+00 6.57083213e-01 7.30201125e-01
-1.12115550e+00 -4.87783015e-01 -1.94121197e-01 2.78880149e-01
-1.36488807e+00 -6.83120370e-01 2.34532565e-01 -3.20438921e-01
1.54718518e+00 3.02256793e-01 -9.31391120e-02 8.25326979e-01
-1.62357718e-01 4.75070804e-01 7.11618841e-01 -8.60810995e-01
-2.76254058e-01 1.33447781e-01 5.46441555e-01 2.25108221e-01
5.69851935e-01 -1.29702643e-01 -4.71730381e-01 -1.59108058e-01
4.31759983e-01 -8.44871521e-01 -4.57076132e-01 -4.94816422e-01
-1.01011574e+00 7.29507983e-01 2.99834579e-01 1.14278007e+00
9.00786966e-02 -5.31032622e-01 9.55875874e-01 3.79520208e-01
6.02521673e-02 8.27604771e-01 -7.46247768e-01 -1.32145122e-01
-4.44267303e-01 1.68620721e-01 1.35065353e+00 1.22956479e+00
5.92162371e-01 -5.79011142e-01 1.82204545e-01 1.04410338e+00
3.56391162e-01 2.28771925e-01 3.12303543e-01 -1.18054068e+00
8.23942661e-01 8.41110349e-01 4.59341377e-01 -5.62430382e-01
-8.43138814e-01 1.89940244e-01 -3.01016979e-02 -1.54266745e-01
5.96231222e-01 -2.28704467e-01 -1.77677926e-02 1.92344248e+00
5.73178649e-01 -3.89768124e-01 7.45729029e-01 8.08517694e-01
1.02849710e+00 1.97801024e-01 4.71226275e-01 -4.89901394e-01
1.59496975e+00 -5.88455677e-01 -1.12476408e+00 1.10378107e-02
1.15418780e+00 -1.07235456e+00 3.92917901e-01 -1.73097745e-01
-1.05515897e+00 -1.71283841e-01 -7.52887666e-01 -4.79776353e-01
-5.99561036e-01 -5.43280959e-01 6.00945473e-01 4.94070202e-01
-9.04357970e-01 2.99532622e-01 -7.79607296e-01 -9.03871357e-01
-2.33257920e-01 4.15418893e-01 -1.06284809e+00 4.81549762e-02
-1.62424231e+00 1.47253191e+00 1.04524148e+00 -1.75672203e-01
4.67761513e-03 -6.20739102e-01 -1.20015311e+00 -1.46908209e-01
3.95222008e-01 -1.96964458e-01 1.40305448e+00 -7.46390462e-01
-9.40049708e-01 1.51699638e+00 -1.55177891e-01 -4.22751009e-01
-8.01988691e-02 -3.58248025e-01 -1.08955085e+00 -1.51182190e-01
4.87937182e-01 7.09514141e-01 -3.93072754e-01 -1.09936953e+00
-1.19110429e+00 -5.65897822e-01 1.57225057e-01 3.51393193e-01
1.30403340e-01 1.12073290e+00 -3.82870585e-01 -3.54210466e-01
-2.24735111e-01 -8.83776605e-01 -1.01924159e-01 -7.56223917e-01
-1.33916095e-01 -6.15702331e-01 6.52744532e-01 -7.13895679e-01
1.40305567e+00 -1.85792005e+00 3.16623032e-01 -2.02105507e-01
-2.20201351e-02 3.03696305e-01 5.16675822e-02 7.22596467e-01
-3.49777251e-01 2.45174214e-01 -2.30807625e-02 -1.25074595e-01
3.54551017e-01 7.15607107e-01 -2.16801599e-01 3.41595590e-01
1.09130725e-01 5.29553890e-01 -1.16187024e+00 -8.95699561e-01
-2.89977528e-02 3.65069479e-01 -3.90310794e-01 2.81114336e-02
-4.51439582e-02 4.20166969e-01 -4.57633734e-01 1.97884947e-01
3.33470047e-01 2.34448358e-01 1.03847039e+00 -2.45405406e-01
-8.93703103e-01 9.14546132e-01 -1.30197906e+00 2.11670494e+00
-3.25467527e-01 5.14566958e-01 3.41352731e-01 -4.98601615e-01
8.38841498e-01 8.93355846e-01 3.65809530e-01 -5.04697680e-01
2.85301562e-02 6.39828920e-01 -8.06904398e-03 -4.67005819e-01
1.09229195e+00 -3.10142301e-02 -7.47482240e-01 3.25179577e-01
6.88401878e-01 3.00304323e-01 6.45829380e-01 2.09902272e-01
8.02395284e-01 3.75026166e-01 8.53723466e-01 -6.42335355e-01
9.49930131e-01 7.15554595e-01 9.98662412e-01 1.64747074e-01
-2.66759247e-01 3.05226855e-02 3.44075680e-01 -3.68832797e-01
-1.04333174e+00 -7.15493321e-01 -8.11162889e-01 1.16170514e+00
1.01752102e-01 -7.19404638e-01 -6.95671678e-01 -7.71181345e-01
-4.23565954e-01 7.23808348e-01 -2.52106428e-01 6.37572706e-01
-1.25485837e+00 -4.48568434e-01 9.63688970e-01 3.82173061e-01
5.34934141e-02 -1.43609691e+00 -4.40636635e-01 3.71704549e-01
-4.41856623e-01 -1.40105867e+00 -3.28394443e-01 4.76597816e-01
-3.09837312e-01 -1.60625541e+00 -3.11009157e-02 -1.30241382e+00
1.46805525e-01 -3.35232407e-01 1.46496260e+00 -1.00564629e-01
1.94742847e-02 2.35920310e-01 -5.77591896e-01 -9.76249725e-02
-6.07240558e-01 3.19428235e-01 8.02950934e-02 -7.12154746e-01
8.94657373e-01 -2.36289203e-01 1.49334401e-01 1.32483020e-01
-7.05964267e-01 -5.45129597e-01 3.09225004e-02 5.49913764e-01
4.88122165e-01 -5.85854530e-01 3.80022258e-01 -1.40119720e+00
5.67091286e-01 -4.78941172e-01 -5.28572142e-01 6.00406110e-01
-3.55141491e-01 3.54029119e-01 2.39742786e-01 2.94360101e-01
-1.58991885e+00 -2.31530502e-01 -4.89599496e-01 3.71352971e-01
-6.27672613e-01 5.85131526e-01 -5.28970480e-01 3.10577322e-02
5.95249653e-01 -5.85268378e-01 -8.34385872e-01 -7.58350909e-01
5.81059992e-01 7.48080492e-01 1.21444809e+00 -1.01780128e+00
2.46156707e-01 -6.04069680e-02 -2.59936422e-01 -6.49439573e-01
-8.69244277e-01 -8.75294447e-01 -1.41748035e+00 2.43758991e-01
1.18288839e+00 -8.82523537e-01 -3.01028699e-01 -1.86381102e-01
-1.51186156e+00 -6.07990697e-02 -2.92097926e-01 5.71227491e-01
-3.95022035e-01 4.32393730e-01 -7.60596871e-01 -2.60007262e-01
-4.67035979e-01 -7.10945427e-01 6.20635867e-01 2.70438135e-01
-9.61612046e-01 -1.35980415e+00 9.43853676e-01 2.74577469e-01
-1.14565782e-01 2.60286540e-01 9.49257612e-01 -1.26528835e+00
1.71671212e-01 1.24721102e-01 -4.18125503e-02 -2.38875046e-01
-7.32768178e-02 3.26034389e-02 -4.89295125e-01 -8.34185556e-02
-4.08354342e-01 -4.48267311e-01 9.99768972e-02 -2.28224993e-01
-4.23890680e-01 -1.31704494e-01 -7.01544285e-01 3.80514532e-01
1.55157602e+00 3.47448111e-01 4.42860007e-01 8.44684720e-01
5.02454460e-01 8.65840197e-01 7.28852212e-01 -1.35269342e-02
6.96714818e-01 9.38002765e-01 -2.01679885e-01 2.90929228e-01
-3.10733467e-01 -1.48809910e-01 8.83628950e-02 1.12811613e+00
-1.16829328e-01 5.86965755e-02 -1.30006516e+00 6.64283812e-01
-1.94351196e+00 -1.09762299e+00 -5.82770348e-01 1.95862579e+00
1.18963885e+00 -1.71332747e-01 -2.49335557e-01 -2.96646297e-01
1.04815567e+00 -2.04817411e-02 1.08225614e-01 -6.27047598e-01
-4.53598827e-01 4.26609904e-01 2.29837090e-01 9.94107962e-01
-1.08734775e+00 1.32882953e+00 6.55356073e+00 1.98581159e-01
-6.09543502e-01 3.49972069e-01 -4.29504871e-01 4.75246847e-01
-2.35577136e-01 4.18371797e-01 -1.37210107e+00 -3.58083961e-03
1.28192401e+00 -2.55497515e-01 5.64365946e-02 6.31760538e-01
-1.80017605e-01 2.12907314e-01 -1.05851936e+00 5.17149270e-01
5.11792675e-02 -1.34001017e+00 -1.52829945e-01 -3.49937618e-01
6.30394697e-01 6.36043310e-01 -8.54245245e-01 3.64751130e-01
1.00551689e+00 -5.15538216e-01 6.41367137e-01 4.03404497e-02
7.35898495e-01 -7.05597162e-01 9.66559947e-01 -3.02655958e-02
-1.49936807e+00 3.42340827e-01 -3.99345964e-01 2.95810550e-01
5.99742532e-01 -1.00892797e-01 -3.36224198e-01 1.04501164e+00
7.68020093e-01 4.19244051e-01 -1.67183757e-01 1.02982700e+00
-5.09099782e-01 5.51567152e-02 -1.75814535e-02 4.12954330e-01
2.28833571e-01 -6.25335798e-02 6.49772465e-01 1.76331532e+00
-1.73260197e-02 3.32786828e-01 3.28494608e-01 2.55768985e-01
-2.03942105e-01 6.29452348e-01 -3.43560755e-01 2.53161132e-01
1.04016685e+00 1.32724977e+00 -4.41122055e-01 -1.51812419e-01
-7.44493604e-01 6.83109283e-01 8.70914042e-01 -6.49912506e-02
-4.43033934e-01 -5.48488259e-01 7.12627709e-01 -2.33241767e-01
2.27850139e-01 -1.19364448e-01 1.92632347e-01 -1.28835106e+00
-4.00365025e-01 -9.76851404e-01 1.20241761e+00 -5.64879477e-01
-1.13962042e+00 9.06786263e-01 6.12266809e-02 -6.53246105e-01
-4.52825010e-01 -5.03983557e-01 -3.99634808e-01 7.82300293e-01
-1.42379951e+00 -1.18982565e+00 1.45796329e-01 9.07553971e-01
1.13484271e-01 1.06815778e-01 1.37163758e+00 5.45819342e-01
-6.93189323e-01 4.29322779e-01 -1.02659948e-01 5.61276376e-01
1.38685060e+00 -1.29739702e+00 1.85662270e-01 8.01164627e-01
3.21291447e-01 8.71743798e-01 9.02962625e-01 -6.82986319e-01
-9.18516517e-01 -7.88853765e-01 1.95900512e+00 -5.51033199e-01
1.23450577e+00 -9.85405035e-03 -1.01362038e+00 1.20700169e+00
7.93585300e-01 -1.37819052e-01 1.13618469e+00 6.47969544e-01
-6.82595849e-01 2.53635675e-01 -1.05601168e+00 1.88033894e-01
1.24336755e+00 -7.03125894e-01 -1.43292081e+00 1.49035692e-01
7.02929318e-01 -6.25001311e-01 -1.66664684e+00 3.56659472e-01
1.61692098e-01 -6.87433660e-01 5.33314347e-01 -1.01021028e+00
-2.10267350e-01 -3.02907705e-01 -4.27373886e-01 -1.05449259e+00
-4.54318494e-01 -8.11742902e-01 3.22349727e-01 2.05626130e+00
7.06523895e-01 -3.60734642e-01 2.46362135e-01 8.35232556e-01
-6.87724054e-01 4.90551531e-01 -1.13978493e+00 -3.44951361e-01
2.60335863e-01 -2.83740848e-01 5.18500805e-01 1.61776483e+00
1.02881098e+00 8.52991462e-01 1.00228012e-01 2.36288622e-01
4.79465395e-01 2.18814656e-01 4.47235107e-01 -1.62715971e+00
1.53191969e-01 -3.08609068e-01 -2.41016954e-01 -4.11507666e-01
7.67053485e-01 -1.07892609e+00 3.81594850e-03 -1.35182762e+00
1.21708862e-01 -6.27378583e-01 -2.18393400e-01 6.50472224e-01
-2.51003541e-02 -4.85890619e-02 1.24823958e-01 6.01988792e-01
-1.05584049e+00 -7.98642337e-02 4.96712476e-01 2.87077963e-01
-2.92651922e-01 -7.43426383e-01 -6.97553933e-01 8.03015828e-01
4.00190920e-01 -7.80315876e-01 2.29726955e-01 -5.25216877e-01
2.86324054e-01 3.05525124e-01 -5.61791897e-01 -7.16842532e-01
5.84706843e-01 -1.78607374e-01 -1.45858079e-01 -3.15948427e-01
-2.66291261e-01 -6.87145770e-01 3.91917616e-01 3.17976028e-01
-3.14562738e-01 3.50454032e-01 2.63887942e-01 -1.17771663e-01
-6.99606776e-01 -6.42150104e-01 5.98320961e-01 -1.85540721e-01
-1.45514977e+00 -1.21528119e-01 -1.51255995e-01 8.10540855e-01
1.01488054e+00 2.47844562e-01 -8.02679002e-01 5.36160767e-01
-1.01564717e+00 4.28736776e-01 5.86154819e-01 6.81303084e-01
-1.95654184e-01 -1.29447389e+00 -7.92975664e-01 -3.77149791e-01
3.35876971e-01 -3.21395606e-01 -2.88503859e-02 6.34648085e-01
-4.03595239e-01 1.08570325e+00 -6.37825966e-01 -1.32213920e-01
-1.41202557e+00 6.95016265e-01 3.30835462e-01 -6.98942721e-01
-5.62740862e-01 3.54902804e-01 -2.20400393e-01 -9.03378546e-01
8.25819522e-02 2.19024092e-01 -8.35160971e-01 4.10670042e-01
6.50389254e-01 1.19682908e-01 1.30886540e-01 -1.43984938e+00
-8.14891219e-01 3.49324375e-01 -1.65704817e-01 -6.80808872e-02
1.33865678e+00 -4.29993987e-01 -5.67982078e-01 4.55608666e-01
9.06236231e-01 4.40747738e-01 -4.47873384e-01 -3.60362947e-01
1.13101959e+00 9.38519910e-02 -3.86470258e-01 -6.92406476e-01
-6.61754131e-01 1.46262333e-01 1.32532477e-01 -8.24827643e-04
5.08494973e-01 5.19809544e-01 4.92976218e-01 4.97298956e-01
6.50378168e-01 -1.11046374e+00 -8.40428233e-01 9.75761354e-01
7.13488340e-01 -9.34249163e-01 -3.40786010e-01 -5.38851798e-01
-5.83361685e-01 1.09612441e+00 6.02238834e-01 2.31670514e-01
3.83270502e-01 7.39069462e-01 4.18167233e-01 -3.90150458e-01
-6.87461436e-01 -5.38568437e-01 7.69066587e-02 8.17922056e-01
1.05671024e+00 3.22676860e-02 -1.02575207e+00 6.82618201e-01
-9.81484950e-02 -2.27437615e-01 2.55524725e-01 8.37739527e-01
-5.56551099e-01 -1.81532407e+00 -4.08558637e-01 -3.50349456e-01
-9.91572380e-01 -2.93026686e-01 -4.96031463e-01 1.29594183e+00
3.89292270e-01 7.90991127e-01 2.67413914e-01 2.13297665e-01
6.80582464e-01 2.63242483e-01 5.53355992e-01 -8.84936988e-01
-8.92195344e-01 -1.52491778e-01 1.10505760e+00 -4.80738670e-01
-1.11199880e+00 -9.88393366e-01 -1.74649155e+00 -2.71468997e-01
-3.10862690e-01 1.02211213e+00 1.84571385e-01 1.01613545e+00
2.85242219e-04 -6.10961057e-02 -9.80395302e-02 -5.41433930e-01
7.58455023e-02 -8.22492898e-01 -4.93671715e-01 6.85623169e-01
-3.06714296e-01 -5.54498076e-01 9.26341340e-02 6.16782084e-02] | [9.397720336914062, 9.559342384338379] |
3cd2d009-2afd-4c20-9b4b-5d946644d92b | how-important-is-weight-symmetry-in | 1510.05067 | null | http://arxiv.org/abs/1510.05067v4 | http://arxiv.org/pdf/1510.05067v4.pdf | How Important is Weight Symmetry in Backpropagation? | Gradient backpropagation (BP) requires symmetric feedforward and feedback
connections -- the same weights must be used for forward and backward passes.
This "weight transport problem" (Grossberg 1987) is thought to be one of the
main reasons to doubt BP's biologically plausibility. Using 15 different
classification datasets, we systematically investigate to what extent BP really
depends on weight symmetry. In a study that turned out to be surprisingly
similar in spirit to Lillicrap et al.'s demonstration (Lillicrap et al. 2014)
but orthogonal in its results, our experiments indicate that: (1) the
magnitudes of feedback weights do not matter to performance (2) the signs of
feedback weights do matter -- the more concordant signs between feedforward and
their corresponding feedback connections, the better (3) with feedback weights
having random magnitudes and 100% concordant signs, we were able to achieve the
same or even better performance than SGD. (4) some
normalizations/stabilizations are indispensable for such asymmetric BP to work,
namely Batch Normalization (BN) (Ioffe and Szegedy 2015) and/or a "Batch
Manhattan" (BM) update rule. | ['Joel Z. Leibo', 'Qianli Liao', 'Tomaso Poggio'] | 2015-10-17 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [-1.97723478e-01 9.94271040e-02 6.54287413e-02 -3.52607340e-01
4.28656220e-01 -3.24735463e-01 7.84156144e-01 5.19722141e-02
-8.83527875e-01 1.01443851e+00 1.89006522e-01 -4.37349886e-01
-2.18256533e-01 -7.16716409e-01 -7.28925109e-01 -1.01439118e+00
-2.65395612e-01 1.81016415e-01 6.74429953e-01 -7.43790329e-01
5.93907297e-01 6.03197157e-01 -1.51733422e+00 4.11557443e-02
3.70614231e-01 8.94670486e-01 2.98684210e-01 7.65126646e-01
-1.30439222e-01 7.92276800e-01 -4.00993198e-01 -5.69977224e-01
5.12168527e-01 -5.94428658e-01 -1.08633947e+00 -5.14021516e-01
2.84030885e-01 9.94529873e-02 -3.17352355e-01 1.14199030e+00
4.67952102e-01 -4.17442853e-03 7.69292414e-01 -8.97980332e-01
-7.34564066e-01 6.53011024e-01 -3.70409369e-01 5.96614182e-01
-1.51745707e-01 2.71408141e-01 1.10351861e+00 -8.18015516e-01
4.99136388e-01 9.16263878e-01 9.15933073e-01 7.57981896e-01
-1.35250700e+00 -4.16961521e-01 3.21129024e-01 2.38163441e-01
-1.20470595e+00 -4.79200423e-01 3.61820400e-01 -3.93533826e-01
8.42531860e-01 3.65465105e-01 1.09608686e+00 9.51712370e-01
3.86539280e-01 3.95322829e-01 1.27864528e+00 -6.87945426e-01
2.29976356e-01 3.53199214e-01 1.90491021e-01 5.32574177e-01
4.32939947e-01 1.72361150e-01 -5.31557977e-01 1.16307244e-01
7.89695859e-01 -2.82255292e-01 -5.64697385e-01 -7.61537328e-02
-1.12680018e+00 7.42156684e-01 7.17677295e-01 7.54307032e-01
-2.76788205e-01 2.39994392e-01 3.35540324e-01 7.00960219e-01
2.25270361e-01 4.58953232e-01 -5.14381111e-01 1.64894268e-01
-9.54030991e-01 5.55535495e-01 8.53956640e-01 4.98416960e-01
6.50756657e-01 2.14957241e-02 -8.35538432e-02 1.01724160e+00
4.23772156e-01 2.38912806e-01 1.02761936e+00 -5.92729211e-01
4.16183412e-01 3.50989401e-01 8.07785317e-02 -1.14543200e+00
-8.78210366e-01 -6.10829473e-01 -8.47435355e-01 4.97717857e-01
9.75194395e-01 -1.88852444e-01 -7.59186149e-01 1.89006090e+00
-3.14310282e-01 -3.81308824e-01 -3.54469299e-01 1.19966352e+00
5.36727071e-01 3.96922797e-01 6.76383302e-02 -1.14022709e-01
1.12069535e+00 -6.25350177e-01 -3.73503864e-01 -4.90088284e-01
5.47047853e-01 -7.01108515e-01 1.10976815e+00 2.74422735e-01
-1.16115546e+00 -3.25420320e-01 -1.34234512e+00 6.58213347e-02
-5.53167880e-01 -1.26560256e-01 9.28213000e-01 6.85352325e-01
-1.27854240e+00 1.17270327e+00 -5.59461951e-01 -4.07460839e-01
2.78073043e-01 2.38678753e-01 -4.17348683e-01 3.19382399e-01
-1.33655190e+00 1.40494847e+00 2.27828726e-01 4.37416106e-01
-3.53035748e-01 -5.53579330e-01 -3.33163857e-01 3.86251658e-02
-3.08354437e-01 -7.37928748e-01 9.54859078e-01 -1.12183201e+00
-1.59428537e+00 1.08360875e+00 -1.14259526e-01 -6.47041321e-01
7.28205144e-01 1.16254821e-01 -1.36172608e-01 -6.89106435e-02
-3.18337321e-01 8.71646106e-01 6.02977395e-01 -1.05666912e+00
-3.95242333e-01 -5.93741119e-01 -1.77269936e-01 1.56742156e-01
-2.71501243e-01 8.59021321e-02 2.23456517e-01 -6.59736991e-01
5.70356727e-01 -7.45775342e-01 7.55837262e-02 2.48247519e-01
-4.09645326e-02 -2.48184633e-02 2.77452201e-01 -4.38725829e-01
1.06131887e+00 -1.93089986e+00 9.01043341e-02 2.45871544e-01
3.48617971e-01 2.01152802e-01 2.03328103e-01 2.78251588e-01
-4.58246678e-01 1.09371860e-02 -3.38244051e-01 -1.00006349e-01
2.62818057e-02 3.16016138e-01 -2.42329851e-01 8.50356281e-01
1.14515111e-01 7.21877754e-01 -6.61164820e-01 -1.60601959e-01
-4.21625786e-02 3.89194429e-01 -3.57163310e-01 -3.56902182e-01
1.98282972e-01 2.94133335e-01 2.48587698e-01 1.23853654e-01
5.96703112e-01 -4.91071083e-02 -2.48696610e-01 -5.27764000e-02
-4.20469701e-01 6.03503644e-01 -1.27662635e+00 9.80421960e-01
-3.37587185e-02 1.02994609e+00 -1.88278873e-02 -1.21124089e+00
1.06409371e+00 3.28680217e-01 1.57894000e-01 -9.18843865e-01
2.82062531e-01 3.71837050e-01 5.29671967e-01 -2.04605177e-01
2.29267135e-01 -5.55839479e-01 6.08117044e-01 3.79094183e-01
3.16331804e-01 -4.10807766e-02 4.95612472e-01 -9.72598940e-02
8.12961459e-01 -8.54292065e-02 1.52764115e-02 -7.84086883e-01
6.61149681e-01 -3.24509054e-01 5.66779435e-01 7.62418866e-01
-3.18754017e-01 9.09558237e-01 7.86979795e-01 -4.20474499e-01
-9.63771164e-01 -8.88763845e-01 -3.81505072e-01 1.11937189e+00
-1.27495721e-01 1.28477797e-01 -6.53181851e-01 -1.38068184e-01
-1.32147374e-03 6.60334527e-01 -1.01281154e+00 -2.54664272e-01
-3.52867573e-01 -1.05648386e+00 8.93719018e-01 4.76332158e-01
4.52520549e-01 -1.15400445e+00 -6.19775653e-01 1.81228831e-01
2.17982069e-01 -3.80488276e-01 6.40191063e-02 7.11108446e-01
-1.20817566e+00 -9.42845643e-01 -1.23149216e+00 -5.08212149e-01
7.90859103e-01 -1.02726035e-02 8.27605903e-01 1.49861142e-01
-6.93109483e-02 -2.21789747e-01 -1.17647044e-01 -6.82797015e-01
-1.11122161e-01 -1.02924414e-01 3.83507721e-02 -2.34164312e-01
2.09251702e-01 -8.23490322e-01 -7.28138268e-01 4.30115730e-01
-7.03622997e-01 -1.23022601e-01 4.61222261e-01 7.57372260e-01
1.84737012e-01 -3.06533605e-01 4.80619997e-01 -8.59965622e-01
8.79254103e-01 -7.76627883e-02 -5.38030207e-01 -3.30521874e-02
-8.86815906e-01 2.57788718e-01 6.26438379e-01 -2.83572614e-01
-7.41846740e-01 -5.16296566e-01 -5.57104468e-01 1.14437722e-01
1.26814644e-03 1.55032024e-01 4.31759268e-01 -4.83557016e-01
1.07055175e+00 2.98166394e-01 -1.40147239e-01 -3.28690797e-01
1.82658628e-01 9.19121057e-02 2.26815030e-01 -4.42207962e-01
4.92246002e-01 5.05807817e-01 -7.15651438e-02 -6.05396152e-01
-3.85792613e-01 -2.48346984e-01 -7.74143636e-01 -3.47302914e-01
7.67809033e-01 -3.80955428e-01 -6.75377429e-01 6.60888076e-01
-1.12419057e+00 -4.30769295e-01 -2.68289000e-01 7.82088518e-01
-4.53543305e-01 1.45396488e-02 -9.01207209e-01 -6.87208772e-01
-2.24067032e-01 -8.74351382e-01 -2.60710008e-02 3.14004600e-01
-3.32331002e-01 -1.23614836e+00 1.98119700e-01 -2.53475428e-01
8.50603342e-01 -2.45731413e-01 8.18385005e-01 -5.18170357e-01
2.56389886e-01 -6.57822937e-02 -4.23026264e-01 5.61662972e-01
-8.29825550e-02 1.12099648e-01 -1.22578132e+00 7.16390461e-02
2.73402482e-01 6.89871088e-02 1.13529766e+00 3.43963474e-01
7.26922035e-01 -3.13636780e-01 6.50332272e-02 4.26206946e-01
1.30625784e+00 -1.09897949e-01 7.93389976e-01 4.12508786e-01
2.95771480e-01 8.12562287e-01 -1.14865460e-01 1.02221347e-01
4.80353609e-02 2.65252173e-01 2.55580872e-01 -4.74259444e-02
-2.93597788e-01 -1.45865187e-01 2.49475554e-01 9.21948969e-01
-5.91714203e-01 2.56394118e-01 -6.67929649e-01 4.21538323e-01
-1.68745172e+00 -1.11989141e+00 -6.28881991e-01 2.27168798e+00
1.06566834e+00 6.25257134e-01 1.64361507e-01 4.38641340e-01
5.77437222e-01 -4.24592197e-03 -2.02554047e-01 -7.29582608e-01
-3.17565411e-01 3.41148704e-01 9.40615594e-01 5.09729028e-01
-5.46763361e-01 5.10862052e-01 6.55357265e+00 3.38670462e-01
-1.36707187e+00 3.84113565e-02 5.25066078e-01 -4.85575087e-02
-1.62387311e-01 -2.68595219e-02 -7.62999475e-01 6.28720820e-01
8.82074714e-01 1.61388256e-02 4.49643672e-01 4.36148822e-01
1.12756319e-01 -3.23591232e-01 -1.06757629e+00 6.44936502e-01
-1.80575788e-01 -1.27291286e+00 -2.58664131e-01 -1.25247717e-01
5.12802958e-01 4.10824597e-01 -1.31533220e-01 1.33487642e-01
2.19261110e-01 -8.97531450e-01 1.10608923e+00 6.46856666e-01
3.27025838e-02 -3.73751819e-01 8.70288014e-01 3.64585102e-01
-5.94364941e-01 -6.72092021e-04 -7.23538697e-01 -5.28739095e-01
-1.47897154e-01 8.61910999e-01 -4.14891630e-01 9.02867168e-02
7.96659648e-01 2.44450480e-01 -6.60075903e-01 1.33580911e+00
-6.22202575e-01 7.89610445e-01 -4.67934668e-01 -3.67245585e-01
2.77011365e-01 -2.35789269e-01 3.61851096e-01 1.57214892e+00
-6.29680008e-02 -2.14754209e-01 -8.69244635e-01 9.70952034e-01
6.34395257e-02 2.57040083e-01 -3.06779981e-01 1.54344395e-01
1.36537880e-01 1.05951941e+00 -1.05111945e+00 -5.69782332e-02
-4.05284375e-01 7.72378266e-01 3.58326465e-01 4.09683228e-01
-6.38300061e-01 -5.22099555e-01 5.17571509e-01 2.39559844e-01
2.31857479e-01 -1.89915940e-01 -8.57473612e-01 -8.85137022e-01
2.78029609e-02 -1.45136997e-01 1.50974065e-01 -7.39833593e-01
-1.31043077e+00 3.76343429e-01 -2.06659123e-01 -5.94166458e-01
1.72840282e-01 -9.39909399e-01 -8.63140464e-01 1.20962250e+00
-1.44591165e+00 -4.64979321e-01 -1.20318895e-02 5.35321057e-01
-2.33920012e-03 2.67444879e-01 8.22688520e-01 5.75966895e-01
-3.82737666e-01 5.17566323e-01 -3.07816621e-02 1.93152428e-01
6.55379474e-01 -1.30155027e+00 2.39187092e-01 5.83099604e-01
1.93896025e-01 8.67511690e-01 9.28178132e-01 -1.72917992e-01
-9.23325598e-01 -5.34929276e-01 1.23887146e+00 -8.04874748e-02
7.68384218e-01 -2.65830636e-01 -1.07751191e+00 5.68991840e-01
1.55400828e-01 6.89267740e-02 4.45828050e-01 1.35879651e-01
-3.75749111e-01 -1.55690297e-01 -1.08195722e+00 7.53783047e-01
9.57148314e-01 -7.12147802e-02 -8.87491524e-01 2.38685936e-01
2.76143759e-01 -4.70407084e-02 -6.02479279e-01 1.25764102e-01
7.67313242e-01 -1.41528475e+00 7.30125606e-01 -5.33936679e-01
4.34195489e-01 -3.00960898e-01 -5.01605961e-03 -1.38116300e+00
-6.26880407e-01 -5.43745421e-02 4.12232220e-01 9.59885776e-01
8.53058398e-01 -1.19524133e+00 4.50327694e-01 5.25226295e-01
-6.87953606e-02 -9.51971650e-01 -9.91509497e-01 -7.36483991e-01
4.00758862e-01 -4.77273881e-01 1.24854572e-01 9.34799850e-01
3.03762257e-01 2.85590172e-01 1.44145504e-01 -2.11006507e-01
1.67433396e-01 -4.61283565e-01 2.16925442e-01 -1.34758985e+00
-2.64903247e-01 -1.24238384e+00 -4.94785786e-01 -7.98992693e-01
-2.64108092e-01 -1.06819999e+00 -1.86913796e-02 -1.55590045e+00
-2.28443474e-01 -4.56865132e-01 -6.52508020e-01 5.42291701e-01
4.67516668e-02 3.31270546e-01 1.54231578e-01 2.81621277e-01
-4.81192432e-02 2.87987173e-01 1.10733867e+00 2.05647945e-01
-2.68619396e-02 1.54126108e-01 -8.03845346e-01 8.65555644e-01
9.05046880e-01 -5.48326850e-01 -2.63078511e-01 -3.36871892e-01
7.51540244e-01 -3.59584123e-01 3.86282712e-01 -1.02113843e+00
3.70090753e-01 2.52540014e-03 7.33470798e-01 -2.22363565e-02
1.03063352e-01 -5.40591836e-01 -1.94890469e-01 8.24653566e-01
-4.88295346e-01 1.35662794e-01 3.44114482e-01 1.97972029e-01
1.41436476e-02 -8.35396588e-01 9.43046689e-01 -3.50960642e-01
-3.41741085e-01 -7.25044236e-02 -6.41466796e-01 1.39883356e-02
5.33458292e-01 -3.39921772e-01 -3.62710685e-01 -1.82110280e-01
-8.08493912e-01 -8.89692083e-02 2.21135929e-01 9.57297236e-02
3.31131667e-01 -1.03274560e+00 -6.60073996e-01 2.54656672e-01
-4.58865166e-02 -3.83484781e-01 -2.58541405e-01 1.31895101e+00
-7.11330950e-01 5.01935422e-01 -3.76967102e-01 -3.08061093e-01
-8.45030427e-01 1.06635377e-01 5.62310934e-01 1.25854582e-01
-6.26782835e-01 1.41093218e+00 -8.09051320e-02 -4.08621788e-01
2.59984970e-01 -4.74268258e-01 -7.16062412e-02 2.27722421e-01
5.66331208e-01 3.63284796e-01 3.95073473e-01 -1.98300570e-01
-2.84708142e-01 2.38227516e-01 -1.58124730e-01 -1.92191735e-01
1.51105094e+00 8.63423571e-02 -2.86761105e-01 8.21108520e-01
8.78084958e-01 -1.23006053e-01 -1.13265538e+00 -7.25558549e-02
3.88935916e-02 -3.12544197e-01 1.49724567e-02 -8.57388854e-01
-1.02054262e+00 1.11682379e+00 5.29927194e-01 4.64125514e-01
7.12687314e-01 -2.43307278e-01 2.24703476e-01 2.99158067e-01
3.15417230e-01 -1.27021456e+00 -4.86485392e-01 6.04475379e-01
1.01044393e+00 -8.78110111e-01 4.74759825e-02 1.47333205e-01
-5.32285631e-01 1.39462996e+00 4.13511723e-01 -3.38521332e-01
6.39515102e-01 1.12834647e-01 -8.85756984e-02 -5.42057902e-02
-7.22821295e-01 -1.30554631e-01 1.88355923e-01 3.13526154e-01
7.82096207e-01 -6.36487380e-02 -9.04691041e-01 4.11788881e-01
-5.03347039e-01 -3.55402716e-02 3.22672755e-01 6.75756037e-01
-7.82793701e-01 -8.80088806e-01 -4.30595517e-01 5.41245520e-01
-2.51903206e-01 -1.48060158e-01 -4.05922055e-01 8.55156064e-01
1.05743073e-01 6.54332757e-01 8.73823166e-02 -2.13957712e-01
3.01432610e-01 3.56613934e-01 8.90773118e-01 -3.61117005e-01
-8.26357901e-01 -3.23344350e-01 -1.62253380e-01 -3.46203417e-01
-1.91768706e-01 -4.85280365e-01 -1.47934973e+00 -6.35366857e-01
-3.74606401e-01 1.10244103e-01 9.99623895e-01 1.14707839e+00
-1.33816198e-01 5.36073148e-01 1.93188399e-01 -7.52551734e-01
-6.36105895e-01 -1.24637151e+00 -7.78995216e-01 3.62882674e-01
1.04438318e-02 -7.73080051e-01 -7.32921958e-01 -4.60790806e-02] | [8.127931594848633, 3.4333786964416504] |
1dcf2052-7c38-46dc-9b44-8026ce212486 | review-guided-helpful-answer-identification | 2003.06209 | null | https://arxiv.org/abs/2003.06209v1 | https://arxiv.org/pdf/2003.06209v1.pdf | Review-guided Helpful Answer Identification in E-commerce | Product-specific community question answering platforms can greatly help address the concerns of potential customers. However, the user-provided answers on such platforms often vary a lot in their qualities. Helpfulness votes from the community can indicate the overall quality of the answer, but they are often missing. Accurately predicting the helpfulness of an answer to a given question and thus identifying helpful answers is becoming a demanding need. Since the helpfulness of an answer depends on multiple perspectives instead of only topical relevance investigated in typical QA tasks, common answer selection algorithms are insufficient for tackling this task. In this paper, we propose the Review-guided Answer Helpfulness Prediction (RAHP) model that not only considers the interactions between QA pairs but also investigates the opinion coherence between the answer and crowds' opinions reflected in the reviews, which is another important factor to identify helpful answers. Moreover, we tackle the task of determining opinion coherence as a language inference problem and explore the utilization of pre-training strategy to transfer the textual inference knowledge obtained from a specifically designed trained network. Extensive experiments conducted on real-world data across seven product categories show that our proposed model achieves superior performance on the prediction task. | ['Wenxuan Zhang', 'Jing Ma', 'Yang Deng', 'Wai Lam'] | 2020-03-13 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [ 7.68273231e-03 1.40917391e-01 -1.20721877e-01 -6.23106718e-01
-1.01556528e+00 -5.48826814e-01 5.70630133e-01 5.26778281e-01
-2.54841805e-01 5.26476383e-01 4.07488704e-01 -2.36607984e-01
-1.72922090e-01 -7.71949708e-01 -3.52351457e-01 -5.59646666e-01
6.63715303e-01 6.29659951e-01 3.46553534e-01 -5.84326446e-01
5.56993783e-01 -2.43684694e-01 -1.37498176e+00 7.44202912e-01
1.51389909e+00 1.26858306e+00 1.72234789e-01 4.36130047e-01
-6.58792496e-01 1.37391019e+00 -5.96469998e-01 -7.69531786e-01
-6.30447865e-02 -4.56941724e-01 -9.31754768e-01 -2.46909540e-03
3.47579002e-01 4.49332036e-02 5.33488989e-01 1.07490110e+00
2.16249347e-01 2.30624676e-02 7.27656066e-01 -9.63671803e-01
-7.71963656e-01 5.05848587e-01 -3.78003448e-01 2.33222604e-01
5.30259848e-01 -1.19769990e-01 1.70546961e+00 -9.90396082e-01
3.57786298e-01 1.13281679e+00 3.96691531e-01 2.87486076e-01
-6.98489189e-01 -2.44531259e-01 1.55278146e-01 5.25935650e-01
-8.63393128e-01 -2.37516269e-01 9.93055701e-01 -5.68978250e-01
4.38206315e-01 2.55345702e-01 2.73885667e-01 4.45519328e-01
3.39922309e-02 8.70434999e-01 9.96312499e-01 -1.83747947e-01
3.28852832e-01 5.55066586e-01 6.61253452e-01 3.59266430e-01
-1.07454352e-01 -1.04990232e+00 -5.49836636e-01 -2.99005717e-01
-2.61231452e-01 -2.02277731e-02 -4.47687626e-01 -1.53044695e-02
-6.75922871e-01 1.05603349e+00 6.08007073e-01 3.72082651e-01
-7.60217845e-01 -3.77007216e-01 4.02926832e-01 4.85886335e-01
7.92928696e-01 5.28916359e-01 -5.61189592e-01 2.60627903e-02
-3.84899408e-01 3.24946880e-01 1.14534664e+00 3.86590481e-01
9.37635839e-01 -5.98733783e-01 -3.46457154e-01 8.58175516e-01
4.27673072e-01 4.85828161e-01 3.26117814e-01 -9.54264700e-01
7.46547580e-01 1.34531772e+00 5.42367935e-01 -1.64765084e+00
-8.97383410e-03 -5.24458528e-01 -7.56596863e-01 -5.99750102e-01
3.43064398e-01 -2.56639689e-01 -1.28335347e-02 1.41577256e+00
6.14340901e-01 -4.81877178e-01 -5.93844615e-02 1.00738692e+00
1.27268255e+00 8.15974176e-01 -6.65529594e-02 -3.55662316e-01
1.40836692e+00 -1.12361741e+00 -9.50524688e-01 -2.79120594e-01
6.99352801e-01 -8.80063415e-01 1.11648893e+00 3.03840369e-01
-8.93315792e-01 -4.95095968e-01 -7.66100168e-01 -1.70734793e-01
-1.71283260e-01 3.84558916e-01 3.52747023e-01 1.47523373e-01
-8.02653074e-01 2.05614582e-01 2.19664961e-01 -1.35545850e-01
9.83377919e-02 2.23751977e-01 1.15386909e-02 -3.64130914e-01
-1.44501317e+00 9.99823689e-01 -1.93592265e-01 4.34553295e-01
-1.05472125e-01 -3.88966709e-01 -4.49932069e-01 9.49070007e-02
6.07815742e-01 -6.92667246e-01 1.35598004e+00 -1.25357306e+00
-1.28028047e+00 4.98288453e-01 -6.90188706e-01 -1.91515312e-01
6.15342595e-02 -2.34414518e-01 -3.99493873e-01 1.89532772e-01
3.98281068e-01 2.25429475e-01 7.59856403e-01 -1.15277052e+00
-7.86664248e-01 -4.89578247e-01 4.06391442e-01 3.89166951e-01
-4.27407414e-01 6.71312809e-02 -1.83329299e-01 9.30583030e-02
1.05834622e-02 -5.30330300e-01 -1.82455048e-01 -3.39641839e-01
-5.63953817e-02 -9.28979993e-01 4.69709516e-01 -8.43912601e-01
1.31248796e+00 -1.80210888e+00 2.23923624e-02 8.74262229e-02
3.20977241e-01 3.75711203e-01 -2.97051370e-01 5.81471562e-01
4.40947086e-01 -1.50534749e-01 -7.71050230e-02 5.62062599e-02
4.14334126e-02 4.17086519e-02 -4.86817896e-01 1.28344086e-03
2.06613094e-01 9.32509482e-01 -9.94825423e-01 -5.18404067e-01
-4.19467121e-01 1.69096202e-01 -3.45707923e-01 4.47047114e-01
-5.96420944e-01 4.94406760e-01 -8.20400834e-01 4.06984955e-01
4.70180750e-01 -6.12555504e-01 -2.80586276e-02 -3.47329825e-01
2.67743826e-01 5.73195636e-01 -8.17561150e-01 9.36813354e-01
-7.31645167e-01 3.11848074e-01 2.98956245e-01 -9.01549876e-01
1.29475582e+00 1.47180244e-01 1.84442967e-01 -1.00898480e+00
1.96819589e-01 5.17134726e-01 3.18029910e-01 -8.34025979e-01
6.59613907e-01 -1.07871078e-01 1.76276565e-01 6.74315155e-01
-4.29665893e-01 -1.65451467e-01 4.00246501e-01 3.80517364e-01
1.05326152e+00 -4.46005285e-01 2.34404638e-01 -2.51756340e-01
1.30948591e+00 1.76774431e-03 4.12021250e-01 4.69530135e-01
-2.31225029e-01 3.07632118e-01 7.45390654e-01 -3.40652257e-01
-4.46019977e-01 -3.55977356e-01 2.76203573e-01 1.16693830e+00
3.32126707e-01 -3.56417656e-01 -6.14144564e-01 -9.00078595e-01
-5.43764420e-02 7.17793286e-01 -6.53339565e-01 1.29054308e-01
-5.09839654e-01 -5.16067505e-01 -2.20658615e-01 1.97348818e-01
4.72450316e-01 -1.02909696e+00 -9.65572596e-02 1.70954421e-01
-9.35872436e-01 -9.98405933e-01 -4.86048758e-01 -9.74164456e-02
-6.41996801e-01 -1.25788784e+00 -6.49306476e-01 -8.12161028e-01
6.53358936e-01 6.15460694e-01 1.54821336e+00 2.71524280e-01
4.44907367e-01 2.55829364e-01 -5.93981922e-01 -2.08026618e-01
-3.18011522e-01 2.31781855e-01 -3.25128138e-01 5.73868692e-01
8.14724267e-01 -3.75208020e-01 -8.78846824e-01 5.58674872e-01
-7.24835634e-01 -3.10039282e-01 4.54099923e-01 6.08620644e-01
2.05178455e-01 -1.36939380e-02 1.31492448e+00 -1.00556731e+00
1.34478283e+00 -7.66170263e-01 -1.29017949e-01 5.83118260e-01
-5.77587545e-01 2.46809367e-02 6.26683712e-01 -1.35782495e-01
-1.37207031e+00 -4.72213686e-01 -3.16887170e-01 4.86041188e-01
2.53602743e-01 8.03869903e-01 -6.90487921e-02 2.07054943e-01
6.66083753e-01 -4.18528989e-02 1.44170132e-02 -3.21342558e-01
5.11061788e-01 9.67130899e-01 -7.43452087e-02 -3.53009939e-01
4.90931660e-01 2.83321857e-01 -3.44923437e-01 -5.99339545e-01
-1.51249635e+00 -1.06412733e+00 -3.64545286e-01 -2.70095199e-01
6.32215858e-01 -6.74826086e-01 -1.00002050e+00 -1.38290180e-02
-1.57453263e+00 3.14021438e-01 -1.98937636e-02 -9.04179439e-02
1.22570232e-01 4.83263403e-01 -5.19355953e-01 -9.71897304e-01
-7.90274978e-01 -1.03141034e+00 9.26390052e-01 2.35580713e-01
-4.82642055e-01 -1.12246037e+00 2.72926688e-01 1.21062541e+00
5.60868025e-01 -3.31464857e-01 1.22689581e+00 -9.30628479e-01
-5.65393567e-01 -4.04848188e-01 -4.73462999e-01 4.33827102e-01
2.46088341e-01 -2.96180874e-01 -7.81581640e-01 3.51311266e-01
5.66646457e-01 -4.72935617e-01 8.28171313e-01 1.69944346e-01
5.36481619e-01 -5.44824004e-01 1.61938772e-01 -7.62376904e-01
9.63844895e-01 -2.11414605e-01 3.39309812e-01 4.75970507e-02
5.62826157e-01 1.41588247e+00 9.56860602e-01 3.07493955e-01
8.36663783e-01 5.71685374e-01 4.85003084e-01 2.99151242e-01
2.13688761e-01 3.11205443e-02 4.45094228e-01 1.47406733e+00
2.66971797e-01 -2.53843337e-01 -5.80504298e-01 6.16334498e-01
-2.08473277e+00 -7.67833054e-01 -8.22325945e-01 1.86429071e+00
1.06361997e+00 -8.45704526e-02 -6.49685338e-02 2.88520884e-02
6.19407773e-01 2.38871321e-01 -4.53243107e-01 -4.81434673e-01
-3.54661830e-02 -1.28943816e-01 -2.76269466e-01 6.98892653e-01
-6.02969170e-01 4.88685638e-01 4.69014072e+00 9.37203288e-01
-5.53404510e-01 1.98773146e-01 7.35594392e-01 3.71001273e-01
-8.55316043e-01 -4.83987369e-02 -7.32848644e-01 3.19543272e-01
5.86385787e-01 -1.07019506e-01 3.26339602e-02 7.84000099e-01
2.94104040e-01 -5.78909695e-01 -1.05845261e+00 7.18983650e-01
4.36404765e-01 -9.38611686e-01 9.73067805e-02 -2.80493379e-01
8.08459044e-01 -3.27374905e-01 5.26010953e-02 3.63696903e-01
8.85792375e-02 -7.19610095e-01 2.30713770e-01 5.66620827e-01
-1.72349125e-01 -6.19795561e-01 1.30115080e+00 7.56380022e-01
-1.01755548e+00 -1.52182624e-01 -4.35214579e-01 -3.14238191e-01
3.50921541e-01 1.19185162e+00 -8.34252119e-01 3.83618027e-01
4.59299952e-01 7.45297134e-01 -7.69584596e-01 6.85566962e-01
-6.28377199e-01 5.93876064e-01 3.39226201e-02 -5.71254909e-01
1.77345410e-01 -3.79092455e-01 1.58140659e-01 6.01337254e-01
1.49776042e-01 1.45565853e-01 -8.34972933e-02 7.54837930e-01
-2.74599642e-01 7.34250307e-01 -4.18377340e-01 7.04448530e-03
1.36041746e-01 1.51715803e+00 -3.63260746e-01 -1.23519279e-01
-4.20138627e-01 6.88027382e-01 4.39722657e-01 2.20866680e-01
-3.51130545e-01 -1.15322277e-01 2.36535549e-01 1.21838644e-01
1.35161653e-01 3.22232008e-01 -4.03083920e-01 -1.01664174e+00
4.70219463e-01 -1.05651677e+00 2.37892449e-01 -7.98977494e-01
-1.66930687e+00 5.42235792e-01 -6.74846947e-01 -1.05641282e+00
-2.85795599e-01 -4.67342287e-01 -7.86844730e-01 1.04361558e+00
-1.75512886e+00 -9.25467491e-01 -3.39595318e-01 2.35802680e-01
6.63240790e-01 7.19122961e-02 4.81832892e-01 2.55502671e-01
-3.81557792e-01 1.19568050e-01 -2.08445162e-01 -2.44601354e-01
7.98220932e-01 -1.15181625e+00 -1.21249214e-01 4.65296894e-01
1.83653813e-02 7.43336141e-01 7.97707856e-01 -5.53846717e-01
-1.31552172e+00 -7.65042424e-01 1.65043092e+00 -6.82516873e-01
8.06780875e-01 2.06588969e-01 -1.20647514e+00 5.49809262e-03
6.47490799e-01 -5.79197228e-01 9.26487982e-01 5.31933963e-01
-2.50835866e-01 -3.83882105e-01 -7.86953092e-01 5.08505583e-01
3.01396161e-01 -5.49719155e-01 -6.67981803e-01 4.36151415e-01
7.88966060e-01 1.44031033e-01 -5.52855134e-01 3.21398348e-01
1.39637128e-01 -1.06702554e+00 5.27758718e-01 -3.31213087e-01
8.11555326e-01 -3.58276337e-01 -1.70342848e-01 -1.23290741e+00
2.75946241e-02 -6.87967688e-02 -1.07598402e-01 1.44128835e+00
6.42129302e-01 -4.85035717e-01 6.35216534e-01 8.42004716e-01
2.89263785e-01 -9.54462528e-01 -5.22430837e-01 2.71305833e-02
-1.54679969e-01 -1.37942806e-01 4.33944583e-01 6.78152561e-01
1.72669604e-01 1.05504298e+00 -1.88162729e-01 -8.26065391e-02
1.95179641e-01 4.71899658e-01 7.14795947e-01 -1.35346889e+00
-2.40364492e-01 -3.68856311e-01 2.02643007e-01 -1.25537086e+00
2.72716075e-01 -6.79152787e-01 1.99429780e-01 -1.88175952e+00
4.37573165e-01 -2.87592411e-01 -4.30021156e-03 -1.19348362e-01
-6.42415583e-01 -4.37839963e-02 -9.71606225e-02 3.53019595e-01
-1.21002829e+00 7.47611165e-01 1.37977087e+00 -2.76668787e-01
-4.47398573e-02 3.63517046e-01 -1.05732179e+00 8.21020007e-01
6.20633602e-01 -2.76670665e-01 -6.88753545e-01 -4.81098026e-01
1.31621325e+00 1.69853047e-01 1.50118724e-01 -3.45596373e-01
4.81037498e-01 -1.12462252e-01 -3.33724409e-01 -7.11081505e-01
1.57553732e-01 -7.84008861e-01 -4.26076382e-01 1.04136132e-01
-7.07190156e-01 7.41768703e-02 -5.47856510e-01 7.38739789e-01
-6.58996344e-01 -6.47870064e-01 1.63086325e-01 -1.69444352e-01
-4.09372538e-01 4.29158062e-02 -2.38084644e-01 3.66704434e-01
4.43646222e-01 2.74227381e-01 -4.80618268e-01 -8.76950145e-01
-3.08244884e-01 6.51995540e-01 -1.63750770e-03 4.10865635e-01
6.42021179e-01 -1.21025050e+00 -9.27083254e-01 -4.89121586e-01
3.64789248e-01 -9.63182822e-02 4.33313042e-01 1.09156203e+00
-7.96229541e-02 5.08061826e-01 5.54994941e-01 -4.54208106e-01
-1.22313643e+00 3.25797051e-01 2.23320380e-01 -6.73058927e-01
2.20414281e-01 8.41485441e-01 2.85655737e-01 -6.23661220e-01
6.59775212e-02 -7.18739405e-02 -1.00411618e+00 6.09420359e-01
7.19886899e-01 3.83455426e-01 1.95915565e-01 -6.61783218e-01
-1.92973524e-01 5.51925957e-01 -1.40415773e-01 9.04054567e-02
1.02359712e+00 -4.98771191e-01 -6.93816006e-01 5.61708212e-01
1.05476713e+00 -1.31567176e-02 -5.45896947e-01 -7.34580219e-01
2.48514012e-01 -1.86039686e-01 -2.52062589e-01 -8.00278008e-01
-9.13730741e-01 1.12556338e+00 -4.42765951e-02 5.59916437e-01
9.38581824e-01 1.27957761e-01 1.02480662e+00 6.17591918e-01
1.81997508e-01 -1.15905750e+00 5.45220733e-01 6.23794675e-01
1.11924970e+00 -1.61616218e+00 -2.13791683e-01 -5.95088601e-01
-9.15511429e-01 9.21895146e-01 6.84615195e-01 2.60961235e-01
5.95312238e-01 -5.08026898e-01 3.11593354e-01 -4.07195717e-01
-9.79221404e-01 -3.22223574e-01 6.12671971e-01 2.11244017e-01
6.36250556e-01 -9.18807015e-02 -7.30057418e-01 8.42369556e-01
4.05680053e-02 -2.04297051e-01 3.58161777e-01 4.62761074e-01
-8.26428652e-01 -1.01992881e+00 -1.14678912e-01 4.21842456e-01
-3.60906810e-01 -1.19235665e-01 -7.07878590e-01 -1.51026577e-01
1.24364212e-01 1.64553595e+00 -4.86724049e-01 -3.10163885e-01
1.75141558e-01 5.09270616e-02 1.13322018e-02 -6.25897467e-01
-9.54833031e-01 -3.45853627e-01 3.72667432e-01 -1.70022681e-01
-8.19183052e-01 -3.66876334e-01 -8.73466015e-01 -2.27376670e-01
-5.30895352e-01 6.28139436e-01 4.31253314e-01 1.42495358e+00
4.69252467e-01 2.72557437e-01 8.72598290e-01 -9.85985696e-02
-8.39500010e-01 -1.07747436e+00 -3.29193234e-01 5.77296436e-01
2.46891961e-01 -2.73478508e-01 -4.64500785e-01 -2.00210437e-01] | [11.545486450195312, 8.049939155578613] |
28bff533-4de5-4119-b77c-e620ad742522 | learning-with-noisy-labels-revisited-a-study-1 | 2110.12088 | null | https://arxiv.org/abs/2110.12088v2 | https://arxiv.org/pdf/2110.12088v2.pdf | Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations | Existing research on learning with noisy labels mainly focuses on synthetic label noise. Synthetic noise, though has clean structures which greatly enabled statistical analyses, often fails to model real-world noise patterns. The recent literature has observed several efforts to offer real-world noisy datasets, yet the existing efforts suffer from two caveats: (1) The lack of ground-truth verification makes it hard to theoretically study the property and treatment of real-world label noise; (2) These efforts are often of large scales, which may result in unfair comparisons of robust methods within reasonable and accessible computation power. To better understand real-world label noise, it is crucial to build controllable and moderate-sized real-world noisy datasets with both ground-truth and noisy labels. This work presents two new benchmark datasets CIFAR-10N, CIFAR-100N, equipping the training datasets of CIFAR-10, CIFAR-100 with human-annotated real-world noisy labels we collected from Amazon Mechanical Turk. We quantitatively and qualitatively show that real-world noisy labels follow an instance-dependent pattern rather than the classically assumed and adopted ones (e.g., class-dependent label noise). We then initiate an effort to benchmarking a subset of the existing solutions using CIFAR-10N and CIFAR-100N. We further proceed to study the memorization of correct and wrong predictions, which further illustrates the difference between human noise and class-dependent synthetic noise. We show indeed the real-world noise patterns impose new and outstanding challenges as compared to synthetic label noise. These observations require us to rethink the treatment of noisy labels, and we hope the availability of these two datasets would facilitate the development and evaluation of future learning with noisy label solutions. Datasets and leaderboards are available at http://noisylabels.com. | ['Yang Liu', 'Gang Niu', 'Tongliang Liu', 'Hao Cheng', 'Zhaowei Zhu', 'Jiaheng Wei'] | 2021-10-22 | learning-with-noisy-labels-revisited-a-study | https://openreview.net/forum?id=TBWA6PLJZQm | https://openreview.net/pdf?id=TBWA6PLJZQm | iclr-2022-4 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 1.94429994e-01 -3.18504840e-01 3.41834486e-01 -6.72043562e-01
-1.21910787e+00 -9.24845517e-01 6.59015179e-01 -1.89178467e-01
-7.66787291e-01 1.05637336e+00 -1.07161524e-02 -1.57183796e-01
-1.63710028e-01 -4.97845322e-01 -5.79826355e-01 -8.03404868e-01
4.80680577e-02 6.93161666e-01 -8.94398466e-02 -1.92427993e-01
-4.51401211e-02 1.34293139e-01 -1.73178792e+00 2.53367543e-01
4.78854448e-01 9.87058103e-01 -3.95689122e-02 6.59456313e-01
1.01046748e-01 1.08789384e+00 -1.05087042e+00 -5.92589676e-01
5.52201033e-01 -3.69807333e-01 -1.03389192e+00 6.13767467e-02
5.60526788e-01 1.00183614e-01 -1.03655152e-01 1.29250836e+00
8.33071232e-01 3.81054789e-01 4.00401115e-01 -1.55304575e+00
-6.98185503e-01 7.66876638e-01 -1.56820536e-01 1.54187679e-01
1.49370864e-01 5.08861482e-01 9.84728396e-01 -7.67478704e-01
7.96334803e-01 1.37134230e+00 9.84337032e-01 8.83272469e-01
-1.46124697e+00 -8.12787294e-01 5.09071648e-02 -2.13633794e-02
-1.39490390e+00 -2.21509382e-01 5.45181930e-01 -2.79316783e-01
4.42632526e-01 3.67885172e-01 2.42448494e-01 1.85928047e+00
-5.47550797e-01 6.60550117e-01 1.89483249e+00 -4.41745400e-01
4.86984164e-01 2.09608570e-01 5.89181900e-01 1.85242862e-01
2.87897736e-01 4.96289313e-01 -4.40526247e-01 -2.29809225e-01
5.13387769e-02 -3.03102851e-01 -3.10626060e-01 1.27371848e-01
-1.21498799e+00 7.23932922e-01 4.07234102e-01 3.27441752e-01
7.45308623e-02 3.95282000e-01 6.19971514e-01 6.44446075e-01
5.99640369e-01 7.53000081e-01 -7.24485636e-01 -3.75443190e-01
-8.54491174e-01 6.07490361e-01 9.67631459e-01 9.37129855e-01
8.38770390e-01 -2.06414983e-02 -1.94373831e-01 9.34855700e-01
8.26728046e-02 8.16575408e-01 6.56471670e-01 -1.32326365e+00
3.53457212e-01 7.63973221e-02 5.72842896e-01 -9.33445573e-01
-6.74478412e-01 -6.11496329e-01 -8.16575408e-01 1.46948680e-01
9.60857928e-01 -3.12341362e-01 -9.43349421e-01 1.98908877e+00
5.13018779e-02 3.34964782e-01 1.90456193e-02 1.04401290e+00
9.20575798e-01 9.90010202e-02 2.80698329e-01 5.14520518e-02
1.19506705e+00 -8.47145796e-01 -7.90834844e-01 -1.40077040e-01
1.08703148e+00 -8.06280732e-01 1.68618631e+00 5.18493295e-01
-5.45977831e-01 -4.76832569e-01 -5.43392003e-01 2.17381883e-02
-5.92296243e-01 -9.29556787e-02 5.84537804e-01 1.00384307e+00
-9.07396853e-01 5.24222612e-01 -4.50252831e-01 -1.85367033e-01
4.34621245e-01 9.47586149e-02 -3.33164722e-01 -4.88193214e-01
-1.41003120e+00 7.96961308e-01 1.21788569e-01 2.34939575e-01
-1.05633318e+00 -5.16464412e-01 -6.77182734e-01 -3.35304499e-01
6.85619175e-01 -2.92179912e-01 1.74175036e+00 -1.01648164e+00
-9.91897941e-01 1.05488467e+00 5.05025871e-02 -2.83172727e-01
9.05881166e-01 1.45089570e-02 -3.83686423e-01 -3.79678786e-01
2.67868936e-01 6.22199535e-01 3.16664577e-01 -1.82353246e+00
-5.04246414e-01 -2.44326964e-01 1.05983332e-01 -8.82964060e-02
9.27576870e-02 8.00747648e-02 1.46590933e-01 -7.87981927e-01
1.96613401e-01 -1.21672416e+00 -2.57737517e-01 -4.32346314e-01
-6.75326347e-01 -2.55458713e-01 5.71005940e-01 1.82306141e-01
8.73001099e-01 -2.00120759e+00 -6.09238505e-01 1.74370959e-01
3.05725515e-01 3.62421095e-01 -6.20218396e-01 2.30944932e-01
-1.95213065e-01 7.05055237e-01 -1.12501562e-01 -7.55770862e-01
2.01628342e-01 5.38824618e-01 -4.26188350e-01 3.92580271e-01
7.97315538e-02 1.04881454e+00 -1.28517056e+00 -2.00914502e-01
-1.61410451e-01 1.61648214e-01 -2.13706642e-01 -1.74535960e-02
-2.76582986e-01 5.21634817e-01 -2.43708089e-01 6.57490969e-01
7.69620657e-01 -9.09662992e-02 -1.59643874e-01 8.60169623e-03
2.96190441e-01 2.02795744e-01 -1.55868328e+00 1.27121043e+00
-2.69732744e-01 4.95648116e-01 -6.77706301e-02 -7.48495281e-01
7.47133791e-01 2.88835585e-01 1.14179805e-01 -8.26834559e-01
2.43247285e-01 4.16547388e-01 -1.05279244e-01 -6.32155299e-01
5.34341276e-01 -5.82691371e-01 -2.83828765e-01 7.28504241e-01
8.37100074e-02 -5.79578876e-01 4.19101626e-01 1.91430956e-01
1.33545005e+00 -2.65328258e-01 -3.12398672e-01 -2.11563215e-01
2.14936715e-02 -6.34720400e-02 6.21097803e-01 1.41439903e+00
-7.17246294e-01 9.80444074e-01 5.16907036e-01 -6.22471213e-01
-8.64206314e-01 -7.83090591e-01 -2.79756695e-01 1.24876142e+00
1.85586140e-02 -2.72924364e-01 -7.44389117e-01 -8.93173873e-01
-1.89326212e-01 6.51689172e-01 -8.13281000e-01 9.95586067e-02
-3.60967010e-01 -1.16498053e+00 1.13174796e+00 3.22703123e-01
4.00279462e-01 -1.31607759e+00 -1.95182636e-01 1.61428005e-01
-5.21892846e-01 -1.22948527e+00 -1.45584613e-01 7.18852222e-01
-3.67112309e-01 -1.32112110e+00 -1.74568161e-01 -6.03917658e-01
4.11904484e-01 3.26815218e-01 1.69671154e+00 4.02489871e-01
-1.07533529e-01 4.53578591e-01 -6.72259331e-01 -5.60038149e-01
-4.96006131e-01 -1.23872245e-02 2.42438003e-01 -2.25403413e-01
5.92892826e-01 -4.98935282e-01 -1.98252931e-01 7.23033428e-01
-1.16752195e+00 -4.55756247e-01 4.79163826e-02 1.23158014e+00
6.71637416e-01 2.47900844e-01 8.25056970e-01 -1.49959886e+00
8.40748429e-01 -6.01272404e-01 -5.48920393e-01 1.69090405e-01
-4.55096245e-01 -5.70974313e-02 7.30270445e-01 -6.40418112e-01
-8.21573913e-01 -1.12441918e-02 -1.46793783e-01 8.79960135e-03
-5.41841924e-01 2.94735700e-01 -9.18110684e-02 1.14081688e-01
1.30604088e+00 -2.37875700e-01 -6.39869213e-01 -4.67525572e-01
5.40611744e-01 7.82169282e-01 5.98785579e-01 -1.03285646e+00
7.36436665e-01 3.49982858e-01 -2.25681633e-01 -2.92118132e-01
-1.37972140e+00 -4.08309847e-01 -4.11749154e-01 3.59948874e-02
4.47998494e-01 -9.66242254e-01 -8.41788411e-01 7.58165419e-01
-9.50181842e-01 -7.12055802e-01 -7.48950422e-01 2.20058605e-01
-5.86386144e-01 1.06397539e-01 -8.58371139e-01 -8.39227021e-01
1.26425058e-01 -1.34163237e+00 1.03768444e+00 3.02002691e-02
-2.86584973e-01 -9.02876973e-01 1.01277031e-01 5.66580474e-01
6.04085028e-01 1.15191452e-01 6.06347740e-01 -9.23825026e-01
-4.63546664e-01 -2.84727395e-01 -3.52166057e-01 6.78590953e-01
-2.26538792e-01 -8.47813189e-02 -1.46199310e+00 -1.41612604e-01
9.72544551e-02 -1.33005321e+00 7.36723959e-01 2.26983298e-02
1.12518609e+00 5.32573164e-02 2.03568816e-01 2.44073853e-01
1.30937970e+00 -1.93852380e-01 3.91990960e-01 3.00319314e-01
5.47351539e-01 5.72657287e-01 5.54672718e-01 2.91448325e-01
3.35542500e-01 5.08749008e-01 4.27647501e-01 1.34489704e-02
-2.68059075e-01 -1.72020704e-01 -1.88032031e-01 8.20167959e-01
2.64573872e-01 -5.83514452e-01 -1.12302756e+00 3.70804518e-01
-1.64067805e+00 -7.98578262e-01 -3.88331383e-01 1.94241846e+00
1.13947642e+00 1.64959654e-01 -2.17416704e-01 4.67274070e-01
5.72847784e-01 -6.81980476e-02 -3.19349855e-01 -6.91522509e-02
-4.79148507e-01 1.57779634e-01 3.55407208e-01 5.95621645e-01
-1.42545831e+00 1.02853000e+00 6.42385483e+00 1.07178736e+00
-8.73450339e-01 5.09176016e-01 9.93648231e-01 -2.14313552e-01
-3.52697134e-01 -3.30187321e-01 -6.25895858e-01 5.78373551e-01
1.11963069e+00 4.79737855e-02 5.98588645e-01 7.88030982e-01
2.05797181e-01 -3.04626286e-01 -1.20908785e+00 1.09592056e+00
-1.94392234e-01 -7.47068524e-01 -3.82572621e-01 -2.21268520e-01
1.20434105e+00 4.50268567e-01 1.16351776e-01 7.08649039e-01
1.11016381e+00 -1.33096802e+00 9.25424993e-01 4.08574581e-01
7.63594270e-01 -4.92553920e-01 1.30188358e+00 7.02828825e-01
-5.24752259e-01 -1.45286858e-01 -5.71665943e-01 -3.53142768e-01
-6.81301951e-02 1.08376253e+00 -3.51938963e-01 1.40152782e-01
8.05653572e-01 2.81045020e-01 -8.81048203e-01 9.79303241e-01
-3.95281315e-01 1.07369483e+00 -4.83213663e-01 6.69641048e-02
3.29627156e-01 1.32006332e-01 8.41526315e-02 1.23034656e+00
1.67511310e-02 5.84205091e-02 4.28935170e-01 7.12863863e-01
-6.20856822e-01 -7.81868026e-02 -5.66318989e-01 1.48558229e-01
6.43347323e-01 1.25162220e+00 -9.12695467e-01 -3.12522203e-01
-2.75005579e-01 6.01215482e-01 4.27459836e-01 5.58405638e-01
-6.82279050e-01 1.90621447e-02 6.24837816e-01 -1.40347958e-01
-3.85921776e-01 -2.50468943e-02 -6.42581999e-01 -1.21909618e+00
-1.47841731e-02 -1.40771389e+00 5.69075570e-02 -7.29916632e-01
-2.02726555e+00 9.35687423e-01 -3.51657987e-01 -9.81161654e-01
-1.34511366e-01 -4.15208131e-01 -1.18057765e-01 8.43627751e-01
-1.40138912e+00 -9.40023601e-01 -5.62884688e-01 3.77243131e-01
4.50890779e-01 1.72909558e-01 1.13433611e+00 4.35826957e-01
-4.20561403e-01 8.41661990e-01 2.32812136e-01 3.07593524e-01
1.03621924e+00 -1.43775427e+00 5.19196868e-01 4.63821411e-01
4.78131115e-01 3.10438514e-01 8.26036990e-01 -5.33161521e-01
-5.57530403e-01 -1.31714571e+00 9.70438302e-01 -1.16285574e+00
7.80375421e-01 -7.06692994e-01 -9.42500770e-01 5.72405100e-01
-2.94106871e-01 7.48426855e-01 5.94307780e-01 1.37433991e-01
-6.86369479e-01 3.56176049e-02 -1.35756326e+00 4.42142844e-01
1.33356917e+00 -6.19701564e-01 -2.53835469e-01 8.02745998e-01
7.59800553e-01 -3.04899961e-01 -4.00774717e-01 3.52443665e-01
1.48051292e-01 -1.04169869e+00 6.01876676e-01 -8.13497424e-01
2.74483800e-01 -2.01329470e-01 -5.60288727e-01 -1.55294836e+00
2.00119074e-02 -2.56462961e-01 5.66744030e-01 1.42084229e+00
4.72780734e-01 -5.40159583e-01 8.52516174e-01 8.62781823e-01
3.66622135e-02 -6.33990228e-01 -1.02964532e+00 -9.18064833e-01
3.12857389e-01 -1.12266243e+00 5.64446032e-01 1.19006836e+00
-6.55444145e-01 1.04533676e-02 -3.99733007e-01 -1.29648939e-01
3.51057351e-01 -5.49061358e-01 7.28213429e-01 -1.19509006e+00
-1.03809461e-01 -8.08154866e-02 -4.03104186e-01 -5.93653440e-01
4.27091688e-01 -7.76217401e-01 4.54553515e-01 -9.87590194e-01
8.85761529e-02 -1.13815749e+00 -4.08473015e-01 7.17479169e-01
-3.06784570e-01 1.03349996e+00 2.52543449e-01 2.31637970e-01
-9.57768619e-01 4.47274685e-01 1.22731340e+00 -9.24904197e-02
2.54134476e-01 8.85424837e-02 -7.43186712e-01 8.27584386e-01
8.23434055e-01 -1.09389949e+00 -4.92635489e-01 -3.80184203e-01
5.04172206e-01 -5.23020685e-01 4.24152792e-01 -1.07065570e+00
3.07054035e-02 -7.09865615e-02 -5.01390314e-03 -2.72523835e-02
1.08497761e-01 -8.75241995e-01 -5.07273823e-02 -1.38341254e-02
-7.90153265e-01 6.59114495e-02 -2.48695582e-01 6.45365477e-01
-3.39929402e-01 -5.97150505e-01 7.63637722e-01 -4.33170348e-01
-4.63221073e-01 -5.72320521e-02 -3.46135259e-01 8.72658432e-01
7.55145192e-01 3.13028157e-01 -6.72878623e-01 -6.16705477e-01
-8.09524894e-01 1.99305341e-01 4.30542737e-01 3.06890786e-01
3.52412388e-02 -1.17086649e+00 -7.58865178e-01 9.84709114e-02
1.88959733e-01 -4.56924662e-02 2.22344920e-01 5.18750072e-01
-4.27249551e-01 1.66764855e-02 3.26756895e-01 -4.85820800e-01
-9.72912550e-01 5.08772969e-01 6.34690344e-01 -5.58147192e-01
-1.19860113e-01 1.23311937e+00 -2.54501343e-01 -1.08243704e+00
6.05700791e-01 -3.78498942e-01 2.57360458e-01 1.38481751e-01
4.73539740e-01 4.49270993e-01 4.04004335e-01 -7.10573912e-01
-7.48482123e-02 3.15882474e-01 2.33846977e-01 -5.74901327e-02
1.01570356e+00 -1.97934121e-01 1.29599858e-03 6.13913476e-01
1.07376361e+00 -8.01780522e-02 -9.78363097e-01 -5.85255384e-01
4.45562184e-01 -5.42082012e-01 -2.36683995e-01 -1.43472195e+00
-1.10082269e+00 7.77882397e-01 7.98390388e-01 3.95922750e-01
9.90940571e-01 -7.73031563e-02 6.07986689e-01 7.95706153e-01
8.25052261e-01 -1.10664082e+00 1.62548363e-01 7.54774332e-01
5.36040068e-01 -1.77286863e+00 -4.55121279e-01 -4.56577182e-01
-5.96215785e-01 4.65572000e-01 6.08649373e-01 4.62026373e-02
7.62333035e-01 5.21178782e-01 9.94283199e-01 -1.67632535e-01
-8.67044687e-01 -3.96544725e-01 -2.11076424e-01 7.99448013e-01
4.47898626e-01 3.22119504e-01 -2.98357189e-01 1.02907622e+00
-5.39659441e-01 4.38275598e-02 7.48736978e-01 6.56627953e-01
-1.35461926e-01 -1.32132125e+00 -4.95728761e-01 4.39871997e-01
-5.59712291e-01 -9.15255472e-02 -3.18225294e-01 7.58762658e-01
7.81759262e-01 1.51874506e+00 -3.70103031e-01 -4.74537462e-01
6.80586696e-01 4.25194293e-01 3.53392027e-02 -8.04659426e-01
-9.06982005e-01 -3.64913195e-01 1.55219793e-01 -5.40067255e-01
-5.34117520e-01 -2.94568002e-01 -9.80485916e-01 -4.18566138e-01
-4.49579090e-01 2.62431026e-01 7.28993952e-01 9.12782490e-01
5.68077043e-02 2.34656855e-01 4.04993206e-01 -8.56020212e-01
-9.92485106e-01 -1.25625134e+00 -9.31898057e-01 1.23291242e+00
1.55655876e-01 -7.22465932e-01 -8.73572052e-01 -5.96248880e-02] | [9.403968811035156, 3.920912265777588] |
479118c4-ab25-46f3-97cd-8c58ce3c9c26 | robust-invertible-image-steganography | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Xu_Robust_Invertible_Image_Steganography_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Xu_Robust_Invertible_Image_Steganography_CVPR_2022_paper.pdf | Robust Invertible Image Steganography | Image steganography aims to hide secret images into a container image, where the secret is hidden from human vision and can be restored when necessary. Previous image steganography methods are limited in hiding capacity and robustness, commonly vulnerable to distortion on container images such as Gaussian noise, Poisson noise, and lossy compression. This paper presents a novel flow-based framework for robust invertible image steganography, dubbed as RIIS. We introduce the conditional normalizing flow to model the distribution of the redundant high-frequency component with the condition of the container image. Moreover, a well-designed container enhancement module (CEM) also contributes to the robust reconstruction. To regulate the network parameters for different distortion levels, we propose a distortion-guided modulation (DGM) over flow-based blocks to make it a one-size-fits-all model. In terms of both clean and distorted image steganography, extensive experiments reveal that the proposed RIIS efficiently improves the robustness while maintaining imperceptibility and capacity. As far as we know, we are the first learning-based scheme to enhance the robustness of image steganography in the literature. The guarantee of steganography robustness significantly broadens the application of steganography in real-world applications. | ['Jian Zhang', 'Jingfen Xie', 'Yujie Hu', 'Chong Mou', 'Youmin Xu'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['image-steganography'] | ['computer-vision'] | [ 8.74398828e-01 -5.84227359e-03 1.74955712e-04 2.11443469e-01
-3.58025916e-02 -2.18667015e-01 3.03098083e-01 -6.61308885e-01
-1.40084505e-01 3.24125201e-01 2.11700693e-01 -4.70576257e-01
-4.88210954e-02 -9.70483303e-01 -6.17190659e-01 -1.34724927e+00
-2.45687306e-01 -5.97018719e-01 2.45661512e-01 -5.33518851e-01
5.08351028e-01 2.13937491e-01 -1.30849683e+00 1.17253847e-01
9.17127550e-01 8.77245724e-01 2.84454972e-01 6.47629082e-01
3.03917348e-01 8.76269281e-01 -6.03097022e-01 -2.51441538e-01
4.91020203e-01 -8.91395748e-01 -2.49321282e-01 4.84610230e-01
-4.32570606e-01 -4.73326802e-01 -7.95902550e-01 1.30041480e+00
4.57403779e-01 -3.12403888e-01 4.95948672e-01 -1.17537653e+00
-9.49833095e-01 5.31044602e-01 -6.61085010e-01 1.45684898e-01
4.08041850e-02 5.15315354e-01 1.98253691e-01 -3.34619522e-01
5.80847919e-01 1.10833657e+00 3.40569884e-01 5.59177577e-01
-7.44560599e-01 -8.76383483e-01 -1.41101077e-01 4.71122861e-01
-1.40011179e+00 -4.71247315e-01 8.83460283e-01 9.83399153e-02
4.05269831e-01 5.94348848e-01 6.24219298e-01 5.48944771e-01
6.33090675e-01 3.73105437e-01 1.28701699e+00 -6.04849100e-01
-2.69545019e-01 7.58653805e-02 -8.20439816e-01 6.13270640e-01
4.69067961e-01 3.88100088e-01 -1.72799695e-02 2.75118560e-01
9.16800976e-01 1.24185219e-01 -8.60020816e-01 -2.52623796e-01
-1.37900341e+00 7.88752317e-01 4.61307526e-01 4.04872119e-01
-5.68336137e-02 2.06875756e-01 2.13496983e-01 6.12502873e-01
1.51635185e-01 1.06414981e-01 1.07374601e-01 3.98840398e-01
-6.15216732e-01 -2.34921142e-01 7.54242897e-01 9.80011821e-01
4.06883001e-01 4.10127401e-01 1.89565554e-01 3.53366971e-01
6.16885662e-01 8.55783165e-01 5.73189020e-01 -7.10319102e-01
4.48943019e-01 3.03775728e-01 -3.11981797e-01 -1.61414814e+00
6.78969994e-02 -4.29529339e-01 -1.25621808e+00 3.84554088e-01
-2.99847007e-01 9.03035328e-02 -8.04440320e-01 1.42597258e+00
1.88066781e-01 3.47345263e-01 3.53084683e-01 7.92244434e-01
6.80440426e-01 1.03856361e+00 -3.47121656e-01 -6.59157872e-01
1.13143265e+00 -7.26800084e-01 -9.94857311e-01 -8.24520539e-04
2.94218451e-01 -1.00709510e+00 4.54156846e-01 3.67881358e-01
-1.00701094e+00 -3.12458813e-01 -1.53600621e+00 3.12385410e-01
-7.98287317e-02 -3.96729976e-01 1.56207129e-01 1.05298686e+00
-8.68163466e-01 1.99472889e-01 -4.66556966e-01 9.42424834e-02
2.36207232e-01 5.00610352e-01 -4.15485799e-01 -3.98999423e-01
-1.40194440e+00 8.13283026e-01 6.95742846e-01 3.29512954e-01
-8.83838296e-01 -1.82225809e-01 -9.50629592e-01 -8.51751342e-02
2.53398567e-01 -4.29294705e-01 3.69767576e-01 -1.03218031e+00
-1.47627831e+00 6.57380641e-01 1.79996237e-01 -4.22724307e-01
5.45282423e-01 6.68789923e-01 -8.09033692e-01 4.30951983e-01
-4.33741927e-01 3.00383985e-01 1.44297624e+00 -1.54243267e+00
-4.92608339e-01 -5.07395938e-02 -1.93430692e-01 1.16877131e-01
-3.64986658e-01 -1.31267561e-02 -4.25483108e-01 -6.82778716e-01
4.07030314e-01 -9.54308987e-01 -1.51819125e-01 -6.59219846e-02
-4.14238393e-01 6.31596625e-01 1.22507691e+00 -9.56290305e-01
1.45520425e+00 -2.37187457e+00 1.09233074e-01 4.56711143e-01
2.20375583e-01 5.64416409e-01 -1.44219905e-01 6.45773411e-01
-1.45425558e-01 2.67358422e-01 -4.52568710e-01 1.24093331e-01
-3.70357573e-01 2.67936260e-01 -1.22812368e-01 1.07524717e+00
-1.19719610e-01 7.47899413e-01 -6.93664908e-01 -6.36331737e-01
4.81722802e-01 9.12452579e-01 -3.26104701e-01 1.47054559e-02
3.61249864e-01 6.54610515e-01 -2.59205848e-01 4.79619831e-01
1.26844370e+00 -3.09596546e-02 1.46247774e-01 -1.52240589e-01
-3.29753943e-02 -3.11036885e-01 -1.11352563e+00 8.64573777e-01
-2.27555409e-01 5.32872200e-01 2.34815240e-01 -8.96122992e-01
1.02956688e+00 4.05800462e-01 3.96603674e-01 -7.69653559e-01
4.73950624e-01 4.36195254e-01 2.17114076e-01 -8.95309865e-01
3.58401716e-01 -7.83277899e-02 1.63663954e-01 2.34138727e-01
-4.02738959e-01 8.11237562e-03 -2.03676134e-01 -8.15352425e-02
8.20485234e-01 -2.33201891e-01 2.80805558e-01 -7.93348029e-02
1.04029381e+00 -4.96383637e-01 4.74176526e-01 3.70267272e-01
-5.45652248e-02 5.83267093e-01 2.44216681e-01 -1.11798033e-01
-1.37861586e+00 -3.58226240e-01 3.81657807e-03 4.35115516e-01
8.38584542e-01 1.13717793e-02 -8.52380991e-01 -2.04821900e-01
-3.17197382e-01 2.67684340e-01 -2.39160717e-01 -5.73857129e-01
-6.94872439e-01 -6.35126650e-01 5.50369442e-01 -2.71207541e-01
1.18195212e+00 -1.08781004e+00 -3.00858796e-01 1.90109864e-01
-3.56811464e-01 -1.15827000e+00 -5.98857403e-01 -3.67318630e-01
-6.87286913e-01 -1.29978502e+00 -7.63140857e-01 -1.15863955e+00
9.81919527e-01 7.72645116e-01 3.85664374e-01 7.35707223e-01
-1.25715137e-01 6.48880452e-02 -5.88837922e-01 -1.77371621e-01
-8.70254874e-01 -2.83828765e-01 -2.66239375e-01 3.43783319e-01
-6.54146001e-02 -4.30915058e-01 -9.49854910e-01 5.71107745e-01
-1.43070865e+00 7.28557855e-02 7.10132301e-01 8.50720704e-01
3.39686155e-01 9.41886544e-01 1.67640988e-02 -5.61368167e-01
3.35168153e-01 -4.42352355e-01 -4.77215111e-01 1.27325580e-01
-8.48115623e-01 -2.28552043e-01 5.34880638e-01 -4.37075317e-01
-8.38165760e-01 -2.70874083e-01 -5.35735413e-02 -1.54720396e-01
3.02916348e-01 2.77207494e-01 -4.77838427e-01 -9.00484145e-01
1.21684261e-01 8.83831501e-01 5.94082654e-01 6.83433861e-02
1.76539019e-01 9.07691479e-01 5.62569201e-01 3.29979748e-01
1.34534037e+00 7.67067194e-01 2.13760898e-01 -8.59865606e-01
1.13411061e-01 -2.70965368e-01 -2.78464258e-02 -2.54447132e-01
7.16297448e-01 -8.66084337e-01 -1.05714321e+00 1.05120385e+00
-9.48093116e-01 5.86897992e-02 2.97037512e-01 4.44843590e-01
-4.54016328e-01 8.22028279e-01 -5.28570771e-01 -8.08470607e-01
-4.14125144e-01 -1.40542018e+00 5.25543153e-01 1.26679704e-01
7.15176404e-01 -1.03805375e+00 -4.71842855e-01 2.73318946e-01
7.30285406e-01 6.72848105e-01 6.53730214e-01 -1.66517543e-03
-1.10929096e+00 -3.04922044e-01 -2.60427326e-01 6.86990380e-01
2.42723912e-01 -3.56649280e-01 -5.08374214e-01 -7.51779795e-01
6.32010341e-01 3.22395861e-01 1.09175265e+00 2.41444990e-01
1.08343518e+00 -8.29454482e-01 -1.73586860e-01 1.20438159e+00
1.83822656e+00 5.05531371e-01 1.51983452e+00 6.16752625e-01
7.45660424e-01 5.19020438e-01 3.69090050e-01 4.68111396e-01
2.54535675e-01 2.39260495e-01 8.96557570e-01 -2.58941621e-01
-1.21514969e-01 -1.91900685e-01 4.60664183e-01 1.05575705e+00
-1.47974372e-01 -8.40433240e-01 -2.28815705e-01 3.66311878e-01
-1.35857117e+00 -1.11861932e+00 -6.18491285e-02 2.08844018e+00
6.37549520e-01 6.11167774e-02 -4.79252964e-01 6.44581318e-01
1.13009834e+00 5.14146090e-01 -3.10151190e-01 -2.04474002e-01
-3.83498132e-01 -3.31698000e-01 1.21797788e+00 6.17657900e-01
-9.70294654e-01 5.02874494e-01 5.76829433e+00 1.11613250e+00
-1.39834857e+00 4.47480008e-02 3.84099603e-01 4.56080109e-01
-6.20552361e-01 2.22526267e-01 -5.18313766e-01 8.79959464e-01
5.59670150e-01 -3.97551171e-02 5.29221237e-01 3.51514190e-01
2.94643998e-01 2.02588104e-02 -1.41819075e-01 1.02614176e+00
6.06372535e-01 -1.14375663e+00 1.33447587e-01 5.88440776e-01
7.46520996e-01 -6.60471022e-01 4.00659293e-01 -4.76211339e-01
-3.67341608e-01 -9.97060835e-01 4.86933321e-01 3.04569662e-01
1.06406486e+00 -8.58169496e-01 8.72699559e-01 2.57071167e-01
-9.96872544e-01 -2.43298575e-01 -4.96526122e-01 3.60145986e-01
2.67296135e-01 3.59434217e-01 -5.10321021e-01 6.34761631e-01
2.95494080e-01 5.90151846e-01 -1.60931200e-01 9.01445210e-01
-3.59619260e-01 6.22953296e-01 -7.20894560e-02 1.85288742e-01
2.26329014e-01 -1.87904596e-01 9.76662278e-01 9.31038916e-01
6.31756186e-01 2.93999702e-01 -3.64520447e-03 3.91271561e-01
7.37448931e-02 1.89467873e-02 -7.35176563e-01 1.89572349e-01
4.44021136e-01 8.25899422e-01 -7.24318743e-01 3.01151141e-03
-3.43292989e-02 1.04975116e+00 -8.56205046e-01 2.96862960e-01
-6.50571525e-01 -7.11599171e-01 1.94711998e-01 3.08429241e-01
7.15734303e-01 -2.13225409e-01 -6.04921803e-02 -1.02795923e+00
-5.71699291e-02 -1.15160811e+00 -3.43422480e-02 -4.43079174e-01
-7.23843515e-01 5.34059823e-01 -1.91545084e-01 -1.74916244e+00
1.63206279e-01 -3.10214937e-01 -4.98299569e-01 5.60630202e-01
-2.24652147e+00 -1.45550203e+00 -3.98310155e-01 9.01331842e-01
1.78733617e-01 -4.68476593e-01 5.56050956e-01 2.91569591e-01
-2.67237216e-01 5.72738171e-01 4.33443546e-01 6.59257248e-02
5.07987678e-01 -4.85910922e-01 2.11253852e-01 1.23778009e+00
-5.42085350e-01 4.41265613e-01 9.66325045e-01 -6.84238672e-01
-1.66908753e+00 -9.77925181e-01 6.03323758e-01 3.86711538e-01
2.73048848e-01 -5.47142625e-02 -8.95102859e-01 4.08510298e-01
4.70054150e-01 2.20627747e-02 3.78651649e-01 -1.14556623e+00
-2.71229297e-01 -1.22013830e-01 -1.54370475e+00 4.04238492e-01
8.24532986e-01 -2.67495304e-01 3.20689343e-02 1.50874510e-01
9.13914979e-01 -5.21641791e-01 -7.44126081e-01 2.71506250e-01
4.79302973e-01 -1.05393541e+00 1.03760731e+00 2.69151688e-01
5.80114305e-01 -5.50761282e-01 -1.23262614e-01 -1.08754539e+00
-3.13905865e-01 -1.16427469e+00 -4.91928384e-02 1.00362182e+00
-9.47477221e-02 -9.76319909e-01 4.54049885e-01 -1.23527586e-01
9.61138159e-02 -4.73175615e-01 -8.35256815e-01 -8.56066883e-01
-2.47744367e-01 1.75853726e-02 9.49461937e-01 7.51052082e-01
-2.09745094e-01 -4.95303780e-01 -1.09837151e+00 5.02720118e-01
1.03263795e+00 -4.14149255e-01 6.04062200e-01 -6.66755557e-01
-7.57185742e-02 -1.54192582e-01 -9.44422662e-01 -9.35496628e-01
-1.79470643e-01 -6.26426101e-01 2.40024626e-02 -1.28702915e+00
-1.92480106e-02 -5.30856669e-01 -1.02562979e-01 -5.55238053e-02
-1.11601464e-01 6.61109388e-01 2.82716960e-01 5.91694653e-01
-2.29619220e-01 3.91257644e-01 1.86644351e+00 -2.33676448e-01
-7.32187927e-02 -1.20017901e-01 -8.02779317e-01 2.90506810e-01
8.24204683e-01 -5.41948259e-01 -5.34643292e-01 -3.39797795e-01
7.65867010e-02 1.93072453e-01 3.30208540e-01 -9.35834348e-01
3.18524808e-01 -2.51757115e-01 2.23996826e-02 -1.57407299e-01
1.05568320e-01 -1.21612215e+00 4.35751826e-01 1.04366076e+00
-7.79519677e-02 -2.84596235e-01 -2.56557226e-01 7.10418463e-01
-2.68784195e-01 -2.59549528e-01 1.05361915e+00 -2.75930077e-01
-7.55122423e-01 3.69837463e-01 -3.35900694e-01 -4.10794705e-01
1.33014619e+00 -7.01242507e-01 -4.76997375e-01 -6.78516150e-01
-1.42054379e-01 -6.66105300e-02 6.03386581e-01 2.15070307e-01
1.15998876e+00 -1.26309037e+00 -9.51037705e-01 7.22895443e-01
2.24356074e-02 -3.42789680e-01 4.94385302e-01 7.95866251e-01
-9.61449385e-01 1.98284015e-01 -2.75408059e-01 -3.76346290e-01
-1.42545962e+00 8.77082705e-01 3.14230174e-01 -1.89703837e-01
-7.89977908e-01 5.54214001e-01 1.85612757e-02 6.50756434e-02
1.11276411e-01 2.02489853e-01 -3.44978184e-01 -3.70422006e-01
8.90942633e-01 3.60900164e-01 -2.84051120e-01 -1.03036451e+00
-4.09643129e-02 7.27292657e-01 7.29138032e-02 -1.33601204e-02
1.09681106e+00 -8.89213622e-01 -5.67656815e-01 -3.29991668e-01
1.41215146e+00 -5.14044389e-02 -1.01797664e+00 -1.64085388e-01
-6.77417815e-01 -1.02132177e+00 2.08940387e-01 -2.46605530e-01
-1.45116782e+00 5.05074143e-01 7.38840044e-01 4.71100271e-01
1.35558569e+00 -5.87264478e-01 1.25605023e+00 -1.96992815e-01
3.87850612e-01 -7.87326455e-01 4.05383669e-02 4.73776646e-02
8.10046971e-01 -1.10898864e+00 1.07985966e-01 -6.17662847e-01
-5.79622567e-01 1.09707046e+00 7.53465220e-02 -2.58018166e-01
7.44150519e-01 2.42308781e-01 2.71971636e-02 9.61995497e-03
-2.66422540e-01 4.32760976e-02 -3.63301151e-02 1.02254796e+00
-1.49731651e-01 -1.83886692e-01 -6.13100708e-01 -2.39463165e-01
-7.69420043e-02 -2.11084671e-02 8.71458530e-01 9.11363482e-01
-7.34913886e-01 -1.03669989e+00 -8.63319099e-01 -3.02345920e-02
-7.02878773e-01 -1.17370769e-01 3.75861704e-01 5.93203247e-01
3.89938295e-01 1.26289773e+00 -2.61517286e-01 -6.00990295e-01
-3.56329754e-02 -6.75458729e-01 3.52578491e-01 -1.42030567e-02
-1.72608942e-01 1.57606944e-01 -4.32100564e-01 -2.38946393e-01
-7.66588688e-01 -1.39748439e-01 -1.08130145e+00 -9.15484250e-01
-6.26644075e-01 -1.85888838e-02 8.45532238e-01 7.96048999e-01
1.05675779e-01 5.19362450e-01 1.34178686e+00 -5.78563750e-01
-3.90540063e-01 -5.33319354e-01 -7.74117231e-01 2.72774905e-01
8.18665266e-01 -1.63071260e-01 -7.29546547e-01 3.20952475e-01] | [4.315708637237549, 8.050440788269043] |
40529e0b-1501-48f3-8543-202604d7887d | boosting-monocular-3d-object-detection-with | 2210.16574 | null | https://arxiv.org/abs/2210.16574v1 | https://arxiv.org/pdf/2210.16574v1.pdf | Boosting Monocular 3D Object Detection with Object-Centric Auxiliary Depth Supervision | Recent advances in monocular 3D detection leverage a depth estimation network explicitly as an intermediate stage of the 3D detection network. Depth map approaches yield more accurate depth to objects than other methods thanks to the depth estimation network trained on a large-scale dataset. However, depth map approaches can be limited by the accuracy of the depth map, and sequentially using two separated networks for depth estimation and 3D detection significantly increases computation cost and inference time. In this work, we propose a method to boost the RGB image-based 3D detector by jointly training the detection network with a depth prediction loss analogous to the depth estimation task. In this way, our 3D detection network can be supervised by more depth supervision from raw LiDAR points, which does not require any human annotation cost, to estimate accurate depth without explicitly predicting the depth map. Our novel object-centric depth prediction loss focuses on depth around foreground objects, which is important for 3D object detection, to leverage pixel-wise depth supervision in an object-centric manner. Our depth regression model is further trained to predict the uncertainty of depth to represent the 3D confidence of objects. To effectively train the 3D detector with raw LiDAR points and to enable end-to-end training, we revisit the regression target of 3D objects and design a network architecture. Extensive experiments on KITTI and nuScenes benchmarks show that our method can significantly boost the monocular image-based 3D detector to outperform depth map approaches while maintaining the real-time inference speed. | ['Dongsuk Kum', 'Jun Won Choi', 'Sangmin Sim', 'Sanmin Kim', 'Youngseok Kim'] | 2022-10-29 | null | null | null | null | ['monocular-3d-object-detection'] | ['computer-vision'] | [ 1.62523210e-01 3.03837121e-01 -4.17475194e-01 -5.36856174e-01
-7.30746627e-01 -3.57380152e-01 5.10344386e-01 -2.06442162e-01
-4.63329792e-01 1.27543300e-01 -3.10237348e-01 -4.73111331e-01
5.71052074e-01 -9.54630911e-01 -1.06006312e+00 -4.59614277e-01
2.02121407e-01 6.03229105e-01 8.58009160e-01 4.43778306e-01
3.14295709e-01 7.99612045e-01 -1.53793728e+00 7.29713887e-02
5.48900306e-01 1.56381488e+00 3.11540455e-01 7.79652596e-01
-2.08427817e-01 7.76816845e-01 -3.59037995e-01 -9.87119414e-03
8.42907190e-01 1.06265299e-01 -1.89104810e-01 2.36897767e-01
1.07374465e+00 -1.28162801e+00 -4.55412626e-01 8.31725001e-01
4.94607329e-01 -1.86743692e-01 6.79445684e-01 -1.25235605e+00
-8.38815868e-02 1.49051398e-01 -1.03814721e+00 4.16300520e-02
1.27971590e-01 3.70636255e-01 7.90250301e-01 -1.23690438e+00
3.94423485e-01 1.49499094e+00 7.23195612e-01 5.83187699e-01
-1.11367774e+00 -8.22244167e-01 5.19910634e-01 -1.58259869e-01
-1.33153963e+00 -1.70410469e-01 8.74436140e-01 -3.61416280e-01
1.22175455e+00 -2.69396961e-01 7.64696896e-01 8.42866182e-01
-1.27807051e-01 1.10506141e+00 9.04127538e-01 -1.91203728e-01
2.37367168e-01 -2.68130517e-03 -2.41591409e-02 9.19992507e-01
4.24182177e-01 4.67119098e-01 -5.48913002e-01 3.73115301e-01
1.13652253e+00 2.61844158e-01 7.04001486e-02 -8.93035173e-01
-8.22771490e-01 8.10251176e-01 1.03154683e+00 -5.31803727e-01
-4.50123586e-02 6.29202366e-01 1.49412006e-01 -1.24801688e-01
8.03890646e-01 -5.21726534e-02 -5.22592664e-01 2.09226564e-01
-9.08393621e-01 2.14101329e-01 3.87345552e-01 1.05825162e+00
1.21954620e+00 -1.64975166e-01 -6.92900345e-02 2.55895019e-01
6.37404025e-01 8.95878434e-01 -2.94319987e-01 -1.26025534e+00
7.14383364e-01 1.16415250e+00 5.63362986e-02 -5.06187499e-01
-3.70377123e-01 -2.85860479e-01 -4.38906401e-01 8.82016659e-01
6.25272691e-01 1.21949114e-01 -1.28211236e+00 1.27887011e+00
8.17858934e-01 1.90492779e-01 -2.56589860e-01 1.16919541e+00
7.62234867e-01 4.54673767e-01 -3.57067943e-01 3.81246597e-01
1.07005107e+00 -8.54290009e-01 1.52324796e-01 -7.50307977e-01
6.42208695e-01 -4.11550283e-01 9.66701746e-01 3.29619169e-01
-1.15838540e+00 -4.53256637e-01 -1.24929416e+00 -6.09566092e-01
-1.01353973e-01 2.05870360e-01 6.76661372e-01 6.61332965e-01
-8.72079313e-01 2.13461116e-01 -1.09310460e+00 -1.06979385e-02
8.75182390e-01 3.52042615e-01 -9.49939061e-03 -3.44067544e-01
-6.07916415e-01 8.83483469e-01 5.45791209e-01 -8.51691514e-02
-1.10295248e+00 -9.76157606e-01 -1.10797203e+00 -1.48422062e-01
5.77446163e-01 -8.92162025e-01 1.29995084e+00 -4.69345599e-01
-1.34621429e+00 1.22895849e+00 -7.53733292e-02 -6.68017507e-01
8.44014704e-01 -4.20756221e-01 6.51277900e-01 3.77738327e-01
2.34155163e-01 1.25632906e+00 7.97655046e-01 -1.16081035e+00
-1.19566667e+00 -6.65169299e-01 3.02621722e-01 2.61798531e-01
1.14993811e-01 -5.06782055e-01 -8.35212111e-01 9.16416720e-02
5.35513103e-01 -7.43524134e-01 -1.77171499e-01 8.91470611e-01
-4.07950133e-01 -2.41658002e-01 9.02028024e-01 2.83139143e-02
3.67258161e-01 -2.00752854e+00 -2.30305344e-01 -9.41126421e-02
4.31712627e-01 4.04860675e-02 1.24272078e-01 -3.28622490e-01
4.77110505e-01 -1.28811985e-01 -1.67078003e-01 -8.43119502e-01
2.71568988e-02 2.91822642e-01 -3.38683546e-01 5.74495792e-01
7.08960235e-01 9.99168932e-01 -9.70850825e-01 -5.33061683e-01
7.74817050e-01 6.88016772e-01 -8.79036605e-01 3.35924715e-01
-5.87135851e-01 1.88251421e-01 -4.63884950e-01 1.03027189e+00
1.06344330e+00 -2.35886186e-01 -3.07772011e-01 -1.85882926e-01
-1.66489527e-01 5.39603591e-01 -9.81953621e-01 1.73365974e+00
-5.50614119e-01 7.20655382e-01 1.46974653e-01 -5.57731092e-01
9.16853964e-01 -3.98036301e-01 2.76778579e-01 -6.98012948e-01
2.59142399e-01 2.92030215e-01 -4.49064553e-01 1.05596140e-01
2.86411762e-01 -6.92345202e-02 1.06808422e-02 2.89051622e-01
-1.05862461e-01 -7.52257288e-01 -2.72476435e-01 2.43854523e-01
1.15330577e+00 5.63352287e-01 -6.74083009e-02 3.34007949e-01
1.85870022e-01 6.18162230e-02 5.70290744e-01 8.62564802e-01
-2.19986215e-01 6.28432274e-01 3.73616695e-01 -4.22074258e-01
-1.02877533e+00 -1.35629010e+00 -3.33666980e-01 5.08060634e-01
5.24843037e-01 7.39878491e-02 -3.86610329e-01 -9.70863104e-01
6.12923920e-01 3.43087733e-01 -5.12043953e-01 -4.63389792e-02
-5.78817427e-01 -3.47566009e-01 4.50082660e-01 8.90633404e-01
6.73918366e-01 -4.29641515e-01 -1.19213426e+00 5.04595153e-02
2.74824440e-01 -1.52383494e+00 -1.74895659e-01 8.03036392e-01
-1.12913799e+00 -1.03558969e+00 -5.30511856e-01 -3.80094528e-01
6.19800091e-01 6.78025186e-01 1.15926492e+00 -1.10082887e-01
-3.66229534e-01 2.65616566e-01 -1.00584120e-01 -6.48708224e-01
2.92881718e-03 1.28824919e-01 -1.36548266e-01 -5.71597397e-01
5.95977962e-01 -5.33724606e-01 -9.92532372e-01 2.35744998e-01
-4.60467964e-01 1.71977952e-01 7.10062563e-01 6.29299581e-01
7.40273237e-01 -3.43854636e-01 5.31542972e-02 -5.88287473e-01
-5.53471565e-01 -8.59590173e-02 -1.30565763e+00 -4.23642635e-01
-7.03059018e-01 1.09828636e-02 -3.05587724e-02 -3.37455690e-01
-8.02193463e-01 6.77333057e-01 -4.16466556e-02 -9.63776350e-01
1.15342237e-01 -2.30974078e-01 -1.82036951e-01 -3.65574121e-01
6.40095830e-01 -3.48509699e-02 -9.95004699e-02 -3.79338801e-01
4.55671728e-01 4.91292179e-01 5.46371102e-01 -3.72638285e-01
1.05904675e+00 9.62660730e-01 3.65429789e-01 -4.16159242e-01
-1.46914077e+00 -7.02305257e-01 -7.14007616e-01 -2.10122705e-01
8.18662584e-01 -1.56477261e+00 -8.82465661e-01 3.94532591e-01
-1.40372384e+00 -5.53949714e-01 -3.08510900e-01 3.97741795e-01
-4.84349161e-01 1.86538920e-01 -5.79199910e-01 -1.08525050e+00
-2.24711925e-01 -9.29112852e-01 1.81605995e+00 -2.59412378e-02
3.12475115e-01 -5.22476256e-01 -4.43667948e-01 3.06350052e-01
-7.30845258e-02 4.32117373e-01 5.63646913e-01 1.47611424e-01
-1.57501996e+00 -2.53230959e-01 -8.44325066e-01 3.50181133e-01
-1.87047675e-01 -2.38661170e-01 -1.44468939e+00 8.89684856e-02
8.01489279e-02 -4.81272668e-01 1.28128576e+00 4.07565981e-01
1.09766090e+00 1.79776788e-01 -4.76534218e-01 9.66379583e-01
1.41969228e+00 -2.67780364e-01 2.89213985e-01 2.79786408e-01
1.06537867e+00 4.52419311e-01 8.50571752e-01 4.04805750e-01
5.13867617e-01 5.38745701e-01 9.91230011e-01 -1.31824076e-01
-4.92020607e-01 -5.51940501e-01 3.65084171e-01 -3.19143273e-02
3.37099850e-01 -9.41757038e-02 -8.96140277e-01 2.71472692e-01
-1.69739652e+00 -5.89496315e-01 -1.31174952e-01 2.13581395e+00
7.77798891e-01 6.93056166e-01 9.52289775e-02 1.36058852e-01
3.28481048e-01 4.54680659e-02 -1.09515405e+00 1.63824961e-01
-1.69850010e-02 1.23111568e-02 1.00266325e+00 6.56082749e-01
-1.10098946e+00 1.14302242e+00 5.41664743e+00 5.03432155e-01
-1.04764116e+00 2.65594814e-02 5.80093503e-01 -6.93319917e-01
-1.76557735e-01 1.85756553e-02 -1.46585095e+00 1.31506026e-01
3.01722556e-01 4.09522116e-01 4.54459004e-02 1.21859181e+00
2.56914020e-01 -4.76211637e-01 -1.51237953e+00 1.32817066e+00
-1.75011680e-01 -1.23434258e+00 -8.27082694e-02 2.14116141e-01
6.34835124e-01 5.17419994e-01 -4.92300130e-02 2.76384622e-01
2.66455173e-01 -7.07618356e-01 9.72964942e-01 -1.76055767e-02
8.57664406e-01 -7.68941343e-01 4.93701011e-01 6.89215243e-01
-1.25698400e+00 -1.35722995e-01 -7.07338810e-01 -3.86524796e-01
2.32874826e-01 1.03572822e+00 -1.16105604e+00 -1.63844991e-02
7.77608931e-01 8.52705598e-01 -5.07915974e-01 8.49219501e-01
-6.25686646e-01 1.15172565e-01 -7.29706705e-01 1.76762100e-02
3.31323713e-01 1.53498769e-01 4.03274119e-01 8.99990320e-01
1.65009990e-01 -7.14328364e-02 3.37510347e-01 1.32181489e+00
-3.31750184e-01 -5.98988175e-01 -5.99801660e-01 3.97145748e-01
6.73267603e-01 1.12626338e+00 -7.58375227e-01 -2.00133100e-01
-4.20121372e-01 8.75863671e-01 4.26047891e-01 -4.01851386e-02
-8.86652052e-01 -1.62840456e-01 7.51569629e-01 5.79301894e-01
5.84056377e-01 -3.08607399e-01 -6.22262299e-01 -1.03293931e+00
2.39639953e-01 -1.17023714e-01 -6.11539744e-02 -7.94261515e-01
-1.11268437e+00 8.98743272e-02 -1.37315109e-01 -1.12774181e+00
-2.40271479e-01 -1.01561284e+00 -2.22957045e-01 8.14266324e-01
-2.24062300e+00 -1.21019650e+00 -7.25356102e-01 3.72579128e-01
3.93481374e-01 4.94828612e-01 7.24980384e-02 3.76648247e-01
-2.34173775e-01 4.16703105e-01 -6.44647717e-01 1.02451950e-01
6.46216512e-01 -1.32986772e+00 6.86841190e-01 7.88668811e-01
-1.31317705e-01 7.48407319e-02 1.51865959e-01 -7.10000455e-01
-1.55135715e+00 -1.32495415e+00 5.55821419e-01 -1.01821065e+00
3.14734936e-01 -8.26757371e-01 -5.60427189e-01 5.30292094e-01
-6.10822976e-01 5.19762993e-01 -1.49480015e-01 -1.64450333e-01
-5.68792582e-01 -2.10611224e-01 -1.12252009e+00 3.72633487e-01
1.62158895e+00 -6.93518996e-01 -3.22711051e-01 2.23516271e-01
1.19328833e+00 -8.33111763e-01 -3.82783681e-01 7.06157029e-01
5.36238849e-01 -1.19546235e+00 1.42941523e+00 1.82292864e-01
4.05148864e-01 -6.07715309e-01 -2.52873212e-01 -4.69622880e-01
1.81925878e-01 -2.51640193e-02 -6.95563614e-01 8.20093513e-01
2.50085115e-01 -3.25583547e-01 1.51768923e+00 5.27519405e-01
-2.86148161e-01 -8.23201239e-01 -1.12331831e+00 -7.47984290e-01
-9.34163108e-02 -8.96046340e-01 3.96387786e-01 3.06790084e-01
-7.54007101e-01 3.93531561e-01 -5.43852150e-02 4.54232514e-01
1.02241325e+00 2.88800716e-01 1.19136000e+00 -1.34525216e+00
-2.90610284e-01 -5.42299330e-01 -6.11758292e-01 -1.99736154e+00
-1.51237305e-02 -8.11365545e-01 3.39732409e-01 -1.44820631e+00
9.04019177e-02 -5.95729351e-01 1.38552919e-01 2.88220227e-01
-1.83212623e-01 4.84972626e-01 1.13389485e-01 1.05702259e-01
-5.83799243e-01 5.97149789e-01 1.18373597e+00 -1.58388376e-01
-2.48269618e-01 -1.02284169e-02 -2.59057790e-01 9.49054897e-01
3.11592042e-01 -6.11430168e-01 -4.39343482e-01 -6.99461401e-01
2.36261979e-01 -1.36698157e-01 8.75999391e-01 -1.01800990e+00
2.68871784e-01 5.34105040e-02 9.30640697e-01 -1.36046839e+00
7.85252333e-01 -7.88883984e-01 -7.37925589e-01 7.02983379e-01
1.35486975e-01 -6.45932555e-01 9.33442637e-02 6.28960848e-01
1.97334558e-01 7.02520460e-02 1.00521827e+00 -1.94725201e-01
-8.57241988e-01 6.64009333e-01 1.58048645e-01 -7.52364919e-02
9.94796455e-01 -6.46061242e-01 -8.80459547e-02 3.38891451e-03
-2.02984467e-01 4.23390508e-01 7.62942016e-01 2.65024006e-01
8.92245650e-01 -1.11535323e+00 -4.26123112e-01 2.62394994e-01
2.85143673e-01 1.09916925e+00 -1.85001388e-01 5.28388202e-01
-6.60144508e-01 3.06480855e-01 2.92712241e-01 -1.31443477e+00
-8.40622663e-01 4.61729974e-01 6.32595301e-01 3.62638757e-02
-7.99201012e-01 1.16273689e+00 6.51616156e-01 -4.60273206e-01
7.32273757e-01 -9.63283956e-01 5.31144202e-01 -1.88939884e-01
4.89977658e-01 1.64331257e-01 1.40405819e-02 9.36206132e-02
-5.47519684e-01 8.65581214e-01 8.71466380e-03 -6.34575859e-02
1.11753166e+00 -2.08537251e-01 3.59852761e-01 4.09309536e-01
1.18797433e+00 -3.10725778e-01 -2.15834665e+00 -3.86802644e-01
-2.11398482e-01 -7.57773459e-01 5.61041117e-01 -6.23283088e-01
-1.06083584e+00 1.27501082e+00 6.96933746e-01 -5.25450408e-01
8.17783833e-01 2.68611789e-01 7.88273156e-01 5.89874744e-01
5.83100677e-01 -9.63989258e-01 4.27262455e-01 6.28039896e-01
3.33903641e-01 -1.87007451e+00 1.58075124e-01 -5.96534729e-01
-3.07240151e-02 9.78743613e-01 1.04296517e+00 -2.70616621e-01
4.94915038e-01 4.77138638e-01 -1.09300852e-01 -1.61119148e-01
-5.35507917e-01 -4.56036985e-01 2.04337478e-01 6.40168250e-01
-5.53463325e-02 -2.38638833e-01 5.22249401e-01 1.73453018e-01
1.56788394e-01 -1.01729415e-01 1.31378055e-01 9.42276597e-01
-8.67695749e-01 -8.44314277e-01 -3.31568599e-01 3.24742436e-01
1.35190925e-02 -3.41895185e-02 -4.58847553e-01 9.27130997e-01
3.33353698e-01 6.87907815e-01 3.62553954e-01 -2.40438551e-01
3.15695792e-01 -3.34489018e-01 8.32874537e-01 -8.81977558e-01
-8.49988461e-02 2.74545066e-02 -2.25364089e-01 -8.78975868e-01
-2.15315595e-01 -3.90568852e-01 -1.46548438e+00 -3.46320629e-01
-5.23839533e-01 -6.64321423e-01 8.55361462e-01 7.33653605e-01
1.96312830e-01 1.90345794e-01 7.18588710e-01 -1.48028922e+00
-4.86079156e-01 -6.16068840e-01 -3.08978170e-01 -1.13172323e-01
5.53426266e-01 -8.57735634e-01 -5.21140397e-01 -4.63941336e-01] | [7.856112003326416, -2.612988233566284] |
d56e4718-a6f7-43ae-bf09-9a756cfd1096 | detrs-with-hybrid-matching | 2207.13080 | null | https://arxiv.org/abs/2207.13080v3 | https://arxiv.org/pdf/2207.13080v3.pdf | DETRs with Hybrid Matching | One-to-one set matching is a key design for DETR to establish its end-to-end capability, so that object detection does not require a hand-crafted NMS (non-maximum suppression) to remove duplicate detections. This end-to-end signature is important for the versatility of DETR, and it has been generalized to broader vision tasks. However, we note that there are few queries assigned as positive samples and the one-to-one set matching significantly reduces the training efficacy of positive samples. We propose a simple yet effective method based on a hybrid matching scheme that combines the original one-to-one matching branch with an auxiliary one-to-many matching branch during training. Our hybrid strategy has been shown to significantly improve accuracy. In inference, only the original one-to-one match branch is used, thus maintaining the end-to-end merit and the same inference efficiency of DETR. The method is named H-DETR, and it shows that a wide range of representative DETR methods can be consistently improved across a wide range of visual tasks, including DeformableDETR, PETRv2, PETR, and TransTrack, among others. The code is available at: https://github.com/HDETR | ['Han Hu', 'Chao Zhang', 'Lei Sun', 'WeiHong Lin', 'Haojun Yu', 'Xiaopei Wu', 'Haodi He', 'Yuhui Yuan', 'Ding Jia'] | 2022-07-26 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jia_DETRs_With_Hybrid_Matching_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jia_DETRs_With_Hybrid_Matching_CVPR_2023_paper.pdf | cvpr-2023-1 | ['set-matching'] | ['computer-vision'] | [ 3.24536592e-01 -2.99387574e-01 -1.08823806e-01 -4.52789038e-01
-1.02117932e+00 -5.16472340e-01 7.21865058e-01 -2.02559620e-01
-5.24412394e-01 5.24455309e-01 -3.76845896e-01 -1.68419927e-01
-1.76289722e-01 -4.44615602e-01 -7.73113728e-01 -7.55860269e-01
1.24966629e-01 2.61124551e-01 7.51932323e-01 -1.37526635e-02
5.74282333e-02 4.05371457e-01 -1.86663508e+00 3.67451727e-01
7.05229163e-01 1.10259056e+00 3.49613219e-01 7.01886594e-01
2.40550578e-01 5.03320992e-01 -3.85167778e-01 -5.95704257e-01
5.45910478e-01 -2.44315505e-01 -4.17279452e-01 -1.97816014e-01
1.05564404e+00 -3.09275538e-01 -3.14719617e-01 9.26752388e-01
7.93977022e-01 1.09364629e-01 5.69785595e-01 -1.22868586e+00
-5.37552953e-01 1.53272584e-01 -9.38134849e-01 4.23795968e-01
-1.46565298e-02 3.99553478e-01 1.19202220e+00 -1.44152558e+00
5.64895034e-01 1.11513710e+00 9.16950822e-01 2.59405822e-01
-1.32858706e+00 -6.89270973e-01 -8.47187415e-02 1.87017769e-01
-1.34210014e+00 -4.73966092e-01 2.76786894e-01 -4.67474937e-01
9.24829543e-01 4.69164997e-01 5.37330389e-01 8.75091434e-01
1.04879387e-01 9.04704571e-01 1.09644556e+00 -4.66043204e-01
8.11503269e-03 6.65612742e-02 3.45862925e-01 5.06327868e-01
2.68599421e-01 2.30310112e-01 -5.89810133e-01 -3.08602095e-01
5.06285429e-01 8.38253126e-02 -1.87129200e-01 -4.13171083e-01
-9.91639853e-01 7.25517035e-01 3.65808725e-01 -4.71008271e-02
-1.64817512e-01 2.23684132e-01 3.97877693e-01 1.63993627e-01
3.70072424e-01 1.97628945e-01 -1.90937221e-01 -7.02358186e-02
-1.13846016e+00 4.36124384e-01 5.08294702e-01 9.22366619e-01
6.67850614e-01 -1.29622877e-01 -7.34768867e-01 9.68629897e-01
4.38567959e-02 6.58740878e-01 1.26840606e-01 -9.95980144e-01
3.37004781e-01 3.15593421e-01 -1.55591685e-02 -8.04782569e-01
-1.21143147e-01 -8.45643938e-01 -7.36142814e-01 3.42219532e-01
3.07457387e-01 1.54261321e-01 -1.04387999e+00 1.84780538e+00
4.67384577e-01 1.79612041e-01 -3.65606070e-01 8.14045191e-01
7.92536318e-01 2.36203462e-01 -1.33375198e-01 4.17235456e-02
1.47712469e+00 -9.37650919e-01 -3.39090526e-01 -5.57098508e-01
3.00553024e-01 -8.94509435e-01 1.01955462e+00 3.53592247e-01
-1.00267982e+00 -5.90678453e-01 -9.94556904e-01 -2.07763806e-01
-1.71637475e-01 2.80409664e-01 5.38216472e-01 4.83361900e-01
-9.38088119e-01 5.62690973e-01 -6.95751667e-01 -3.29642057e-01
5.90281606e-01 4.04288709e-01 -4.05592084e-01 -4.27633151e-02
-8.85211825e-01 7.54184067e-01 2.15027645e-01 1.48846298e-01
-4.82568145e-01 -5.77689826e-01 -5.88401079e-01 -4.56701368e-02
5.82580507e-01 -8.43571663e-01 1.45268095e+00 -5.55527091e-01
-8.23235631e-01 9.33897078e-01 -5.14295936e-01 -5.20275235e-01
8.30726624e-01 -2.44731516e-01 -9.12653208e-02 2.20744431e-01
3.32193077e-01 7.55930662e-01 8.98875654e-01 -1.14289880e+00
-8.13925922e-01 -3.48927498e-01 -2.80534744e-01 7.62758777e-02
-3.52073699e-01 -1.70595068e-02 -9.93556261e-01 -7.28433609e-01
1.10418245e-01 -9.59158421e-01 7.65103772e-02 2.95107871e-01
-2.61947930e-01 -1.31996885e-01 8.80468786e-01 -5.27877092e-01
1.23614597e+00 -2.29309511e+00 -3.15157712e-01 1.26905262e-01
5.38852513e-01 5.17157018e-01 -2.07913727e-01 5.00617146e-01
-5.20941354e-02 -9.51217562e-02 -2.92364180e-01 -7.76448250e-01
-1.71421319e-01 4.80435528e-02 -1.72636151e-01 5.91071606e-01
4.04714406e-01 8.58613968e-01 -9.48705494e-01 -5.13311684e-01
2.80826122e-01 3.86244059e-01 -3.89140576e-01 -5.18287644e-02
6.27787784e-02 5.03912196e-02 -2.09143028e-01 5.84483385e-01
8.50787222e-01 -4.61849868e-01 -1.44071519e-01 -3.17682147e-01
-1.05138972e-01 7.26507604e-02 -1.25660360e+00 1.04119658e+00
1.49098694e-01 8.33351016e-01 -6.75285533e-02 -6.88409925e-01
7.39734113e-01 -1.09007299e-01 3.72021377e-01 -7.69150853e-01
7.44292811e-02 2.53659040e-01 3.90873104e-02 -2.87170947e-01
5.63548207e-01 1.02329507e-01 2.86175519e-01 6.70800060e-02
2.37057675e-02 2.67249972e-01 5.14172554e-01 4.91914868e-01
1.04345047e+00 -7.41343126e-02 4.47909564e-01 -8.92929137e-02
1.52072906e-01 -1.10386536e-02 7.22305775e-01 1.11152482e+00
-2.35561594e-01 9.36204672e-01 3.21685135e-01 8.09144415e-03
-9.59585965e-01 -9.79943156e-01 -5.02516389e-01 9.64211345e-01
2.03991264e-01 -3.28922749e-01 -4.68083322e-01 -4.01473314e-01
3.32998633e-01 6.32640004e-01 -7.06898332e-01 -1.83214784e-01
-3.72292161e-01 -6.26523972e-01 6.49711609e-01 7.02768743e-01
4.50074732e-01 -8.90057087e-01 -6.48907959e-01 7.56032020e-02
-9.04747695e-02 -1.16683555e+00 -7.38347948e-01 2.75980175e-01
-7.05058455e-01 -1.13291645e+00 -6.72383845e-01 -6.21528268e-01
5.31238794e-01 8.48236442e-01 1.05450797e+00 8.14865008e-02
-4.81202841e-01 3.76523525e-01 -4.01194245e-01 -5.61550617e-01
-6.96047172e-02 -2.57863075e-01 4.70687859e-02 1.77987665e-02
5.01748800e-01 -3.24757755e-01 -7.18789279e-01 5.98744392e-01
-5.93558013e-01 1.08510926e-02 6.64012790e-01 1.09672570e+00
9.61007655e-01 -4.04660344e-01 3.66172671e-01 -7.10979104e-01
2.76337475e-01 -2.10760280e-01 -5.87255359e-01 1.27969369e-01
-7.26230621e-01 -2.48122662e-01 4.12866771e-01 -5.78378499e-01
-6.59270883e-01 9.08433795e-02 5.32053970e-03 -7.88555026e-01
3.78080308e-01 3.17716122e-01 1.84238866e-01 -1.62021279e-01
7.30705857e-01 2.56294310e-01 2.78404206e-01 -4.12735224e-01
3.05536330e-01 6.70112073e-01 9.04516459e-01 -1.82839379e-01
7.68695951e-01 5.93566418e-01 -1.55139416e-02 -7.98427641e-01
-7.87160635e-01 -9.75099206e-01 -2.90991843e-01 -2.18526706e-01
5.39850295e-01 -1.01519477e+00 -7.29353845e-01 6.83799207e-01
-8.39791358e-01 -1.56838104e-01 -2.43407309e-01 4.82645035e-01
-2.83371866e-01 4.49270308e-01 -5.52097321e-01 -9.19135511e-01
-6.30080700e-01 -9.33683395e-01 1.09684992e+00 2.44688168e-01
-8.85973647e-02 -4.00508523e-01 -1.52405649e-01 2.51775205e-01
2.10940272e-01 6.74072206e-02 4.24554527e-01 -6.36831105e-01
-6.82301044e-01 -3.67567599e-01 -5.35143316e-01 4.93541270e-01
-9.05455798e-02 1.39896423e-01 -1.19270396e+00 -6.80843234e-01
-8.88131857e-02 -2.47896507e-01 1.35873234e+00 4.51735973e-01
1.06534982e+00 7.32907653e-02 -4.07303214e-01 5.26825786e-01
1.43210912e+00 -4.30915877e-02 7.24641562e-01 2.63325393e-01
6.29392803e-01 4.24071044e-01 8.01110208e-01 3.32463622e-01
2.66824603e-01 7.90493369e-01 4.96967793e-01 -4.25306082e-01
-3.23879302e-01 -4.20865230e-03 1.93430126e-01 2.51116991e-01
1.31032154e-01 -2.22807184e-01 -6.42795682e-01 6.30038023e-01
-1.90771770e+00 -1.33417606e+00 -4.64243889e-01 2.54121590e+00
6.86923504e-01 3.08876932e-01 2.34139398e-01 -5.80147132e-02
7.99026489e-01 1.12102933e-01 -6.32439137e-01 1.45599842e-01
-2.49672502e-01 1.03450894e-01 5.47599971e-01 2.35631183e-01
-1.02669752e+00 6.51402414e-01 6.19648552e+00 1.22927916e+00
-9.78596091e-01 7.80933499e-02 4.42883909e-01 -2.19405830e-01
7.12580681e-02 3.44509855e-02 -1.19790769e+00 3.94091457e-01
4.67795372e-01 8.74802023e-02 2.93753862e-01 9.25631702e-01
1.87407970e-01 -3.71545613e-01 -1.22286820e+00 1.12122607e+00
-2.63836719e-02 -1.30294609e+00 -1.26422837e-01 1.06518045e-01
3.90085548e-01 2.27352053e-01 7.72881955e-02 2.74807781e-01
-8.04705471e-02 -5.86230874e-01 9.89133179e-01 2.92017519e-01
9.46826935e-01 -4.84211475e-01 7.16167569e-01 3.56975108e-01
-1.34209299e+00 8.26548487e-02 -3.71959388e-01 2.69945741e-01
2.44771913e-02 8.23732257e-01 -8.37234020e-01 5.76180577e-01
9.94620919e-01 5.98824620e-01 -7.83135533e-01 1.45493865e+00
1.48578912e-01 5.66280723e-01 -3.97262275e-01 -4.23646718e-02
6.61559924e-02 -6.28127530e-02 7.65670180e-01 1.46755469e+00
2.81938553e-01 -2.00913534e-01 2.10661396e-01 6.90340400e-01
-9.06157345e-02 -2.85909057e-01 -4.37743336e-01 1.62467316e-01
7.30100632e-01 1.15552115e+00 -6.73051953e-01 -3.81826401e-01
-4.36971813e-01 8.51271749e-01 3.33557010e-01 1.88697696e-01
-8.82702708e-01 -6.02723241e-01 4.76238310e-01 1.26446396e-01
7.92081475e-01 3.75716202e-03 -4.34487224e-01 -6.95987046e-01
4.76607978e-01 -9.99950707e-01 5.01813948e-01 -6.99203849e-01
-1.60365808e+00 3.69501591e-01 8.17918777e-02 -1.47819340e+00
-1.98600322e-01 -5.98213196e-01 -5.14008045e-01 9.08666015e-01
-1.42590141e+00 -1.02905321e+00 -5.41904092e-01 5.44895947e-01
5.07364511e-01 6.67519420e-02 4.56907272e-01 4.57278967e-01
-5.27281284e-01 9.36051607e-01 2.39630952e-01 2.05254719e-01
9.69276190e-01 -1.23384225e+00 5.27153492e-01 1.17481804e+00
-4.85426225e-02 8.49745512e-01 7.03574598e-01 -5.92805326e-01
-1.31732464e+00 -1.07173467e+00 6.98089957e-01 -5.35237789e-01
3.43071431e-01 -4.12950665e-01 -1.09927619e+00 5.27739882e-01
-4.00685258e-02 1.55491397e-01 3.56559783e-01 1.88694417e-01
-4.63774294e-01 -3.23697984e-01 -9.95853961e-01 5.73981762e-01
1.19260287e+00 -3.40549439e-01 -4.11440432e-01 2.29315042e-01
4.97368187e-01 -6.09381437e-01 -4.81265724e-01 5.39676845e-01
7.56681502e-01 -1.13138616e+00 1.10778785e+00 -1.69399232e-01
3.35008800e-01 -5.34588397e-01 -1.55400068e-01 -7.03708768e-01
-5.64624846e-01 -5.04725337e-01 -3.02858293e-01 1.25676918e+00
3.88062060e-01 -1.00937104e+00 6.62419677e-01 3.65814239e-01
-2.20569372e-01 -9.59447622e-01 -9.00901437e-01 -1.24550188e+00
-3.31909180e-01 -3.45239252e-01 1.27291322e-01 6.49056017e-01
-7.79644608e-01 3.58628154e-01 -5.49901068e-01 1.91560924e-01
8.36573243e-01 1.24121778e-01 1.01297641e+00 -1.31447887e+00
-7.63839722e-01 -3.04433107e-01 -2.89822400e-01 -1.28860533e+00
-4.73688364e-01 -7.54280210e-01 1.30016640e-01 -1.44226682e+00
5.53655684e-01 -5.26254654e-01 -2.87233174e-01 6.20108485e-01
-4.96953607e-01 5.61324298e-01 2.60464668e-01 5.20125628e-01
-4.37010467e-01 4.88968015e-01 1.24077201e+00 3.92775871e-02
-1.80404022e-01 1.46612689e-01 -7.78590620e-01 4.59301382e-01
5.99883080e-01 -7.47087538e-01 -3.66399050e-01 -2.62038559e-01
6.06109090e-02 -2.47100323e-01 8.01963747e-01 -9.93083715e-01
3.88318598e-01 6.75089238e-03 4.36041504e-01 -8.67537022e-01
6.54206991e-01 -4.51749682e-01 1.35448724e-01 4.54090536e-01
4.13335115e-03 6.03924766e-02 3.21131885e-01 7.94000506e-01
-1.13037318e-01 -2.02883452e-01 1.01906812e+00 9.31147486e-02
-8.44336629e-01 2.94883192e-01 1.13668844e-01 2.40924045e-01
1.06385899e+00 -7.79385328e-01 -4.14680898e-01 -2.22261980e-01
-3.10802400e-01 4.68764365e-01 2.91548640e-01 2.08945334e-01
4.02520418e-01 -1.17809820e+00 -8.84881377e-01 9.97288823e-02
3.50883961e-01 -8.70524943e-02 2.70664632e-01 1.26389456e+00
-1.17565170e-01 1.60549149e-01 3.66786346e-02 -1.09296811e+00
-1.82266498e+00 3.43420208e-01 2.72322297e-01 -1.97393119e-01
-9.34471607e-01 1.12709188e+00 2.45769262e-01 -1.80598229e-01
3.61267716e-01 -2.87987702e-02 1.83165789e-01 2.86992509e-02
7.70551443e-01 4.37574387e-01 2.91313589e-01 -3.65661949e-01
-5.02864122e-01 5.66318989e-01 -3.58067691e-01 -5.14428094e-02
1.13323307e+00 8.65819305e-02 1.89797021e-02 3.25356722e-01
1.02554262e+00 1.42106488e-01 -1.31947505e+00 -3.61151218e-01
-4.57590461e-01 -8.61793339e-01 -7.97672272e-02 -6.31939650e-01
-8.49353313e-01 7.13417828e-01 7.79195666e-01 5.33532277e-02
1.19077063e+00 3.61931790e-03 6.63038969e-01 3.33892882e-01
3.66936326e-01 -9.13151681e-01 9.56257731e-02 4.30831760e-01
8.75877500e-01 -1.31981611e+00 1.62870228e-01 -6.41269565e-01
-5.88576198e-01 7.15456486e-01 6.39098048e-01 3.10825594e-02
3.48136991e-01 3.18108082e-01 -2.37654492e-01 -1.23552859e-01
-7.10117936e-01 -4.01016951e-01 4.51367319e-01 5.55854678e-01
7.88078383e-02 -2.03341290e-01 -3.10718894e-01 5.70971407e-02
3.09619922e-02 7.12028295e-02 8.59065354e-02 9.33529019e-01
-5.01805902e-01 -1.04723835e+00 -3.22518885e-01 8.94107103e-01
-3.02995294e-01 -2.21821412e-01 -3.39751840e-01 1.01370049e+00
3.41781601e-02 1.01099682e+00 -4.52391505e-02 -4.16087568e-01
4.58164155e-01 -1.07302785e-01 3.98583531e-01 -4.26048040e-01
-6.84511900e-01 8.42006058e-02 1.37484550e-01 -6.27188444e-01
-3.05975825e-01 -7.12432384e-01 -1.08786607e+00 -4.82686013e-01
-7.15055645e-01 -3.76339734e-01 2.80972838e-01 7.16357410e-01
6.75134182e-01 3.42482656e-01 3.05013835e-01 -8.53318334e-01
-9.61701632e-01 -9.69666958e-01 -4.20669436e-01 4.84786749e-01
5.36857367e-01 -8.98392916e-01 -3.92960668e-01 -2.70700660e-02] | [9.030004501342773, 0.3464679419994354] |
3a3cd82c-f136-43ff-8ae0-4f7829cf5d28 | causal-identification-with-matrix-equations | null | null | http://proceedings.neurips.cc/paper/2021/hash/4ea06fbc83cdd0a06020c35d50e1e89a-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/4ea06fbc83cdd0a06020c35d50e1e89a-Paper.pdf | Causal Identification with Matrix Equations | Causal effect identification is concerned with determining whether a causal effect is computable from a combination of qualitative assumptions about the underlying system (e.g., a causal graph) and distributions collected from this system. Many identification algorithms exclusively rely on graphical criteria made of a non-trivial combination of probability axioms, do-calculus, and refined c-factorization (e.g., Lee & Bareinboim, 2020). In a sequence of increasingly sophisticated results, it has been shown how proxy variables can be used to identify certain effects that would not be otherwise recoverable in challenging scenarios through solving matrix equations (e.g., Kuroki & Pearl, 2014; Miao et al., 2018). In this paper, we develop a new causal identification algorithm which utilizes both graphical criteria and matrix equations. Specifically, we first characterize the relationships between certain graphically-driven formulae and matrix multiplications. With such characterizations, we broaden the spectrum of proxy variable based identification conditions and further propose novel intermediary criteria based on the pseudoinverse of a matrix. Finally, we devise a causal effect identification algorithm, which accepts as input a collection of marginal, conditional, and interventional distributions, integrating enriched matrix-based criteria into a graphical identification approach. | ['Elias Bareinboim', 'Sanghack Lee'] | 2021-12-01 | null | https://openreview.net/forum?id=kt_s_ZbYvtP | https://openreview.net/pdf?id=kt_s_ZbYvtP | neurips-2021-12 | ['causal-identification'] | ['reasoning'] | [ 3.58213603e-01 1.06283017e-01 -4.40117985e-01 -4.19591088e-03
-2.68810034e-01 -9.95567739e-01 9.06958401e-01 1.88781351e-01
1.66560456e-01 8.42996478e-01 2.76185632e-01 -6.51218474e-01
-8.61653864e-01 -8.25995505e-01 -6.99818313e-01 -4.19873565e-01
-3.70174497e-01 8.76495987e-02 -3.41609925e-01 2.70673838e-02
2.07720086e-01 4.37471658e-01 -1.27708995e+00 -1.61455914e-01
1.01151526e+00 6.82729363e-01 -1.48768246e-01 5.21323621e-01
3.05327594e-01 8.24168503e-01 -1.58528581e-01 -2.98557401e-01
-2.50104889e-02 -6.64088547e-01 -7.55181074e-01 -4.14912730e-01
5.23123033e-02 -3.42459708e-01 -3.12438250e-01 1.04385424e+00
2.10204482e-01 -1.17893502e-01 1.00649750e+00 -1.47186446e+00
-5.79319775e-01 1.15573227e+00 -4.34329271e-01 -8.91656354e-02
6.08740866e-01 1.46948412e-01 1.26970363e+00 -9.20970738e-01
5.03845215e-01 1.47413635e+00 6.95640743e-01 7.21499603e-03
-1.76298153e+00 -6.94581568e-01 7.41861165e-02 8.65093321e-02
-1.32812071e+00 -9.72185954e-02 7.54579306e-01 -8.24679494e-01
1.85186610e-01 7.20586419e-01 6.85788751e-01 1.12025750e+00
9.56343710e-02 6.17076814e-01 1.17936099e+00 -4.45298523e-01
2.78523922e-01 -1.45916924e-01 2.08625257e-01 6.17635667e-01
5.60791433e-01 4.18338120e-01 -5.00651956e-01 -5.53904951e-01
8.84337842e-01 -2.39253506e-01 -2.40525097e-01 -5.25309622e-01
-1.16929352e+00 9.19401467e-01 1.59965456e-01 2.54385352e-01
-3.31641614e-01 3.19939375e-01 1.03305787e-01 1.05439402e-01
1.22588418e-01 5.55940270e-01 -2.14041099e-01 8.71217474e-02
-6.83014870e-01 2.93060988e-01 7.18371332e-01 6.93529010e-01
8.25881541e-01 -1.50168374e-01 -2.47600362e-01 3.29982519e-01
4.50624585e-01 7.32665956e-01 -1.73369035e-01 -8.41827691e-01
2.23529097e-02 6.95868313e-01 1.74709141e-01 -1.20876300e+00
-5.14579117e-01 -4.11237776e-01 -9.97771800e-01 -1.57584310e-01
6.22709513e-01 -2.54302710e-01 -4.50690329e-01 2.31213260e+00
4.54783976e-01 2.66539156e-01 -3.66895288e-01 7.80234694e-01
4.81554359e-01 2.44454682e-01 1.53025597e-01 -5.41276217e-01
9.33557987e-01 -1.11758346e-02 -7.45157123e-01 2.94565469e-01
3.87871057e-01 -3.63921106e-01 1.04302418e+00 2.77870357e-01
-8.36649597e-01 -2.45947108e-01 -8.52270126e-01 3.35301608e-01
-1.13331959e-01 1.62648752e-01 1.00624967e+00 9.30952311e-01
-8.92189980e-01 6.01936996e-01 -6.87796175e-01 -3.23022544e-01
1.37695652e-02 2.78593540e-01 -2.20507309e-01 2.17899188e-01
-1.54512882e+00 7.71880031e-01 1.00996919e-01 1.92086995e-01
-1.01879573e+00 -1.12122321e+00 -5.68638504e-01 3.23345289e-02
5.91901064e-01 -1.00752974e+00 9.26797092e-01 -7.95198679e-01
-1.34161830e+00 3.14563304e-01 1.69883683e-01 -2.18494982e-01
5.61643898e-01 -3.62557657e-02 -2.93250024e-01 6.41636103e-02
1.55494243e-01 3.41860503e-02 8.02677989e-01 -1.29628801e+00
-2.32994258e-01 -4.16111708e-01 4.68772978e-01 -1.19023830e-01
-1.89800516e-01 1.04414010e-02 -6.86133355e-02 -4.79957163e-01
4.77822088e-02 -7.67167091e-01 -2.70410568e-01 -2.30302095e-01
-1.07003272e+00 3.01120966e-03 2.65565276e-01 -6.13354206e-01
1.61409867e+00 -2.04321241e+00 3.62842411e-01 8.13049078e-01
5.99185109e-01 -3.91505808e-01 5.33620082e-02 7.26570666e-01
-5.52688956e-01 6.38055980e-01 -6.75086141e-01 2.85199910e-01
4.55801070e-01 -2.83487469e-01 -4.61769044e-01 7.57816076e-01
2.32150108e-01 8.62167835e-01 -9.83308077e-01 -3.56925070e-01
3.89834791e-01 3.19192499e-01 -6.44616008e-01 7.52410889e-02
-1.05189703e-01 6.43916368e-01 -4.51988637e-01 4.36384052e-01
4.67628896e-01 -2.15817943e-01 6.72597528e-01 -2.28329509e-01
-3.54979575e-01 1.79030046e-01 -1.56202030e+00 1.28079534e+00
-3.30577552e-01 2.60178685e-01 -1.24463446e-01 -1.08895946e+00
4.77424681e-01 1.94334209e-01 6.58524275e-01 -1.31764427e-01
3.22961211e-01 5.20793721e-02 1.81474954e-01 -2.91668624e-01
1.68768491e-03 -1.31492704e-01 -2.89506942e-01 5.75391293e-01
-1.44403502e-01 3.41910161e-02 4.01166320e-01 4.30866033e-01
1.28992486e+00 -2.63331980e-02 8.04930091e-01 -5.69634974e-01
4.32554960e-01 -1.21510699e-01 3.87857169e-01 9.92649555e-01
9.35165808e-02 2.29216114e-01 1.01838660e+00 8.21286216e-02
-7.16442585e-01 -1.47102726e+00 -5.28450549e-01 7.28660107e-01
-5.56418188e-02 -4.79076415e-01 -6.87669277e-01 -3.52458149e-01
3.49778652e-01 6.70100033e-01 -1.09640837e+00 -3.47830713e-01
-1.15923360e-01 -9.11335409e-01 7.26999521e-01 2.36845002e-01
1.39802834e-02 -4.15655017e-01 -3.64685267e-01 -7.80123239e-03
-1.04284771e-01 -4.76096690e-01 -1.96283773e-01 2.03802623e-02
-7.20237017e-01 -1.42049837e+00 -2.48230383e-01 9.13787633e-03
5.42142868e-01 -1.62049547e-01 9.82083678e-01 -1.66585490e-01
-2.59499729e-01 7.03394473e-01 -1.38931572e-01 -1.54726669e-01
-4.26996648e-01 -3.39073062e-01 4.10018206e-01 3.13708305e-01
-2.22046375e-01 -8.91711831e-01 -3.75995785e-01 2.19215542e-01
-8.54750574e-01 1.68557629e-01 8.20361137e-01 8.17826509e-01
5.10203183e-01 7.71641433e-02 5.97369790e-01 -8.27759624e-01
7.32431352e-01 -8.44815493e-01 -6.66467488e-01 4.04452085e-01
-6.52664244e-01 3.39086831e-01 6.31277978e-01 -4.72849518e-01
-1.05964470e+00 1.23645857e-01 3.13175112e-01 -2.52503991e-01
1.71219721e-01 1.18053412e+00 -6.07350707e-01 2.17925161e-01
6.59885049e-01 -6.30114824e-02 -2.50857919e-01 -3.87263000e-01
9.16394532e-01 3.04124117e-01 5.41891336e-01 -1.01101303e+00
1.05148363e+00 3.77216816e-01 3.29986393e-01 -3.96449655e-01
-5.85328639e-01 -1.14919931e-01 -7.89162934e-01 -5.01349866e-01
6.38671279e-01 -5.10163188e-01 -1.10617840e+00 7.95044452e-02
-9.11522508e-01 -3.43204945e-01 -6.44793138e-02 6.47662163e-01
-3.99143904e-01 2.93701887e-01 -2.59288132e-01 -1.13667738e+00
2.83859313e-01 -7.43528664e-01 6.13383234e-01 -1.95086926e-01
-5.81623733e-01 -1.10197055e+00 3.66438031e-01 -2.79374421e-01
-8.27923417e-02 5.74708998e-01 1.22203028e+00 -2.46738464e-01
-5.29287875e-01 3.09488140e-02 -3.57460380e-01 -1.45231828e-01
1.84607685e-01 5.01198471e-01 -7.48407304e-01 1.06120398e-02
-1.47967204e-01 1.94998890e-01 6.14231288e-01 7.16464162e-01
9.61745024e-01 -4.56953704e-01 -3.94736886e-01 5.00605524e-01
1.35823667e+00 9.26508680e-02 3.18616092e-01 5.73139125e-03
8.08328688e-01 8.10731292e-01 3.30959201e-01 6.68062627e-01
2.80166954e-01 5.68740249e-01 3.53134871e-01 3.56155224e-02
1.86831251e-01 -6.12281263e-01 2.70246804e-01 6.14999652e-01
-1.81156635e-01 2.09667184e-03 -1.03576684e+00 3.37524354e-01
-1.86271238e+00 -9.58436370e-01 -7.10331380e-01 2.37506008e+00
1.12005961e+00 -1.46862358e-01 3.31355304e-01 8.56241286e-02
9.42646563e-01 -3.22451562e-01 -6.02542758e-01 1.28789902e-01
-2.21381277e-01 2.14147463e-01 6.39500558e-01 6.14157140e-01
-9.01364505e-01 3.84359956e-01 6.51917696e+00 6.31457925e-01
-7.25553453e-01 -5.68936281e-02 4.16700453e-01 2.24071220e-01
-9.25315917e-01 4.80991036e-01 -1.76592365e-01 4.47262138e-01
8.58598709e-01 -6.92416489e-01 5.80978513e-01 4.32163358e-01
5.02958775e-01 -1.51601717e-01 -1.46198666e+00 8.06767941e-01
-2.68060207e-01 -1.09501505e+00 -3.38208117e-02 2.02215597e-01
7.52332985e-01 -6.20089710e-01 2.81657457e-01 -4.93685864e-02
7.89518833e-01 -1.15894032e+00 9.20792818e-01 7.53955007e-01
1.02202809e+00 -5.03819227e-01 2.14080721e-01 7.90336207e-02
-1.02680588e+00 -1.82057962e-01 7.78926164e-02 -3.93918961e-01
4.88270298e-02 1.14228439e+00 -4.91082072e-01 1.00498652e+00
3.26877028e-01 7.84532785e-01 -3.85704875e-01 8.14033091e-01
-5.41614354e-01 1.01398420e+00 -4.51806694e-01 1.04511485e-01
-3.22489768e-01 -4.09127146e-01 6.23205960e-01 9.71437335e-01
3.34062517e-01 1.01546310e-01 -2.60116994e-01 1.45739317e+00
1.60889059e-01 1.53460965e-01 -6.29404604e-01 -2.27263331e-01
5.82905293e-01 1.09492123e+00 -5.96379936e-01 -9.84389558e-02
-3.13804328e-01 2.38255426e-01 -1.44088990e-03 4.75804329e-01
-9.67422783e-01 1.89016890e-02 6.91146255e-01 -6.98641166e-02
-2.51554310e-01 -7.80938938e-02 -7.36503541e-01 -1.13502789e+00
-2.09617212e-01 -8.69821608e-01 3.61064136e-01 -5.66502452e-01
-1.31594610e+00 -2.38373041e-01 4.31814611e-01 -1.10796547e+00
-4.38379437e-01 -4.88721073e-01 -5.29247105e-01 9.32453096e-01
-7.70702183e-01 -9.18305993e-01 6.88979849e-02 6.50362968e-01
-4.34297949e-01 3.21533144e-01 6.99691415e-01 3.40758383e-01
-1.05778909e+00 3.59559208e-01 1.61536738e-01 -7.24672377e-02
5.32863915e-01 -1.55211735e+00 5.06777130e-02 1.08559561e+00
-8.01391229e-02 1.04494083e+00 8.60642552e-01 -9.20036256e-01
-1.87129080e+00 -8.08665812e-01 5.22664368e-01 -7.02557743e-01
1.28607249e+00 -5.38985312e-01 -5.02685308e-01 8.63898814e-01
-2.43060485e-01 -4.96432811e-01 5.36755323e-01 6.26664102e-01
-4.31904048e-01 -3.40006384e-03 -6.80848360e-01 8.03661764e-01
1.27695823e+00 -6.59535229e-01 -2.82621294e-01 3.20031084e-02
4.56504136e-01 -3.23976874e-02 -1.00542247e+00 5.27844787e-01
7.70889163e-01 -8.33646297e-01 1.07901502e+00 -6.66706026e-01
4.55990881e-01 -4.97974813e-01 -1.11427277e-01 -1.24119186e+00
-4.58023489e-01 -7.13625610e-01 -4.85532470e-02 1.34792686e+00
4.88695234e-01 -6.08207047e-01 3.36702652e-02 6.56927586e-01
2.15157736e-02 -2.43355274e-01 -6.71215594e-01 -7.82343805e-01
7.09060282e-02 -7.94626117e-01 5.97679913e-01 1.24908566e+00
4.09605771e-01 4.88266677e-01 -4.71704125e-01 3.93531919e-01
6.98179841e-01 1.36723012e-01 7.57690430e-01 -1.44030917e+00
-4.23889637e-01 -7.14883506e-01 -2.02993855e-01 -6.44909263e-01
1.47915721e-01 -1.02370274e+00 -2.04822794e-01 -1.37749910e+00
6.18717313e-01 -5.60330451e-01 -3.46660256e-01 4.75391179e-01
-2.96454191e-01 -1.70747533e-01 1.30311742e-01 2.18820378e-01
-1.35127038e-01 4.24946457e-01 9.91978824e-01 -8.81861150e-02
-2.19962269e-01 -1.79575786e-01 -8.99578810e-01 6.36552632e-01
5.91394663e-01 -4.40469772e-01 -5.85779846e-01 1.58955038e-01
9.02683437e-01 4.12144721e-01 7.93312311e-01 -4.75737572e-01
7.57476315e-02 -4.90656167e-01 8.14203471e-02 -3.21038216e-01
-2.66183496e-01 -4.60960150e-01 7.14278936e-01 4.07508254e-01
-4.14764881e-01 -1.92900926e-01 1.59342214e-01 6.57299280e-01
4.18710895e-03 -9.40652192e-02 2.08521187e-01 3.71122181e-01
-3.34081233e-01 2.06809379e-02 -4.10309851e-01 -4.45765108e-02
6.55626774e-01 1.84161678e-01 -2.69896150e-01 -5.19778490e-01
-7.28495419e-01 5.68233356e-02 1.91007599e-01 -4.00934443e-02
1.90896884e-01 -1.43657422e+00 -8.16899955e-01 -1.91841692e-01
-1.27017051e-01 -5.19000471e-01 3.54411751e-01 1.41556919e+00
-3.52960415e-02 4.88931566e-01 4.28334512e-02 -4.30742979e-01
-8.60476196e-01 6.78060114e-01 1.00074381e-01 -1.92126453e-01
-1.42977268e-01 4.46145982e-01 7.92508721e-01 -3.76492649e-01
-2.16310188e-01 -5.11915445e-01 -5.76653555e-02 1.57063976e-01
2.97097504e-01 4.55201507e-01 -4.03581202e-01 -2.68900484e-01
-5.44116795e-01 2.90718108e-01 7.44289339e-01 -3.85242403e-01
1.02985871e+00 -1.94379196e-01 -4.40796793e-01 7.67113328e-01
9.14461017e-01 3.57582092e-01 -1.16701579e+00 4.10990119e-02
3.69728841e-02 -2.94553459e-01 -5.07587902e-02 -1.01197624e+00
-7.06628799e-01 6.02629960e-01 3.44023049e-01 6.14179134e-01
1.22929513e+00 4.96128723e-02 -2.49894485e-01 1.40734479e-01
1.37689456e-01 -6.54243588e-01 -3.63672465e-01 1.02040991e-01
1.06601989e+00 -8.21202695e-01 5.81733882e-02 -6.48024201e-01
-7.58369789e-02 9.49530125e-01 5.95732294e-02 1.70441940e-01
6.84714139e-01 2.05845028e-01 -5.85729122e-01 -1.41972035e-01
-5.66162288e-01 -3.94570231e-01 4.40031260e-01 4.17430103e-01
3.63529593e-01 6.70324564e-01 -6.80533111e-01 1.08217812e+00
-4.22758937e-01 -2.27497995e-01 5.30331671e-01 2.85525113e-01
9.17598419e-03 -9.58199441e-01 -6.49430990e-01 5.25883496e-01
-6.10314384e-02 -3.91061425e-01 -7.02138126e-01 1.07592463e+00
-7.30043650e-03 1.06237209e+00 -9.46275592e-02 -4.69434470e-01
2.80541211e-01 -8.27228799e-02 4.78066027e-01 -3.97489101e-01
-1.66317567e-01 3.56770605e-02 1.88775212e-01 -4.45709646e-01
-4.87588465e-01 -1.16473269e+00 -9.44221318e-01 -5.60002089e-01
-3.55214328e-01 -4.04950976e-02 3.82445693e-01 1.06386876e+00
6.72817752e-02 5.96353054e-01 6.02664471e-01 -2.59654492e-01
-3.87377888e-01 -9.62515175e-01 -7.86387920e-01 1.53088450e-01
3.39600183e-02 -1.05461705e+00 -5.42875648e-01 1.24516942e-01] | [7.858068466186523, 5.325561046600342] |
f7e767aa-67aa-4a68-8b26-ee5bc64fc2d3 | bayesian-inference-for-neighborhood-filters | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Huang_Bayesian_Inference_for_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Huang_Bayesian_Inference_for_2015_CVPR_paper.pdf | Bayesian Inference for Neighborhood Filters With Application in Denoising | Range-weighted neighborhood filters are useful and popular for their edge-preserving property and simplicity, but they are originally proposed as intuitive tools. Previous works needed to connect them to other tools or models for indirect property reasoning or parameter estimation. In this paper, we introduce a unified empirical Bayesian framework to do both directly. A neighborhood noise model is proposed to reason and infer the Yaroslavsky, bilateral, and modified non-local means filters. An EM+ algorithm is devised to estimate the essential parameter, range variance, via the model fitting to empirical distributions. Finally, we apply this framework to color-image denoising. Experimental results show that the proposed model fits noisy images well and the range variance is estimated successfully. The image quality can also be improved by a proposed recursive fitting and filtering scheme. | ['Chao-Tsung Huang'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['color-image-denoising'] | ['computer-vision'] | [-1.99241377e-02 -2.45007128e-01 7.03671724e-02 -4.29283619e-01
-4.50862408e-01 -1.29669383e-01 4.61323351e-01 -3.58780324e-01
-3.67912591e-01 6.96189642e-01 1.29830986e-01 2.67558359e-02
-5.99636078e-01 -8.20801079e-01 -3.93482029e-01 -1.04617548e+00
2.88437843e-01 -4.43964005e-02 4.48598862e-01 3.22877839e-02
6.34964824e-01 5.45150280e-01 -1.67566073e+00 -1.02520011e-01
9.84483719e-01 1.08859491e+00 3.82628798e-01 5.93763709e-01
-1.11769773e-01 5.50598025e-01 -2.06021592e-01 -3.52612257e-01
1.22743949e-01 -4.80663866e-01 -4.17049974e-01 2.69687265e-01
1.86662048e-01 -4.03612554e-01 -2.91528404e-01 1.73589861e+00
4.49810594e-01 3.90042156e-01 8.65826249e-01 -1.09510565e+00
-7.95769811e-01 5.47156513e-01 -8.49545121e-01 -1.98175102e-01
1.78533137e-01 -1.58280462e-01 6.07850671e-01 -7.73292661e-01
5.01701951e-01 1.30982912e+00 7.48163581e-01 3.24464321e-01
-1.31840301e+00 -4.13426787e-01 -3.23753767e-02 6.22621775e-01
-1.55855477e+00 -1.44031614e-01 9.80301082e-01 -1.09628975e-01
3.50881279e-01 4.11323845e-01 6.19007587e-01 7.35545993e-01
2.23357633e-01 6.08090699e-01 1.50099397e+00 -5.42032957e-01
1.07579067e-01 2.74755538e-01 2.92467117e-01 6.28534436e-01
5.06738663e-01 8.30054730e-02 -4.42555606e-01 -1.43681735e-01
1.04802549e+00 -5.37738316e-02 -5.52729487e-01 -2.58819491e-01
-8.36097360e-01 5.53454041e-01 1.96747661e-01 3.83304536e-01
-3.16913933e-01 3.04240078e-01 -6.31685853e-02 2.93241888e-02
5.05905151e-01 -1.09549448e-01 -3.05403799e-01 6.11011200e-02
-1.01125085e+00 6.93939328e-02 6.03137612e-01 9.40969467e-01
7.77753532e-01 -9.03152674e-02 -1.80570796e-01 7.89326072e-01
7.82251716e-01 8.97444367e-01 -7.51442015e-02 -1.34372842e+00
-3.30500782e-01 1.43672794e-01 3.59402865e-01 -1.29449093e+00
-3.03115666e-01 -3.20173919e-01 -1.01541889e+00 5.82657695e-01
4.91816819e-01 2.67767698e-01 -8.24828029e-01 1.38224804e+00
3.79319519e-01 6.69027328e-01 -3.98764402e-01 8.74487519e-01
7.15225637e-01 5.56738794e-01 -1.86841995e-01 -4.93420124e-01
1.41056979e+00 -4.99199539e-01 -1.37979829e+00 2.98257947e-01
-1.23205304e-01 -1.05394578e+00 9.15114105e-01 9.54298854e-01
-1.19771922e+00 -6.97842419e-01 -7.61523843e-01 -3.69592384e-02
-2.04892725e-01 2.98426121e-01 3.76060396e-01 8.07781219e-01
-8.26177597e-01 8.57939541e-01 -8.19688976e-01 -2.35066697e-01
3.63752276e-01 -1.28906488e-01 -6.30915910e-02 2.73077488e-02
-8.91742051e-01 9.66925621e-01 3.18306237e-01 4.87047583e-01
-3.65985364e-01 -4.20547307e-01 -5.25991261e-01 6.55534640e-02
2.02121928e-01 -7.56769657e-01 5.96152127e-01 -4.39351618e-01
-1.70146871e+00 5.64718425e-01 -2.35694021e-01 -1.91374019e-01
5.94779670e-01 -4.58621562e-01 -4.33504283e-01 3.19683701e-01
-1.40538737e-01 2.29321525e-01 1.19942439e+00 -1.55817914e+00
-5.38092792e-01 -2.82410324e-01 -5.42950451e-01 -2.88773894e-01
-2.13929847e-01 -1.58478603e-01 -5.30461431e-01 -8.00859749e-01
7.59112239e-01 -3.79093558e-01 -2.04149812e-01 3.94782931e-01
-4.34986204e-01 1.63080432e-02 6.94576025e-01 -8.86555016e-01
1.30619109e+00 -2.31601572e+00 -9.44134295e-02 8.09413552e-01
8.58871862e-02 -3.22039239e-02 8.75507593e-02 1.69576243e-01
1.34744734e-01 -3.25931534e-02 -4.61614460e-01 -1.10702157e-01
9.43469107e-02 2.07115844e-01 -2.80667901e-01 8.62829804e-01
-2.66742349e-01 4.25067186e-01 -6.02717996e-01 -5.83152890e-01
5.75521767e-01 9.92102444e-01 -1.48858845e-01 -1.77376028e-02
1.80725619e-01 2.73070067e-01 -2.72838503e-01 4.85395581e-01
1.52264845e+00 1.48252442e-01 -8.92001837e-02 -8.49652112e-01
-3.00390134e-03 -4.21347201e-01 -1.85127330e+00 1.49750543e+00
-9.95922089e-02 5.19585967e-01 4.02157784e-01 -8.83944035e-01
1.21048093e+00 1.49638774e-02 2.11536810e-01 -4.13719654e-01
4.07423675e-01 1.36832803e-01 -3.88597608e-01 -8.06045413e-01
3.00519645e-01 -1.95091814e-01 6.17119491e-01 1.07657172e-01
-1.36393279e-01 -4.08409894e-01 5.02756760e-02 -3.19030061e-02
4.91053969e-01 5.51069081e-01 3.55517596e-01 -6.49744987e-01
9.58222747e-01 -2.25514263e-01 7.62974739e-01 9.41510379e-01
-5.87781556e-02 6.62891209e-01 1.41944155e-01 -1.48455128e-01
-8.95720661e-01 -1.32734895e+00 -5.48892856e-01 4.40248668e-01
6.20061457e-01 -2.06348464e-01 -7.73198307e-01 -3.02375227e-01
-1.40914664e-01 9.09072101e-01 -4.54646289e-01 -1.94181912e-02
-2.10320383e-01 -7.38563955e-01 1.92126393e-01 4.16727871e-01
9.14505959e-01 -5.70063233e-01 -2.35236734e-01 9.28742960e-02
-2.41025537e-01 -9.16261554e-01 -1.62090376e-01 -2.55578816e-01
-1.06994581e+00 -1.20458901e+00 -5.99376798e-01 -5.94427466e-01
7.59524047e-01 2.43791148e-01 7.89787769e-01 2.42816806e-01
-2.52782464e-01 6.16359115e-01 -2.50915945e-01 -9.86094326e-02
-1.72177404e-01 -8.57955337e-01 -6.27559889e-03 2.79272020e-01
3.26935619e-01 -6.80254281e-01 -6.51560485e-01 6.31826758e-01
-7.91926682e-01 -2.38713905e-01 5.43501973e-01 9.09164548e-01
8.57418418e-01 6.64887428e-01 2.14310840e-01 -7.48038948e-01
7.26803958e-01 1.15702316e-01 -8.44516456e-01 3.53668630e-01
-7.68971264e-01 -9.84373130e-03 2.82891333e-01 -5.39220214e-01
-1.80304468e+00 -1.18067540e-01 -1.05093166e-01 -2.77276337e-01
-4.63941962e-01 3.09655428e-01 -4.13670599e-01 -2.31486708e-01
5.06594837e-01 3.41585219e-01 1.20494984e-01 -7.23508060e-01
6.10433459e-01 5.32185376e-01 7.60304153e-01 -7.14612961e-01
1.01399517e+00 9.91529465e-01 3.02753687e-01 -9.30198312e-01
-4.65222359e-01 -4.29174036e-01 -3.80910873e-01 -3.94015402e-01
8.44088197e-01 -5.73378980e-01 -1.09691954e+00 6.78620994e-01
-1.07241511e+00 -4.04929817e-02 -1.99254304e-01 8.16626668e-01
-6.15703464e-01 8.14476609e-01 -7.28425086e-01 -1.20402980e+00
-7.56165311e-02 -8.43580723e-01 6.58022881e-01 4.50609207e-01
1.69613734e-01 -1.25462234e+00 -2.60181963e-01 1.79127842e-01
5.30127227e-01 -2.61015110e-02 5.56536555e-01 -4.89991307e-02
-5.94629943e-01 -4.35835421e-02 -5.59307218e-01 6.30880594e-01
6.23478517e-02 3.10026169e-01 -1.05306220e+00 1.59750879e-02
5.55408001e-01 4.56391424e-01 9.87864196e-01 9.41106200e-01
1.37369812e+00 1.14355326e-01 -3.24975729e-01 7.59873152e-01
1.63124430e+00 -9.30678565e-03 1.13046873e+00 2.49979660e-01
2.80491769e-01 4.35980409e-01 7.46696293e-01 5.66470206e-01
-9.57522020e-02 2.32829660e-01 3.65192592e-01 2.26602051e-02
-3.38854551e-01 7.66341668e-03 3.41208391e-02 7.81071424e-01
-3.99205923e-01 -6.02249913e-02 -3.34000200e-01 3.02670419e-01
-2.03574181e+00 -1.10604036e+00 -8.16931367e-01 2.12200546e+00
5.64303517e-01 1.89198941e-01 -4.22340870e-01 1.53194845e-01
9.53406513e-01 -1.90950364e-01 -1.62638739e-01 -4.00916021e-03
-3.82308602e-01 -1.58328414e-02 5.11038005e-01 7.93031693e-01
-9.24272180e-01 6.12329423e-01 6.75210285e+00 1.40456414e+00
-5.37900031e-01 4.12977226e-02 4.42220598e-01 5.28386712e-01
-5.25549293e-01 2.69889981e-01 -6.30203128e-01 3.73295397e-01
2.11757034e-01 2.44306475e-01 4.95347351e-01 5.85883975e-01
5.48764408e-01 -6.40296638e-01 -4.92709816e-01 1.29054773e+00
-4.55557406e-02 -1.13571489e+00 -1.02545619e-01 -4.64441888e-02
6.83552980e-01 -7.71168768e-01 -1.77528709e-02 -4.24634069e-01
6.92590028e-02 -6.60843730e-01 6.61939979e-01 1.46078575e+00
2.19211608e-01 -7.20312536e-01 9.59224224e-01 2.85720825e-01
-1.01229584e+00 1.54366251e-02 -7.33704209e-01 -2.22406432e-01
3.23893309e-01 1.20128036e+00 -1.11221872e-01 5.81649005e-01
9.45539296e-01 7.49266505e-01 -3.95334512e-01 1.47105002e+00
-5.75633705e-01 5.80400884e-01 -4.71130520e-01 -1.15390137e-01
-2.71187454e-01 -1.00168252e+00 7.29719758e-01 1.01296020e+00
5.01311600e-01 5.20921052e-02 -3.32168102e-01 1.29057765e+00
4.08742338e-01 4.64031175e-02 -2.88595498e-01 5.05910516e-01
4.87827837e-01 1.39023471e+00 -1.14944458e+00 -3.63831341e-01
-4.26453680e-01 8.25295806e-01 -3.63808870e-01 7.29375720e-01
-8.90623868e-01 -5.43079734e-01 4.83667314e-01 -1.07708812e-01
3.54168624e-01 -1.70490593e-01 -7.14088082e-01 -1.12363648e+00
-2.36506671e-01 -5.53555608e-01 1.99342072e-01 -1.10964561e+00
-1.67869985e+00 1.30297527e-01 2.52986580e-01 -1.27142680e+00
4.16541189e-01 -8.16991627e-01 -6.55507445e-01 7.96378493e-01
-1.36444271e+00 -1.04271317e+00 -3.48693520e-01 6.83274269e-01
7.39303082e-02 5.01246341e-02 4.17828500e-01 2.30724484e-01
-3.28608334e-01 1.79410011e-01 4.06367481e-01 -4.90346574e-04
7.00400174e-01 -1.11760485e+00 -3.97751391e-01 1.08360922e+00
7.65099525e-02 8.31947327e-01 1.24453568e+00 -6.43542171e-01
-1.18283665e+00 -3.78922433e-01 3.64183336e-01 -1.53484896e-01
6.38214231e-01 2.51521710e-02 -9.72767711e-01 3.60570133e-01
2.76249170e-01 9.91493016e-02 4.11667436e-01 -1.28459290e-01
-1.95441648e-01 -3.10439080e-01 -1.46285748e+00 5.93338132e-01
8.54043186e-01 -3.93722236e-01 -6.69440627e-01 -4.88799401e-02
2.27725118e-01 2.87439814e-03 -9.46895957e-01 3.95010650e-01
6.43204868e-01 -1.28854108e+00 1.19003904e+00 4.22456339e-02
-2.16886431e-01 -8.34672749e-01 -3.84818494e-01 -9.17939782e-01
-5.59496284e-01 -6.43890262e-01 1.61145534e-02 1.58882117e+00
1.11827493e-01 -7.94004560e-01 4.79029655e-01 4.56219286e-01
1.68927923e-01 -2.49720976e-01 -9.36403394e-01 -7.84455776e-01
-3.35578769e-01 -7.94116497e-01 3.96216631e-01 7.46872962e-01
-3.18012387e-01 -8.43453258e-02 -5.47862053e-01 4.29526418e-01
1.33982682e+00 -6.04182947e-03 5.95453143e-01 -1.44345200e+00
-1.95190489e-01 -6.71509981e-01 -3.41990680e-01 -1.18158340e+00
-9.97667983e-02 -2.56983399e-01 1.55942038e-01 -1.68498766e+00
1.78981379e-01 -2.33617537e-02 -3.38289350e-01 -1.02421701e-01
-1.85907856e-01 3.48015726e-01 -1.45320967e-01 -9.91401821e-02
-2.11283728e-01 5.28221726e-01 1.49772561e+00 -3.49960811e-02
-9.26500112e-02 1.54665306e-01 -4.15576369e-01 1.15717542e+00
6.88886285e-01 -4.36771005e-01 -4.07896340e-01 -8.22046027e-02
3.76238823e-01 -1.87653124e-01 6.05337262e-01 -8.90884995e-01
3.72034788e-01 -1.18197918e-01 5.93829513e-01 -7.72158444e-01
3.24612051e-01 -1.24925113e+00 2.54473478e-01 2.37546638e-01
4.24856655e-02 -5.76709211e-01 -3.99731696e-02 8.73412907e-01
-2.13756353e-01 -6.19681835e-01 9.95199800e-01 2.32361443e-02
-7.80967355e-01 -1.33194968e-01 -4.18068111e-01 -4.00581181e-01
7.44105935e-01 -5.60218215e-01 -1.35479063e-01 -6.37391806e-01
-1.11212993e+00 -2.16069505e-01 2.31332973e-01 -2.31942326e-01
9.76860225e-01 -1.36812663e+00 -5.52960694e-01 2.59609252e-01
-3.40246081e-01 -4.23452288e-01 5.79218209e-01 1.25973964e+00
-6.20680928e-01 -4.13009733e-01 -4.27037142e-02 -6.43533170e-01
-1.14532197e+00 5.81323743e-01 1.81450009e-01 3.27855527e-01
-6.77347064e-01 7.66261816e-01 -1.54295772e-01 -2.49159649e-01
2.50344872e-01 -4.41661775e-01 -1.57345027e-01 -1.23056844e-01
6.33872986e-01 1.05798888e+00 -4.73756701e-01 -5.00941098e-01
-1.60904810e-01 1.18880916e+00 4.36065853e-01 -1.62154362e-01
1.28744280e+00 -6.82349026e-01 -6.50244534e-01 2.54753649e-01
8.48288536e-01 2.67267317e-01 -1.25908256e+00 -1.84915870e-01
-7.99510777e-02 -8.56556535e-01 4.63593781e-01 -5.54242432e-01
-1.04237235e+00 8.60858142e-01 9.36360955e-01 4.36734587e-01
1.46718335e+00 -1.93832204e-01 4.48532194e-01 2.14094475e-01
1.07843876e-01 -1.31379294e+00 -1.98494360e-01 1.25310883e-01
6.56296849e-01 -1.05464530e+00 3.45498174e-01 -8.56653214e-01
-1.83915317e-01 1.47598994e+00 2.81792670e-01 -2.12821215e-01
1.16114426e+00 3.78869236e-01 -2.85167731e-02 1.68319046e-03
-1.18499734e-01 -3.95537943e-01 4.65026915e-01 8.45264733e-01
1.57508641e-01 -1.37878060e-01 -6.48790300e-01 5.92459977e-01
2.56978989e-01 1.30446583e-01 4.14702982e-01 4.87199634e-01
-9.11112666e-01 -9.45824504e-01 -9.20925617e-01 2.05358550e-01
-2.84670830e-01 6.21196255e-02 3.18035603e-01 6.75784409e-01
2.83807099e-01 1.22792685e+00 -4.90129329e-02 -1.71850145e-01
1.76363140e-01 -2.57489920e-01 6.30744636e-01 2.54075140e-01
1.43110424e-01 8.80456924e-01 -1.43162459e-01 -5.84168553e-01
-8.26820195e-01 -4.69198227e-01 -1.18909228e+00 -2.90706694e-01
-6.70569301e-01 1.80536225e-01 6.80656254e-01 8.06645751e-01
-3.77019867e-02 3.30187321e-01 3.28313529e-01 -5.91159165e-01
-4.62928772e-01 -9.09015775e-01 -1.14032197e+00 3.95356178e-01
-4.92633097e-02 -6.92481995e-01 -6.96888208e-01 9.75336209e-02] | [11.319449424743652, -2.514735221862793] |
eb6c7b08-d26f-49fb-8da4-48fad30f78e6 | collateral-facilitation-in-humans-and | 2211.05198 | null | https://arxiv.org/abs/2211.05198v1 | https://arxiv.org/pdf/2211.05198v1.pdf | Collateral facilitation in humans and language models | Are the predictions of humans and language models affected by similar things? Research suggests that while comprehending language, humans make predictions about upcoming words, with more predictable words being processed more easily. However, evidence also shows that humans display a similar processing advantage for highly anomalous words when these words are semantically related to the preceding context or to the most probable continuation. Using stimuli from 3 psycholinguistic experiments, we find that this is also almost always also the case for 8 contemporary transformer language models (BERT, ALBERT, RoBERTa, XLM-R, GPT-2, GPT-Neo, GPT-J, and XGLM). We then discuss the implications of this phenomenon for our understanding of both human language comprehension and the predictions made by language models. | ['Benjamin K. Bergen', 'James A. Michaelov'] | 2022-11-09 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [ 2.38219760e-02 3.27916205e-01 5.24682887e-02 -6.46042228e-01
6.93961456e-02 -4.79512274e-01 7.32120931e-01 6.64501786e-01
-6.01059973e-01 3.48492414e-01 5.82212329e-01 -8.66741836e-01
1.99492469e-01 -7.35945165e-01 -2.46528521e-01 1.76214613e-02
3.69762369e-02 4.19460356e-01 3.90552223e-01 -5.02096474e-01
9.21576262e-01 3.91195506e-01 -1.22740841e+00 3.13062310e-01
6.15163684e-01 3.52695644e-01 3.99057299e-01 3.52417439e-01
-2.19864845e-01 7.34448433e-01 -2.34522983e-01 -6.72943413e-01
-1.10436864e-01 -6.08662784e-01 -8.11278284e-01 -1.68218315e-01
2.24386640e-02 -4.27918918e-02 4.31208424e-02 8.83011520e-01
3.67369652e-02 3.82539123e-01 7.10405231e-01 -6.62876785e-01
-8.34722638e-01 1.08897686e+00 -4.00917679e-01 5.85275292e-01
8.99173081e-01 3.98372442e-01 9.46647465e-01 -8.63301218e-01
5.42934716e-01 1.83834815e+00 5.57104230e-01 5.59725761e-01
-1.34884369e+00 -4.71801966e-01 6.09474301e-01 1.44176677e-01
-1.41423750e+00 -6.10632896e-01 2.90816575e-01 -3.72251838e-01
1.54325175e+00 -1.54277179e-02 7.96509981e-01 1.14109194e+00
1.18327510e+00 1.78630188e-01 1.42282546e+00 -6.49549901e-01
-5.27318195e-02 2.85467684e-01 4.75663304e-01 3.43396932e-01
4.20101941e-01 6.84701920e-01 -9.19725358e-01 -2.56305605e-01
1.83567137e-01 -4.76081401e-01 -2.63988107e-01 5.90651214e-01
-1.27556407e+00 7.04367042e-01 2.31600665e-02 8.78356516e-01
-6.98049188e-01 -1.39444575e-01 2.76633888e-01 5.08883178e-01
4.01570052e-01 6.21923804e-01 -6.95370018e-01 7.44725950e-03
-4.99700695e-01 3.79284173e-01 8.97403836e-01 6.42176032e-01
4.80037063e-01 2.50242472e-01 5.56043610e-02 4.90277767e-01
8.43082845e-01 2.59552330e-01 9.34743822e-01 -5.60541272e-01
1.67502686e-01 7.26884827e-02 1.60148650e-01 -1.22987998e+00
-9.52304423e-01 -8.10480639e-02 4.47946824e-02 -2.91421469e-02
2.36267805e-01 1.59990287e-03 -5.21758795e-01 2.11999130e+00
-1.02418408e-01 -5.24448514e-01 3.01636577e-01 4.23158586e-01
2.08865047e-01 9.75414455e-01 1.16416967e+00 -8.26768935e-01
1.21651912e+00 -3.40344816e-01 -6.20831668e-01 -1.07029057e+00
1.02023172e+00 -8.98563087e-01 1.26544929e+00 2.29343563e-01
-1.05669439e+00 -7.94225872e-01 -9.09859300e-01 -2.88582921e-01
-1.38043627e-01 -6.34161830e-01 7.67399728e-01 6.25649393e-01
-1.15987277e+00 3.92787606e-01 -5.15479863e-01 -7.79574096e-01
-3.33391726e-01 -1.54809102e-01 -1.67545155e-01 2.53079623e-01
-1.33983624e+00 1.64392304e+00 1.13387048e+00 -4.33694571e-03
-5.98964512e-01 -1.89759076e-01 -8.17390263e-01 3.55886705e-02
-3.95915173e-02 -4.62828517e-01 1.38737190e+00 -1.23994589e+00
-1.20268989e+00 1.32073510e+00 -7.85248876e-01 -5.55164874e-01
-1.30844682e-01 -2.71313012e-01 -1.01231742e+00 -1.53162464e-01
2.02552289e-01 6.47393525e-01 7.46306002e-01 -8.99773598e-01
-3.26223135e-01 -5.28629124e-01 -3.47100079e-01 3.62228066e-01
3.37552249e-01 5.87237418e-01 6.90550685e-01 -7.47786462e-01
5.57053864e-01 -5.91166258e-01 -7.30177388e-02 -1.87375367e-01
-4.50583488e-01 -8.41694176e-01 -1.92603599e-02 -8.15000534e-01
1.26396823e+00 -2.28453565e+00 -4.47050303e-01 3.17762315e-01
-2.90081557e-02 -1.22614607e-01 -1.29250422e-01 7.37536609e-01
-5.56498885e-01 6.20128751e-01 2.01656520e-02 3.06605041e-01
7.42570534e-02 1.67169288e-01 -6.78439915e-01 3.29049438e-01
-4.08842079e-02 9.26836431e-01 -5.01326323e-01 -4.11389977e-01
-5.10254875e-02 -5.30231930e-02 -4.72853452e-01 -8.91670138e-02
-4.10686046e-01 1.45965621e-01 -2.67222613e-01 1.69538140e-01
3.36402893e-01 -2.40205079e-01 4.58913743e-01 4.66807485e-01
-4.12223428e-01 1.02710700e+00 -4.68455732e-01 9.03416753e-01
-2.18434215e-01 5.61198235e-01 -5.43103218e-01 -6.36715770e-01
8.99889946e-01 5.91163814e-01 -7.67388880e-01 -8.29853475e-01
1.58090711e-01 2.71679968e-01 9.24031556e-01 -5.47759533e-01
3.74461323e-01 -1.08179617e+00 -1.62454113e-01 8.20472836e-01
-3.68934184e-01 -3.08602363e-01 1.63591534e-01 1.73983306e-01
2.80444503e-01 -4.27462067e-03 1.09674299e+00 -5.53881824e-01
6.28975987e-01 -1.27628431e-01 7.95216084e-01 1.04763770e+00
-3.08211625e-01 -8.57574791e-02 5.57451189e-01 -5.10881543e-01
-9.58818316e-01 -1.24854660e+00 -2.96909869e-01 1.43108010e+00
4.56949398e-02 -4.13504481e-01 -5.73533297e-01 9.02327802e-03
-3.05639952e-01 2.48542976e+00 -4.35807019e-01 -5.95585525e-01
-5.34317076e-01 -4.86267328e-01 2.27481693e-01 5.52920520e-01
-1.46804070e-02 -1.44185340e+00 -6.66542947e-01 4.55860972e-01
-2.27619838e-02 -9.68735695e-01 -2.23059818e-01 -3.59624475e-02
-1.03147709e+00 -3.88547957e-01 4.32324141e-01 -7.19239056e-01
3.87112200e-01 -2.97617912e-02 1.19944048e+00 6.98115051e-01
2.71629453e-01 3.32326561e-01 -3.70227933e-01 -7.86915481e-01
-9.89559829e-01 -4.83255148e-01 2.79213518e-01 -6.53669417e-01
8.54067981e-01 -4.26700801e-01 1.63720623e-02 5.91099560e-02
-6.74891770e-01 2.08012953e-01 3.30220938e-01 3.65392864e-01
1.76343381e-01 -2.31165841e-01 5.68984568e-01 -8.82508516e-01
9.48054194e-01 -6.16524756e-01 -1.62385955e-01 3.25563997e-01
-8.26282978e-01 -1.27287522e-01 7.46479928e-01 -4.22846675e-01
-1.55056310e+00 -7.11802363e-01 -4.16638106e-01 6.56608820e-01
-6.79789007e-01 5.95935464e-01 6.91883042e-02 3.82354051e-01
7.88915455e-01 4.93603826e-01 -1.16526850e-01 -8.27370062e-02
1.47646576e-01 1.42042667e-01 4.44982313e-02 -5.15105546e-01
6.57414377e-01 1.84814900e-03 -4.98902678e-01 -1.13386726e+00
-1.01715541e+00 3.14197928e-01 -3.55443567e-01 2.20513958e-02
9.60724056e-01 -8.00470889e-01 -4.98244941e-01 1.42111480e-01
-1.32347655e+00 -9.92164016e-02 -4.79251184e-02 7.49233067e-01
-7.40967989e-01 4.17810977e-01 -7.46624649e-01 -9.69300568e-01
-2.02432379e-01 -8.17293942e-01 2.82197326e-01 1.61033362e-01
-1.21756864e+00 -1.20518577e+00 -3.11026663e-01 -9.91927832e-02
4.05369937e-01 -2.96707869e-01 1.79912257e+00 -1.56278706e+00
-1.93146411e-02 1.28032848e-01 1.25817060e-01 -1.14087373e-01
-2.04333678e-01 -1.32620022e-01 -3.96740794e-01 1.38250932e-01
7.05995917e-01 -4.51379001e-01 6.85179770e-01 2.15796441e-01
6.11278057e-01 -3.51084769e-01 -4.94499803e-01 1.22287214e-01
1.33403134e+00 6.50175333e-01 4.32870418e-01 -1.32156294e-02
3.24616358e-02 9.09614742e-01 7.13113844e-01 8.20056275e-02
3.53196889e-01 1.64105862e-01 -2.61730969e-01 7.81968951e-01
3.63614678e-01 -6.28001094e-01 6.27791882e-01 8.64079177e-01
4.53246236e-01 -7.54853070e-01 -1.14895415e+00 5.43951452e-01
-1.11629367e+00 -9.65769947e-01 -1.99667424e-01 2.06745744e+00
7.24676073e-01 5.92539608e-01 -5.20958126e-01 -3.16575095e-02
6.45478249e-01 4.27108675e-01 -4.28048640e-01 -9.87298131e-01
-1.64848223e-01 1.10336646e-01 -2.37978518e-01 8.78734887e-01
-2.99145371e-01 1.55666780e+00 7.71897984e+00 3.45874697e-01
-8.38937044e-01 3.93546745e-02 9.26753521e-01 3.21895689e-01
-5.98192930e-01 4.45080787e-01 -6.75154984e-01 1.14999622e-01
1.48109853e+00 -9.38741267e-01 1.95111126e-01 5.10394752e-01
5.24051547e-01 -5.92336059e-01 -1.31861901e+00 2.75634259e-01
6.35661557e-02 -5.04004955e-01 5.09896398e-01 -4.09035861e-01
1.16747469e-02 -3.09569180e-01 9.59629714e-02 2.70958394e-01
-9.47365463e-02 -1.03265345e+00 1.10186219e+00 5.25415480e-01
1.08453847e-01 -4.94872034e-01 2.82139987e-01 8.38063776e-01
-4.45291281e-01 6.30177259e-02 -8.22486579e-01 -7.79181302e-01
6.71300769e-01 2.54180282e-01 -9.01940763e-01 -2.46873617e-01
2.80073136e-01 2.46080294e-01 -5.91630995e-01 4.41624820e-01
-7.76681960e-01 9.99565721e-01 -1.87407464e-01 -5.13237894e-01
7.06151128e-02 -1.28490075e-01 8.22694063e-01 9.43570852e-01
2.78552949e-01 5.90809345e-01 -1.02117039e-01 8.92052889e-01
5.81247985e-01 5.22073507e-01 -6.03454471e-01 -3.02208900e-01
5.51351368e-01 5.13763607e-01 -9.54227149e-01 -4.76275027e-01
-4.05707538e-01 7.83994019e-01 3.41501921e-01 3.15900028e-01
-3.49201530e-01 1.99629530e-01 4.48634803e-01 2.37844929e-01
-9.69629586e-02 -3.32171023e-01 -3.57203692e-01 -8.69429171e-01
-2.34598532e-01 -7.68150330e-01 1.06668375e-01 -1.12539530e+00
-1.43102050e+00 5.89860082e-01 2.78836310e-01 -2.68729091e-01
-7.27563202e-01 -8.98078918e-01 -6.21672094e-01 1.24613273e+00
-7.75084317e-01 -5.15138388e-01 6.80030048e-01 2.93124318e-01
7.46312737e-01 1.73887774e-01 1.07574224e+00 -7.97902763e-01
-8.45201015e-02 1.17397383e-01 -6.14283323e-01 -2.83745497e-01
4.95539457e-01 -8.19866598e-01 9.33609009e-01 8.75700355e-01
4.78911661e-02 1.15865600e+00 1.26037323e+00 -1.04964674e+00
-6.46510005e-01 -2.68604070e-01 1.82844055e+00 -4.26838130e-01
7.61544764e-01 1.63665023e-02 -1.22302282e+00 1.22056222e+00
5.67906678e-01 -7.55229533e-01 1.02761269e+00 4.45627570e-02
-2.64000505e-01 4.84390199e-01 -1.02729404e+00 9.78070974e-01
1.30075586e+00 -5.16713381e-01 -1.59205222e+00 4.78427291e-01
1.04754460e+00 8.44140574e-02 -5.19788206e-01 -3.03173400e-02
3.37537915e-01 -1.30844247e+00 5.94307721e-01 -9.80236828e-01
2.61696100e-01 2.89184917e-02 -2.10488364e-01 -1.40612626e+00
-7.42498815e-01 -1.90493271e-01 5.52615523e-01 8.57047856e-01
8.83199036e-01 -1.19870365e+00 -1.25656769e-01 9.65305746e-01
-2.11524516e-02 -2.90966660e-01 -9.36408579e-01 -3.95270586e-01
5.60367525e-01 -8.25410962e-01 3.59978169e-01 7.58713901e-01
2.91196674e-01 4.07812864e-01 -3.35121676e-02 1.48739427e-01
2.80093998e-01 -4.53620553e-02 -5.37310764e-02 -1.36733210e+00
-2.92773873e-01 -4.33734119e-01 -1.30183488e-01 -1.12764490e+00
4.81374502e-01 -7.24521220e-01 6.73090890e-02 -1.28680718e+00
-2.42560096e-02 4.14136201e-02 1.72584534e-01 2.66343147e-01
-3.94585967e-01 -4.31040823e-01 5.23816407e-01 2.11206064e-01
-1.54698202e-02 3.65476012e-01 1.07452178e+00 7.45013595e-01
-1.71883762e-01 -1.41441315e-01 -1.22484994e+00 1.26262665e+00
8.80640447e-01 -5.39662540e-01 -4.16181266e-01 -4.17898715e-01
7.37566710e-01 2.21699685e-01 1.20578557e-01 -7.12186754e-01
9.04643722e-03 -5.70734859e-01 7.11027503e-01 -4.52076435e-01
2.08567202e-01 -5.76650679e-01 1.99824385e-02 8.80394936e-01
-6.59913898e-01 8.58444214e-01 5.35386264e-01 3.97460520e-01
1.09990194e-01 -6.52368605e-01 8.12028050e-01 -3.74296814e-01
-1.03104281e+00 -2.68369824e-01 -1.36750793e+00 1.94761187e-01
1.01406002e+00 -2.56826609e-01 -2.16466859e-01 -5.82412362e-01
-9.48065937e-01 -5.94344996e-02 3.43827188e-01 6.37502849e-01
8.45754743e-01 -7.38151729e-01 -6.77370965e-01 1.64742500e-01
4.85206433e-02 -8.40645432e-01 5.08656316e-02 7.74167240e-01
-5.15873075e-01 5.17087579e-01 -2.63951551e-02 2.95334533e-02
-6.76608682e-01 9.93262053e-01 3.00283849e-01 3.03048790e-01
-4.16729897e-01 9.01780665e-01 6.45052671e-01 -1.80271149e-01
-5.82668960e-01 -2.70409733e-01 -3.41545105e-01 -1.11421488e-01
7.81841218e-01 -8.19091722e-02 -5.44920683e-01 -1.19446051e+00
-2.89879650e-01 2.03670889e-01 -6.71350896e-01 -3.14325333e-01
8.10735345e-01 -4.98367399e-01 -5.25294662e-01 1.24246514e+00
5.20245671e-01 1.90883130e-01 -4.48397428e-01 -1.00142546e-01
4.37975168e-01 -3.33695650e-01 -4.49994206e-01 -9.51915562e-01
-2.20855847e-02 6.78631842e-01 1.38289794e-01 2.63267308e-01
8.29595149e-01 2.54698277e-01 3.17901164e-01 4.29420948e-01
6.09728396e-01 -9.83694613e-01 -4.04218882e-01 7.61844516e-01
1.16137397e+00 -6.63547993e-01 -6.73932955e-02 -3.98061007e-01
-7.44221985e-01 1.04410636e+00 9.26700115e-01 -3.32460590e-02
6.64667010e-01 -1.81285828e-01 1.36027336e-01 -3.40972319e-02
-1.38509047e+00 2.97468960e-01 -2.14411616e-01 4.23015535e-01
1.11325562e+00 1.98887140e-01 -1.12891304e+00 5.48657715e-01
-8.09896410e-01 -5.92254758e-01 6.89633667e-01 7.13333130e-01
-9.20252681e-01 -8.99034441e-01 -6.12249732e-01 6.43926799e-01
-5.11355162e-01 -6.64530575e-01 -5.93002975e-01 1.05371833e+00
-1.62303895e-01 1.04606605e+00 3.41723174e-01 8.04892704e-02
2.49254122e-01 8.32431018e-01 1.74165487e-01 -8.46366584e-01
-6.72135949e-01 8.77224728e-02 2.59399801e-01 -4.48712379e-01
-1.35684133e-01 -1.00318718e+00 -1.61943424e+00 -4.60550547e-01
-4.30548698e-01 1.36500284e-01 1.26531556e-01 1.39041269e+00
1.65365860e-02 -6.05780035e-02 -7.85678104e-02 -1.77072942e-01
-6.58452511e-01 -1.22424173e+00 -7.20447183e-01 2.55562276e-01
-9.01746228e-02 -4.43183601e-01 -5.80427766e-01 -1.41681746e-01] | [10.262497901916504, 8.681357383728027] |
55c5c5fd-229e-43cb-8685-d29f1c67362a | general-audio-tagging-with-ensembling | 1810.12832 | null | http://arxiv.org/abs/1810.12832v1 | http://arxiv.org/pdf/1810.12832v1.pdf | General audio tagging with ensembling convolutional neural network and statistical features | Audio tagging aims to infer descriptive labels from audio clips. Audio
tagging is challenging due to the limited size of data and noisy labels. In
this paper, we describe our solution for the DCASE 2018 Task 2 general audio
tagging challenge. The contributions of our solution include: We investigated a
variety of convolutional neural network architectures to solve the audio
tagging task. Statistical features are applied to capture statistical patterns
of audio features to improve the classification performance. Ensemble learning
is applied to ensemble the outputs from the deep classifiers to utilize
complementary information. a sample re-weight strategy is employed for ensemble
training to address the noisy label problem. Our system achieves a mean average
precision (mAP@3) of 0.958, outperforming the baseline system of 0.704. Our
system ranked the 1st and 4th out of 558 submissions in the public and private
leaderboard of DCASE 2018 Task 2 challenge. Our codes are available at
https://github.com/Cocoxili/DCASE2018Task2/. | ['Dezhi Wang', 'Huaimin Wang', 'Haibo Mi', 'Bo Ding', 'Kele Xu', 'Qiuqiang Kong', 'Boqing Zhu'] | 2018-10-30 | null | null | null | null | ['audio-tagging'] | ['audio'] | [ 8.89247060e-02 -9.50363576e-02 -7.06944689e-02 -2.58164138e-01
-1.57484770e+00 -7.78837740e-01 1.63536221e-01 9.32907909e-02
-4.27387118e-01 5.41354597e-01 6.13314688e-01 3.36417019e-01
1.17154412e-01 -1.38933972e-01 -5.50534844e-01 -4.86384928e-01
-3.87050271e-01 -7.08446652e-02 -1.00970313e-01 2.07172632e-01
-5.93692847e-02 -8.65187421e-02 -1.71932733e+00 9.44025159e-01
4.39384341e-01 1.77445018e+00 -6.17799535e-02 8.83571506e-01
2.27757618e-01 8.24006140e-01 -6.11781776e-01 -1.79324955e-01
-8.30088090e-03 -2.07484171e-01 -9.94183183e-01 -4.59986925e-01
5.97324729e-01 -9.85482782e-02 -4.03298497e-01 7.03988969e-01
1.02978396e+00 2.92172670e-01 3.85132641e-01 -1.51628923e+00
-3.57789993e-01 1.13257694e+00 -2.08635584e-01 3.79183024e-01
4.15529907e-01 -4.33451742e-01 1.48375249e+00 -1.15529311e+00
2.71531016e-01 6.17290199e-01 9.66358840e-01 6.27052128e-01
-9.22504902e-01 -1.24355102e+00 1.98938102e-01 5.17287016e-01
-1.78732419e+00 -8.94504249e-01 7.68770695e-01 -6.11665130e-01
9.15312827e-01 2.40771428e-01 4.32859421e-01 1.21369195e+00
-2.70764649e-01 9.27338660e-01 8.39852273e-01 -3.64553273e-01
4.79459576e-02 -5.43629639e-02 1.52911752e-01 4.49219108e-01
-4.96712685e-01 -2.69217789e-01 -1.12019253e+00 -3.91267002e-01
6.40170425e-02 -3.75422359e-01 -3.16784203e-01 3.23949873e-01
-1.16885793e+00 5.49705684e-01 3.43414187e-01 4.00484294e-01
-2.10676491e-01 2.81612217e-01 8.47044468e-01 2.57595181e-01
8.43137741e-01 6.39672458e-01 -6.58856034e-01 -5.74189782e-01
-8.55753899e-01 2.97717988e-01 7.97536433e-01 8.87758255e-01
1.31788343e-01 -1.08576559e-01 -5.79185724e-01 1.29235303e+00
-1.24735944e-01 2.61940688e-01 4.21709150e-01 -1.38867009e+00
4.40719247e-01 5.04518263e-02 7.25690722e-02 -6.69495225e-01
-5.16392291e-01 -6.63322866e-01 -8.28828096e-01 -4.44079489e-01
3.74894999e-02 -3.91869783e-01 -7.67362475e-01 1.89354539e+00
-3.22371498e-02 6.13448143e-01 -2.75183141e-01 8.07554007e-01
1.41312921e+00 7.10831761e-01 3.02104771e-01 6.20506965e-02
1.45336175e+00 -9.76863325e-01 -8.23565185e-01 1.12624519e-01
7.71699905e-01 -1.01035070e+00 9.64844525e-01 5.50074697e-01
-8.81281674e-01 -6.78131700e-01 -9.36331451e-01 2.92534716e-02
-2.07624376e-01 6.41382873e-01 2.61046857e-01 1.52969420e-01
-1.10358274e+00 4.75620061e-01 -6.77384138e-01 -7.24865794e-02
5.69932759e-01 4.07082587e-01 -3.08781385e-01 1.75513417e-01
-1.39059901e+00 2.76619107e-01 4.53873843e-01 9.46425870e-02
-1.00543034e+00 -9.71915781e-01 -4.14181709e-01 1.61265835e-01
3.38175148e-01 -2.70810753e-01 1.86994398e+00 -9.70709085e-01
-1.42825687e+00 8.69136274e-01 -7.71810710e-02 -5.18341839e-01
2.22232267e-01 -6.66149676e-01 -4.99963820e-01 3.88724962e-03
1.78857222e-01 9.16600883e-01 4.68912154e-01 -8.09539795e-01
-1.05980098e+00 1.58528849e-01 2.65117940e-02 1.14753209e-01
-5.59672892e-01 2.73483694e-01 -2.94688970e-01 -7.94638991e-01
-7.01374039e-02 -1.15770435e+00 2.86130399e-01 -5.17681897e-01
-4.12223786e-01 -6.66423738e-01 4.99099374e-01 -7.58775473e-01
1.50894880e+00 -2.58482194e+00 -4.32701446e-02 -2.29931533e-01
2.98412234e-01 7.43963048e-02 -3.97141755e-01 6.00656331e-01
-2.66273230e-01 1.47928327e-01 2.20322385e-01 -6.36714399e-01
2.62610883e-01 -4.54250067e-01 -6.07511282e-01 2.51729358e-02
1.92511529e-01 5.36673725e-01 -9.15948629e-01 -2.99170256e-01
-3.59623671e-01 5.96790671e-01 -3.90973955e-01 5.06026208e-01
2.87041999e-02 3.08412760e-01 -8.86752754e-02 7.10598171e-01
2.74182916e-01 -5.25528565e-02 -9.69917253e-02 -4.36201543e-01
-2.38760952e-02 9.44879115e-01 -1.07013273e+00 1.99741316e+00
-6.19791269e-01 7.83209860e-01 9.33172256e-02 -8.21616590e-01
8.11355829e-01 8.68869960e-01 8.44006062e-01 -3.62955421e-01
7.25489333e-02 2.63070107e-01 -1.41169220e-01 -4.94240195e-01
4.24439609e-01 1.75589159e-01 -5.04544199e-01 5.10199487e-01
4.36833531e-01 1.25842616e-01 1.06979281e-01 2.92245060e-01
1.26593614e+00 6.92356005e-02 5.98187521e-02 -4.31472016e-03
2.75842994e-01 -3.46620381e-01 8.87287974e-01 7.35485196e-01
-2.94110119e-01 8.84039342e-01 4.97820735e-01 -5.08018374e-01
-6.85145319e-01 -7.93686569e-01 -6.66340068e-02 1.79189003e+00
-4.45718288e-01 -1.01105666e+00 -5.04960060e-01 -6.28629208e-01
-1.88791752e-01 3.38795573e-01 -7.37503350e-01 -1.78166598e-01
-3.74030232e-01 -3.58171374e-01 1.08898544e+00 7.07157969e-01
4.26770538e-01 -1.20152712e+00 -2.48622984e-01 3.28839511e-01
-8.87646496e-01 -1.09084451e+00 -5.78437626e-01 4.77857590e-01
-2.64575779e-01 -8.39113235e-01 -4.32609499e-01 -8.60584378e-01
-1.59529686e-01 1.36510015e-01 1.15496540e+00 -9.70492586e-02
3.56249586e-02 3.23472470e-01 -8.43217850e-01 -7.35675454e-01
-4.50144289e-03 7.24231422e-01 2.47718915e-01 8.85602385e-02
4.97152895e-01 -6.55448198e-01 -5.29108167e-01 1.33263707e-01
-5.18206298e-01 -1.83806017e-01 1.79782048e-01 9.05446172e-01
6.32273376e-01 -3.31486762e-01 8.72909606e-01 -4.40270275e-01
7.06546485e-01 -5.69479525e-01 -1.70333311e-01 -5.04925475e-02
-2.21426353e-01 -4.91973042e-01 4.49329287e-01 -6.30955160e-01
-5.74062705e-01 3.45392078e-01 -3.19461286e-01 -4.25767779e-01
-3.38945448e-01 5.28726459e-01 9.13007110e-02 3.66638988e-01
3.74948978e-01 -1.93169668e-01 -5.02145410e-01 -8.44290555e-01
5.46696112e-02 1.22432899e+00 6.96620107e-01 -7.01374710e-01
9.27392319e-02 2.48726364e-02 -4.86369163e-01 -3.14639300e-01
-1.73178804e+00 -6.86954618e-01 -4.73937273e-01 -4.17701513e-01
7.07441926e-01 -1.42265677e+00 -7.28196561e-01 4.40426409e-01
-1.08784068e+00 -3.54883015e-01 -2.69891232e-01 5.85856915e-01
-4.60705757e-01 -3.12440425e-01 -6.71923697e-01 -7.25912571e-01
-6.11904681e-01 -8.38481426e-01 1.23781312e+00 1.39667643e-02
-6.11738205e-01 -3.68352562e-01 3.93228471e-01 4.24098194e-01
2.83036113e-01 9.93354097e-02 3.25802058e-01 -1.05405033e+00
-5.89957610e-02 -2.84112424e-01 -4.03613225e-02 5.17222226e-01
-4.35767733e-02 -1.97712705e-02 -1.69580877e+00 -3.21313024e-01
-5.23690045e-01 -8.85386050e-01 1.17613351e+00 3.77168924e-01
1.69045889e+00 -1.61881879e-01 -2.18164399e-01 4.16355044e-01
8.08784902e-01 1.10472448e-01 1.04446366e-01 3.87094676e-01
5.10502934e-01 4.79982793e-01 7.12872446e-01 6.11046195e-01
4.08516556e-01 1.00342178e+00 2.42205366e-01 2.11846218e-01
-2.55347252e-01 -3.76101822e-01 4.18235272e-01 1.27509975e+00
-6.80046603e-02 -2.70898998e-01 -1.12868297e+00 7.65176415e-01
-1.76307499e+00 -8.23301077e-01 2.19624005e-02 1.86694920e+00
8.51789236e-01 -1.12847313e-01 3.76934767e-01 3.06480020e-01
7.43585050e-01 1.02743313e-01 -1.55752301e-01 -5.54660410e-02
1.09517001e-01 2.86265284e-01 -9.32607725e-02 1.68791696e-01
-1.77219105e+00 6.92099094e-01 5.71121025e+00 1.04399991e+00
-9.85028982e-01 6.31700754e-01 4.54097390e-01 -6.98385775e-01
2.91215539e-01 -2.57517427e-01 -7.47345269e-01 7.02999651e-01
1.26651180e+00 -5.91336377e-02 1.86159194e-01 5.85003257e-01
-1.16669759e-01 2.65258431e-01 -1.03928399e+00 1.19721997e+00
8.69443044e-02 -1.31177342e+00 -5.82081795e-01 -1.03010096e-01
8.27812374e-01 4.92753595e-01 2.79219657e-01 5.36365509e-01
1.46788314e-01 -9.40795481e-01 9.36402678e-01 5.50140977e-01
9.52190340e-01 -8.08571398e-01 1.14399087e+00 -7.26094320e-02
-1.39512920e+00 -3.58714044e-01 -6.79464936e-02 -1.89412326e-01
6.64076284e-02 6.48035228e-01 -8.21447790e-01 3.12222809e-01
1.37750673e+00 8.35970521e-01 -4.01655644e-01 1.22827673e+00
-1.23606250e-01 1.07860446e+00 -4.85665709e-01 1.04219981e-01
1.53608099e-01 5.47683597e-01 4.18002486e-01 1.45517027e+00
5.59787631e-01 -9.35838465e-03 2.09495544e-01 2.90558070e-01
-6.98468447e-01 2.33079895e-01 -2.78176844e-01 -7.22537488e-02
9.83769536e-01 1.45097101e+00 -3.99009049e-01 -2.55211025e-01
-7.87313208e-02 5.37730217e-01 4.30096626e-01 3.96272503e-02
-8.14816535e-01 -6.16401613e-01 5.86400747e-01 -1.69510320e-01
3.66242260e-01 5.73786870e-02 -1.20137349e-01 -1.08663023e+00
2.81815324e-03 -7.04317510e-01 5.94069064e-01 -7.88100123e-01
-1.27954936e+00 8.98158848e-01 -1.47048816e-01 -1.53864348e+00
-5.84723316e-02 -8.85874033e-02 -5.31970441e-01 4.51507717e-01
-1.11095023e+00 -1.12384856e+00 -3.81490827e-01 2.62815475e-01
4.61065710e-01 -5.02254128e-01 1.14706421e+00 8.22753251e-01
-3.93133163e-01 9.42572355e-01 1.67711973e-01 3.28969687e-01
1.23569369e+00 -1.05370915e+00 1.46460548e-01 3.54915053e-01
5.42303920e-01 5.66826500e-02 3.96204472e-01 -1.98801130e-01
-5.57147264e-01 -1.21370518e+00 1.37266338e+00 -4.38395441e-01
9.43429232e-01 -6.50507450e-01 -7.44969845e-01 5.76804042e-01
3.50965410e-01 1.62057593e-01 1.28408861e+00 6.22953713e-01
-6.71290815e-01 -4.09220278e-01 -5.31844795e-01 -1.53375287e-02
1.05341709e+00 -8.55198920e-01 -2.73330927e-01 3.92037123e-01
8.56689453e-01 -4.39020216e-01 -1.24006486e+00 4.40033764e-01
9.34657514e-01 -4.50008005e-01 7.00150192e-01 -4.84210670e-01
5.54341674e-01 -1.69816196e-01 -4.17958170e-01 -1.30630624e+00
-4.56336707e-01 -5.53473890e-01 -2.32090920e-01 1.78516674e+00
4.71504748e-01 -9.98864248e-02 5.43325305e-01 1.71505548e-02
-4.79326397e-01 -7.81398237e-01 -1.27402985e+00 -5.99975228e-01
-2.13679001e-01 -7.82817483e-01 4.61273372e-01 9.69483435e-01
1.57251045e-01 3.82313848e-01 -6.00038826e-01 -8.83059725e-02
2.46557266e-01 9.66723729e-03 5.08419633e-01 -1.37251329e+00
-1.54168889e-01 -2.37487212e-01 -3.67890477e-01 -6.44973397e-01
4.74179566e-01 -1.07295203e+00 1.51637897e-01 -1.07285261e+00
3.00360680e-01 -6.00773633e-01 -1.08887255e+00 1.00475037e+00
9.93270352e-02 9.37720001e-01 5.35543919e-01 2.67364323e-01
-1.16682851e+00 4.68203127e-01 4.72665519e-01 -3.75765920e-01
-1.83097333e-01 1.56382114e-01 -7.55470514e-01 4.41444397e-01
1.24430799e+00 -7.51968145e-01 -1.26899868e-01 -5.38605809e-01
3.98204744e-01 -3.43548059e-01 3.23690236e-01 -1.23857224e+00
1.25520080e-01 4.33982670e-01 2.04130396e-01 -5.56842744e-01
5.81939995e-01 -4.64913428e-01 2.68370807e-01 2.75400747e-02
-9.63602364e-01 -1.81525424e-01 2.95591384e-01 2.53573358e-01
-5.71239471e-01 -1.81454778e-01 4.30389345e-01 2.69328892e-01
-4.66648459e-01 2.47144267e-01 -4.59124029e-01 1.96665719e-01
5.54132760e-01 4.75100100e-01 -2.12479189e-01 -6.32453322e-01
-1.05619037e+00 1.62724197e-01 -2.47799158e-01 7.85267651e-01
3.39984208e-01 -1.87384498e+00 -8.79946887e-01 -1.57894224e-01
3.60330611e-01 -2.94423640e-01 3.94347340e-01 9.22057033e-01
1.84802059e-02 4.79635000e-01 -1.06108874e-01 -5.13711035e-01
-1.49007642e+00 -8.66392255e-02 2.06555933e-01 -1.94009364e-01
-3.25780928e-01 1.30888844e+00 -1.45978227e-01 -3.15449834e-01
7.50329554e-01 -2.93347612e-02 -3.51021737e-01 4.09112602e-01
7.01548755e-01 3.65469366e-01 2.87244380e-01 -7.01574564e-01
-4.95783687e-01 3.16543967e-01 -2.25333661e-01 -1.04563937e-01
1.58784819e+00 -3.09349857e-02 3.44482400e-02 7.87697375e-01
1.33215618e+00 -7.39286020e-02 -9.89533842e-01 -3.91382277e-01
-1.15544740e-02 -1.24486357e-01 3.10816944e-01 -1.00310028e+00
-1.01952493e+00 9.51275349e-01 8.72735262e-01 4.25988376e-01
1.28562117e+00 2.56509811e-01 8.33615839e-01 3.53053629e-01
8.17721989e-03 -1.34568191e+00 -1.55924529e-01 7.11624682e-01
1.02441168e+00 -1.18474865e+00 -1.90094829e-01 -1.12946637e-01
-6.81900561e-01 9.49543059e-01 5.94561100e-01 -7.78286979e-02
7.00503409e-01 4.24579889e-01 2.20215663e-01 6.56890422e-02
-1.14159334e+00 -1.42745465e-01 6.53385878e-01 5.55472195e-01
9.76417780e-01 2.00770125e-01 -1.05070002e-01 1.22608447e+00
-3.50994945e-01 1.52079493e-01 1.52807057e-01 7.74658263e-01
-1.92619458e-01 -1.00265515e+00 -9.73358974e-02 4.31600094e-01
-9.43789661e-01 -3.65006812e-02 -6.87627196e-01 3.25714171e-01
4.52496469e-01 1.04657948e+00 2.33587131e-01 -9.63371098e-01
3.09823006e-01 4.34506297e-01 7.47025311e-02 -5.74877083e-01
-7.79322922e-01 4.06762093e-01 3.89249176e-01 -6.61199689e-01
-7.96759546e-01 -5.59610367e-01 -8.99687469e-01 1.59668305e-03
-1.11643210e-01 5.66839457e-01 6.49910092e-01 5.53140938e-01
6.78622723e-01 6.95007622e-01 6.83606505e-01 -9.85377729e-01
-2.48140767e-01 -1.38499343e+00 -5.78686893e-01 1.57457709e-01
4.67606872e-01 -8.09672892e-01 -3.80320430e-01 2.90444255e-01] | [15.227123260498047, 5.112461090087891] |
90451133-e388-4b2f-a5cb-4328d81ef6d7 | ntire-2021-challenge-on-high-dynamic-range | 2106.01439 | null | https://arxiv.org/abs/2106.01439v1 | https://arxiv.org/pdf/2106.01439v1.pdf | NTIRE 2021 Challenge on High Dynamic Range Imaging: Dataset, Methods and Results | This paper reviews the first challenge on high-dynamic range (HDR) imaging that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2021. This manuscript focuses on the newly introduced dataset, the proposed methods and their results. The challenge aims at estimating a HDR image from one or multiple respective low-dynamic range (LDR) observations, which might suffer from under- or over-exposed regions and different sources of noise. The challenge is composed by two tracks: In Track 1 only a single LDR image is provided as input, whereas in Track 2 three differently-exposed LDR images with inter-frame motion are available. In both tracks, the ultimate goal is to achieve the best objective HDR reconstruction in terms of PSNR with respect to a ground-truth image, evaluated both directly and with a canonical tonemapping operation. | ['Radu Timofte', 'Aleš Leonardis', 'Sibi Catley-Chandar', 'Eduardo Pérez-Pellitero'] | 2021-06-02 | null | null | null | null | ['hdr-reconstruction'] | ['computer-vision'] | [ 8.75970006e-01 -3.20266962e-01 2.93436140e-01 -1.82887867e-01
-1.08290684e+00 -3.22085351e-01 4.77648824e-01 -3.34825426e-01
-1.51907399e-01 7.35461056e-01 2.94135779e-01 1.14544094e-01
-8.61616582e-02 -4.76730525e-01 -4.72585112e-01 -9.13676560e-01
-1.47559106e-01 -1.81731924e-01 2.20139965e-01 -3.89462858e-01
7.35983029e-02 6.03512943e-01 -1.61328888e+00 4.05449212e-01
6.68446779e-01 9.92539525e-01 6.70882523e-01 1.08630443e+00
4.74065125e-01 1.17719257e+00 -4.05335128e-01 6.04997575e-02
5.24124980e-01 -7.51572490e-01 -7.38525748e-01 3.95914257e-01
6.62775099e-01 -6.20450974e-01 -6.18341506e-01 1.19386446e+00
9.23799038e-01 1.34699970e-01 3.47991854e-01 -5.80187321e-01
-5.91584325e-01 1.33685112e-01 -7.40026236e-01 5.93199134e-01
5.82693875e-01 1.25650808e-01 5.08481205e-01 -9.23777044e-01
7.78631449e-01 9.81660247e-01 7.13104904e-01 2.06749290e-01
-1.70525026e+00 -2.33955055e-01 -3.76378685e-01 4.31813985e-01
-1.38868761e+00 -7.04703629e-01 7.11254776e-01 -2.37523094e-01
6.80868924e-01 5.06948888e-01 4.25588936e-01 6.79928601e-01
1.70423076e-01 4.57600266e-01 1.84944761e+00 -5.41513503e-01
-5.84078953e-03 -5.00872172e-02 -2.04244256e-01 2.32142732e-02
-1.33347511e-01 5.05191982e-01 -4.99893755e-01 9.86391231e-02
8.53932023e-01 -5.55942714e-01 -8.52905750e-01 -1.83263943e-01
-1.29699159e+00 3.23491275e-01 3.29356819e-01 5.29126763e-01
-3.24659586e-01 -1.58929631e-01 1.52015388e-02 6.41270995e-01
5.15867114e-01 1.61575988e-01 -1.89017832e-01 2.68254071e-01
-1.28728795e+00 3.20640611e-05 1.70688510e-01 5.87236822e-01
6.16150796e-01 2.06366643e-01 -2.18294203e-01 1.11740148e+00
1.47553936e-01 7.11860657e-01 1.80096298e-01 -1.08855283e+00
1.58823133e-01 -3.96546900e-01 2.46455237e-01 -9.11833167e-01
-1.50061443e-01 -4.37431425e-01 -1.12651026e+00 4.81169820e-01
1.88589469e-01 2.19071180e-01 -1.00271153e+00 1.34549153e+00
1.95279390e-01 2.05137745e-01 1.89714521e-01 1.29885960e+00
8.71197522e-01 9.34954524e-01 -1.20174035e-01 -7.44090796e-01
1.15171528e+00 -6.36106670e-01 -1.05848932e+00 -2.45730072e-01
-2.44924545e-01 -1.29880750e+00 7.82609046e-01 7.03471363e-01
-1.44421089e+00 -9.38645244e-01 -1.13493907e+00 -3.62951487e-01
6.76805675e-02 1.54110789e-01 -1.67687252e-01 7.07575321e-01
-1.52687025e+00 5.93386054e-01 -2.73665875e-01 -1.86696321e-01
3.78437643e-03 -7.26002231e-02 -3.87975693e-01 -8.18122387e-01
-1.30241907e+00 1.15693879e+00 1.41028762e-01 3.24955493e-01
-1.06752348e+00 -9.29797947e-01 -7.17384279e-01 -5.06312311e-01
2.18834788e-01 -2.87171811e-01 6.74375653e-01 -7.86549091e-01
-1.42416680e+00 1.15565550e+00 3.68245281e-02 -3.89620572e-01
7.39402413e-01 -8.40034187e-02 -8.42374802e-01 3.68178487e-01
-1.80143148e-01 3.76953363e-01 1.03634810e+00 -1.71767628e+00
-4.19329733e-01 -3.50358546e-01 -3.50692630e-01 3.02994221e-01
4.23256159e-01 3.36763293e-01 -5.14021814e-01 -1.01280749e+00
3.20650011e-01 -6.83648288e-01 2.06552073e-02 2.70265737e-03
-3.85677099e-01 7.79611647e-01 9.54420507e-01 -1.34346461e+00
1.06544304e+00 -2.08295178e+00 9.92151126e-02 1.53255987e-03
1.82375446e-01 1.64714783e-01 -2.69161522e-01 1.68122098e-01
-4.63685781e-01 -2.86144644e-01 -4.99887794e-01 -1.00246660e-01
-3.97307277e-01 -5.23186550e-02 -5.25941372e-01 1.10091341e+00
-9.76342857e-02 5.07085502e-01 -6.62982285e-01 -4.54873711e-01
8.33063185e-01 8.64557922e-01 2.10672498e-01 4.12302196e-01
1.70302853e-01 9.42473352e-01 2.17240661e-01 6.14990056e-01
1.27160299e+00 9.45930481e-02 1.51901573e-01 -8.70475590e-01
-4.41273183e-01 -3.05082083e-01 -1.40612829e+00 1.40060496e+00
-5.52243054e-01 1.11966109e+00 2.55087107e-01 -6.97375894e-01
9.89892900e-01 3.20795864e-01 7.40821183e-01 -1.16262317e+00
-2.42859378e-01 4.17846143e-01 -4.03199732e-01 -4.57860380e-01
6.08241439e-01 -3.11984509e-01 4.09021139e-01 2.61101067e-01
-8.25356767e-02 -2.06824303e-01 1.01168811e-01 -9.28932354e-02
9.42488134e-01 6.82939067e-02 4.31102514e-01 -3.30116898e-01
8.07326376e-01 -2.64308721e-01 3.84964913e-01 7.63596356e-01
-3.25511396e-01 1.38925195e+00 5.13777621e-02 -2.69891918e-01
-1.52505362e+00 -1.37879932e+00 -5.65398753e-01 5.23882985e-01
4.58485514e-01 1.43628463e-01 -3.28171015e-01 5.40448688e-02
-6.73873365e-01 4.86188650e-01 -4.59818274e-01 2.18730018e-01
-7.55117655e-01 -1.01830149e+00 3.16395879e-01 -3.88479717e-02
8.92881632e-01 -8.49458754e-01 -5.54933906e-01 3.71871167e-03
-7.64536798e-01 -1.35998619e+00 -4.83713180e-01 1.22158840e-01
-7.12770283e-01 -9.34325993e-01 -1.34112918e+00 -6.66516960e-01
4.64167058e-01 5.22272587e-01 1.48824644e+00 -7.82218575e-02
-6.94450498e-01 1.39370918e-01 -4.76543337e-01 3.08231115e-01
-5.18968344e-01 -7.58699834e-01 -4.01101887e-01 2.21927106e-01
-3.12561482e-01 -3.32467735e-01 -9.05172825e-01 6.78368270e-01
-1.18742144e+00 7.89210647e-02 6.53719127e-01 8.28539968e-01
1.03251278e+00 6.11392200e-01 1.05512492e-01 -5.67088902e-01
2.11162820e-01 -1.49615854e-01 -6.52604103e-01 4.19708550e-01
-6.45301700e-01 -3.18740785e-01 3.88179988e-01 -4.06287074e-01
-1.44662261e+00 7.45785311e-02 -1.66247070e-01 -3.87247473e-01
-9.14482623e-02 -1.10381886e-01 -3.34761828e-01 -4.53568071e-01
8.84546697e-01 6.46420121e-01 -7.14395791e-02 -4.93996173e-01
3.18338186e-01 6.55755401e-01 1.06735504e+00 -2.44189635e-01
9.91502821e-01 6.91662312e-01 1.09041356e-01 -1.07741964e+00
-5.67917228e-01 -6.17462516e-01 -4.38039780e-01 -6.63131595e-01
8.80391657e-01 -1.20994520e+00 -1.34518474e-01 8.80620301e-01
-8.63274455e-01 -6.21743143e-01 -3.90205055e-01 4.03352082e-01
-7.81593919e-01 5.84486008e-01 -7.07221806e-01 -6.61856651e-01
-1.96793437e-01 -1.23796809e+00 1.03499508e+00 2.25779206e-01
3.91237229e-01 -8.61959100e-01 2.09721938e-01 5.24712086e-01
8.80284190e-01 2.97052979e-01 6.07707679e-01 2.17643887e-01
-6.31939292e-01 1.53796822e-01 -5.02775252e-01 7.29915261e-01
3.20560448e-02 -4.19805288e-01 -1.21541893e+00 -5.89547813e-01
2.77366549e-01 -2.07506269e-01 7.98222423e-01 8.42332244e-01
9.80047643e-01 5.96030056e-02 1.70691103e-01 7.71673799e-01
2.07445216e+00 -9.33962315e-03 1.61602557e+00 4.22178984e-01
2.87082225e-01 6.86973691e-01 8.41650665e-01 2.32275918e-01
-1.44051090e-01 1.15990889e+00 2.31009543e-01 -5.38498521e-01
-9.82952058e-01 1.84079170e-01 4.46788132e-01 3.26401532e-01
-1.38118282e-01 -3.17198336e-01 -6.01077855e-01 4.69720513e-01
-1.17709458e+00 -9.63355839e-01 -5.41258037e-01 2.51119637e+00
8.52909088e-01 -5.47841012e-01 9.56115220e-03 3.56417090e-01
9.66801047e-01 4.84732002e-01 -4.66720551e-01 7.84920305e-02
-8.55615854e-01 9.93401036e-02 7.50331819e-01 6.72800183e-01
-1.04566610e+00 3.81497175e-01 7.28220987e+00 9.13369119e-01
-1.09951603e+00 2.55805343e-01 1.00114369e+00 1.10065877e-01
-1.72127053e-01 -8.72077420e-02 -3.82103264e-01 3.46505463e-01
9.56544220e-01 4.27095890e-02 7.39219069e-01 1.64498255e-01
4.04128611e-01 -4.08130825e-01 -5.52203715e-01 1.13403547e+00
4.59199399e-01 -1.03145897e+00 -4.14211601e-01 1.09060585e-01
9.38606262e-01 -1.67177655e-02 4.27003145e-01 -2.70302951e-01
-2.93954089e-02 -1.15126956e+00 6.63396776e-01 6.07492387e-01
1.43563092e+00 -5.71304142e-01 6.73437297e-01 -1.87587783e-01
-1.21453357e+00 1.06441654e-01 -4.25555021e-01 7.06090987e-01
5.43480515e-01 8.83422494e-01 -1.00777619e-01 7.45953619e-01
1.03912342e+00 7.35517561e-01 -4.85110939e-01 1.07132936e+00
-1.49315104e-01 1.74465701e-01 3.81662361e-02 1.14501452e+00
-5.23475647e-01 -2.24893570e-01 7.68266439e-01 1.24847245e+00
3.55557114e-01 2.22125307e-01 -8.29037875e-02 7.07441568e-01
1.18433349e-01 -1.87456459e-01 -3.22314709e-01 4.71502215e-01
1.01585433e-01 1.27471077e+00 -6.22578740e-01 -1.42720878e-01
-2.16929898e-01 1.31890643e+00 -5.90053082e-01 5.68691194e-01
-7.75174737e-01 -1.86238617e-01 3.40346873e-01 4.09466207e-01
2.19456896e-01 -3.28233116e-03 -9.95790362e-02 -1.02887619e+00
-1.71132132e-01 -1.08426499e+00 2.42152095e-01 -1.43869269e+00
-1.20412648e+00 8.31977844e-01 -8.56646076e-02 -1.41311479e+00
-3.90722491e-02 -3.25376987e-01 -1.12177901e-01 1.22175860e+00
-2.06009150e+00 -9.51375604e-01 -6.19359434e-01 7.08969951e-01
5.92416942e-01 1.10147521e-01 4.45391029e-01 8.64958167e-01
-1.74216270e-01 2.15203181e-01 3.64185542e-01 -2.96934247e-01
9.43618178e-01 -1.14508307e+00 1.00544289e-01 1.31455731e+00
-1.85848400e-01 9.43308473e-02 1.16177619e+00 -6.40836358e-01
-1.32632649e+00 -1.15761995e+00 6.49887621e-01 1.00162983e-01
1.89173207e-01 8.11164975e-02 -9.14169133e-01 3.33907068e-01
1.02069169e-01 5.03239989e-01 2.55902588e-01 -7.33069479e-01
-8.40938166e-02 -2.38638416e-01 -1.68231201e+00 1.84256300e-01
7.91061163e-01 -6.62605226e-01 -8.46845433e-02 1.33983135e-01
5.19300103e-01 -6.16972864e-01 -1.20216250e+00 5.33689380e-01
3.63049299e-01 -1.28386378e+00 1.59715962e+00 2.93389708e-01
5.31162381e-01 -8.99051309e-01 -7.00697362e-01 -1.06252420e+00
-2.17701957e-01 -5.74495375e-01 -3.18715535e-02 1.04806232e+00
-4.82896939e-02 -2.12616488e-01 3.22644800e-01 1.31312639e-01
6.16144501e-02 -1.50898442e-01 -8.03359628e-01 -8.19520116e-01
-4.06418890e-01 -2.15198025e-01 4.28675786e-02 8.25620890e-01
-9.43238974e-01 -8.60540122e-02 -1.07183278e+00 3.57912809e-01
1.33241832e+00 5.91727123e-02 4.82390851e-01 -6.79630756e-01
-4.68973249e-01 2.00069621e-01 -4.65801477e-01 -9.03525651e-01
-4.44069982e-01 -4.16612804e-01 5.00246882e-01 -1.61864269e+00
2.51078933e-01 -2.98802137e-01 -1.74626812e-01 -1.44563079e-01
4.09009345e-02 9.90849555e-01 2.31839612e-01 3.79290015e-01
-3.28753978e-01 2.35278457e-01 1.28816402e+00 -9.45234895e-02
2.55995449e-02 -2.06313357e-01 -3.78966212e-01 2.29182333e-01
5.04121006e-01 -1.90461636e-01 -3.43806386e-01 -2.97586471e-01
2.12569628e-02 4.74662036e-01 6.15033031e-01 -1.09272611e+00
5.45846932e-02 -6.63557323e-03 7.13260293e-01 -6.65920019e-01
4.81512308e-01 -8.25473964e-01 9.04513717e-01 2.28711292e-01
-4.41772252e-01 -1.74604625e-01 -1.61796771e-02 5.75637996e-01
-4.31635767e-01 2.76821125e-02 1.64472270e+00 -7.47787133e-02
-9.00404334e-01 9.24932659e-02 -2.07225963e-01 -1.46268174e-01
8.87356281e-01 -3.95079732e-01 -5.64114392e-01 -3.35984528e-01
-7.89068580e-01 -4.37860727e-01 7.47281611e-01 8.47302005e-02
8.91071320e-01 -1.09813702e+00 -1.17956114e+00 1.10660888e-01
-2.39177421e-02 -4.71228570e-01 1.02792764e+00 1.03852177e+00
-7.73226202e-01 -5.18152080e-02 -4.27823722e-01 -5.11793137e-01
-1.52115631e+00 5.90498388e-01 6.68500483e-01 -4.60966676e-01
-1.04349661e+00 4.48141396e-01 4.83113155e-02 2.49981675e-02
-5.73878512e-02 3.45814466e-01 -2.50056505e-01 -3.58392715e-01
1.13234293e+00 5.96530259e-01 1.50181547e-01 -1.03311837e+00
-5.94683737e-02 8.85198772e-01 2.04896361e-01 -3.61290395e-01
1.34880269e+00 -9.37609076e-01 -1.85030207e-01 3.99766505e-01
1.29478157e+00 -1.10168517e-01 -1.33568418e+00 -3.50712210e-01
-3.48510563e-01 -1.12898588e+00 7.22437799e-01 -1.07786691e+00
-1.36752546e+00 5.40660620e-01 1.47959471e+00 3.62120122e-02
1.82532275e+00 -1.65308475e-01 8.00813854e-01 -3.78403962e-01
3.03626567e-01 -1.00359774e+00 1.12313829e-01 -1.21873513e-01
1.14256716e+00 -1.27841771e+00 3.41204405e-01 -5.38027227e-01
-5.02968967e-01 1.13248229e+00 5.03196009e-02 1.72591314e-01
4.35971260e-01 4.01262283e-01 1.88471168e-01 8.93899128e-02
-3.83340597e-01 -3.00451308e-01 3.19858968e-01 1.16370273e+00
4.45835531e-01 -4.27086711e-01 -1.32639185e-01 -3.64898682e-01
2.85732538e-01 1.43095866e-01 6.95368707e-01 5.05991638e-01
-5.00092447e-01 -8.08193147e-01 -8.29002321e-01 1.40167505e-01
-5.81044972e-01 -1.27428800e-01 2.30079740e-01 4.48510140e-01
1.79494470e-01 1.22762966e+00 -1.07203588e-01 -3.12513322e-01
3.75152409e-01 -8.19232523e-01 6.96412623e-01 2.20959987e-02
-3.15318614e-01 3.08561713e-01 8.90179724e-02 -8.35898340e-01
-7.08076060e-01 -5.29536843e-01 -6.63424313e-01 -4.88615274e-01
-2.08861101e-02 -3.42021674e-01 8.40098560e-01 4.35104549e-01
-8.79274681e-02 6.08231127e-01 8.39105606e-01 -9.06357050e-01
-1.31024346e-01 -6.98928475e-01 -1.08402431e+00 3.67041171e-01
7.80803025e-01 -3.01217973e-01 -6.39705181e-01 6.04459584e-01] | [11.00524616241455, -2.1685707569122314] |
91cc2db5-5bf5-4f4c-83a2-89020cfa9b97 | energy-management-for-a-dm-i-plug-in-hybrid | 2306.08823 | null | https://arxiv.org/abs/2306.08823v1 | https://arxiv.org/pdf/2306.08823v1.pdf | Energy Management for a DM-i Plug-in Hybrid Electric Vehicle via Continuous-Discrete Reinforcement Learning | Energy management strategy (EMS) is a key technology for plug-in hybrid electric vehicles (PHEVs). The energy management of PHEVs needs to output continuous variables such as engine torque, as well as discrete variables such as clutch engagement or disengagement. This type of problem is a mixed-integer programming problem. In addition, the hybrid powertrain system is highly nonlinear and complex. Designing an efficient EMS is a challenging task. We establish a control-oriented mathematical model for a BYD DM-i hybrid powertrain system from the perspective of mixed-integer programming. Then, an EMS based on continuous-discrete reinforcement learning is introduced, which can output both continuous and discrete variables simultaneously. Finally, the effectiveness of the proposed control strategy is verified by comparing EMS based on charge-depleting charge-sustaining (CD-CS) and Dynamic Programming (DP). The simulation results show that the reinforcement learning EMS can improve energy efficiency by 10.08% compared to the CD-CS EMS, and the fuel economy gap is about 6.4% compared with the benchmark global optimum based on DP. | ['Yuan Lin', 'Jinming Xu', 'Changfu Gong'] | 2023-06-15 | null | null | null | null | ['management', 'energy-management'] | ['miscellaneous', 'time-series'] | [-2.38066807e-01 1.38808146e-01 -4.76997524e-01 9.47525576e-02
-2.63672173e-01 -4.81670946e-01 1.62772581e-01 1.03385620e-01
-3.94458115e-01 1.10047925e+00 -6.48676991e-01 -2.54155219e-01
-6.89832568e-01 -9.28479075e-01 -7.02956200e-01 -1.20796907e+00
7.39564449e-02 5.11005878e-01 6.55304715e-02 -1.78535298e-01
1.66187048e-01 3.75868112e-01 -1.33521008e+00 -6.70204937e-01
1.28142893e+00 1.04166532e+00 7.90004194e-01 2.36706018e-01
4.87869326e-03 2.94247240e-01 -4.05505896e-01 1.60879955e-01
1.62311822e-01 -3.71401638e-01 -2.39591762e-01 -9.45567898e-03
-8.78934503e-01 -2.15111837e-01 -4.94720414e-02 1.15041518e+00
3.83823961e-01 4.92776781e-01 4.94106472e-01 -2.25705338e+00
-8.96735713e-02 2.57698238e-01 -5.95516741e-01 3.37792896e-02
-4.90300208e-01 3.31687182e-01 5.40412545e-01 -4.00648654e-01
1.84990302e-01 9.60229099e-01 -3.13796178e-02 4.24076080e-01
-1.01768959e+00 -8.54551196e-01 1.40108436e-01 5.24820328e-01
-1.32920647e+00 2.70016491e-02 8.58945847e-01 -2.53299236e-01
1.11442327e+00 3.59003365e-01 1.33959401e+00 6.37557507e-02
6.50184751e-01 7.16490149e-01 1.08625603e+00 -2.25604758e-01
5.14648259e-01 2.43574858e-01 -9.93707180e-02 2.61286110e-01
5.51371694e-01 2.30897516e-01 2.89670974e-01 1.94194674e-01
2.83754587e-01 1.65704507e-02 -4.04531322e-03 -5.26494861e-01
-7.01492488e-01 8.07878792e-01 2.56500632e-01 1.38148025e-03
-4.53265458e-01 -4.26043198e-02 2.63854235e-01 2.54181237e-03
7.28359371e-02 5.66825509e-01 -4.29493755e-01 -1.21926665e-01
-6.15799367e-01 4.03891951e-01 7.29377866e-01 1.24536359e+00
5.40401161e-01 4.11989182e-01 -2.33954206e-01 6.66870236e-01
3.77999276e-01 8.26734543e-01 1.31521881e-01 -1.06782305e+00
1.54278591e-01 6.01378918e-01 4.43149149e-01 -5.29675364e-01
-4.94443834e-01 -1.16851196e-01 -8.76021922e-01 4.26123619e-01
-4.12920803e-01 -6.08022153e-01 -5.32296002e-01 1.50624633e+00
3.23660195e-01 -2.17601374e-01 1.81624621e-01 7.11900175e-01
2.69322336e-01 1.41167057e+00 1.44048139e-01 -1.14455521e+00
1.09289420e+00 -9.99693394e-01 -1.42994320e+00 -6.00733198e-02
2.13998765e-01 -2.07425252e-01 1.88714057e-01 9.75235179e-02
-1.53210354e+00 -3.03909332e-01 -1.35785878e+00 3.18947464e-01
-5.15984356e-01 1.72737196e-01 3.20258960e-02 1.65520161e-01
-8.78793180e-01 2.35198185e-01 -6.61114991e-01 4.29262742e-02
-1.30452022e-01 6.48595929e-01 8.35751444e-02 4.00995404e-01
-1.32713759e+00 1.25787437e+00 9.75833833e-01 3.41481745e-01
-3.80565226e-01 -8.29806745e-01 -8.07702422e-01 4.35499661e-02
8.89005363e-01 -3.57155144e-01 1.38404930e+00 -5.09946704e-01
-1.94328725e+00 1.05233498e-01 8.52242708e-02 -1.44317046e-01
4.12245303e-01 2.69956380e-01 -4.15014416e-01 2.36128226e-01
-2.67434895e-01 4.38995093e-01 2.69481242e-01 -1.05387533e+00
-1.00939095e+00 9.26279742e-03 -9.75185856e-02 4.60469007e-01
-1.82549283e-01 -2.43000314e-01 -2.35832203e-02 -3.70700657e-02
-6.55329645e-01 -1.00642562e+00 -1.09737806e-01 -2.39830926e-01
-1.09866194e-01 -8.46625626e-01 1.16802561e+00 -5.79259217e-01
1.29506171e+00 -1.83092368e+00 6.12765133e-01 3.21917266e-01
-1.44024149e-01 2.69722909e-01 4.58933264e-01 4.68489200e-01
1.61728308e-01 1.72284126e-01 -1.10388994e-01 2.05240160e-01
4.13486958e-01 5.08186340e-01 2.27729425e-01 1.69223994e-01
5.03923059e-01 8.81486475e-01 -9.09145355e-01 -5.84294856e-01
5.77210009e-01 9.82727949e-03 -4.13055390e-01 4.96245086e-01
-2.78645128e-01 -1.23567455e-01 -7.12925553e-01 4.15131986e-01
1.05046833e+00 1.52308181e-01 4.34077740e-01 -4.61501122e-01
-9.99878347e-01 -6.15707934e-01 -1.10074306e+00 1.01023579e+00
-4.60599750e-01 4.38398629e-01 7.04388916e-01 -1.45917189e+00
8.54955494e-01 1.32099465e-01 7.78256118e-01 -1.24064338e+00
3.32206249e-01 2.00803474e-01 3.26781394e-03 -5.61992466e-01
6.05729520e-01 -3.43877524e-01 -5.45384996e-02 2.81345733e-02
1.03070717e-02 -4.74067420e-01 5.52707553e-01 -2.17001781e-01
6.58371747e-01 -1.29636452e-01 1.15775444e-01 -8.63403678e-01
8.20590556e-01 3.20832223e-01 1.18047595e+00 -1.07150517e-01
-4.57343571e-02 -6.11764371e-01 6.93983495e-01 3.76219481e-01
-1.27622378e+00 -7.43886650e-01 -1.62614956e-01 2.80255049e-01
9.77703631e-01 2.58702248e-01 -5.01088917e-01 -1.07998408e-01
2.24147305e-01 1.09477210e+00 1.45261630e-01 -5.50713897e-01
-5.86367726e-01 -7.63225853e-01 -3.82729679e-01 5.78615725e-01
6.33037984e-01 -6.48151100e-01 -7.27078974e-01 2.94249207e-01
1.83445245e-01 -1.00345457e+00 -3.67023677e-01 4.66961652e-01
-4.27469939e-01 -9.55829918e-01 -5.28695285e-01 -1.02918243e+00
7.85113871e-01 6.89432546e-02 6.36325836e-01 1.12389922e-01
-1.26313537e-01 7.30510578e-02 -6.16826070e-03 -7.07448661e-01
-2.67536521e-01 -1.97563674e-02 2.28421107e-01 -3.68560046e-01
1.60039634e-01 4.93654571e-02 -5.78807712e-01 5.84198952e-01
-8.29686463e-01 4.86710280e-01 6.82856083e-01 1.11871350e+00
8.97686303e-01 9.56713974e-01 1.06519306e+00 -2.01091036e-01
6.18683398e-01 -5.66674232e-01 -1.35776114e+00 4.02349889e-01
-9.06251431e-01 -5.57229519e-02 9.32615340e-01 -3.14554036e-01
-1.18519044e+00 2.49720477e-02 2.58840416e-02 -5.96588552e-01
5.09199262e-01 2.28010818e-01 -7.02786803e-01 -5.34860007e-02
-8.04188073e-01 4.02443141e-01 5.13113439e-01 -5.77286538e-03
-1.90809563e-01 6.62571609e-01 3.76193941e-01 -6.17442131e-01
8.77472997e-01 -3.79838139e-01 4.81825888e-01 -5.62192619e-01
1.15240254e-01 -1.29034564e-01 -1.42316222e-01 -4.77762640e-01
1.01169395e+00 -7.27635086e-01 -1.51350212e+00 6.80664241e-01
-8.02520096e-01 -5.21820843e-01 -8.69508162e-02 5.56493402e-01
-7.06594408e-01 -1.80755943e-01 -2.77196884e-01 -1.01180470e+00
-2.58742183e-01 -1.31861818e+00 3.78888130e-01 8.82783353e-01
5.25277913e-01 -8.30286086e-01 -1.61836788e-01 4.61738138e-03
4.25474226e-01 4.31336731e-01 1.23737156e+00 -1.14575349e-01
-6.70274794e-01 1.86435312e-01 3.85710336e-02 5.27736962e-01
-1.44718483e-01 1.84386760e-01 -1.42236993e-01 -7.48819232e-01
-7.28233904e-02 -1.81481674e-01 -6.15619496e-02 4.12600607e-01
1.15050447e+00 -5.01994312e-01 -6.70757413e-01 3.25023800e-01
2.04459572e+00 1.28699243e+00 4.40781325e-01 4.78111058e-01
2.59740084e-01 3.59640598e-01 1.18446898e+00 4.17199641e-01
5.53484559e-01 6.56627119e-01 6.92311406e-01 -2.75037438e-01
8.12568843e-01 -8.42005163e-02 1.54880971e-01 1.01345062e+00
1.07371502e-01 -3.71546268e-01 -4.86093581e-01 4.38347757e-01
-1.92126489e+00 -7.91493833e-01 -8.17251205e-02 2.00727487e+00
8.99305761e-01 9.67968330e-02 -8.61831941e-03 3.36988777e-01
9.47882950e-01 -4.24270153e-01 -8.99435043e-01 -8.63891065e-01
1.88658461e-01 -7.36484528e-02 9.13534760e-01 2.72340477e-01
-5.36136150e-01 2.25712322e-02 4.89515400e+00 1.01984680e+00
-1.09356892e+00 -1.38219342e-01 3.82209361e-01 -1.79628745e-01
-2.15182275e-01 -6.34738803e-02 -6.93595052e-01 1.11686385e+00
9.91537750e-01 -1.17399597e+00 7.00362086e-01 5.73098660e-01
5.86528957e-01 -4.72384304e-01 -9.12631333e-01 8.47063661e-01
-3.49319726e-01 -1.07299352e+00 -5.50983846e-01 1.34036496e-01
7.73438811e-01 -5.53880095e-01 -4.32395786e-01 8.10474694e-01
-9.78509039e-02 -6.51823878e-01 7.43835390e-01 6.14718556e-01
5.80941737e-01 -1.39221060e+00 8.51585388e-01 6.78942800e-01
-1.36540973e+00 -5.33554018e-01 -2.68268645e-01 2.28851929e-01
5.04414737e-01 2.76342869e-01 -2.56425232e-01 8.61698687e-01
5.16198277e-01 4.62054104e-01 5.74833527e-02 1.11898935e+00
2.59243578e-01 4.37458344e-02 -3.92808825e-01 -7.14170098e-01
1.26518264e-01 -8.55863035e-01 3.83388847e-01 7.11187661e-01
4.67326909e-01 5.21364510e-01 4.34702545e-01 1.04716003e+00
1.12494983e-01 -1.70186564e-01 -2.48968333e-01 -2.80198425e-01
7.05608249e-01 1.59698164e+00 -4.82056141e-01 -1.89792305e-01
-2.51208574e-01 4.28553313e-01 -3.31574172e-01 2.98499972e-01
-1.36800361e+00 -1.06331348e+00 6.83897614e-01 -1.28653184e-01
1.73116177e-01 -1.11033343e-01 2.40785524e-01 -4.55977112e-01
1.36730792e-02 -1.34935275e-01 -6.46390021e-03 -8.27544630e-01
-1.04743397e+00 -6.56455383e-02 4.59057510e-01 -1.28186703e+00
-2.38383606e-01 -7.29122698e-01 -8.81201327e-01 8.88592720e-01
-2.10655379e+00 -5.18154800e-01 -3.34075272e-01 2.02697247e-01
6.46833837e-01 -8.12403113e-03 1.82391509e-01 7.51946032e-01
-1.21626258e+00 1.89385161e-01 7.03997254e-01 -5.05662441e-01
1.15859278e-01 -1.29395044e+00 -7.42337048e-01 3.87999952e-01
-1.58479655e+00 -1.69747114e-01 7.61810482e-01 -3.50882024e-01
-2.41170764e+00 -9.83074605e-01 3.93998802e-01 6.36487663e-01
6.20567441e-01 -1.09668836e-01 -7.56954610e-01 1.87638506e-01
8.67110789e-01 -9.76547599e-02 -6.83889315e-02 -8.70080888e-01
1.11495161e+00 -5.15864253e-01 -1.38829875e+00 3.23983133e-01
5.21661818e-01 3.61102410e-02 -2.87849486e-01 1.66824579e-01
7.78367579e-01 -6.06231570e-01 -1.09240496e+00 7.70893097e-01
2.71183282e-01 7.16123206e-04 5.89587748e-01 -2.40978211e-01
8.77218917e-02 -5.51446378e-01 3.62212360e-01 -1.74925077e+00
-3.43950540e-01 -6.01772487e-01 -3.04487735e-01 1.45308483e+00
2.83483099e-02 -6.58073306e-01 2.41930172e-01 4.74194050e-01
-4.19488400e-01 -1.15569139e+00 -1.07983959e+00 -1.09271550e+00
2.21849397e-01 2.75814414e-01 5.76439023e-01 7.62317777e-01
2.61924714e-01 1.42754510e-01 -4.55801561e-02 3.34908336e-01
5.83218873e-01 -2.92412974e-02 3.43642235e-01 -9.29684520e-01
2.59762239e-02 -5.34467101e-01 -5.41686937e-02 -5.00817657e-01
5.06991446e-01 -8.49981725e-01 3.75171721e-01 -1.90825546e+00
1.91401139e-01 -2.64569610e-01 -5.12356639e-01 3.02141249e-01
3.76504436e-02 -5.44032216e-01 2.39784792e-01 -2.70846993e-01
-4.79816377e-01 1.20516181e+00 1.45010006e+00 -4.66043383e-01
-2.64668614e-01 2.00674664e-02 -2.92646408e-01 3.34824979e-01
1.00184286e+00 -1.59637600e-01 -7.36260414e-01 -9.08033084e-03
7.73803666e-02 6.82579994e-01 1.23820297e-01 -7.60951459e-01
4.66426790e-01 -8.54909301e-01 3.86778042e-02 -9.83355165e-01
1.77524194e-01 -1.28670335e+00 5.11350751e-01 8.49393070e-01
1.60714462e-01 3.43696266e-01 4.34785575e-01 3.94604027e-01
-1.98105603e-01 -1.76791057e-01 7.56094396e-01 3.38871151e-01
-9.32108343e-01 1.98522836e-01 -5.13058782e-01 -6.07641675e-02
2.00227046e+00 -1.79274321e-01 -5.95482826e-01 1.81226507e-01
-2.33256742e-01 1.49557340e+00 6.93483353e-02 5.30005455e-01
3.61034065e-01 -1.52595735e+00 -2.52729803e-01 3.12778279e-02
-2.23727390e-01 -8.90927240e-02 4.15576398e-01 9.76787865e-01
-3.44434470e-01 5.36292553e-01 -6.91849291e-01 -5.35389781e-01
-1.13385069e+00 5.08765161e-01 4.94720399e-01 -7.69701675e-02
-2.02783480e-01 -3.20295878e-02 -2.05847189e-01 -2.78439194e-01
-9.39793438e-02 -4.26965237e-01 -1.46274403e-01 1.47488087e-01
-6.75446987e-02 9.24784839e-01 -1.14824414e-01 -2.02981487e-01
-3.55681926e-01 5.10731637e-01 5.17374456e-01 1.98274776e-01
1.35868347e+00 -4.55748960e-02 -1.72850937e-01 2.06786036e-01
1.24244308e+00 -5.65755963e-01 -1.25169194e+00 3.33145887e-01
-3.48369122e-01 -2.03219160e-01 4.39035743e-01 -7.49193490e-01
-1.18290615e+00 4.02249306e-01 7.02023983e-01 5.36272347e-01
1.39762342e+00 -4.11548674e-01 8.21791768e-01 2.32083291e-01
4.43150848e-01 -1.92506516e+00 -1.79282978e-01 3.76694679e-01
7.57037997e-01 -8.63619983e-01 9.10099372e-02 -4.77641225e-02
-5.80247879e-01 1.09311759e+00 1.10605681e+00 7.01576704e-03
6.55465722e-01 5.71217120e-01 -7.60206163e-01 3.19214553e-01
-9.20032084e-01 9.09021944e-02 -3.91932726e-02 2.51651146e-02
-2.71640718e-01 3.13471287e-01 -9.58583891e-01 7.64117241e-01
4.24097061e-01 1.73274398e-01 4.81892526e-01 1.29298592e+00
-5.37896395e-01 -9.80407059e-01 -1.93110898e-01 3.62982929e-01
-7.51853436e-02 6.99911416e-01 5.40509820e-01 1.11481297e+00
3.45555067e-01 9.09920990e-01 5.51076293e-01 -2.05581561e-01
7.27137685e-01 -1.91179961e-01 2.22670525e-01 -2.93253697e-02
-8.76191184e-02 -4.89929244e-02 -1.70220062e-01 -1.63561478e-01
-3.46266210e-01 -3.86620462e-01 -1.92470562e+00 -4.60160464e-01
-7.96294153e-01 6.07875824e-01 1.00375187e+00 9.46078300e-01
1.65805027e-01 9.45090294e-01 1.27540660e+00 -6.72598541e-01
-6.81154668e-01 -4.65886384e-01 -8.59616697e-01 -2.91043073e-01
2.75445879e-01 -1.08914840e+00 -5.24570882e-01 -5.15801191e-01] | [5.553229808807373, 2.288217306137085] |
64a01920-346c-41b6-9564-1bcb245de8a8 | enhanced-frame-and-event-based-simulator-and | 2112.09379 | null | https://arxiv.org/abs/2112.09379v1 | https://arxiv.org/pdf/2112.09379v1.pdf | Enhanced Frame and Event-Based Simulator and Event-Based Video Interpolation Network | Fast neuromorphic event-based vision sensors (Dynamic Vision Sensor, DVS) can be combined with slower conventional frame-based sensors to enable higher-quality inter-frame interpolation than traditional methods relying on fixed motion approximations using e.g. optical flow. In this work we present a new, advanced event simulator that can produce realistic scenes recorded by a camera rig with an arbitrary number of sensors located at fixed offsets. It includes a new configurable frame-based image sensor model with realistic image quality reduction effects, and an extended DVS model with more accurate characteristics. We use our simulator to train a novel reconstruction model designed for end-to-end reconstruction of high-fps video. Unlike previously published methods, our method does not require the frame and DVS cameras to have the same optics, positions, or camera resolutions. It is also not limited to objects a fixed distance from the sensor. We show that data generated by our simulator can be used to train our new model, leading to reconstructed images on public datasets of equivalent or better quality than the state of the art. We also show our sensor generalizing to data recorded by real sensors. | ['Kynan Eng', 'Hyunsurk Eric Ryu', 'Paul K. J. Park', 'Chang-Woo Shin', 'Moosung Kwak', 'Minwon Seo', 'Luca Longinotti', 'Chenghan Li', 'Thomas Debrunner', 'Andreas Georgiou', 'Adam Radomski'] | 2021-12-17 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 3.49505723e-01 -2.78984576e-01 7.38442957e-01 -2.80612469e-01
-5.04188359e-01 -5.26526630e-01 3.83895993e-01 -3.89694534e-02
-9.46927667e-01 6.81756556e-01 -7.45154843e-02 2.72119045e-01
2.78036684e-01 -4.93986905e-01 -1.41123986e+00 -4.26550865e-01
9.66086388e-02 1.60447419e-01 9.83808219e-01 1.34899005e-01
2.44456187e-01 5.17112494e-01 -1.98031342e+00 2.54084796e-01
2.67877847e-01 1.03132594e+00 7.28695333e-01 1.09275234e+00
3.91778976e-01 7.51745224e-01 -7.76821136e-01 -6.22027814e-02
4.55055416e-01 -3.52371752e-01 -4.50200588e-02 3.96781377e-02
7.33337522e-01 -9.62771058e-01 -4.98010308e-01 7.77865648e-01
6.72249198e-01 -1.07811846e-01 -3.85404490e-02 -1.04289269e+00
-1.81747675e-01 1.21819004e-01 -3.67570549e-01 3.48182470e-01
6.28177106e-01 5.20331800e-01 1.70422822e-01 -6.17672026e-01
9.77227867e-01 1.06250763e+00 8.26199412e-01 8.86127770e-01
-1.47696042e+00 -5.05572081e-01 -1.13019966e-01 3.04243118e-01
-1.09778416e+00 -6.48349702e-01 3.52853090e-01 -1.51739001e-01
1.09485495e+00 -4.61209845e-03 1.13980687e+00 1.44529831e+00
3.68695587e-01 2.62071133e-01 1.07890594e+00 -2.01642513e-01
6.77493155e-01 -1.79533660e-01 -1.43467421e-02 4.64069098e-01
2.64579833e-01 2.89146692e-01 -9.03018057e-01 -1.60045866e-02
1.30624366e+00 1.10607399e-02 -6.39776647e-01 -2.01605231e-01
-1.52933264e+00 3.42472702e-01 2.34112427e-01 -7.05352575e-02
-4.41516042e-01 7.32138216e-01 1.69650257e-01 7.27505684e-02
-1.46689773e-01 1.25806481e-01 -1.44001782e-01 -5.39693832e-01
-1.04344308e+00 2.14424431e-01 8.60056937e-01 1.05159724e+00
5.35109639e-01 1.82644397e-01 -2.20788836e-01 1.31880000e-01
1.89867392e-01 5.15322685e-01 5.21207452e-01 -1.68558264e+00
-1.65473074e-01 8.77014101e-02 3.82746309e-01 -3.67789596e-01
-3.77540678e-01 -1.36354089e-01 -4.43432480e-01 6.81080699e-01
5.30186713e-01 -8.63489211e-02 -8.23012114e-01 1.62965405e+00
6.42100861e-03 9.37992096e-01 -3.49631980e-02 1.04695237e+00
5.64481735e-01 7.74508893e-01 -3.65720838e-01 -2.61781842e-01
1.16327548e+00 -5.69351137e-01 -6.32876575e-01 5.26350997e-02
2.44558174e-02 -4.55066860e-01 9.96917069e-01 8.34627271e-01
-1.52311385e+00 -4.16654140e-01 -1.29399526e+00 -2.57045805e-01
5.76815009e-02 -1.38751611e-01 4.89091277e-01 6.48136795e-01
-1.53732634e+00 7.85414577e-01 -1.22155845e+00 -2.91515380e-01
3.84904861e-01 5.90018153e-01 -1.40199527e-01 1.28827110e-01
-5.84530056e-01 7.68553019e-01 6.42659664e-02 -2.71928400e-01
-1.36612821e+00 -9.13620830e-01 -5.22593796e-01 -1.25505865e-01
-1.98314115e-01 -9.45600092e-01 1.41970909e+00 -9.97857928e-01
-2.02756047e+00 6.43620133e-01 -1.60241544e-01 -1.04862320e+00
5.11190832e-01 -2.19934005e-02 -3.08534480e-03 5.45918524e-01
-4.39159960e-01 1.00317168e+00 7.17109919e-01 -1.06390369e+00
-2.97212154e-01 -1.17647827e-01 2.58607745e-01 -3.53153911e-03
-2.85532027e-01 2.49482289e-01 -4.58934128e-01 -2.78095454e-01
-4.13130164e-01 -7.09282279e-01 -1.96946099e-01 8.30664039e-01
9.41893011e-02 4.34630752e-01 8.12314272e-01 -3.46603572e-01
4.47355479e-01 -2.01981425e+00 -2.38910336e-02 -4.00113314e-01
2.02346206e-01 4.25040066e-01 -1.28466219e-01 1.59797877e-01
2.21965462e-01 -5.82929850e-01 -2.67109096e-01 -7.34413505e-01
-3.86418879e-01 4.06118572e-01 -2.27276325e-01 6.21263325e-01
1.10679376e-03 5.40745735e-01 -9.37594533e-01 -1.73077732e-01
6.18415475e-01 9.78391171e-01 -7.46873796e-01 1.29493237e-01
-3.29947293e-01 7.45795965e-01 1.36661321e-01 1.97531387e-01
6.84345484e-01 -1.71869934e-01 -3.03819090e-01 -1.73962966e-01
-3.57832402e-01 2.05197006e-01 -1.34025121e+00 2.22923899e+00
-4.01214540e-01 1.06787527e+00 1.52458146e-01 -4.19831127e-01
7.41658270e-01 2.22696617e-01 5.50459921e-01 -7.89009809e-01
1.84416264e-01 4.08343896e-02 -1.81309655e-01 -3.71044278e-01
3.90047014e-01 1.62213624e-01 4.17055070e-01 1.86220646e-01
2.25082800e-01 -9.49036703e-02 4.81296889e-02 1.47829637e-01
1.54193389e+00 4.16655093e-01 -1.72888383e-01 -6.57000691e-02
7.68403336e-02 -1.48060471e-01 5.90043128e-01 6.90961063e-01
-8.78118128e-02 1.14647615e+00 4.77880165e-02 -2.73512334e-01
-1.37646878e+00 -1.34125125e+00 -3.86680305e-01 2.83144474e-01
4.89884108e-01 -5.17404914e-01 -9.61280584e-01 2.29329109e-01
-4.12514210e-01 5.86309254e-01 -2.38794073e-01 5.67752942e-02
-4.11997139e-01 -5.07607460e-01 7.05779910e-01 3.90773267e-01
6.22026622e-01 -8.29504251e-01 -1.67318761e+00 3.40385526e-01
1.33397117e-01 -1.49343622e+00 -2.68803686e-01 -2.50068102e-02
-9.35206056e-01 -9.01254117e-01 -6.96275473e-01 -3.86129111e-01
5.68509281e-01 2.46916339e-01 1.00649917e+00 -3.54669541e-01
-2.70465493e-01 7.85886049e-01 -2.19127968e-01 -3.63360852e-01
-2.40226641e-01 -7.16372848e-01 1.08009353e-01 1.00146988e-02
-6.25917092e-02 -8.64785492e-01 -9.73652184e-01 1.16519094e-01
-1.37120867e+00 3.82990390e-01 4.01693583e-02 4.97310281e-01
8.35445583e-01 -5.17877340e-01 2.14526355e-02 -3.97072256e-01
1.67138413e-01 -1.40826240e-01 -1.09238970e+00 -1.16057336e-01
-2.81743348e-01 1.28547415e-01 9.09227550e-01 -8.26018870e-01
-7.60778069e-01 6.02050900e-01 -1.24223173e-01 -8.47586334e-01
-1.01446845e-01 -1.90562993e-01 1.30970642e-01 -4.83501494e-01
8.78549755e-01 3.16424429e-01 -3.66530567e-02 -1.74562171e-01
2.50395890e-02 3.49424034e-01 9.48930740e-01 -1.07237488e-01
3.44305456e-01 9.98251736e-01 5.14974967e-02 -7.84589827e-01
1.11431301e-01 -1.18125409e-01 -2.08118796e-01 -4.50184107e-01
6.52527392e-01 -1.04621363e+00 -1.05476511e+00 9.54606473e-01
-1.61179543e+00 -5.14943242e-01 -6.22312188e-01 8.02533567e-01
-7.66570985e-01 2.18647286e-01 -7.48460233e-01 -7.24072754e-01
-1.93984851e-01 -1.23914361e+00 1.28247917e+00 5.81770599e-01
1.42128661e-01 -6.82117164e-01 2.98218399e-01 -1.71023682e-01
4.45686728e-01 4.16058689e-01 -1.96280003e-01 1.08711876e-01
-9.90106761e-01 8.53479877e-02 -7.98322558e-02 2.06927314e-01
-3.84901822e-01 2.10635796e-01 -1.16530859e+00 -3.53215158e-01
3.89967263e-01 -1.71608418e-01 7.86249399e-01 6.63740754e-01
1.23993444e+00 -9.24679544e-03 -1.05115898e-01 1.09374750e+00
1.79222763e+00 2.50447154e-01 1.22586048e+00 2.59892821e-01
4.34803754e-01 -3.54719325e-03 5.42435758e-02 7.03529537e-01
5.46098709e-01 9.15819585e-01 6.89623415e-01 1.65646970e-01
-3.31442475e-01 -5.24345674e-02 9.38573241e-01 3.36913526e-01
-1.34684056e-01 -2.45050639e-01 -4.74991232e-01 4.90435451e-01
-1.70317507e+00 -1.00162625e+00 -4.36541051e-01 2.70700979e+00
6.23334706e-01 -1.47407148e-02 1.53447330e-01 -1.65098649e-03
6.84091389e-01 -2.86008358e-01 -8.68402302e-01 -3.21601540e-01
-3.17167699e-01 5.00858247e-01 9.03227329e-01 2.77029961e-01
-4.16007966e-01 4.14216876e-01 6.48379469e+00 2.40604803e-01
-1.19736612e+00 3.54626805e-01 -1.93282198e-02 -7.35097766e-01
-3.29823583e-01 2.60034829e-01 -8.87497306e-01 7.69031763e-01
1.58666432e+00 -2.03845408e-02 7.05887198e-01 3.11018437e-01
4.93968278e-01 -5.11319637e-01 -1.19348621e+00 1.43806493e+00
1.40653089e-01 -1.57478499e+00 -2.46148124e-01 9.84658897e-02
4.16531891e-01 2.94392616e-01 -2.32194737e-01 -3.93735141e-01
2.43209884e-01 -7.10481822e-01 1.03211308e+00 8.86381269e-01
8.16034794e-01 -3.83010864e-01 2.09697828e-01 3.09275240e-01
-8.16449583e-01 1.33043960e-01 -6.35207176e-01 -1.49248853e-01
3.94879669e-01 6.54018104e-01 -4.62822974e-01 1.33134484e-01
9.01833415e-01 8.24939728e-01 -5.79720259e-01 1.48136473e+00
2.58660167e-01 4.81205165e-01 -8.07993829e-01 -1.49245471e-01
-2.40417540e-01 9.13081318e-02 6.62125647e-01 8.47484648e-01
6.16861582e-01 -3.78072225e-02 -4.18228686e-01 9.31536496e-01
5.37014008e-02 -6.33020997e-01 -6.15283310e-01 6.00039542e-01
4.10866171e-01 1.12385368e+00 -4.66235936e-01 -2.27026537e-01
-6.29091322e-01 1.36675811e+00 8.14048648e-02 1.91729099e-01
-1.17123008e+00 -2.35204726e-01 6.35716319e-01 3.35574776e-01
3.37508559e-01 -4.98609334e-01 4.77296039e-02 -1.42902637e+00
1.00500055e-01 -4.64456737e-01 -6.34135902e-02 -1.38726568e+00
-8.41214418e-01 6.47681415e-01 -1.06936425e-01 -1.37223983e+00
-2.86842644e-01 -7.03758180e-01 -4.16443676e-01 6.22969925e-01
-1.46130633e+00 -6.34411097e-01 -6.61084056e-01 9.65883970e-01
3.00463617e-01 2.90842772e-01 8.04582179e-01 3.63860816e-01
-3.03539217e-01 3.60664338e-01 1.95862368e-01 -2.17028916e-01
6.67284310e-01 -1.00086606e+00 3.82125914e-01 1.15479732e+00
8.95893276e-02 6.90969750e-02 8.90471458e-01 -3.02002966e-01
-1.87719095e+00 -9.82680619e-01 1.13030799e-01 -4.07146394e-01
4.75224495e-01 -5.50625384e-01 -6.94744289e-01 6.62726641e-01
2.84032971e-01 4.96442318e-01 2.65473783e-01 -8.96323383e-01
-4.37504239e-02 -4.77560461e-01 -1.39041734e+00 4.43582088e-01
1.20877063e+00 -3.50578815e-01 -1.70979023e-01 2.59027004e-01
6.49577200e-01 -7.75026739e-01 -6.10791564e-01 3.49233113e-02
6.70752823e-01 -1.28942132e+00 9.43668365e-01 2.18815401e-01
3.67050111e-01 -6.37297928e-01 -1.41032100e-01 -1.08897841e+00
1.08580269e-01 -8.53164613e-01 -3.06670278e-01 9.42911565e-01
-1.48766652e-01 -6.90690339e-01 5.85702479e-01 5.33852339e-01
-2.81928331e-01 -2.04400301e-01 -1.30329347e+00 -9.03190851e-01
-4.81864750e-01 -5.49208343e-01 1.89588413e-01 9.33032185e-02
-2.48427361e-01 -2.15880740e-02 -2.96206683e-01 3.14143926e-01
8.73092294e-01 -2.84526855e-01 5.89645207e-01 -1.09361076e+00
-6.91455126e-01 -1.08659640e-02 -9.84779656e-01 -1.19614220e+00
-3.12048942e-01 -4.59083676e-01 7.18903467e-02 -1.32711673e+00
5.45997620e-02 7.77440071e-02 -4.16487968e-03 4.74041477e-02
3.98152679e-01 4.26254094e-01 1.08232833e-01 1.32394508e-01
-6.43086374e-01 4.75685090e-01 1.08859420e+00 4.37567383e-01
-1.65074214e-01 -3.51647168e-01 -1.91293269e-01 6.25538588e-01
4.29770023e-01 -4.17250305e-01 -4.40603673e-01 -8.14375997e-01
2.98734874e-01 3.38475406e-01 1.08080781e+00 -1.88714969e+00
7.96000004e-01 3.28704447e-01 5.36589026e-01 -2.12126523e-01
7.17342615e-01 -9.77712512e-01 7.64288843e-01 5.64317286e-01
-1.87394395e-01 3.37525725e-01 4.56352353e-01 5.89422047e-01
2.80774117e-01 -3.26487690e-01 9.41361547e-01 -1.66155517e-01
-6.08650506e-01 1.75436810e-01 -6.18739009e-01 -3.23253393e-01
1.19079125e+00 -4.14143950e-01 -5.95350266e-01 -4.11710799e-01
-4.05932546e-01 -2.58229941e-01 1.05006945e+00 5.33471182e-02
9.46733177e-01 -1.03106558e+00 -5.30052841e-01 2.06041902e-01
-1.15225159e-01 1.23156540e-01 1.11832656e-01 6.42576694e-01
-9.65670586e-01 -3.50623354e-02 -5.68833053e-01 -9.55420136e-01
-1.03484070e+00 4.75725532e-01 5.62521815e-01 2.57524133e-01
-9.41221297e-01 5.09999514e-01 -1.83535784e-01 1.95660338e-01
2.82321274e-01 -6.21511877e-01 3.15172195e-01 -4.41538215e-01
9.23001051e-01 2.27645516e-01 1.38521314e-01 -2.05763072e-01
-3.09277683e-01 5.84010601e-01 3.33839417e-01 -6.43158972e-01
1.48654163e+00 -1.05332427e-01 2.57584423e-01 6.84788287e-01
7.33538866e-01 -1.50130719e-01 -2.17136216e+00 2.10002884e-01
-7.18732715e-01 -4.11573440e-01 2.56913573e-01 -5.25909543e-01
-1.11233795e+00 8.48357439e-01 1.16161835e+00 -1.14814326e-01
1.46608734e+00 -3.03335130e-01 1.00577986e+00 2.65361518e-01
1.00000441e+00 -8.05492818e-01 -1.04739137e-01 2.55442441e-01
4.98377472e-01 -7.23298728e-01 -3.45305353e-01 -9.80090499e-02
-2.00322852e-01 1.17343700e+00 3.93251896e-01 -6.45219088e-01
2.76713908e-01 9.58401501e-01 -3.77538472e-01 1.71092033e-01
-1.02733409e+00 -5.35883196e-02 -4.08279181e-01 9.29098666e-01
-2.17825770e-02 -4.02345210e-01 -1.15371622e-01 1.83520362e-01
9.24814958e-03 6.85586333e-01 1.22179663e+00 7.53535271e-01
-2.09449172e-01 -9.14568901e-01 -3.72698694e-01 2.70405978e-01
-2.97105849e-01 -1.27203524e-01 5.26972972e-02 2.95427859e-01
1.60023034e-01 8.26320529e-01 4.89598483e-01 -3.63720894e-01
3.33227187e-01 -3.65295678e-01 1.14427471e+00 -2.62351394e-01
-6.73218906e-01 -2.89238602e-01 -2.66058475e-01 -1.01229715e+00
-5.98731756e-01 -8.15269232e-01 -1.45175338e+00 -4.59210724e-01
3.39074470e-02 -4.76612002e-01 9.81285393e-01 5.73807359e-01
7.96003282e-01 5.97249985e-01 4.19239908e-01 -1.24010861e+00
-1.65960938e-02 -5.54047167e-01 -4.04900730e-01 2.25513965e-01
5.07670760e-01 -3.69761348e-01 -2.56472349e-01 4.79192764e-01] | [8.69745922088623, -1.2734531164169312] |
31644622-0056-4c1b-ad7d-72690f7f11d3 | outfit-compatibility-prediction-and-diagnosis | 1907.11496 | null | https://arxiv.org/abs/1907.11496v2 | https://arxiv.org/pdf/1907.11496v2.pdf | Outfit Compatibility Prediction and Diagnosis with Multi-Layered Comparison Network | Existing works about fashion outfit compatibility focus on predicting the overall compatibility of a set of fashion items with their information from different modalities. However, there are few works explore how to explain the prediction, which limits the persuasiveness and effectiveness of the model. In this work, we propose an approach to not only predict but also diagnose the outfit compatibility. We introduce an end-to-end framework for this goal, which features for: (1) The overall compatibility is learned from all type-specified pairwise similarities between items, and the backpropagation gradients are used to diagnose the incompatible factors. (2) We leverage the hierarchy of CNN and compare the features at different layers to take into account the compatibilities of different aspects from the low level (such as color, texture) to the high level (such as style). To support the proposed method, we build a new type-specified outfit dataset named Polyvore-T based on Polyvore dataset. We compare our method with the prior state-of-the-art in two tasks: outfit compatibility prediction and fill-in-the-blank. Experiments show that our approach has advantages in both prediction performance and diagnosis ability. | ['Bo Wu', 'Yun Ye', 'Yueqi Zhong', 'Xin Wang'] | 2019-07-26 | null | null | null | null | ['fashion-compatibility-learning'] | ['computer-vision'] | [-2.88607106e-02 -2.18572810e-01 -4.35572922e-01 -9.06765521e-01
-7.52973408e-02 -2.92078972e-01 3.39825749e-01 1.19630978e-01
-9.40142572e-02 2.08278716e-01 5.67288518e-01 1.71364933e-01
-3.08610439e-01 -7.73379147e-01 -7.72728086e-01 -3.74666154e-01
4.06561017e-01 2.12518468e-01 -8.63791630e-02 -4.03667122e-01
1.98628306e-01 -1.90413892e-01 -1.47580874e+00 7.58640110e-01
1.10844111e+00 1.46772480e+00 1.73590675e-01 3.62515040e-02
-2.74885055e-02 8.25547099e-01 -1.65499374e-02 -7.99872518e-01
1.80566370e-01 -4.49638784e-01 -3.77015710e-01 1.89079911e-01
7.69259632e-01 -4.06834304e-01 -1.77424639e-01 1.18021011e+00
2.68201470e-01 -2.98394024e-01 6.01040483e-01 -1.22275043e+00
-1.44307292e+00 8.47844481e-01 -5.66221178e-01 -1.14583947e-01
2.50865191e-01 2.72588968e-01 1.20601344e+00 -8.34798813e-01
4.84320432e-01 1.23462093e+00 6.39964938e-01 2.39887759e-01
-9.87640321e-01 -7.59663641e-01 7.60246575e-01 4.64581788e-01
-9.48886454e-01 -6.86449558e-02 1.12513602e+00 -6.05098665e-01
3.34338456e-01 3.92117083e-01 1.11384165e+00 1.21294689e+00
2.98463047e-01 1.10168684e+00 1.64788783e+00 -1.25231683e-01
5.68209216e-02 1.21282727e-01 4.60948735e-01 8.32250893e-01
2.30791643e-02 1.06916940e-02 -6.03515983e-01 2.39722386e-01
8.28385592e-01 9.17204916e-02 3.69372964e-02 -2.06310853e-01
-1.15157115e+00 6.34838581e-01 8.65255773e-01 8.14388841e-02
-3.16717833e-01 -9.92927328e-02 1.04868896e-01 3.50682497e-01
4.90215391e-01 5.00735939e-01 -5.55320859e-01 2.89474607e-01
-3.83163959e-01 4.95675087e-01 5.67694366e-01 9.47631419e-01
3.89253139e-01 -4.14603502e-01 -5.67034066e-01 9.95703757e-01
6.31647408e-01 3.86695087e-01 2.10891798e-01 -6.30145013e-01
5.60855865e-01 7.11021781e-01 -4.52427194e-02 -1.36240184e+00
-5.89146733e-01 -7.96126485e-01 -9.95040536e-01 -4.46795858e-02
3.10307831e-01 1.08824603e-01 -7.55327761e-01 2.11163855e+00
2.95038819e-01 3.22503150e-02 -4.58969176e-01 1.31631231e+00
1.13631201e+00 2.70310789e-01 2.90511519e-01 1.23094721e-02
1.62858176e+00 -1.30094290e+00 -8.32913101e-01 -3.36520582e-01
3.32746714e-01 -8.63526642e-01 1.18099678e+00 5.46824515e-01
-1.15681469e+00 -1.03535461e+00 -1.11544585e+00 -2.40757823e-01
-4.06172901e-01 6.17185652e-01 9.21866238e-01 4.02885526e-01
-8.07450652e-01 6.50299966e-01 -3.45359862e-01 -1.73782393e-01
4.29850042e-01 4.10102367e-01 -2.88638622e-01 -2.09229141e-01
-1.39418387e+00 9.18884695e-01 2.09437519e-01 4.63470459e-01
-3.55934590e-01 -5.70121109e-01 -8.10235739e-01 1.10111848e-01
3.51177275e-01 -1.18123269e+00 7.56466150e-01 -1.31077302e+00
-1.32332289e+00 8.68727744e-01 2.56113946e-01 3.63492638e-01
4.74896908e-01 -4.47368383e-01 -5.85489929e-01 -4.72311169e-01
1.72660127e-01 7.92511761e-01 5.83695531e-01 -1.16726649e+00
-6.58162355e-01 -5.25055289e-01 4.62698340e-01 4.76855040e-01
-5.95315695e-01 -3.58783752e-02 -6.68792069e-01 -8.48383248e-01
2.05250114e-01 -1.09440970e+00 -7.97790959e-02 3.69126111e-01
-6.89277291e-01 -2.38044068e-01 7.92922974e-02 -9.19111252e-01
1.08422363e+00 -2.15416861e+00 3.82677644e-01 1.70686632e-01
2.69902408e-01 -8.95819664e-02 -1.64696053e-01 4.47632074e-02
7.25706741e-02 -1.10455327e-01 2.89428264e-01 -3.89100641e-01
2.62482703e-01 1.59305206e-03 -3.31599005e-02 1.96903735e-01
1.49153844e-01 9.72471774e-01 -7.40670145e-01 -3.26080054e-01
1.85329869e-01 4.20014471e-01 -8.23508680e-01 2.91979402e-01
-2.28747338e-01 6.13530040e-01 -6.12177193e-01 6.60009563e-01
8.52426171e-01 -2.35875130e-01 2.74141669e-01 -8.55173647e-01
1.53776824e-01 2.89114684e-01 -1.12199461e+00 1.76282930e+00
-2.70341009e-01 -3.39293256e-02 -2.90975928e-01 -4.91699308e-01
9.32776809e-01 -2.04652756e-01 1.81278050e-01 -9.97683108e-01
1.81574628e-01 1.74207360e-01 1.59273863e-01 -7.44650364e-01
4.52518731e-01 4.57946062e-02 -1.48319438e-01 1.27605647e-01
1.16451427e-01 4.70485568e-01 -4.13800776e-02 -5.38927875e-02
4.58707958e-01 4.22914922e-01 1.48740575e-01 -3.47181141e-01
2.54748821e-01 -2.47422427e-01 6.76114082e-01 6.00668728e-01
-1.84914768e-01 4.21174049e-01 5.34306049e-01 -5.00570953e-01
-8.26333940e-01 -8.86569738e-01 1.39359504e-01 1.19039655e+00
6.54741764e-01 -5.69303870e-01 -6.02599442e-01 -7.76910245e-01
4.25778702e-02 4.79035735e-01 -1.09640956e+00 -3.82576913e-01
-4.29630995e-01 -7.38142610e-01 5.31203151e-02 1.04641163e+00
7.90431261e-01 -8.22439313e-01 1.52192071e-01 9.21114534e-03
-2.28964910e-01 -1.07602859e+00 -5.36980927e-01 -8.42713937e-02
-6.73105478e-01 -7.97900677e-01 -3.25258106e-01 -9.61026430e-01
7.00727284e-01 1.93747133e-01 1.22827923e+00 2.07313791e-01
2.44220495e-01 -2.33282149e-01 -7.93656409e-02 -1.87878042e-01
5.57287261e-02 -1.37128189e-01 1.03683159e-01 3.18124294e-01
4.01865512e-01 -4.90777850e-01 -8.94137681e-01 5.74007154e-01
-5.48736095e-01 7.73472011e-01 6.47578835e-01 7.59453714e-01
7.14451313e-01 1.98158734e-02 3.02319318e-01 -9.31160390e-01
6.28005445e-01 -4.98272568e-01 1.01023003e-01 4.67147768e-01
-5.00579357e-01 -2.24906489e-01 5.96936941e-01 -5.79421461e-01
-1.10279131e+00 -3.13623436e-02 -3.71217698e-01 -3.76630217e-01
-2.39477143e-01 6.52966380e-01 -6.09290361e-01 1.79357544e-01
1.89223260e-01 -9.77745559e-03 -4.09251094e-01 -6.79822028e-01
5.48998415e-01 3.42697710e-01 4.59595770e-01 -6.68830991e-01
4.48237598e-01 2.36791700e-01 -2.31561795e-01 -2.72547044e-02
-1.47971010e+00 7.42086163e-03 -5.76353431e-01 -3.92889649e-01
8.60776842e-01 -1.11975837e+00 -7.48428822e-01 6.09773457e-01
-8.11694980e-01 -2.89222188e-02 2.08321847e-02 4.83530521e-01
-3.85437489e-01 -5.33553027e-02 -1.01850331e+00 -3.80725980e-01
-2.32371002e-01 -1.22642314e+00 9.81963634e-01 4.10840124e-01
-3.60652357e-01 -9.18871164e-01 -2.10859254e-01 7.29726255e-01
5.07099867e-01 -6.34675473e-03 1.28336334e+00 -5.73655307e-01
-2.69386917e-01 -1.41671579e-02 -4.17645305e-01 1.22551270e-01
6.71386644e-02 -2.46281072e-01 -9.81756032e-01 1.26202390e-01
4.50533107e-02 -4.83604848e-01 1.07249129e+00 5.03609538e-01
1.48247159e+00 -1.98220730e-01 -1.61414310e-01 6.51154220e-01
1.22227728e+00 -9.96133462e-02 6.56377554e-01 3.23217720e-01
1.13201773e+00 7.64163852e-01 8.40499043e-01 2.27762192e-01
1.01498866e+00 9.45996463e-01 3.36043954e-01 -4.17432129e-01
-3.55927378e-01 -6.50324345e-01 3.21548074e-01 1.05217195e+00
-2.06368417e-01 -9.15706232e-02 -1.31822124e-01 1.60573289e-01
-2.23591995e+00 -5.88474393e-01 -3.23943287e-01 1.72634125e+00
8.62300336e-01 2.85370320e-01 1.38321862e-01 -1.60075590e-01
6.89058721e-01 1.58197191e-02 -5.85113168e-01 -3.22370678e-01
-2.38280371e-01 -3.58278692e-01 -1.35232121e-01 2.20436513e-01
-1.28391993e+00 8.65363717e-01 5.55899620e+00 8.67873728e-01
-1.03516054e+00 -6.13595312e-03 1.05399978e+00 6.74919114e-02
-6.73840642e-01 -7.42667094e-02 -6.98597550e-01 6.24054193e-01
3.21408473e-02 6.30455792e-01 4.08687502e-01 6.04064286e-01
-3.98692675e-02 1.23152636e-01 -1.37413216e+00 8.18092167e-01
3.45252395e-01 -8.66747200e-01 4.72594053e-02 -2.15685904e-01
5.62961638e-01 -5.97869098e-01 4.54129368e-01 5.47791004e-01
2.11705342e-01 -9.19522822e-01 1.07675231e+00 7.27819383e-01
5.82605004e-01 -4.62913394e-01 6.51463151e-01 2.24863529e-01
-1.07847095e+00 -5.61627559e-02 -2.97148287e-01 -3.83988500e-01
-2.00369097e-02 6.92419887e-01 -1.76621303e-01 5.53000748e-01
6.98711157e-01 9.99122262e-01 -9.23670113e-01 8.05500686e-01
-5.59518397e-01 3.78508538e-01 -7.11331889e-02 -4.40467075e-02
-4.26343121e-02 -2.05362737e-01 -4.77029681e-02 9.36878085e-01
2.96129167e-01 5.80179552e-03 5.24131954e-01 1.08793628e+00
-1.95323061e-02 2.74221510e-01 -9.68917310e-02 1.91721618e-01
1.87275603e-01 1.34730256e+00 -5.87483108e-01 -3.37440133e-01
-5.02936423e-01 9.58990276e-01 5.85854352e-01 4.27128561e-02
-9.88419294e-01 3.47105652e-01 5.47648549e-01 -7.80242449e-03
3.80349576e-01 9.64714363e-02 -8.50987434e-01 -1.35200250e+00
1.25336498e-01 -8.52072179e-01 5.44641435e-01 -1.05177367e+00
-2.11148190e+00 4.40952957e-01 -2.38397792e-01 -1.25463331e+00
3.38627636e-01 -7.22705126e-01 -4.96950001e-01 7.57862866e-01
-1.35964274e+00 -1.94415808e+00 -4.66859668e-01 5.23244917e-01
4.42369759e-01 -4.16334793e-02 6.84540272e-01 3.73097599e-01
-5.97152650e-01 7.34917045e-01 -2.95352459e-01 1.83710486e-01
8.12349260e-01 -1.16733301e+00 2.67388642e-01 4.38640147e-01
-3.77058797e-02 8.06703687e-01 6.57994509e-01 -9.27994549e-01
-1.43504274e+00 -8.99155855e-01 1.00635564e+00 -2.16656685e-01
5.85404754e-01 -4.61843729e-01 -5.06283462e-01 4.87456083e-01
1.71663821e-01 -6.30149171e-02 1.03393948e+00 7.98090219e-01
-7.69189894e-01 -2.26132020e-01 -8.94748688e-01 8.43183398e-01
1.29574883e+00 -2.23414034e-01 -4.42958891e-01 2.61300236e-01
4.65188175e-01 -4.28064406e-01 -1.04310191e+00 4.63709921e-01
1.10976648e+00 -1.07800114e+00 8.64561677e-01 -6.90538943e-01
1.26548564e+00 -2.38701805e-01 -3.97619873e-01 -1.50334513e+00
-1.13131881e+00 8.72293562e-02 -4.94870991e-02 1.48283088e+00
3.74438375e-01 -2.35461041e-01 3.58586222e-01 8.18945885e-01
-2.27058515e-01 -1.26371276e+00 -4.06901449e-01 -1.78038925e-01
-1.03753641e-01 -3.41116518e-01 7.74751008e-01 1.21296346e+00
1.24604180e-02 6.75120175e-01 -9.74267304e-01 8.45994055e-02
4.26613212e-01 6.50345504e-01 4.03436273e-01 -1.10014307e+00
-6.52132332e-01 -4.24230129e-01 -2.61942267e-01 -1.05962217e+00
-1.56720325e-01 -9.40109253e-01 -2.17617124e-01 -1.51515508e+00
7.65835941e-01 -7.07416534e-01 -7.65751362e-01 5.62411904e-01
-6.59954786e-01 3.08213353e-01 4.60829496e-01 1.74205065e-01
-8.46819043e-01 5.36737740e-01 1.65782034e+00 -3.07250947e-01
1.35922343e-01 -3.20079952e-01 -1.35237420e+00 9.07280922e-01
5.91063261e-01 -8.72486979e-02 -4.47524577e-01 -9.02844608e-01
5.88006079e-01 -7.58803915e-03 4.13162172e-01 -8.31971705e-01
3.85407847e-03 -3.88258964e-01 1.03641307e+00 -3.72284442e-01
4.15676355e-01 -7.77176976e-01 9.93341058e-02 2.45702699e-01
-9.20932591e-01 8.94742981e-02 -1.73595399e-01 4.96050209e-01
5.36608957e-02 2.06594303e-01 4.06976283e-01 -4.01385501e-02
-7.23505914e-01 4.19342875e-01 2.01154515e-01 -2.54464805e-01
6.26093864e-01 1.77284122e-01 -6.36877060e-01 -2.87434280e-01
-7.04895020e-01 2.39215076e-01 3.75192910e-01 8.68233979e-01
6.07249856e-01 -1.88616490e+00 -6.47823155e-01 3.94706726e-01
3.64738107e-01 -5.34953892e-01 6.45403743e-01 9.29186046e-01
3.17858756e-02 -2.28860211e-02 -6.60406291e-01 -5.46725571e-01
-1.16046715e+00 7.53874123e-01 1.08014613e-01 -3.99632156e-01
-3.58708769e-01 9.95342135e-01 7.19399214e-01 -5.62254727e-01
1.34284094e-01 -2.68779218e-01 -6.88422084e-01 -3.50925745e-03
4.74189699e-01 -2.22461089e-01 -1.99310526e-01 -5.47410429e-01
-2.67058671e-01 7.20517159e-01 -4.01309758e-01 2.37177178e-01
1.28616846e+00 -1.84364334e-01 -1.00983217e-01 5.55225790e-01
8.78663480e-01 -3.69029976e-02 -1.23581064e+00 -3.29557598e-01
-4.20198560e-01 -7.50911772e-01 4.24575545e-02 -1.39532065e+00
-1.41334212e+00 7.24425554e-01 8.53899717e-01 -9.94186848e-02
1.21335256e+00 3.02020889e-02 9.88953352e-01 -9.01886448e-02
2.19269350e-01 -1.11523592e+00 3.16467136e-02 2.09862232e-01
9.20597792e-01 -1.27972853e+00 6.65431693e-02 -8.88846219e-01
-1.00670481e+00 8.27792823e-01 1.09244013e+00 -2.43164629e-01
7.20401585e-01 1.28618062e-01 3.67401615e-02 -2.20087022e-01
-8.12783062e-01 -7.22310022e-02 7.35365748e-01 3.56404632e-01
7.37831056e-01 4.08970594e-01 -4.55928862e-01 1.38796008e+00
-1.68718681e-01 2.16049090e-01 -1.18641526e-01 5.91251135e-01
-1.59484446e-01 -9.83713627e-01 -1.03147015e-01 6.20501101e-01
-2.02048510e-01 -1.40660584e-01 -6.04871213e-01 3.02362978e-01
8.05676043e-01 9.10605967e-01 -9.26692933e-02 -9.49990988e-01
4.45404679e-01 -4.21801269e-01 7.25858331e-01 -1.31067634e-01
-8.42355490e-01 2.59707481e-01 4.86679077e-01 -6.34096086e-01
-4.91369069e-01 -4.06079918e-01 -6.76574528e-01 -2.90717244e-01
-2.43126139e-01 -5.15112042e-01 5.89228868e-01 1.21843290e+00
2.98349768e-01 6.29176855e-01 5.35287499e-01 -6.20271027e-01
-4.34912622e-01 -8.56232047e-01 -7.05063939e-01 1.07629395e+00
-1.69937924e-01 -8.99778903e-01 3.10924768e-01 -1.96371630e-01] | [11.02662467956543, 0.09846800565719604] |
9ab69622-f061-472e-9708-cd3011425d4e | key-information-extraction-from-documents | 2106.14624 | null | https://arxiv.org/abs/2106.14624v1 | https://arxiv.org/pdf/2106.14624v1.pdf | Key Information Extraction From Documents: Evaluation And Generator | Extracting information from documents usually relies on natural language processing methods working on one-dimensional sequences of text. In some cases, for example, for the extraction of key information from semi-structured documents, such as invoice-documents, spatial and formatting information of text are crucial to understand the contextual meaning. Convolutional neural networks are already common in computer vision models to process and extract relationships in multidimensional data. Therefore, natural language processing models have already been combined with computer vision models in the past, to benefit from e.g. positional information and to improve performance of these key information extraction models. Existing models were either trained on unpublished data sets or on an annotated collection of receipts, which did not focus on PDF-like documents. Hence, in this research project a template-based document generator was created to compare state-of-the-art models for information extraction. An existing information extraction model "Chargrid" (Katti et al., 2019) was reconstructed and the impact of a bounding box regression decoder, as well as the impact of an NLP pre-processing step was evaluated for information extraction from documents. The results have shown that NLP based pre-processing is beneficial for model performance. However, the use of a bounding box regression decoder increases the model performance only for fields that do not follow a rectangular shape. | ['Constantin Spille', 'Mirela Popa', 'Oliver Bensch'] | 2021-06-09 | null | null | null | null | ['key-information-extraction'] | ['natural-language-processing'] | [ 1.70409337e-01 4.65464830e-01 6.04287051e-02 -4.10603702e-01
-4.10083413e-01 -7.99755335e-01 1.00137711e+00 8.39083254e-01
-6.30850732e-01 7.72829711e-01 3.97032887e-01 -5.82319438e-01
-3.25637490e-01 -8.80086243e-01 -7.08586752e-01 -1.41948491e-01
-9.04601216e-02 6.13135934e-01 1.15851037e-01 -1.77679121e-01
5.52691698e-01 8.31641912e-01 -1.44721210e+00 8.66097510e-01
5.95926702e-01 7.14458227e-01 5.20701706e-01 7.97055960e-01
-8.84192407e-01 4.96021956e-01 -7.89097846e-01 -1.46384433e-01
1.91498026e-01 -7.88333490e-02 -8.49663436e-01 -1.38706952e-01
1.34971172e-01 -3.89945030e-01 -4.85132858e-02 7.18210876e-01
3.29722643e-01 -4.88859802e-01 8.89687061e-01 -7.89348483e-01
-2.96065688e-01 7.77467191e-01 -3.41097355e-01 9.55862850e-02
6.47080183e-01 -1.70626357e-01 9.07672763e-01 -7.95803726e-01
1.08396196e+00 1.16471195e+00 6.90205157e-01 -5.27716801e-02
-1.14512074e+00 -3.29172969e-01 -1.46263689e-01 -4.39157570e-03
-1.00018024e+00 -1.38019338e-01 5.42616785e-01 -6.18741512e-01
1.42201030e+00 1.38116732e-01 5.71994245e-01 9.01841879e-01
3.82505655e-01 9.10516262e-01 9.93723571e-01 -9.48895931e-01
5.27415890e-03 4.27957773e-01 2.00647160e-01 1.79939210e-01
5.23389339e-01 -8.20529014e-02 -2.12639049e-01 3.33537221e-01
4.38001394e-01 -5.01787126e-01 1.35059491e-01 -2.31413677e-01
-8.98693025e-01 7.37814069e-01 1.35801658e-01 8.43746781e-01
-5.24318635e-01 -3.86501372e-01 5.35430253e-01 1.71050638e-01
3.90326828e-01 7.90229797e-01 -4.99460101e-01 -2.45563284e-01
-1.41372204e+00 4.88460600e-01 1.07220852e+00 8.82430911e-01
6.00722134e-01 -3.81154329e-01 -3.63794357e-01 4.53506649e-01
4.06937271e-01 3.71909775e-02 4.12762374e-01 -1.48813695e-01
9.20755923e-01 1.12989008e+00 1.95480943e-01 -1.18269527e+00
-1.00770879e+00 -1.53812125e-01 -7.70658791e-01 1.07079700e-01
5.88964999e-01 -2.82609761e-01 -1.25036645e+00 9.71030951e-01
2.18690690e-02 -6.47444606e-01 9.21389461e-02 5.83561599e-01
6.69501781e-01 6.53949142e-01 -1.04284957e-01 4.13077585e-02
1.17626512e+00 -2.04514116e-01 -8.65565538e-01 -8.02443475e-02
8.40664387e-01 -1.01700497e+00 7.01337337e-01 5.71789324e-01
-8.28874886e-01 -5.32587409e-01 -1.21934938e+00 -3.28473747e-01
-1.11648655e+00 3.95553261e-01 4.95877504e-01 7.17913866e-01
-7.82092154e-01 7.10870683e-01 -4.83697116e-01 -3.54721785e-01
5.39530277e-01 3.53523880e-01 -4.42842185e-01 -1.17125854e-01
-1.37612605e+00 1.08921051e+00 7.83451557e-01 1.43112242e-01
-3.40501070e-01 -6.01397097e-01 -7.30290830e-01 1.66662052e-01
2.83671856e-01 -2.38355294e-01 1.00701237e+00 -7.22633004e-01
-1.01313782e+00 7.08820641e-01 1.22560542e-02 -7.82517374e-01
6.61801934e-01 -3.72506499e-01 -3.33352625e-01 1.44619912e-01
1.81106701e-01 6.09875739e-01 7.35975385e-01 -1.18522620e+00
-6.77084565e-01 -4.06141043e-01 -9.58125442e-02 -7.25694746e-02
-2.50726998e-01 2.81231254e-01 -3.08214247e-01 -4.62960035e-01
-2.13205367e-01 -8.34381759e-01 -8.50770622e-02 -4.14999515e-01
-6.05160952e-01 -9.44748335e-03 8.96433532e-01 -1.16131282e+00
1.32713223e+00 -1.85497141e+00 -2.28616595e-01 4.20339912e-01
-7.73002282e-02 5.15383422e-01 7.41312951e-02 6.00924909e-01
-1.85175419e-01 4.09787625e-01 -1.29070282e-01 -7.55081549e-02
6.75445274e-02 -1.20401895e-02 -2.06065521e-01 -7.46273994e-03
6.23682916e-01 7.06859469e-01 -4.54375774e-01 -6.46969199e-01
2.23093152e-01 5.11635602e-01 -3.80873710e-01 4.18266132e-02
-5.70545852e-01 2.36289188e-01 -3.10016066e-01 2.93360770e-01
6.97860241e-01 1.19175866e-01 1.75637528e-01 -1.11395344e-01
-5.24694145e-01 6.12836719e-01 -1.35710168e+00 1.71248806e+00
-3.53228867e-01 1.06209743e+00 -1.01240441e-01 -9.11978126e-01
1.05755341e+00 3.93370479e-01 5.97471356e-01 -9.33700264e-01
1.23059414e-01 3.39702368e-02 9.75919291e-02 -5.55980027e-01
9.46121573e-01 2.44910553e-01 -5.31138405e-02 4.54277039e-01
-3.05426233e-02 -2.72084355e-01 7.52611935e-01 3.09523791e-01
1.04806626e+00 5.42039633e-01 3.58603716e-01 -2.61248529e-01
5.05242765e-01 4.02288407e-01 7.49770477e-02 5.76078892e-01
4.69063193e-01 8.20802212e-01 7.87264705e-01 -4.32689399e-01
-1.34886265e+00 -3.64541888e-01 -2.47018144e-01 6.53665781e-01
-5.05593061e-01 -6.35799885e-01 -9.13632095e-01 -5.36269903e-01
-6.64827973e-02 1.02687705e+00 -4.87221777e-01 2.46547863e-01
-5.98296463e-01 -6.39514089e-01 6.41541064e-01 3.97802114e-01
2.53325075e-01 -1.15066397e+00 -7.61581957e-01 5.68003476e-01
1.84265196e-01 -1.17817867e+00 3.12054396e-01 5.87992489e-01
-7.28239119e-01 -1.02033722e+00 -6.51190281e-01 -3.68350416e-01
5.35833776e-01 -2.29489803e-01 1.04460990e+00 -8.12594220e-02
-4.45130020e-01 6.21142611e-03 -5.18898249e-01 -1.06685889e+00
-6.63496137e-01 5.69294095e-01 -2.41937280e-01 -3.09454113e-01
7.66572773e-01 -1.96983531e-01 -5.84210791e-02 -8.10233578e-02
-1.35010910e+00 2.04989910e-01 7.87486911e-01 4.76683050e-01
2.58122236e-01 1.75103754e-01 4.39907253e-01 -8.18793774e-01
1.12443531e+00 -2.56300539e-01 -5.78887939e-01 1.85965821e-01
-7.49740422e-01 5.63516974e-01 7.33883381e-01 -1.53768852e-01
-1.26078129e+00 1.06315151e-01 -1.83794677e-01 1.79557011e-01
-6.67225540e-01 8.18286657e-01 -3.15064847e-01 6.48086429e-01
7.50330806e-01 9.56296101e-02 -1.29101276e-01 -6.26255929e-01
2.94783026e-01 1.04047000e+00 1.07394136e-01 -4.65063602e-01
6.67155206e-01 1.47863925e-01 7.89297074e-02 -1.06443739e+00
-4.64137793e-01 -4.38143104e-01 -1.32343829e+00 7.12233707e-02
8.62088799e-01 -4.32907194e-01 -2.04046071e-01 2.81644315e-01
-1.59104466e+00 -4.50098254e-02 -2.56005019e-01 4.78569806e-01
-2.96090275e-01 1.77705005e-01 -3.36347133e-01 -7.13860333e-01
-3.83808464e-01 -9.25273180e-01 9.59344089e-01 2.60253459e-01
-5.71798801e-01 -7.98999548e-01 8.56831744e-02 1.12656616e-01
3.48817557e-01 1.49316013e-01 1.26680171e+00 -1.08248031e+00
-4.92389858e-01 -5.44890463e-01 -4.91391242e-01 1.82337552e-01
6.03340529e-02 1.12021610e-01 -7.29487360e-01 2.39899248e-01
-3.66030961e-01 1.43095972e-02 7.23680139e-01 1.47855029e-01
7.59015560e-01 -2.98042834e-01 -4.94628459e-01 1.67466387e-01
1.20849788e+00 4.77433145e-01 1.01143515e+00 7.68826246e-01
5.97712219e-01 9.58864331e-01 7.52969623e-01 3.95823240e-01
1.41834021e-01 4.40134436e-01 3.12347651e-01 3.19074430e-02
-2.49572843e-02 -1.31090194e-01 1.22929156e-01 3.17929000e-01
8.57284144e-02 -3.42688948e-01 -1.21028531e+00 6.27958417e-01
-1.46773219e+00 -7.75795996e-01 -2.88248181e-01 1.84034300e+00
9.22590733e-01 7.25274146e-01 -7.87118897e-02 4.79256064e-01
3.18778545e-01 -6.85132518e-02 -1.36771351e-01 -8.32953155e-01
-7.68283233e-02 2.73532182e-01 5.27844310e-01 1.64770067e-01
-1.09088242e+00 8.32053244e-01 5.60974312e+00 5.43185472e-01
-9.79276001e-01 -6.27731204e-01 2.90460765e-01 1.38901517e-01
-2.15651259e-01 5.91088198e-02 -1.28213859e+00 3.72845650e-01
1.08205450e+00 1.61257729e-01 1.77247729e-02 6.73266649e-01
4.80924159e-01 -4.78785127e-01 -1.20376861e+00 5.36954761e-01
-5.45470007e-02 -1.53316331e+00 1.96448833e-01 2.35952213e-01
3.09153348e-01 -2.84614563e-01 -5.21673739e-01 8.91980976e-02
1.82041556e-01 -1.34222257e+00 5.65465033e-01 7.82002270e-01
3.22151065e-01 -6.79049969e-01 1.05360663e+00 4.84759510e-01
-8.64537299e-01 9.25416276e-02 -1.14152916e-01 -3.40900756e-02
2.41789401e-01 8.17745447e-01 -1.52904379e+00 7.42897868e-01
7.26019561e-01 3.85832459e-01 -1.01154625e+00 9.05713499e-01
-1.70407966e-01 3.33887428e-01 -6.15789175e-01 -4.14220303e-01
5.12397289e-01 -1.81488782e-01 3.47392172e-01 1.74760187e+00
3.07732314e-01 -2.41912499e-01 -3.72430235e-01 8.46837759e-01
2.22208008e-01 3.86400372e-01 -7.60811865e-01 -4.68691349e-01
1.12502888e-01 1.15607858e+00 -1.02212930e+00 -1.53181687e-01
-4.60792810e-01 5.81099033e-01 -1.19551830e-01 7.66156837e-02
-3.56215179e-01 -8.02693665e-01 1.22891508e-01 6.96339071e-01
5.84955275e-01 -5.46859562e-01 -6.82273328e-01 -5.09630144e-01
1.43887311e-01 -9.47960138e-01 6.85123280e-02 -9.53314066e-01
-5.65320194e-01 5.22804201e-01 2.21537873e-01 -9.29605961e-01
-7.77062476e-01 -8.86285543e-01 -3.30643654e-01 1.09810019e+00
-1.07940745e+00 -1.15581965e+00 -3.54673900e-02 2.13081971e-01
4.06794101e-01 -3.10378581e-01 7.71455109e-01 1.55423895e-01
-1.68862849e-01 9.89069417e-02 1.60303250e-01 5.49924254e-01
4.64921236e-01 -1.43986559e+00 5.26360035e-01 8.21778357e-01
4.05743331e-01 7.31433392e-01 7.77903616e-01 -9.07212436e-01
-1.17717266e+00 -7.36196935e-01 1.09946036e+00 -2.71091998e-01
5.01285434e-01 -5.29904842e-01 -1.11871111e+00 3.81772518e-01
5.56355774e-01 -6.64856076e-01 5.33701420e-01 -1.24086216e-02
-1.39859200e-01 1.18939318e-01 -1.03465641e+00 5.47889948e-01
4.40920562e-01 -3.96124333e-01 -8.02101374e-01 1.72200173e-01
5.14474213e-01 -2.72235930e-01 -8.33199322e-01 2.04030037e-01
6.08456075e-01 -8.36350560e-01 8.18504035e-01 -6.30718470e-01
6.10391498e-01 -1.41802192e-01 2.09352881e-01 -1.18031645e+00
1.20576017e-01 -3.98002714e-01 -1.27355933e-01 1.40098226e+00
9.03800845e-01 -7.53997788e-02 7.17428863e-01 6.60038471e-01
2.14797810e-01 -3.28631639e-01 -5.87117255e-01 -4.74869788e-01
5.76436706e-02 -6.92360938e-01 5.60804963e-01 6.97557211e-01
-7.11183399e-02 7.67341495e-01 -1.76998619e-02 -1.70748383e-01
8.19296837e-02 -8.92401338e-02 8.64169180e-01 -1.61466730e+00
1.47682533e-01 -5.09185076e-01 -2.03633457e-01 -6.86685324e-01
-3.80279124e-02 -7.92612374e-01 -1.02129489e-01 -2.12883711e+00
-3.62940311e-01 -3.95206809e-01 1.86698928e-01 4.15253878e-01
3.63071382e-01 -4.06775713e-01 1.28835618e-01 9.29597206e-03
-1.60809904e-02 -1.10719185e-02 9.28775549e-01 -2.32264578e-01
-5.49942851e-01 -5.57126738e-02 -5.40107727e-01 5.88694513e-01
9.78619635e-01 -6.35967791e-01 -1.94358468e-01 -1.50762469e-01
8.09333563e-01 -2.21065193e-01 4.20003459e-02 -9.15147126e-01
2.89486945e-01 8.68950561e-02 9.43553686e-01 -1.12943363e+00
5.69746494e-02 -9.30580139e-01 -9.88605991e-02 2.31785834e-01
-3.15229923e-01 6.56561330e-02 6.00932956e-01 1.24881074e-01
-2.77157217e-01 -7.68014431e-01 1.43120885e-01 -2.27569178e-01
-4.72125769e-01 -3.85484219e-01 -7.09112525e-01 -3.56430560e-01
7.26983726e-01 -4.36079562e-01 -3.14431429e-01 -2.07226679e-01
-3.71542305e-01 -7.88920522e-02 1.31653145e-01 5.91892362e-01
4.97684479e-01 -8.30139697e-01 -6.69128776e-01 2.70811856e-01
-9.78794917e-02 3.27304393e-01 -3.57703090e-01 6.12364173e-01
-9.38207567e-01 1.10659134e+00 -5.25568426e-01 -3.36377382e-01
-1.24712121e+00 7.09104538e-01 -1.81597829e-01 -7.57792056e-01
-5.07852018e-01 1.33569524e-01 -5.48607528e-01 -3.36592674e-01
3.25050019e-02 -8.18158567e-01 -6.91168845e-01 7.73363650e-01
4.84983772e-01 3.27873640e-02 5.50840914e-01 -4.38195974e-01
-2.85565883e-01 2.98583537e-01 -3.84365082e-01 -4.37991112e-01
1.55422497e+00 2.15351224e-01 -4.41259965e-02 2.12729484e-01
1.12689841e+00 4.60198522e-02 -8.53467584e-01 -2.71293018e-02
8.49880874e-01 -1.04110159e-01 2.08086707e-02 -9.12677646e-01
-4.84051645e-01 9.60944355e-01 4.55312192e-01 5.94198227e-01
9.10024107e-01 -2.12828040e-01 2.71201909e-01 6.10869825e-01
3.01186722e-02 -1.60233104e+00 -1.24490499e-01 7.69629717e-01
1.03734171e+00 -1.20113397e+00 4.98615772e-01 -2.54219562e-01
-4.20421600e-01 1.46171713e+00 1.77382380e-01 2.70503730e-01
7.25493371e-01 7.56353915e-01 -1.45431161e-01 -2.38602862e-01
-2.71753758e-01 -3.35940570e-01 4.20218736e-01 7.45490432e-01
7.13142991e-01 -1.50400862e-01 -5.12546301e-01 7.11497605e-01
-3.46034676e-01 3.54818672e-01 5.36112785e-01 1.16577041e+00
-3.38019341e-01 -1.36011755e+00 -6.52847052e-01 6.89291775e-01
-6.97323084e-01 -2.93165535e-01 -9.36344564e-01 1.16930449e+00
1.45146966e-01 8.09455633e-01 2.12738812e-01 -2.63651073e-01
2.38028973e-01 3.95786464e-01 2.89372772e-01 -6.20867670e-01
-8.22834551e-01 1.00672366e-02 3.96670461e-01 -1.09899305e-01
-3.73271674e-01 -7.38591135e-01 -1.30419159e+00 1.28577009e-01
-1.54193312e-01 -4.20427322e-02 1.35256088e+00 1.22416043e+00
3.98095369e-01 6.89326882e-01 1.07886456e-02 -9.24157441e-01
-1.51670814e-01 -1.39303696e+00 -1.67940393e-01 1.87187552e-01
2.57215253e-03 -2.93863177e-01 2.47950718e-01 2.75264591e-01] | [11.65796184539795, 2.8955605030059814] |
48b8be05-b118-4a09-aa41-a96cda900ffa | selectivity-of-protein-interactions | 2207.01572 | null | https://arxiv.org/abs/2207.01572v1 | https://arxiv.org/pdf/2207.01572v1.pdf | Selectivity of Protein Interactions Stimulated by Terahertz Signals | It has been established that Terahertz (THz) band signals can interact with biomolecules through resonant modes. Specifically, of interest here, protein activation. Our research goal is to show how directing the mechanical signaling inside protein molecules using THz signals can control changes in their structure and activate associated biochemical and biomechanical events. To establish that, we formulate a selectivity metric that quantifies the system performance and captures the capability of the nanoantenna to induce a conformational change in the desired protein molecule/population. The metric provides a score between -1 and 1 that indicates the degree of control we have over the system to achieve targeted protein interactions. To develop the selectivity measure, we first use the Langevin stochastic equation driven by an external force to model the protein behavior. We then determine the probability of protein folding by computing the steady-state energy of the driven protein and then generalize our model to account for protein populations. Our numerical analysis results indicate that a maximum selectivity score is attained when only the targeted population experiences a folding behavior due to the impinging THz signal. From the achieved selectivity values, we conclude that the system response not only depends on the resonant frequency but also on the system controlling parameters namely, the nanoantenna force, the damping constant, and the abundance of each protein population. The presented work sheds light on the potential associated with the electromagnetic-based control of protein networks, which could lead to a plethora of applications in the medical field ranging from bio-sensing to targeted therapy. | ['Raviraj Adve', 'Andrew W. Eckford', 'Hadeel Elayan'] | 2022-07-04 | null | null | null | null | ['protein-folding'] | ['natural-language-processing'] | [ 5.99554539e-01 2.92250872e-01 -8.82119238e-02 -2.25869454e-02
-1.23916224e-01 -6.36881232e-01 1.92437232e-01 2.41925612e-01
-2.36329839e-01 7.83298314e-01 -7.62227997e-02 -2.32362803e-02
-1.85293123e-01 -1.02789807e+00 -7.88804412e-01 -1.59547913e+00
3.46734785e-02 2.11139143e-01 1.13165602e-01 -3.60741705e-01
3.33069801e-01 7.53435552e-01 -1.20443308e+00 -3.43638211e-02
9.30799603e-01 8.96430790e-01 1.92905247e-01 3.96645367e-01
7.93445855e-02 2.61408806e-01 -4.15338039e-01 1.83946133e-01
-2.01692149e-01 -2.70705879e-01 -3.24579924e-01 -4.14213270e-01
-2.08378628e-01 3.70239824e-01 -2.31896386e-01 7.53965616e-01
6.14341915e-01 3.01317591e-02 9.29051161e-01 -1.50342852e-01
-3.47064883e-01 4.77325439e-01 -1.52394354e-01 1.18975624e-01
3.22839439e-01 2.68380821e-01 5.49584508e-01 -5.57302535e-01
9.45557952e-01 8.47628474e-01 2.90450782e-01 7.64915764e-01
-1.40905964e+00 -5.10756850e-01 -5.21620750e-01 -2.76906013e-01
-8.01343083e-01 -2.86696732e-01 9.22233760e-01 -6.93624973e-01
5.89674413e-01 4.51358259e-01 7.84362137e-01 1.00352812e+00
9.44766283e-01 -3.42511773e-01 1.05967259e+00 -5.36753953e-01
4.15324688e-01 -2.46358942e-02 3.80203396e-01 6.35231555e-01
2.94421315e-01 3.29708695e-01 -5.64349771e-01 -4.14897978e-01
4.62738603e-01 -1.06010638e-01 -3.95672858e-01 -1.89506575e-01
-9.81953323e-01 4.01012003e-01 1.96627453e-01 7.79504120e-01
-7.14343309e-01 3.91572475e-01 -9.22971517e-02 6.75795451e-02
2.51970410e-01 8.24641764e-01 -7.41963461e-02 -2.13605203e-02
-4.71765608e-01 2.95304596e-01 5.95477700e-01 -2.07531080e-01
6.23672247e-01 -5.21712780e-01 -5.40008247e-01 3.15158188e-01
4.64616083e-02 9.16856110e-01 1.19853369e-03 -8.11646700e-01
1.39883131e-01 4.80352759e-01 4.07817483e-01 -9.02873099e-01
-4.86973673e-01 -3.75919908e-01 -4.35155034e-01 -1.91625252e-01
4.19782043e-01 -5.37395716e-01 -4.95135427e-01 1.90516651e+00
4.88095522e-01 -4.71913099e-01 1.64854482e-01 8.86493325e-01
7.04412937e-01 8.00636292e-01 2.08605736e-01 -6.97752535e-01
1.34988272e+00 1.08233966e-01 -8.24344873e-01 2.61029899e-01
4.97681230e-01 -5.22054851e-01 8.20267141e-01 -1.19465619e-01
-9.24775898e-01 -3.78563181e-02 -8.55459392e-01 4.22822088e-01
-1.64106861e-01 2.47549955e-02 4.16532189e-01 5.65629721e-01
-4.92674023e-01 7.40771830e-01 -1.08223689e+00 -6.57556593e-01
-6.95810094e-02 4.96944308e-01 3.94181497e-02 2.79522002e-01
-1.22352600e+00 6.79264903e-01 1.40946750e-02 1.09372668e-01
-2.70966053e-01 -9.41090226e-01 1.06700458e-01 1.66073948e-01
-3.61709654e-01 -9.46018934e-01 5.17941773e-01 -2.74524629e-01
-1.91465187e+00 8.60409200e-01 -2.74623156e-01 -3.33645463e-01
1.40454799e-01 3.55071753e-01 -1.45027101e-01 4.72980469e-01
-1.71121925e-01 2.14428335e-01 2.86672264e-01 -8.96183610e-01
3.11645091e-01 -6.35261536e-01 -3.09051901e-01 -2.89716899e-01
-5.85818470e-01 9.54821613e-03 5.36948621e-01 -1.26649126e-01
4.32867765e-01 -1.04755974e+00 -1.38854846e-01 -3.67072225e-01
-4.90353614e-01 -1.85863823e-01 8.00840080e-01 1.34287253e-01
8.19837928e-01 -2.08236384e+00 4.23802733e-01 5.06011903e-01
3.35991442e-01 -1.47859054e-02 1.76523224e-01 1.14543509e+00
6.80416301e-02 -6.19761273e-02 -2.17955440e-01 6.32178068e-01
-4.62210715e-01 -4.75849062e-01 -2.74967968e-01 6.73100710e-01
1.01478808e-01 8.21346462e-01 -7.60038972e-01 -4.56537418e-02
-1.12678930e-01 8.19833577e-01 -4.90433633e-01 1.13167070e-01
-4.06759620e-01 9.18221653e-01 -1.12647831e+00 5.07280529e-01
7.75413454e-01 -3.19617748e-01 5.61687231e-01 -7.55514741e-01
-6.24006152e-01 6.15776107e-02 -3.55704099e-01 8.31799686e-01
6.92200214e-02 2.50104755e-01 1.26746580e-01 -1.09486210e+00
1.25914717e+00 3.24670911e-01 1.05699253e+00 -5.50693452e-01
1.73428088e-01 4.23884779e-01 1.72748715e-01 -8.63591135e-01
-8.41750279e-02 -6.11426950e-01 -9.99800414e-02 3.01973879e-01
-2.74182290e-01 3.06497458e-02 -4.32549044e-03 -1.82909042e-01
1.36616683e+00 -1.58454135e-01 -2.93386191e-01 -7.77949750e-01
3.26769590e-01 1.83318570e-01 -3.46831791e-02 5.79498172e-01
-4.79525812e-02 -1.40836641e-01 6.99472070e-01 -3.62080306e-01
-7.58802712e-01 -9.94369864e-01 -5.06326377e-01 6.80761755e-01
5.67226470e-01 2.23589584e-01 -8.83210719e-01 1.26418293e-01
2.30149478e-01 1.14525549e-01 -5.46317875e-01 -5.84819853e-01
-5.77955782e-01 -1.27118015e+00 3.35702419e-01 -3.44139993e-01
2.02166364e-01 -9.60331559e-01 -8.51661801e-01 3.01583767e-01
-2.44042695e-01 -1.15987039e+00 3.65564376e-02 1.26162753e-01
-8.67423594e-01 -8.92438352e-01 -5.46263814e-01 -2.62854367e-01
7.52007186e-01 -1.18549407e-01 3.47991377e-01 -3.30975533e-01
-2.81798571e-01 4.37003970e-01 -8.50279331e-02 -9.69208032e-02
-5.91365814e-01 2.77895778e-01 2.13566199e-01 7.25974888e-02
2.39807129e-01 -9.09122050e-01 -9.94309425e-01 2.46647850e-01
-5.75200856e-01 -2.77491122e-01 4.13956285e-01 2.18800381e-01
5.78052461e-01 -3.81679058e-01 6.60099924e-01 -6.30773783e-01
5.12624681e-01 -3.44660342e-01 -4.69141603e-01 3.75133678e-02
-8.83203670e-02 3.82164121e-01 4.76099640e-01 -5.68930328e-01
-8.80686760e-01 1.56492963e-01 -1.32569037e-02 2.57169664e-01
3.31218652e-02 6.43422067e-01 8.16728100e-02 -4.64642465e-01
7.72442043e-01 1.56204507e-01 2.07336664e-01 -7.29747266e-02
2.32166983e-02 4.41775382e-01 2.42952153e-01 -9.05210197e-01
4.65258896e-01 7.71661043e-01 5.80043733e-01 -1.22662890e+00
-4.14121151e-01 -1.21668451e-01 8.58813524e-02 -4.99106139e-01
9.48597908e-01 -2.91437835e-01 -1.68655169e+00 1.60532609e-01
-1.03018391e+00 -2.93688864e-01 -3.08188666e-02 5.89176416e-01
-3.56801361e-01 -1.93731617e-02 -8.15958977e-01 -1.10963762e+00
-5.89414060e-01 -8.52611661e-01 8.88441026e-01 2.89625615e-01
-4.75501865e-02 -5.97172320e-01 3.70775014e-01 2.20951244e-01
5.96924543e-01 9.08189714e-01 1.27901232e+00 9.48199257e-02
-7.99957275e-01 -2.10947677e-01 2.65228182e-01 -5.56325912e-01
1.25059500e-01 6.39695227e-02 -6.79168046e-01 -1.49154156e-01
1.90292701e-01 5.16835526e-02 9.75378335e-01 9.26617622e-01
7.05349982e-01 -1.12103656e-01 -9.93474245e-01 3.05444032e-01
1.47650826e+00 2.34565005e-01 5.56772947e-01 -9.50874463e-02
4.61154789e-01 8.82866859e-01 3.85676682e-01 3.23413610e-01
-3.17281425e-01 8.00619423e-01 2.84402162e-01 -2.89581250e-02
4.02792752e-01 9.90810916e-02 5.13206780e-01 6.74943864e-01
-4.70163882e-01 -3.59753966e-01 -7.87143111e-01 -2.83038914e-02
-1.46383238e+00 -1.05585206e+00 -3.93538088e-01 2.38112450e+00
7.39780545e-01 -7.55671188e-02 5.60867451e-02 -4.40665662e-01
8.09888542e-01 -4.14945446e-02 -6.11655116e-01 -1.61449805e-01
-1.14592046e-01 5.02938628e-01 6.27625823e-01 6.45943940e-01
-4.81448799e-01 5.98544061e-01 6.69022322e+00 3.78089815e-01
-1.75890648e+00 3.04849148e-02 3.31579179e-01 -2.90258288e-01
-5.52733600e-01 -7.19932392e-02 -6.60734236e-01 5.37215531e-01
9.59094763e-01 -1.74300745e-01 -7.91354512e-04 1.49669692e-01
7.83593118e-01 -1.15002230e-01 -6.84702635e-01 2.38370508e-01
-6.29604936e-01 -1.41032708e+00 -6.09193323e-03 5.05442858e-01
3.35403860e-01 -1.33947298e-01 5.06841205e-02 -4.69567388e-01
-5.93556583e-01 -7.86230683e-01 1.36619195e-01 1.00049663e+00
8.63341391e-01 -5.38346827e-01 3.31382632e-01 4.39786196e-01
-9.03986871e-01 -2.49805506e-02 -3.21060419e-01 -4.87851314e-02
1.40889242e-01 1.30583203e+00 -6.25200927e-01 -3.08457427e-02
1.82701632e-01 1.09004296e-01 3.75576138e-01 1.94143787e-01
3.86114508e-01 5.71484983e-01 -5.38692176e-01 -6.86499357e-01
-2.05729336e-01 -6.17735088e-01 7.53392875e-01 8.14584017e-01
5.64392745e-01 6.03529274e-01 -3.01190376e-01 1.52997577e+00
5.31545132e-02 7.53693581e-02 -4.80244666e-01 -4.62717295e-01
6.32361591e-01 1.26262081e+00 -8.66906881e-01 1.35864288e-01
3.89918953e-01 4.62056935e-01 1.08653605e-01 5.36191404e-01
-6.70635939e-01 -4.68094617e-01 8.60851645e-01 7.18640625e-01
1.48406699e-01 -4.89671618e-01 -5.20425476e-02 -9.10202682e-01
1.56823974e-02 -8.27735513e-02 -4.33309227e-01 -2.24737570e-01
-7.06011653e-01 6.09327517e-02 -2.67271578e-01 -4.17467386e-01
3.04812431e-01 -5.13209462e-01 -2.89600998e-01 7.76707113e-01
-1.20733070e+00 -5.65202117e-01 -2.31658235e-01 5.89387342e-02
-4.92739737e-01 4.86631632e-01 7.53149748e-01 1.90897405e-01
-6.32209003e-01 1.60518870e-01 4.10456896e-01 -3.35426420e-01
4.97868508e-01 -5.89421153e-01 -1.48395211e-01 3.59144032e-01
-6.87973797e-01 9.68022823e-01 1.15286410e+00 -7.74520099e-01
-1.86165750e+00 -5.99565446e-01 9.16516304e-01 -3.65214825e-01
3.85763586e-01 -5.93728721e-01 -5.99297702e-01 -1.53633237e-01
-3.46367121e-01 1.65180087e-01 5.20957828e-01 -3.55875611e-01
1.92369744e-01 -2.25488991e-01 -1.34234738e+00 3.16219628e-01
9.35697436e-01 -4.86900955e-01 3.11778225e-02 6.08773470e-01
6.19789362e-01 -4.31979328e-01 -1.07110822e+00 7.08705902e-01
9.08796132e-01 -8.94002259e-01 7.54729390e-01 -5.54005861e-01
4.61740136e-01 -3.83764617e-02 -2.16544494e-02 -7.54264176e-01
-3.45046222e-01 -6.86698675e-01 5.33099920e-02 6.95988774e-01
4.10621554e-01 -9.50252950e-01 8.24305356e-01 7.13318169e-01
1.71017110e-01 -1.01283073e+00 -9.81757462e-01 -3.69513690e-01
1.71300769e-01 3.97887439e-01 2.79980063e-01 7.53832996e-01
4.35106069e-01 -2.93190009e-03 2.20857680e-01 3.11680317e-01
5.79654634e-01 1.36726782e-01 1.23725563e-01 -1.26096845e+00
-9.98963267e-02 -9.71801057e-02 -2.20946684e-01 -7.08540022e-01
2.45981831e-02 -6.95214033e-01 3.14534851e-03 -1.10912597e+00
2.05196515e-01 -3.68118942e-01 -2.61292398e-01 -1.27966821e-01
2.26133496e-01 -3.76115739e-02 -2.24168360e-01 3.02892715e-01
1.10488385e-01 5.34131408e-01 1.24472809e+00 2.72198766e-01
-6.22514606e-01 -6.87883422e-02 -3.81643474e-01 8.98199913e-04
6.70789123e-01 -5.92474401e-01 7.49806389e-02 4.38323885e-01
5.03243327e-01 2.81160563e-01 4.19515014e-01 -1.03467238e+00
3.90763804e-02 -3.20920825e-01 -8.78667161e-02 2.70523161e-01
2.97557503e-01 -4.66665030e-01 3.73067170e-01 9.74214971e-01
-4.00761306e-01 -7.04342782e-01 -2.90758640e-01 5.47375023e-01
3.19789231e-01 5.49877547e-02 1.06170881e+00 8.61179084e-02
4.59035248e-01 2.61495411e-02 -9.16897535e-01 -2.90461987e-01
9.59937871e-01 -1.09467395e-01 -5.37414551e-01 -2.42720917e-02
-7.45359838e-01 -2.69451141e-01 4.22178507e-01 -3.51374000e-01
1.13822833e-01 -9.57268715e-01 -3.62082064e-01 1.73452303e-01
-1.04473218e-01 -7.95591116e-01 3.63096535e-01 1.03067827e+00
-4.68988538e-01 4.72250998e-01 -9.82314870e-02 -8.18064213e-01
-1.04385114e+00 1.15693122e-01 8.36111486e-01 -1.27343327e-01
-3.24224949e-01 4.71151173e-01 7.57135004e-02 -3.78081650e-02
-4.45117503e-01 -4.92158085e-01 -1.94738075e-01 -2.08696738e-01
-1.18335094e-02 2.21224144e-01 7.84004480e-02 -3.37316722e-01
-5.30570984e-01 1.24670815e+00 3.58324111e-01 3.94683741e-02
1.13121295e+00 6.25169575e-02 -4.16715473e-01 3.11569065e-01
8.90022933e-01 2.75591850e-01 -8.91307592e-01 1.87711209e-01
-3.02960277e-01 2.46406838e-01 -7.77576938e-02 -4.59196776e-01
-6.23823225e-01 4.98638690e-01 7.41030753e-01 5.57442605e-01
8.03833246e-01 5.14532626e-01 6.47836208e-01 5.50148129e-01
3.86522710e-01 -9.20845270e-01 2.48063114e-02 1.57451376e-01
5.04055977e-01 -5.93455255e-01 -2.69933432e-01 -8.00971329e-01
2.03818351e-01 1.16208243e+00 2.24507749e-01 -2.69490063e-01
7.09492683e-01 3.10847580e-01 -1.34962901e-01 -5.75604200e-01
-5.10352433e-01 2.14165241e-01 -7.89823532e-02 3.33045125e-01
9.81544614e-01 3.24046195e-01 -9.28012073e-01 4.37052697e-01
-1.21681362e-01 1.65530086e-01 4.25938219e-01 5.57965577e-01
-8.39539111e-01 -1.03318942e+00 -2.75904328e-01 3.18084061e-01
-5.92248797e-01 3.76488388e-01 -5.28436065e-01 2.71301657e-01
6.93274215e-02 7.80284822e-01 -1.35887429e-01 -1.76580325e-01
4.78005081e-01 1.27966359e-01 7.79408157e-01 -4.15269822e-01
-2.72285819e-01 -2.94867977e-02 -2.36960515e-01 -5.87744117e-01
-7.85910726e-01 -4.00632381e-01 -1.59489393e+00 -2.68316746e-01
-4.45900708e-01 4.95793909e-01 7.19074905e-01 1.03229070e+00
8.26846480e-01 3.76528203e-01 5.37033558e-01 -7.32952833e-01
-3.03517371e-01 -7.19615042e-01 -6.88190877e-01 1.57403335e-01
3.79562259e-01 -6.08117104e-01 -5.18158793e-01 -4.27295536e-01] | [4.99808406829834, 5.105839729309082] |
cd166f5e-3f89-4c1a-8285-8116e8835264 | interactive-image-synthesis-with-panoptic | 2203.02104 | null | https://arxiv.org/abs/2203.02104v3 | https://arxiv.org/pdf/2203.02104v3.pdf | Interactive Image Synthesis with Panoptic Layout Generation | Interactive image synthesis from user-guided input is a challenging task when users wish to control the scene structure of a generated image with ease.Although remarkable progress has been made on layout-based image synthesis approaches, in order to get realistic fake image in interactive scene, existing methods require high-precision inputs, which probably need adjustment several times and are unfriendly to novice users. When placement of bounding boxes is subject to perturbation, layout-based models suffer from "missing regions" in the constructed semantic layouts and hence undesirable artifacts in the generated images. In this work, we propose Panoptic Layout Generative Adversarial Networks (PLGAN) to address this challenge. The PLGAN employs panoptic theory which distinguishes object categories between "stuff" with amorphous boundaries and "things" with well-defined shapes, such that stuff and instance layouts are constructed through separate branches and later fused into panoptic layouts. In particular, the stuff layouts can take amorphous shapes and fill up the missing regions left out by the instance layouts. We experimentally compare our PLGAN with state-of-the-art layout-based models on the COCO-Stuff, Visual Genome, and Landscape datasets. The advantages of PLGAN are not only visually demonstrated but quantitatively verified in terms of inception score, Fr\'echet inception distance, classification accuracy score, and coverage. | ['Peng Du', 'Minfeng Zhu', 'Tao Wu', 'Bo wang'] | 2022-03-04 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Interactive_Image_Synthesis_With_Panoptic_Layout_Generation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Interactive_Image_Synthesis_With_Panoptic_Layout_Generation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['layout-to-image-generation'] | ['computer-vision'] | [ 5.39151967e-01 2.76983827e-01 3.72973472e-01 6.84732571e-03
-1.60864547e-01 -1.00384474e+00 6.09445632e-01 -4.14834648e-01
2.18041852e-01 7.57550240e-01 -1.19802244e-01 -3.76404941e-01
2.04369262e-01 -1.13787007e+00 -1.10713255e+00 -7.78783798e-01
3.62576455e-01 1.60502344e-01 1.65491745e-01 -3.84099275e-01
2.13896587e-01 5.08540154e-01 -1.30431652e+00 3.95468920e-01
1.19080937e+00 7.25081265e-01 9.17479917e-02 5.68978429e-01
-2.77851731e-01 5.78571737e-01 -9.44964826e-01 -5.57306290e-01
5.49873114e-01 -7.49424756e-01 -3.56793582e-01 4.04743016e-01
4.43456411e-01 -1.96389973e-01 -4.82921183e-01 1.30957043e+00
2.51874924e-01 -2.59215057e-01 8.11463475e-01 -1.30200148e+00
-1.18726110e+00 3.24536622e-01 -7.33230770e-01 -4.51927066e-01
2.79892594e-01 6.28561974e-01 4.95620131e-01 -5.43573797e-01
7.48528063e-01 1.32255471e+00 5.36859214e-01 5.17212629e-01
-1.57562733e+00 -8.54831457e-01 1.02973968e-01 -2.35504523e-01
-1.41758156e+00 -1.89820945e-01 1.12081861e+00 -5.48711061e-01
3.17178249e-01 6.64814174e-01 7.34141111e-01 1.20079386e+00
3.74904931e-01 6.20199978e-01 1.23296511e+00 -3.30432743e-01
1.67905465e-01 4.16235626e-01 -5.59202433e-01 7.17503369e-01
3.83367419e-01 2.38233864e-01 5.22950180e-02 2.87028700e-01
1.19158661e+00 3.60304043e-02 -4.73568022e-01 -5.27840376e-01
-1.05893254e+00 6.40762508e-01 8.49107385e-01 1.72329992e-01
-6.93130791e-02 1.94818676e-01 -9.03580617e-03 1.58239365e-01
2.74781018e-01 8.63125145e-01 -2.81436890e-02 4.05555874e-01
-8.90696466e-01 3.96293104e-01 4.49967712e-01 1.24078608e+00
5.76456904e-01 3.47559065e-01 -2.67753690e-01 6.70587361e-01
-3.96434180e-02 7.74762988e-01 1.88288674e-01 -6.50396645e-01
4.50358301e-01 8.40083838e-01 2.18966439e-01 -1.54981339e+00
6.42893985e-02 -2.83966750e-01 -1.38876712e+00 5.18980920e-01
3.55285347e-01 -1.77390963e-01 -1.24673033e+00 1.48378563e+00
9.76703987e-02 -1.99576408e-01 -1.02962188e-01 8.50838661e-01
7.93401480e-01 9.78290141e-01 -1.55288771e-01 1.07155085e-01
1.36411393e+00 -1.05571640e+00 -7.26287305e-01 -2.34602273e-01
1.75441548e-01 -8.12165141e-01 1.36359417e+00 3.53450179e-01
-9.83773351e-01 -7.40003407e-01 -1.20867968e+00 9.41528603e-02
-4.58385378e-01 1.38861269e-01 5.11177659e-01 8.55050623e-01
-9.62315381e-01 3.04745436e-01 -2.05935448e-01 1.36021310e-02
6.47338986e-01 7.71182254e-02 -3.30933660e-01 3.72922458e-02
-1.02822280e+00 5.36600113e-01 2.86943704e-01 2.20168009e-01
-9.43704069e-01 -6.47527993e-01 -7.40309000e-01 7.51331449e-02
4.97418791e-01 -5.34339905e-01 7.46560812e-01 -1.29024589e+00
-1.45396113e+00 5.46837568e-01 3.38171273e-01 -2.30276495e-01
9.24000978e-01 3.24166328e-01 -4.12052423e-01 7.94526115e-02
-1.81340370e-02 8.77000093e-01 1.26481581e+00 -1.78725517e+00
-2.19404042e-01 5.30484393e-02 2.64311373e-01 9.53300670e-02
3.98052111e-02 -5.31488776e-01 -2.31332287e-01 -9.52384412e-01
2.46397421e-01 -8.90997767e-01 -2.36978740e-01 2.10538492e-01
-8.45456541e-01 4.60718632e-01 1.30649841e+00 -5.21935105e-01
1.11221170e+00 -2.13637996e+00 -8.59188288e-02 1.95419401e-01
1.27024576e-01 3.68880510e-01 -1.36117145e-01 4.69200760e-01
-1.83783710e-01 6.43087924e-01 -4.28325862e-01 1.64216056e-01
-9.33592618e-02 9.21550393e-02 -5.69862425e-01 2.68876702e-01
1.46657154e-01 1.08958578e+00 -8.00887704e-01 -2.76272595e-01
3.33613902e-01 2.95027167e-01 -7.19901085e-01 3.03361207e-01
-4.60141897e-01 7.41399050e-01 -3.01978439e-01 6.77648723e-01
1.01171219e+00 -4.93736267e-02 3.19687761e-02 -2.64494359e-01
2.09006481e-02 -1.70193449e-01 -8.85270953e-01 1.37271833e+00
-3.77922654e-01 6.91519082e-01 -9.32976007e-02 -6.15191877e-01
9.74467695e-01 5.56122735e-02 -2.02277675e-03 -8.05384934e-01
-2.54194401e-02 3.55865732e-02 -5.98042011e-02 -5.39638758e-01
6.21680140e-01 -1.35782897e-01 -3.35738987e-01 1.53302193e-01
-4.68947440e-01 -7.51961887e-01 -1.43368140e-01 1.01709992e-01
8.65320444e-01 2.77244180e-01 4.98519912e-02 -2.49091908e-01
3.63052338e-01 -8.47053006e-02 2.28534237e-01 6.64880216e-01
2.58280933e-01 1.07645261e+00 9.66233969e-01 -4.65245992e-01
-1.44793642e+00 -1.13903749e+00 2.32099798e-02 2.37677857e-01
4.85825181e-01 -6.62663132e-02 -1.09268653e+00 -7.35594869e-01
-2.74719536e-01 7.87173331e-01 -6.96317196e-01 -2.00675502e-01
-6.05184853e-01 -4.10025388e-01 6.29774094e-01 1.42091259e-01
9.57701981e-01 -1.37595713e+00 -6.10190749e-01 2.54276805e-02
-2.07775950e-01 -8.98611307e-01 -6.16919041e-01 -3.23666900e-01
-4.62920606e-01 -9.33372498e-01 -8.12085092e-01 -6.77824020e-01
1.22267044e+00 2.77251661e-01 1.02287734e+00 1.37397692e-01
-3.38372529e-01 -2.88910776e-01 -3.46408576e-01 -2.66043365e-01
-5.63960135e-01 -1.44670755e-01 -3.80053014e-01 1.91800162e-01
-5.53880572e-01 -6.43296838e-01 -9.16298449e-01 5.86700022e-01
-1.43190002e+00 5.63706219e-01 6.36463940e-01 9.92954731e-01
4.15932417e-01 3.98239523e-01 3.01559567e-01 -1.07852852e+00
4.41609740e-01 -2.16595247e-01 -8.02577853e-01 2.56745130e-01
-3.70526046e-01 -3.85121704e-04 9.88306403e-01 -5.53349853e-01
-1.00467134e+00 -1.79809466e-01 2.80690077e-03 -6.00096703e-01
-1.30830109e-01 -6.23445362e-02 -7.15243399e-01 -1.63694412e-01
7.40564048e-01 4.01949525e-01 -1.02967247e-01 -7.33041391e-02
5.83212554e-01 4.75774169e-01 5.62413633e-01 -6.27091587e-01
1.21597123e+00 5.61964691e-01 -5.97275980e-02 -8.24519873e-01
-2.78135985e-01 3.57831955e-01 -3.86104047e-01 -2.49302223e-01
7.80812562e-01 -5.11425674e-01 -4.87170517e-01 6.30098343e-01
-1.24355948e+00 -3.90064776e-01 -3.64593744e-01 -3.81044060e-01
-4.24984008e-01 3.95815283e-01 -3.07517111e-01 -6.72384679e-01
-1.13030002e-01 -1.35102320e+00 9.87497628e-01 3.16269994e-01
-9.13363397e-02 -7.25446224e-01 -4.29509521e-01 2.22996891e-01
3.49478215e-01 8.86353076e-01 1.24006605e+00 1.12170815e-01
-1.13981414e+00 -3.41447890e-01 -4.47551191e-01 3.55574340e-01
1.31018862e-01 1.90265357e-01 -7.63073981e-01 -1.65722474e-01
-1.38082087e-01 6.86869994e-02 5.14919162e-01 1.68328583e-01
1.28714681e+00 -8.09611142e-01 -2.87187070e-01 8.27092111e-01
1.60278904e+00 6.01950526e-01 1.32996058e+00 1.26472965e-01
9.37513292e-01 5.58148563e-01 3.55685323e-01 -2.84211040e-02
-1.14109002e-01 5.15762806e-01 5.24115086e-01 -2.69530803e-01
-2.69026250e-01 -7.99925148e-01 7.03114420e-02 2.49882340e-01
1.57473028e-01 -8.38139892e-01 -5.48622310e-01 5.03604233e-01
-1.52227676e+00 -9.00813282e-01 -2.65626520e-01 2.01019144e+00
5.59289098e-01 2.72724301e-01 -3.02678078e-01 7.60847405e-02
8.46478820e-01 4.27896291e-01 -4.35111135e-01 -5.83672643e-01
-2.77082592e-01 2.04275250e-01 6.21616662e-01 2.45229349e-01
-9.02911186e-01 1.12919724e+00 5.73561478e+00 1.21809137e+00
-1.14648962e+00 -1.51873946e-01 1.01469159e+00 2.74152040e-01
-6.78848445e-01 1.88986555e-01 -2.92255223e-01 7.79488087e-01
2.02613428e-01 3.04253638e-01 4.97171879e-01 6.69869304e-01
1.21282637e-01 -2.09749222e-01 -7.73092687e-01 9.35265899e-01
-1.32984042e-01 -1.46507752e+00 3.50465477e-01 8.68889242e-02
1.08880889e+00 -8.18852544e-01 5.53360760e-01 9.89920497e-02
1.78668350e-01 -1.35769498e+00 1.00368035e+00 4.40016329e-01
1.35440576e+00 -7.17289209e-01 3.98335069e-01 3.32190722e-01
-9.26479399e-01 2.93485401e-03 -3.88991892e-01 1.50616840e-01
4.63760719e-02 5.10574460e-01 -6.23374403e-01 5.04270196e-01
4.86619592e-01 1.71836376e-01 -5.68955421e-01 7.39611626e-01
-5.66712022e-01 3.17481190e-01 6.72222208e-03 -4.50484157e-02
3.74743283e-01 -5.25439382e-01 6.44118965e-01 8.60496998e-01
4.63249981e-01 9.49551240e-02 -2.02134296e-01 1.46846688e+00
-1.55798286e-01 -1.35684609e-01 -1.02202296e+00 -6.52604997e-02
2.82890856e-01 1.10255706e+00 -9.48605299e-01 -2.89222747e-01
9.60255414e-02 1.20185375e+00 -1.22105069e-01 4.71862912e-01
-1.20468724e+00 -6.35563135e-01 4.12524045e-01 6.03113353e-01
2.52600312e-01 -7.71936402e-02 -7.42451787e-01 -1.08296907e+00
2.78909933e-02 -1.09019244e+00 -3.67886811e-01 -1.14635897e+00
-9.85901177e-01 6.88484490e-01 -1.13079756e-01 -1.35640466e+00
1.73692003e-01 -3.49286109e-01 -6.26691282e-01 8.22449565e-01
-8.71046841e-01 -1.31554008e+00 -4.93178904e-01 2.30605572e-01
7.03504443e-01 -7.80282617e-02 6.31106973e-01 8.62009525e-02
-4.73933905e-01 7.13753819e-01 9.58371460e-02 2.97866672e-01
4.37234312e-01 -9.85902190e-01 6.32327378e-01 1.01509833e+00
5.25107719e-02 3.91305447e-01 7.53572464e-01 -8.34008694e-01
-1.17275953e+00 -1.31458461e+00 3.55299145e-01 -4.91406471e-01
1.61631986e-01 -7.03419089e-01 -7.81562269e-01 3.33811849e-01
3.85600448e-01 -2.43708745e-01 1.66693819e-03 -5.92435420e-01
-4.61599737e-01 -1.05868705e-01 -1.38694453e+00 1.03725052e+00
1.07932746e+00 -2.24417046e-01 -2.20749065e-01 2.63659775e-01
7.37358689e-01 -5.12333512e-01 -2.78378725e-01 3.99625510e-01
6.03201509e-01 -1.26658177e+00 1.00285041e+00 -1.67485923e-01
7.82510877e-01 -5.91557860e-01 1.65464029e-01 -1.34792006e+00
-1.57797411e-01 -8.60471249e-01 4.90294665e-01 1.10265386e+00
3.65468383e-01 -5.06785870e-01 7.72320092e-01 3.28203976e-01
-1.61009699e-01 -4.89330471e-01 -4.84876782e-01 -6.75064921e-01
2.51022782e-02 -8.61743614e-02 9.09434080e-01 8.40701401e-01
-3.30230743e-01 2.25764841e-01 -6.70203686e-01 1.15688227e-01
5.12145221e-01 1.91055238e-01 9.60462809e-01 -6.89286172e-01
-2.51607001e-01 -4.85738754e-01 -3.29331070e-01 -9.59888577e-01
-2.66900241e-01 -4.74573940e-01 3.22126560e-02 -1.24024153e+00
-1.52558535e-01 -6.15106165e-01 2.98367858e-01 2.76882917e-01
7.53887445e-02 5.71221769e-01 2.79642045e-01 1.47012636e-01
-1.69440001e-01 4.88748431e-01 1.89950097e+00 -4.39940333e-01
-2.70162016e-01 -2.03227654e-01 -7.49358356e-01 6.84130430e-01
7.05356896e-01 -3.62624764e-01 -6.61072969e-01 -3.97605449e-01
2.35785022e-01 2.45008603e-01 6.56095564e-01 -1.12954748e+00
-2.15677202e-01 -2.36547962e-01 6.41721070e-01 -4.92402792e-01
2.03231707e-01 -7.93600321e-01 7.99612641e-01 5.33352494e-01
-1.89488858e-01 -6.70046508e-02 1.18851699e-01 5.35759032e-01
1.68673210e-02 -7.46228769e-02 1.04713857e+00 -4.47917312e-01
-3.34608734e-01 1.57727107e-01 -2.20664203e-01 -1.81435257e-01
1.18743634e+00 -4.48873937e-01 -4.69872296e-01 -5.16401350e-01
-3.86344284e-01 -2.34503388e-01 8.77346218e-01 3.89975578e-01
6.73368692e-01 -1.35413420e+00 -3.67250741e-01 6.09520614e-01
-6.90222252e-03 3.85408193e-01 5.27647614e-01 3.16879243e-01
-1.18209887e+00 3.39420766e-01 -4.31443661e-01 -4.88189578e-01
-8.86491358e-01 8.92439902e-01 1.55665249e-01 -1.50929883e-01
-5.81356406e-01 7.71118045e-01 9.45309699e-01 -3.89597923e-01
-1.39376640e-01 -4.35994506e-01 2.21594274e-01 -2.26855248e-01
1.72633871e-01 8.07311237e-02 -3.23117971e-01 -4.61337984e-01
2.03072920e-01 5.66279888e-01 1.20266095e-01 4.06081462e-03
8.69584978e-01 -3.98015156e-02 -1.04159974e-01 -6.80850148e-02
9.63380635e-01 1.95229307e-01 -1.47215950e+00 3.08055609e-01
-6.09178841e-01 -7.88729131e-01 -4.28795546e-01 -9.30144966e-01
-1.13238692e+00 9.26635504e-01 4.79404122e-01 4.60939467e-01
1.15192246e+00 -4.00617450e-01 7.87470937e-01 -1.05882570e-01
4.85855758e-01 -7.53656566e-01 3.44498098e-01 -7.43424892e-02
1.10783386e+00 -7.45152414e-01 -1.71481237e-01 -6.71053708e-01
-6.81949675e-01 7.19207406e-01 7.47810662e-01 -4.12587851e-01
2.24524722e-01 2.44694889e-01 -7.23123029e-02 -5.77012710e-02
-2.45789170e-01 2.22091481e-01 3.52052003e-01 7.56063581e-01
-9.03092772e-02 1.64067045e-01 -5.26079759e-02 3.60160410e-01
-3.82319987e-01 -4.11174387e-01 5.95346451e-01 5.48193812e-01
-2.59166300e-01 -9.86363053e-01 -6.51131094e-01 2.03324199e-01
-2.02374667e-01 -1.68541223e-01 -4.07979518e-01 1.02537119e+00
6.38898969e-01 7.10374773e-01 8.45924299e-03 -4.80838001e-01
3.07572603e-01 -2.31500000e-01 4.72303838e-01 -3.56963247e-01
-4.60849762e-01 1.93986535e-01 -1.88339546e-01 -2.66765594e-01
-1.62111875e-03 -1.89644128e-01 -8.04271042e-01 -4.03515130e-01
-1.94273457e-01 -8.24725404e-02 5.10395169e-01 5.52209437e-01
3.06513697e-01 7.60542035e-01 7.37857103e-01 -7.47594595e-01
-1.00825772e-01 -8.30870092e-01 -4.40864921e-01 4.91868258e-01
1.57219872e-01 -4.80965048e-01 -2.72694856e-01 1.86870351e-01] | [11.617511749267578, -0.5264559984207153] |
16c67933-d608-45ef-a448-dd243069ce09 | automated-essay-scoring-based-on-two-stage | 1901.07744 | null | https://arxiv.org/abs/1901.07744v2 | https://arxiv.org/pdf/1901.07744v2.pdf | Automated Essay Scoring based on Two-Stage Learning | Current state-of-art feature-engineered and end-to-end Automated Essay Score (AES) methods are proven to be unable to detect adversarial samples, e.g. the essays composed of permuted sentences and the prompt-irrelevant essays. Focusing on the problem, we develop a Two-Stage Learning Framework (TSLF) which integrates the advantages of both feature-engineered and end-to-end AES models. In experiments, we compare TSLF against a number of strong baselines, and the results demonstrate the effectiveness and robustness of our models. TSLF surpasses all the baselines on five-eighths of prompts and achieves new state-of-the-art average performance when without negative samples. After adding some adversarial essays to the original datasets, TSLF outperforms the feature-engineered and end-to-end baselines to a great extent, and shows great robustness. | ['Yaguang Zhu', 'Jiawei Liu', 'Yang Xu'] | 2019-01-23 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [-1.91296469e-02 -2.60000855e-01 -3.85330198e-03 -4.12382901e-01
-1.44085872e+00 -1.10026920e+00 8.81376565e-01 -1.73118070e-01
-4.74255770e-01 8.07747722e-01 5.54787755e-01 -3.13588679e-01
2.76173074e-02 -4.60631430e-01 -5.74970961e-01 -2.01069891e-01
3.68607461e-01 3.25726539e-01 1.61592975e-01 -4.65303510e-01
5.09015083e-01 -9.98757780e-02 -1.16729784e+00 3.52250695e-01
1.10718703e+00 6.46116734e-01 -3.64008874e-01 1.02805948e+00
-5.56551740e-02 1.01251125e+00 -1.15168488e+00 -1.07968855e+00
3.50180596e-01 -4.37064320e-01 -6.79559886e-01 -4.59411711e-01
1.08361638e+00 -8.51227462e-01 -8.48160863e-01 9.63063359e-01
7.79654264e-01 1.07578665e-01 7.08566785e-01 -1.17514944e+00
-1.05242896e+00 5.30797660e-01 -3.81905347e-01 1.83500461e-02
8.30588996e-01 5.96226931e-01 1.34730518e+00 -1.13949156e+00
4.01614785e-01 1.10355318e+00 7.51369536e-01 7.75172114e-01
-9.70921874e-01 -7.94889629e-01 1.36014551e-01 1.70370027e-01
-6.57565236e-01 -4.25902843e-01 8.76895070e-01 -3.97860676e-01
5.96212745e-01 5.39039910e-01 -6.47969693e-02 1.80646610e+00
2.90486097e-01 1.48843122e+00 1.44413531e+00 -4.22922283e-01
1.64697036e-01 1.54154468e-02 5.72967470e-01 6.01248205e-01
-5.07095940e-02 3.87982726e-01 -8.76023829e-01 -4.31328714e-01
7.82282278e-03 -4.04222965e-01 -2.49861732e-01 2.04716533e-01
-1.11722445e+00 9.08350706e-01 -7.81464279e-02 -2.58531496e-02
-1.64166167e-01 -3.94226387e-02 6.03300035e-01 7.61988282e-01
6.36052132e-01 9.65215802e-01 -5.56689918e-01 -4.35033768e-01
-1.33626771e+00 7.39836514e-01 9.66894507e-01 7.67664671e-01
2.24190131e-01 1.36587545e-01 -1.16120088e+00 5.82857966e-01
-3.58736008e-01 4.99727935e-01 6.37169719e-01 -8.62842143e-01
6.46146119e-01 4.05501962e-01 4.02535945e-01 -6.46066248e-01
-3.53542119e-01 -5.27787089e-01 -4.39786792e-01 2.96368182e-01
7.34412014e-01 -4.25144196e-01 -6.61572695e-01 1.83188164e+00
1.29160574e-02 -5.75078391e-02 -1.74362898e-01 9.14663494e-01
1.01586699e+00 2.90967911e-01 4.11246493e-02 9.24435481e-02
1.31133437e+00 -1.32436669e+00 -6.57765031e-01 -3.49165648e-01
5.71941376e-01 -8.19631517e-01 1.81493640e+00 6.49877608e-01
-1.29100096e+00 -6.36311173e-01 -1.02575111e+00 2.02491265e-02
-1.04086660e-01 2.03912124e-01 4.14368600e-01 9.58669603e-01
-8.84837508e-01 5.57875097e-01 -2.73225427e-01 3.44252110e-01
3.79516691e-01 1.52914941e-01 -2.53591001e-01 7.57937727e-04
-1.48864031e+00 8.05234730e-01 -1.98939249e-01 -5.87320626e-01
-9.19139802e-01 -1.22780704e+00 -4.94568348e-01 1.64826497e-01
3.19593728e-01 -6.29944265e-01 1.92419243e+00 -9.65290010e-01
-1.83605278e+00 7.63581455e-01 -2.60188505e-02 -1.72920436e-01
1.06539023e+00 -8.54427636e-01 -3.19950968e-01 9.65237916e-02
6.23048730e-02 1.30767569e-01 8.77183616e-01 -8.30103278e-01
-4.02307212e-01 -1.36373878e-01 1.90425068e-01 7.00205117e-02
-8.53184938e-01 2.07414135e-01 -6.35052845e-02 -1.00547802e+00
-5.83835125e-01 -7.81736493e-01 1.01628780e-01 -2.01955840e-01
-5.99757075e-01 -5.41267216e-01 6.19718909e-01 -7.07764387e-01
1.32536077e+00 -1.85743523e+00 -2.51028270e-01 -2.25249842e-01
4.32668000e-01 5.25379896e-01 -6.92655027e-01 3.70930851e-01
-1.38465583e-01 -5.38447201e-02 5.30864112e-02 -5.24641573e-01
5.98792434e-01 -6.10304356e-01 -5.36834002e-01 2.57061750e-01
2.02857673e-01 1.09811163e+00 -1.14951789e+00 -7.74962977e-02
-1.60336569e-01 -1.53437093e-01 -4.24030781e-01 5.62166512e-01
-4.10204470e-01 2.42215306e-01 -4.65192348e-01 4.67690319e-01
5.33673048e-01 9.21005104e-03 -3.81789953e-01 4.05958861e-01
2.02011436e-01 7.00449526e-01 -7.56343246e-01 1.57528877e+00
-2.40958944e-01 6.67301476e-01 -1.50995210e-01 -2.48528481e-01
9.03771400e-01 3.66885990e-01 8.55208337e-02 -8.43564212e-01
-1.02846727e-01 3.54950845e-01 -1.54014245e-01 -3.85063648e-01
8.22400510e-01 1.47168905e-01 -5.73962867e-01 1.15951884e+00
8.66181925e-02 -1.99108362e-01 1.71443030e-01 7.33282685e-01
1.63767290e+00 2.43463516e-02 4.65133190e-02 -9.26352739e-02
6.71254337e-01 -3.02441776e-01 1.53975919e-01 1.31267738e+00
-4.89913940e-01 7.14392483e-01 7.34276295e-01 -3.67506891e-01
-8.39855015e-01 -1.15056086e+00 3.29189986e-01 1.59056211e+00
-2.78361171e-01 -3.57085764e-01 -9.89608288e-01 -1.36983252e+00
1.55330330e-01 1.08949387e+00 -7.07024574e-01 -4.05054510e-01
-3.14388990e-01 -4.91917312e-01 9.57891285e-01 3.70610446e-01
4.98840094e-01 -1.08359945e+00 -3.29604864e-01 1.00813314e-01
-2.89353162e-01 -9.08169985e-01 -9.80412602e-01 -1.96262881e-01
-2.73500860e-01 -1.04312432e+00 -7.49621928e-01 -2.74340987e-01
2.92125165e-01 1.71234146e-01 1.61008036e+00 -5.24426810e-02
9.35489461e-02 5.45923293e-01 -4.33072746e-01 -5.83037615e-01
-6.32131517e-01 3.30432236e-01 1.10256135e-01 -1.35640621e-01
6.17504120e-01 -4.07039315e-01 -7.15841591e-01 4.12478261e-02
-7.64950931e-01 -4.60898042e-01 6.02923751e-01 1.11434650e+00
-1.76879615e-01 -6.03365004e-01 1.06767046e+00 -1.22110975e+00
1.43539560e+00 -1.73410267e-01 -3.31225753e-01 3.09231907e-01
-7.28385448e-01 1.37650013e-01 1.12427795e+00 -6.59062088e-01
-8.82774651e-01 -1.58473253e-01 -1.39349192e-01 -3.18089634e-01
-6.07170351e-02 1.68859482e-01 -1.66898683e-01 -6.92949891e-02
1.05549788e+00 2.33293086e-01 -1.76494434e-01 -2.13404268e-01
3.96295577e-01 8.85023415e-01 9.23132539e-01 -8.49801660e-01
9.53282058e-01 -3.00527126e-01 -4.50659722e-01 -4.61971089e-02
-1.21722257e+00 -3.10160995e-01 -3.61918837e-01 -2.08139434e-01
1.09326482e-01 -8.64937782e-01 -6.53087437e-01 7.02615201e-01
-1.26718593e+00 -3.50041866e-01 -4.36035901e-01 2.11209282e-01
-4.88409817e-01 5.70828676e-01 -8.76817107e-01 -7.65669286e-01
-9.05817747e-01 -9.97342229e-01 9.66938734e-01 2.00360313e-01
-7.27389157e-01 -8.47911417e-01 4.05322492e-01 4.73546684e-01
3.75990599e-01 1.22938834e-01 7.82176495e-01 -1.16412258e+00
-8.14659372e-02 -6.88301384e-01 2.06260700e-02 5.41091263e-01
-2.46430680e-01 -1.46101117e-01 -1.37866974e+00 -3.31197321e-01
6.84341714e-02 -9.17906880e-01 1.08104408e+00 -7.51883443e-03
1.20799863e+00 -5.10180593e-01 3.96650195e-01 3.46565932e-01
8.68821979e-01 -3.64475101e-01 5.82056940e-01 4.25203532e-01
4.35935646e-01 4.06773359e-01 5.83335161e-01 4.95991766e-01
3.02520633e-01 4.30503726e-01 2.88599759e-01 1.09638825e-01
-1.59708813e-01 -4.15653467e-01 9.13769364e-01 7.02328920e-01
4.82458264e-01 -4.71913695e-01 -4.84986603e-01 5.54555058e-01
-1.54032636e+00 -1.01724279e+00 -2.80206800e-01 2.11932898e+00
1.27446306e+00 3.94675165e-01 3.87493253e-01 2.97814727e-01
3.19493651e-01 5.32513857e-01 -7.73776174e-01 -7.65089810e-01
-1.63967013e-01 6.20442390e-01 1.52965993e-01 5.01993477e-01
-1.28287113e+00 8.53088856e-01 7.26747704e+00 1.05115604e+00
-6.87587082e-01 1.94110543e-01 6.00712240e-01 -4.21836376e-01
-5.60466588e-01 -3.95721436e-01 -6.41102314e-01 7.02035487e-01
1.06412292e+00 -2.94489473e-01 3.96833837e-01 8.32851231e-01
-1.42279685e-01 2.03534752e-01 -1.13629854e+00 5.10298312e-01
1.49452060e-01 -8.06505799e-01 -1.07362434e-01 -4.87286061e-01
1.04574358e+00 -1.53106600e-01 4.77380276e-01 8.39416981e-01
8.33848000e-01 -1.00274813e+00 9.41789806e-01 4.02853608e-01
1.16385257e+00 -7.07957983e-01 8.20425987e-01 4.99772936e-01
-3.43437523e-01 -1.83689281e-01 -1.83478981e-01 -3.58597100e-01
-3.58667612e-01 7.32049882e-01 -6.64594352e-01 4.65079129e-01
2.11477488e-01 3.12855631e-01 -8.64886284e-01 8.44701052e-01
-8.63251150e-01 1.08427036e+00 1.00609377e-01 -5.30359685e-01
4.39664960e-01 3.31628948e-01 5.83926082e-01 1.33549106e+00
-2.60456037e-02 1.29995532e-02 -3.74552533e-02 8.77688289e-01
-6.83323622e-01 -8.58582556e-02 -2.22309038e-01 1.76734142e-02
5.83576500e-01 1.36909354e+00 1.89449355e-01 -2.68994033e-01
-4.33287650e-01 1.18365669e+00 5.06472468e-01 3.47184718e-01
-8.44866574e-01 -9.15705800e-01 5.41068137e-01 -1.25444278e-01
-5.92212081e-02 2.36796081e-01 -6.64769530e-01 -1.14042854e+00
1.89210951e-01 -1.52625775e+00 2.67221868e-01 -5.09669542e-01
-1.81688607e+00 3.20055902e-01 -5.84052324e-01 -1.09664834e+00
-4.55057591e-01 -7.04542220e-01 -1.20639586e+00 9.66228962e-01
-1.37000060e+00 -8.98576438e-01 -3.65141124e-01 6.03982866e-01
6.95351124e-01 -5.18424392e-01 1.08129430e+00 -1.23788789e-02
-2.94019133e-01 1.68128026e+00 5.04730821e-01 4.39961731e-01
1.38540065e+00 -1.61806083e+00 8.35009038e-01 9.93306577e-01
8.19132179e-02 3.35221261e-01 7.88341284e-01 -4.36082244e-01
-1.23921657e+00 -9.31393504e-01 1.00831902e+00 -1.16608715e+00
7.51523793e-01 -5.46688735e-01 -8.12246263e-01 4.80119348e-01
2.90880561e-01 -2.02099338e-01 6.20954871e-01 3.27147633e-01
-9.07686651e-01 6.80109933e-02 -1.13841248e+00 8.05291772e-01
7.45058596e-01 -7.56133378e-01 -7.62251854e-01 5.73398173e-01
6.91937447e-01 -6.05177939e-01 -4.91797298e-01 1.23174191e-01
8.83167028e-01 -1.03060007e+00 8.84126842e-01 -9.40394282e-01
1.07464385e+00 4.25190270e-01 2.21898586e-01 -1.62767744e+00
-5.39152503e-01 -1.17688000e+00 -4.11376625e-01 1.30785549e+00
2.60962576e-01 -3.74993801e-01 9.83206868e-01 7.35245347e-01
-1.11500449e-01 -7.09755123e-01 -7.93330252e-01 -9.23740983e-01
7.08650768e-01 -1.83984816e-01 7.07535088e-01 9.07651544e-01
2.19238281e-01 5.29887974e-01 -4.66150641e-01 -2.70927668e-01
6.49315000e-01 4.15355712e-02 1.09261525e+00 -1.09179282e+00
-3.49889219e-01 -8.17561507e-01 2.24021375e-02 -1.01508069e+00
5.11848748e-01 -7.01059878e-01 4.43231687e-02 -1.09966123e+00
3.03912967e-01 -5.71003966e-02 -5.09445310e-01 4.30922568e-01
-1.13590050e+00 1.67670831e-01 3.23388845e-01 -1.24544613e-01
-9.59151089e-01 6.83465362e-01 1.21026576e+00 -3.90162110e-01
-4.65544537e-02 1.59643322e-01 -1.18115151e+00 3.77229333e-01
8.31448197e-01 -6.36978328e-01 -2.07652807e-01 -4.03189480e-01
5.08111082e-02 -1.58609018e-01 1.81300148e-01 -7.61233091e-01
3.44261914e-01 -9.17595718e-03 4.08671349e-01 -4.69028324e-01
2.95238961e-02 -1.42375261e-01 -7.67505646e-01 3.96088183e-01
-8.01196337e-01 2.20062193e-02 9.01902467e-02 4.30604845e-01
3.73597592e-02 -4.19282377e-01 5.25037467e-01 -4.42890152e-02
-7.45020062e-02 1.41930282e-01 -2.18564466e-01 6.59981728e-01
6.95399642e-01 7.10809976e-02 -7.75875330e-01 -7.71288693e-01
4.36272994e-02 2.96262830e-01 1.50102615e-01 6.74698293e-01
4.70857739e-01 -1.20904577e+00 -1.36691582e+00 2.63784617e-01
3.83213721e-02 -5.43428242e-01 2.29670957e-01 5.43337643e-01
-1.88906044e-01 5.39025247e-01 5.85865043e-02 -6.85807616e-02
-1.26642919e+00 2.75215566e-01 6.13321783e-03 -9.32999134e-01
-1.01683289e-01 1.09572875e+00 7.56607577e-02 -9.58415747e-01
4.23596740e-01 2.87413448e-01 6.29869923e-02 -3.20284247e-01
8.79152477e-01 4.44104701e-01 2.61857897e-01 -3.25892940e-02
-9.18216538e-03 7.95161948e-02 -4.67808902e-01 -1.34435698e-01
1.17184711e+00 4.40243810e-01 3.39790225e-01 1.52343661e-01
8.42441618e-01 4.99442905e-01 -1.36121821e+00 -3.28418702e-01
5.61526194e-02 -5.87646186e-01 -1.54186308e-01 -1.49485946e+00
-7.20429122e-01 1.00964236e+00 9.44500789e-02 -1.30435869e-01
9.56359863e-01 -5.85470557e-01 1.04310000e+00 3.93127680e-01
2.06899479e-01 -1.08578539e+00 6.91445053e-01 8.53314996e-01
9.17883933e-01 -1.12699008e+00 1.93234265e-01 6.33473471e-02
-6.88487351e-01 1.12075198e+00 7.77050257e-01 -3.75275522e-01
5.49615249e-02 2.42480397e-01 9.54516456e-02 9.82259884e-02
-1.01304519e+00 4.21845704e-01 5.08862615e-01 4.53025401e-01
4.25956279e-01 8.08660313e-02 -4.31219071e-01 1.34410286e+00
-6.79471135e-01 -1.32918984e-01 6.18743837e-01 5.30438423e-01
-2.37252280e-01 -1.08134031e+00 -4.28803742e-01 6.44098938e-01
-9.37949002e-01 -2.50732541e-01 -9.65098500e-01 6.34787619e-01
-4.43040818e-01 1.13448346e+00 -4.44043338e-01 -7.90692449e-01
5.75756550e-01 4.16715533e-01 3.60421538e-01 -2.40211248e-01
-1.45398724e+00 -6.15119994e-01 1.28355116e-01 -4.68353063e-01
3.39815557e-01 -5.98215699e-01 -6.51785612e-01 -6.03669763e-01
-4.74563032e-01 1.42778635e-01 2.93819308e-01 8.13378215e-01
3.09338599e-01 6.38020873e-01 1.14363468e+00 -5.12951434e-01
-1.68531334e+00 -1.23331630e+00 -4.60841149e-01 5.53807735e-01
5.04520237e-01 -2.33483657e-01 -6.99465990e-01 -3.50024849e-01] | [11.25451946258545, 9.324196815490723] |
9c5ed696-4a08-4516-ae1a-95e5dbb70d8b | scalable-neural-probabilistic-answer-set | 2306.08397 | null | https://arxiv.org/abs/2306.08397v1 | https://arxiv.org/pdf/2306.08397v1.pdf | Scalable Neural-Probabilistic Answer Set Programming | The goal of combining the robustness of neural networks and the expressiveness of symbolic methods has rekindled the interest in Neuro-Symbolic AI. Deep Probabilistic Programming Languages (DPPLs) have been developed for probabilistic logic programming to be carried out via the probability estimations of deep neural networks. However, recent SOTA DPPL approaches allow only for limited conditional probabilistic queries and do not offer the power of true joint probability estimation. In our work, we propose an easy integration of tractable probabilistic inference within a DPPL. To this end, we introduce SLASH, a novel DPPL that consists of Neural-Probabilistic Predicates (NPPs) and a logic program, united via answer set programming (ASP). NPPs are a novel design principle allowing for combining all deep model types and combinations thereof to be represented as a single probabilistic predicate. In this context, we introduce a novel $+/-$ notation for answering various types of probabilistic queries by adjusting the atom notations of a predicate. To scale well, we show how to prune the stochastically insignificant parts of the (ground) program, speeding up reasoning without sacrificing the predictive performance. We evaluate SLASH on a variety of different tasks, including the benchmark task of MNIST addition and Visual Question Answering (VQA). | ['Kristian Kersting', 'Devendra Singh Dhami', 'Daniel Ochs', 'Arseny Skryagin'] | 2023-06-14 | null | null | null | null | ['visual-question-answering-1', 'probabilistic-programming'] | ['computer-vision', 'methodology'] | [ 5.21136075e-02 4.89692569e-01 -8.72754082e-02 -6.63571835e-01
-8.68574500e-01 -7.60500550e-01 5.96477628e-01 1.83320045e-01
-3.65460366e-01 6.74920917e-01 -2.04980358e-01 -5.56383967e-01
-4.10092682e-01 -1.18125188e+00 -1.25975406e+00 -5.87760150e-01
-1.04055583e-01 9.11123455e-01 6.29630983e-01 2.75949035e-02
-2.11919948e-01 5.60130179e-01 -1.74797380e+00 6.79111481e-01
7.52476871e-01 1.17078698e+00 -9.65313315e-02 3.85926813e-01
-4.43391532e-01 1.06813431e+00 -5.86342931e-01 -7.25861847e-01
-1.56902105e-01 1.12978429e-01 -6.95547640e-01 -8.80495071e-01
4.30143416e-01 -2.51050800e-01 -2.94450760e-01 1.12315714e+00
2.16933146e-01 5.50591089e-02 6.68773532e-01 -1.53376877e+00
-3.99120986e-01 1.54561269e+00 -1.67801213e-02 4.69897352e-02
3.07620674e-01 8.33371207e-02 1.43704700e+00 -4.67888832e-01
3.75175685e-01 1.61944902e+00 6.91585720e-01 3.94494474e-01
-1.45544004e+00 -3.47981066e-01 5.06474115e-02 5.77540457e-01
-1.61490870e+00 -1.48387505e-02 7.88349450e-01 -3.28553110e-01
1.25150692e+00 3.69728804e-01 3.50303471e-01 9.40692306e-01
1.82965949e-01 1.08044338e+00 8.85850430e-01 -3.77285808e-01
6.17077112e-01 1.62291527e-01 4.22108620e-01 8.68960261e-01
1.97406873e-01 1.44881696e-01 -5.30833602e-01 -4.26936954e-01
2.33799756e-01 -3.69125962e-01 9.29498859e-03 -4.76965666e-01
-7.53772616e-01 9.76035357e-01 3.27561677e-01 4.91887070e-02
-1.79463118e-01 7.21858442e-01 4.76793796e-01 -1.00312017e-01
-8.58306363e-02 3.70804250e-01 -6.44947052e-01 -8.34736824e-02
-8.77991557e-01 6.76473439e-01 1.16749346e+00 8.40563536e-01
5.97512007e-01 -1.51877314e-01 -6.30484819e-01 5.24094105e-01
4.68881339e-01 6.23920500e-01 5.67140654e-02 -1.18286085e+00
2.17268825e-01 5.65112352e-01 -6.35319725e-02 -9.50245917e-01
-2.67460644e-01 -1.64872929e-01 -4.66682017e-01 1.71070591e-01
5.13262153e-01 9.59352404e-03 -6.45058453e-01 2.16551113e+00
2.20869407e-01 -1.15223505e-01 6.93983212e-02 3.59528244e-01
6.68435752e-01 8.68036568e-01 2.75907397e-01 2.50468671e-01
1.42534077e+00 -4.65403348e-01 -3.76370430e-01 -8.99106264e-02
5.24670422e-01 -1.37503576e-02 1.16485655e+00 6.19469345e-01
-1.08514500e+00 -2.00517610e-01 -9.31618750e-01 -1.72579363e-01
-4.46802288e-01 -1.01217553e-01 7.70736039e-01 7.54936993e-01
-1.02049088e+00 6.26750946e-01 -9.18472052e-01 3.94551635e-01
6.14693284e-01 4.99870121e-01 -1.54276267e-01 -1.68429911e-01
-1.42309165e+00 9.35069978e-01 9.86121058e-01 6.64303005e-02
-6.33843720e-01 -7.94789076e-01 -1.05912697e+00 5.31873643e-01
5.58539987e-01 -7.75780737e-01 1.32975912e+00 -6.51302755e-01
-1.57732093e+00 5.97035050e-01 -1.29911527e-01 -9.18447673e-01
4.09332603e-01 -2.74917722e-01 4.58484776e-02 1.15020961e-01
-3.20132196e-01 8.70083332e-01 4.63877767e-01 -9.57256258e-01
-4.67070401e-01 -2.01801106e-01 6.09723628e-01 -1.63675204e-01
2.04769015e-01 1.50982708e-01 -3.06367248e-01 -1.62694827e-01
-9.31637660e-02 -8.28617692e-01 1.12432418e-02 3.29964161e-02
-7.01162636e-01 -7.01255620e-01 4.48259175e-01 -3.39003146e-01
8.11487019e-01 -2.14972711e+00 2.67089009e-01 2.87104964e-01
1.48272544e-01 1.43988505e-01 1.93381339e-01 -3.82315256e-02
5.44239767e-02 2.22127140e-01 -5.71750164e-01 -2.94570684e-01
7.45879114e-01 6.86599791e-01 -8.23441625e-01 1.54434130e-01
5.79452038e-01 1.07694221e+00 -7.75530994e-01 -7.82792926e-01
1.14099151e-02 3.78679723e-01 -9.63373065e-01 5.46873957e-02
-1.16725838e+00 -3.17053795e-01 -3.33392620e-01 5.56132436e-01
5.67517221e-01 -3.71471159e-02 1.50499880e-01 -1.51974723e-01
3.11857879e-01 4.19607371e-01 -1.28420341e+00 1.36502171e+00
-2.83194423e-01 3.89193147e-01 -3.19071680e-01 -7.20575213e-01
6.83484733e-01 -6.71481490e-02 1.59501918e-02 -3.12149644e-01
1.99807495e-01 1.14664890e-01 -3.25829675e-03 -2.90664673e-01
3.61189455e-01 -3.59182417e-01 -3.28660786e-01 2.46005595e-01
2.51954913e-01 -4.30490822e-01 3.70285600e-01 7.55402669e-02
1.01829636e+00 5.81744432e-01 1.09907255e-01 -2.29971349e-01
4.69338357e-01 6.69538081e-02 5.52735686e-01 1.02522802e+00
-2.58793533e-02 2.28245556e-01 1.11958528e+00 -2.35302433e-01
-7.18855619e-01 -1.49864495e+00 -1.71511456e-01 1.17220140e+00
-2.21035853e-01 -4.00324196e-01 -5.77901065e-01 -6.10484838e-01
1.52868211e-01 1.25583971e+00 -4.66507226e-01 -2.55636901e-01
-5.48278093e-01 -7.91495681e-01 1.08940470e+00 6.04877114e-01
2.04938456e-01 -1.16224182e+00 -8.88665974e-01 1.73671424e-01
-7.12958202e-02 -1.05394208e+00 1.79604009e-01 6.47473633e-01
-6.17608964e-01 -7.87858844e-01 -2.50562549e-01 -3.93331647e-01
9.69739929e-02 -7.28009999e-01 1.34336364e+00 -5.10855436e-01
8.90261084e-02 1.75929606e-01 1.56173613e-02 -5.63084185e-01
-5.16573668e-01 9.87054221e-03 6.61184173e-03 -3.11213076e-01
3.99712265e-01 -8.15439284e-01 1.01107538e-01 -8.65436792e-02
-1.12586105e+00 -1.03278400e-03 3.63532662e-01 6.84910893e-01
8.37405980e-01 2.58042037e-01 2.31859669e-01 -9.14716601e-01
3.54037613e-01 -4.51713145e-01 -1.05932832e+00 4.52326328e-01
-1.14444859e-01 7.89935768e-01 7.17556715e-01 -5.42213857e-01
-9.30437028e-01 4.88155410e-02 -3.53107214e-01 -5.00846922e-01
-1.95734158e-01 8.55861545e-01 -4.64076608e-01 2.95521587e-01
7.56676972e-01 1.82973355e-01 -2.61448860e-01 -1.30231246e-01
7.53314734e-01 5.56703517e-03 7.14926302e-01 -1.28160083e+00
5.22025108e-01 3.74507815e-01 2.71928310e-01 -2.25788683e-01
-1.01117086e+00 2.09166110e-01 -2.57081777e-01 2.11505339e-01
5.31221271e-01 -5.37872612e-01 -1.14430106e+00 2.87373036e-01
-1.42154241e+00 -4.48798776e-01 -6.42696261e-01 2.05802500e-01
-7.25194752e-01 2.75225490e-01 -5.68302274e-01 -9.65891302e-01
-1.59818500e-01 -1.34395039e+00 9.24016833e-01 1.46958604e-01
-2.70608038e-01 -6.63605809e-01 -1.03946310e-02 -1.82173342e-01
2.89815098e-01 2.41313919e-01 1.48030508e+00 -9.76385951e-01
-6.89554393e-01 -2.15117991e-01 -3.25354815e-01 4.02611285e-01
-6.41854346e-01 3.06661755e-01 -1.27509940e+00 5.27963281e-01
-5.41953370e-02 -3.49833935e-01 9.62501585e-01 3.93695623e-01
1.42933333e+00 -5.48179388e-01 -1.62792563e-01 4.77353424e-01
1.37129056e+00 2.13380926e-03 7.04628229e-01 1.22307211e-01
5.64730585e-01 6.09510422e-01 9.71700996e-02 3.64982545e-01
5.63298881e-01 4.62462068e-01 6.93028152e-01 6.99740767e-01
2.20996723e-01 -3.84881705e-01 4.49745119e-01 1.56850711e-01
2.96841741e-01 -3.17425162e-01 -1.18021929e+00 3.68449241e-01
-1.76203763e+00 -9.53692257e-01 -8.43731314e-03 2.03367162e+00
1.32687378e+00 4.53117251e-01 -3.71571667e-02 2.24956244e-01
4.67642605e-01 1.11138999e-01 -4.34680492e-01 -5.05223155e-01
-3.58673215e-01 4.98491615e-01 2.63686597e-01 6.03267252e-01
-1.01653671e+00 8.73947382e-01 5.97296000e+00 1.14120984e+00
-8.59983802e-01 -5.08737117e-02 5.42743087e-01 -1.32455811e-01
-7.21787095e-01 5.80946766e-02 -1.00093615e+00 2.69926161e-01
1.24052250e+00 3.61385942e-02 3.70894194e-01 1.24671602e+00
-4.80257541e-01 -2.49177858e-01 -1.56501257e+00 9.06236529e-01
-2.09334984e-01 -1.39875269e+00 3.04271489e-01 -3.33455533e-01
1.80099964e-01 1.64837494e-01 1.03683427e-01 7.46305227e-01
6.50427878e-01 -1.19391096e+00 1.24452043e+00 7.88688421e-01
4.13977265e-01 -8.31914604e-01 7.27880597e-01 4.70153153e-01
-6.55729592e-01 -2.24310353e-01 -2.19605610e-01 4.97099496e-02
1.42398730e-01 9.16515589e-01 -6.85038984e-01 4.03768718e-01
6.53079212e-01 2.13149078e-02 -3.35231692e-01 8.23739588e-01
-5.93963742e-01 6.05942965e-01 -9.55723226e-01 -2.12279648e-01
1.18743427e-01 1.93906620e-01 5.35334587e-01 1.21760333e+00
1.59552723e-01 -1.21314496e-01 -1.76641867e-01 1.61714375e+00
-1.55323213e-02 -2.75830179e-01 -3.78578603e-01 -1.50076777e-01
3.83599937e-01 8.32715988e-01 -3.57993513e-01 -3.66858780e-01
1.07883781e-01 5.00590980e-01 6.23838305e-01 1.10833205e-01
-1.18314219e+00 -3.26818258e-01 3.21764827e-01 -1.83614254e-01
3.53590131e-01 -2.22304136e-01 -4.76662129e-01 -9.09634769e-01
1.87660620e-01 -7.21172154e-01 5.22113681e-01 -8.61549675e-01
-1.17559361e+00 4.73269939e-01 5.82833707e-01 -2.47608483e-01
-4.43277389e-01 -8.92047107e-01 -3.39771330e-01 9.63865757e-01
-1.52861619e+00 -9.72705901e-01 2.56220967e-01 3.41639996e-01
-1.03478618e-01 2.40829483e-01 9.53082681e-01 4.70598564e-02
-5.19044936e-01 6.35838866e-01 -3.28836739e-01 -1.17260441e-01
1.77917242e-01 -1.39293110e+00 1.01800792e-01 8.20576966e-01
1.77535847e-01 6.60563469e-01 1.02088559e+00 -3.61722320e-01
-1.50052333e+00 -9.25679088e-01 1.05667162e+00 -5.90813220e-01
7.69650400e-01 -2.84664333e-01 -9.69969511e-01 6.74700379e-01
-2.91765362e-01 3.95269692e-02 4.12177324e-01 2.36165047e-01
-9.39621210e-01 -1.52409628e-01 -1.30607164e+00 5.89803636e-01
4.98352647e-01 -6.82534158e-01 -9.28603113e-01 1.81778893e-01
1.09529126e+00 -5.49952686e-01 -7.03272045e-01 5.49073339e-01
6.53808951e-01 -9.13023353e-01 1.03409350e+00 -4.95031178e-01
6.24529243e-01 -3.81555527e-01 -7.53725767e-01 -7.01170921e-01
4.14724387e-02 -3.18110555e-01 -7.65349984e-01 1.08143139e+00
4.56864595e-01 -6.39954090e-01 7.56040156e-01 9.84989822e-01
-1.17497131e-01 -8.39352548e-01 -1.17383003e+00 -6.79606497e-01
1.95497453e-01 -1.12576616e+00 7.60525286e-01 2.97159225e-01
-5.15951868e-03 -6.50252029e-02 1.75080150e-01 4.35867041e-01
4.50940162e-01 1.56841010e-01 2.76127189e-01 -1.23471236e+00
-9.26391840e-01 -7.69422174e-01 -4.32779521e-01 -8.05130124e-01
5.15656471e-01 -9.20705855e-01 3.34985852e-01 -1.22317910e+00
1.37869954e-01 -6.32789910e-01 -1.77437782e-01 9.07255650e-01
6.13624789e-02 -1.34877339e-01 1.69511467e-01 -1.95937425e-01
-4.65898216e-01 6.03024900e-01 5.10348141e-01 -2.86963612e-01
1.68656297e-02 -5.38301095e-02 -5.94078839e-01 9.45120454e-01
6.62231088e-01 -6.56658292e-01 -4.39652681e-01 -4.63622004e-01
9.81856883e-01 9.17236730e-02 7.12355018e-01 -1.01396132e+00
3.43455076e-01 -2.19918400e-01 -1.05922759e-01 -7.32199371e-01
5.41646481e-01 -6.52508438e-01 9.24513713e-02 3.44004929e-01
-4.57291186e-01 -1.86179996e-01 5.29080391e-01 5.05743623e-01
-2.19699860e-01 -4.12482649e-01 6.62666976e-01 7.30775818e-02
-5.94110310e-01 8.64302814e-02 -1.12011835e-01 8.18938091e-02
7.37556100e-01 2.50129372e-01 -3.07114422e-01 6.55578151e-02
-7.18900204e-01 1.18627064e-01 4.18923460e-02 -2.63546016e-02
3.81997436e-01 -8.83684278e-01 -5.06012261e-01 -1.80266932e-01
4.81682867e-02 4.57750678e-01 3.77052128e-01 5.49130380e-01
-6.57024801e-01 5.23261786e-01 -1.27390800e-02 -6.27523482e-01
-6.98291779e-01 6.53624654e-01 5.48087060e-01 -5.21745265e-01
-4.26565081e-01 1.17786050e+00 1.26688644e-01 -6.73742771e-01
5.06031930e-01 -1.01479650e+00 -7.12392256e-02 -1.37965739e-01
4.31741089e-01 -3.39649245e-02 -8.84547383e-02 -1.47371754e-01
-5.56779504e-01 -3.84743586e-02 1.84271500e-01 -2.47334912e-01
1.21700430e+00 4.47181314e-01 -3.20188373e-01 7.27280498e-01
7.49158204e-01 -3.69558334e-02 -1.09946275e+00 -3.61130893e-01
1.82714432e-01 2.52641648e-01 -1.84313774e-01 -1.01844776e+00
-6.20712042e-01 9.76445973e-01 2.68747956e-01 7.21278340e-02
8.49953115e-01 3.59475672e-01 4.63949978e-01 8.61015618e-01
4.63185966e-01 -5.35803556e-01 -5.92400968e-01 8.80997777e-01
7.18794763e-01 -8.45651686e-01 -2.60399729e-01 -3.08833450e-01
-4.25597131e-01 1.14234483e+00 1.91798329e-01 -1.74203455e-01
4.50595886e-01 5.15335560e-01 -6.87164903e-01 -1.45679325e-01
-7.79848158e-01 -6.41141012e-02 3.39029133e-01 5.85721016e-01
-6.29737601e-03 2.83104211e-01 2.21071035e-01 1.23315525e+00
-4.06995326e-01 1.72534794e-01 2.57198989e-01 7.49752998e-01
-3.77493262e-01 -1.03812099e+00 -2.70757943e-01 3.43487829e-01
-5.47637939e-01 -5.01322985e-01 9.63895768e-03 7.04733849e-01
2.44903728e-01 5.02500057e-01 -1.89940900e-01 -1.86989307e-01
7.35548064e-02 4.95476693e-01 7.01659262e-01 -5.95182419e-01
-2.84672797e-01 -6.59679770e-01 2.48179883e-01 -4.64805007e-01
-1.94734901e-01 -4.94508386e-01 -1.35808635e+00 -1.79799601e-01
-1.10756606e-03 -1.28140196e-03 7.07349718e-01 1.19065404e+00
2.47034281e-01 5.42426586e-01 -1.83677733e-01 -5.28102636e-01
-9.58544016e-01 -5.35245180e-01 -6.01156771e-01 -2.19668031e-01
6.14401624e-02 -6.97356462e-01 -2.18568161e-01 -1.47396266e-01] | [8.679677963256836, 7.002292156219482] |
8337fec9-ed29-44bc-8bea-10f39df8122f | geonet-benchmarking-unsupervised-adaptation | 2303.15443 | null | https://arxiv.org/abs/2303.15443v1 | https://arxiv.org/pdf/2303.15443v1.pdf | GeoNet: Benchmarking Unsupervised Adaptation across Geographies | In recent years, several efforts have been aimed at improving the robustness of vision models to domains and environments unseen during training. An important practical problem pertains to models deployed in a new geography that is under-represented in the training dataset, posing a direct challenge to fair and inclusive computer vision. In this paper, we study the problem of geographic robustness and make three main contributions. First, we introduce a large-scale dataset GeoNet for geographic adaptation containing benchmarks across diverse tasks like scene recognition (GeoPlaces), image classification (GeoImNet) and universal adaptation (GeoUniDA). Second, we investigate the nature of distribution shifts typical to the problem of geographic adaptation and hypothesize that the major source of domain shifts arise from significant variations in scene context (context shift), object design (design shift) and label distribution (prior shift) across geographies. Third, we conduct an extensive evaluation of several state-of-the-art unsupervised domain adaptation algorithms and architectures on GeoNet, showing that they do not suffice for geographical adaptation, and that large-scale pre-training using large vision models also does not lead to geographic robustness. Our dataset is publicly available at https://tarun005.github.io/GeoNet. | ['Manmohan Chandraker', 'Wangdong Xu', 'Tarun Kalluri'] | 2023-03-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kalluri_GeoNet_Benchmarking_Unsupervised_Adaptation_Across_Geographies_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kalluri_GeoNet_Benchmarking_Unsupervised_Adaptation_Across_Geographies_CVPR_2023_paper.pdf | cvpr-2023-1 | ['scene-recognition'] | ['computer-vision'] | [ 6.06213883e-03 -1.14464372e-01 1.61199048e-01 -4.01531219e-01
-3.70705038e-01 -7.54125297e-01 9.10644472e-01 2.98171993e-02
-6.34534478e-01 6.49562776e-01 4.11037624e-01 -2.13536054e-01
1.68350805e-02 -5.08679450e-01 -9.77140665e-01 -3.94473583e-01
1.44585609e-01 3.96100312e-01 2.90749848e-01 -3.86279672e-01
3.86437654e-01 5.93815029e-01 -1.51371491e+00 -1.37093216e-01
9.97879028e-01 4.20850933e-01 9.38422009e-02 6.68510616e-01
1.90790340e-01 3.66471231e-01 -6.39625669e-01 -3.60499233e-01
5.75558543e-01 -2.63674259e-01 -9.32438374e-01 -5.85330129e-02
1.10922623e+00 -2.10073680e-01 -6.34777188e-01 1.10190606e+00
6.90622032e-01 3.30325991e-01 9.48576450e-01 -1.18442786e+00
-1.30197644e+00 2.85900146e-01 -5.79017699e-01 3.80296707e-01
6.13208264e-02 4.99261081e-01 5.09544313e-01 -8.69880676e-01
8.65910649e-01 9.77303982e-01 1.06654036e+00 4.78151679e-01
-1.06434143e+00 -3.67121190e-01 3.88805538e-01 4.40140158e-01
-1.66215885e+00 -7.32489824e-01 7.18580842e-01 -6.32372141e-01
1.02907968e+00 8.21835548e-02 3.69476736e-01 1.27695930e+00
-3.00411433e-02 3.79288495e-01 1.11534464e+00 -6.04545116e-01
2.85688132e-01 3.08514118e-01 -1.45196512e-01 3.15511316e-01
3.13876361e-01 2.43688852e-01 -3.57290506e-01 1.21955387e-01
6.50157332e-01 -1.95731089e-01 -3.28917056e-01 -6.05994165e-01
-1.06791723e+00 6.98088467e-01 7.14612126e-01 2.02915013e-01
-2.23999202e-01 2.37150602e-02 2.63874173e-01 2.61593431e-01
4.84173030e-01 7.87118852e-01 -5.06958961e-01 3.66287213e-03
-6.85177445e-01 2.65448928e-01 6.37402534e-01 1.15850782e+00
9.31899428e-01 1.76636696e-01 2.09491909e-01 1.09806490e+00
-2.14756168e-02 5.95667243e-01 6.29991889e-01 -7.16773510e-01
3.15177560e-01 3.44114780e-01 -8.94610118e-03 -1.14948285e+00
-5.70918381e-01 -3.72721046e-01 -5.19475162e-01 1.35531247e-01
7.57249475e-01 -1.49810642e-01 -1.20198977e+00 1.92634463e+00
3.16359937e-01 3.84142786e-01 1.03159091e-02 7.98159957e-01
8.67079616e-01 2.22640455e-01 3.33933175e-01 5.36837816e-01
1.05951595e+00 -8.64805222e-01 1.22008391e-01 -7.49011040e-01
4.95930165e-01 -8.79833579e-01 1.17697740e+00 -9.71133709e-02
-5.38638949e-01 -6.28483117e-01 -9.98854160e-01 1.87611535e-01
-9.29078579e-01 1.28649808e-02 4.76093680e-01 5.63715816e-01
-1.24692774e+00 2.05933571e-01 -2.09122673e-01 -1.39442265e+00
2.76894450e-01 1.57202840e-01 -5.19039035e-01 -3.04076940e-01
-1.04980755e+00 1.18471634e+00 5.41944504e-01 -1.69549018e-01
-7.82517195e-01 -8.25432956e-01 -9.08455312e-01 -3.27192724e-01
1.62521169e-01 -6.68475151e-01 1.12660348e+00 -1.10653520e+00
-1.07685256e+00 1.10996175e+00 1.58975542e-01 -3.69708627e-01
4.55487549e-01 -2.73439121e-02 -8.30754519e-01 -1.43908439e-02
3.38847399e-01 8.02976072e-01 7.84984887e-01 -1.32900310e+00
-7.41996348e-01 -4.43008631e-01 9.73511413e-02 3.97029430e-01
-2.85218537e-01 -6.66918606e-02 -4.56041098e-01 -8.03234696e-01
-2.62562800e-02 -1.20203233e+00 -2.78408706e-01 -3.20020437e-01
-2.03358144e-01 9.43412930e-02 3.69886070e-01 -7.84981072e-01
9.31420028e-01 -2.26602578e+00 -1.19419537e-01 1.50148019e-01
-1.32990219e-02 2.71383345e-01 -3.68655235e-01 4.26224709e-01
-7.25644603e-02 1.38556898e-01 -2.64325291e-01 4.22349200e-02
-2.09290721e-02 1.14614926e-01 -4.70148891e-01 6.57875836e-01
5.36233820e-02 9.09056425e-01 -8.36359680e-01 -2.22955123e-01
3.58242780e-01 1.04360253e-01 -5.64041853e-01 -5.93626425e-02
-2.52593271e-02 6.66555822e-01 -1.14786886e-01 8.08459401e-01
8.16282690e-01 -6.68050498e-02 -1.49978966e-01 3.03658620e-02
-2.38608092e-01 -1.20957322e-01 -1.02024341e+00 1.69788468e+00
-3.57579172e-01 7.46730745e-01 -3.81865710e-01 -7.66015470e-01
7.54468322e-01 -3.04453462e-01 1.85171455e-01 -9.32723701e-01
-1.12812147e-01 1.82699755e-01 -2.95358896e-01 -5.40051103e-01
9.66930091e-01 6.71233907e-02 -3.99645507e-01 2.68873990e-01
3.06781739e-01 -2.31803715e-01 3.88360396e-02 3.05560976e-02
8.93654406e-01 1.15115911e-01 4.03311670e-01 -3.15492213e-01
1.86150491e-01 6.48977160e-01 5.56475461e-01 1.19115376e+00
-6.64172649e-01 9.10671651e-01 -4.49095219e-02 -5.35770297e-01
-1.34936523e+00 -1.13159978e+00 -1.40408456e-01 1.32804763e+00
2.99226522e-01 8.41673762e-02 -7.40376234e-01 -7.39638209e-01
1.23292327e-01 7.20787704e-01 -8.14400494e-01 -3.90402526e-01
-4.09340650e-01 -8.91095102e-01 9.32995439e-01 6.46886170e-01
6.30322814e-01 -9.42604065e-01 -4.44295377e-01 1.27280205e-01
-8.13215822e-02 -1.05673981e+00 -5.15192747e-01 -1.45561963e-01
-5.08599699e-01 -1.04163730e+00 -8.00244212e-01 -7.26257920e-01
6.69327736e-01 5.89636385e-01 1.30083847e+00 -1.34144962e-01
-3.28973621e-01 8.79965603e-01 -4.43485856e-01 -5.98440945e-01
-1.75200358e-01 4.52936381e-01 2.83129871e-01 -1.41794100e-01
5.93978047e-01 -5.18393457e-01 -6.82934761e-01 5.70159078e-01
-7.96864748e-01 -3.59131962e-01 5.30416965e-01 6.05898738e-01
2.97303259e-01 -2.10360393e-01 6.49446309e-01 -9.25584495e-01
4.80867714e-01 -7.64701188e-01 -6.21093988e-01 5.52981734e-01
-5.64534366e-01 -1.95802614e-01 5.17452955e-01 -3.44766825e-01
-1.24456203e+00 1.56914681e-01 5.65868132e-02 -2.71103144e-01
-7.15898752e-01 5.15339196e-01 -4.67901006e-02 -4.41785485e-01
1.40510762e+00 2.30358899e-01 -3.15980434e-01 -3.93042207e-01
5.99491894e-01 4.82805759e-01 8.47891390e-01 -5.67148030e-01
1.12521255e+00 6.32835150e-01 -4.08549875e-01 -1.15057623e+00
-5.80749750e-01 -5.59412479e-01 -9.60122883e-01 -1.57336563e-01
7.33737946e-01 -1.07006776e+00 1.56164080e-01 8.02826762e-01
-6.98078811e-01 -7.02818632e-01 -3.31880569e-01 3.91024500e-01
-3.90136868e-01 3.10726583e-01 7.03076944e-02 -2.56153375e-01
1.57478571e-01 -6.31325543e-01 6.14520907e-01 6.66480362e-01
-1.03535987e-01 -1.42765677e+00 4.26314294e-01 1.37230948e-01
5.69453359e-01 2.66260713e-01 5.88803053e-01 -6.27770841e-01
-4.48487163e-01 2.01238245e-02 -3.26838046e-01 -2.53507923e-02
1.37831911e-01 2.99026929e-02 -1.15890884e+00 -3.25027019e-01
-4.78664607e-01 -2.25777358e-01 8.70008349e-01 5.10965288e-01
8.34209621e-01 -7.16621056e-02 -3.15268427e-01 1.19525480e+00
1.59635425e+00 7.89340511e-02 6.74318671e-01 8.81754816e-01
8.54946256e-01 5.90347171e-01 7.09661543e-01 2.23054156e-01
6.56729162e-01 6.36641264e-01 1.65450245e-01 -2.73954332e-01
-4.42792118e-01 -3.31553489e-01 6.34105876e-02 1.60982206e-01
1.78609770e-02 -9.43203047e-02 -1.41500115e+00 1.21341813e+00
-1.95513237e+00 -8.52895558e-01 -5.20001948e-02 2.25234485e+00
5.21964252e-01 -2.40820393e-01 1.94112837e-01 -5.92503071e-01
8.55566084e-01 -6.05040863e-02 -8.73559296e-01 -1.66428223e-01
-5.18268049e-01 -1.55687436e-01 1.10994184e+00 2.65013307e-01
-1.38546515e+00 1.35121560e+00 6.84988070e+00 5.19365191e-01
-1.25057495e+00 -3.91089416e-04 4.35586452e-01 1.91117570e-01
-2.86212023e-02 6.43192232e-02 -6.97102129e-01 3.67358208e-01
7.30947316e-01 -2.38022998e-01 4.25588161e-01 1.09081495e+00
-1.22580126e-01 -1.79401860e-01 -7.24876583e-01 9.29657698e-01
2.94610322e-01 -1.12389767e+00 -1.19988441e-01 2.97512226e-02
1.12964737e+00 5.15251160e-01 4.07907546e-01 2.90907145e-01
7.42045462e-01 -8.38200688e-01 8.43719482e-01 4.50146526e-01
1.05238426e+00 -4.87731814e-01 5.28104663e-01 -1.78714424e-01
-1.02537119e+00 -1.99368030e-01 -7.47141302e-01 2.69933939e-02
-1.20190702e-01 1.37889862e-01 -1.16856527e+00 4.11748737e-01
1.10481572e+00 6.36905134e-01 -1.34463596e+00 1.41898739e+00
-2.60450006e-01 5.34350634e-01 -3.65539372e-01 4.81143445e-01
2.98283309e-01 6.63266852e-02 6.29711628e-01 1.28532827e+00
3.33869129e-01 -1.97994679e-01 3.05392034e-02 6.20548129e-01
-3.41312587e-02 -1.18825525e-01 -1.08728850e+00 2.73391515e-01
7.27804601e-01 8.82925987e-01 -5.67937851e-01 -1.74952149e-01
-4.76642191e-01 1.05952489e+00 4.41172034e-01 6.99788928e-01
-7.70812750e-01 -3.04417431e-01 8.34973931e-01 -6.63805706e-03
2.14367405e-01 -2.03850776e-01 -5.71347535e-01 -1.24136233e+00
-1.24446131e-01 -7.70865142e-01 8.67800713e-01 -8.09330642e-01
-1.57939470e+00 3.72533262e-01 1.08284242e-01 -1.20797884e+00
-2.17903167e-01 -4.62045491e-01 -6.00025833e-01 9.70316410e-01
-1.69120300e+00 -1.55001783e+00 -7.87156880e-01 8.70456934e-01
4.57816422e-01 -5.01070976e-01 5.90685964e-01 2.97002226e-01
-5.49984932e-01 9.15959537e-01 4.21026289e-01 1.20741986e-01
1.41874826e+00 -1.24913108e+00 9.21945572e-01 1.27259910e+00
6.66308179e-02 7.63784587e-01 8.26876342e-01 -6.05821490e-01
-1.05086935e+00 -1.60962999e+00 6.98833466e-01 -9.17530835e-01
6.24092638e-01 -8.80488083e-02 -9.16943550e-01 9.22352850e-01
5.65648042e-02 -1.17318526e-01 5.90849459e-01 9.65684801e-02
-5.78503966e-01 -4.87534739e-02 -1.24957156e+00 8.36838245e-01
1.32043576e+00 -5.41688442e-01 -6.96559906e-01 2.34757751e-01
5.13405859e-01 -3.60629827e-01 -5.39391577e-01 1.66485429e-01
2.91508198e-01 -9.04837787e-01 1.11292732e+00 -6.27676547e-01
2.19812199e-01 -5.04707575e-01 -1.88214004e-01 -1.69302189e+00
-7.21973836e-01 -3.13171923e-01 4.78270769e-01 1.34867251e+00
5.03010571e-01 -9.24890518e-01 5.60382605e-01 7.31760442e-01
-2.87083775e-01 2.06280380e-01 -8.24225485e-01 -9.91452694e-01
3.15893412e-01 -1.70851886e-01 7.35766530e-01 1.39282513e+00
-4.54646975e-01 -9.69613157e-03 -3.34727466e-01 3.45929295e-01
5.24888396e-01 -2.09576786e-01 1.17782652e+00 -1.14944077e+00
-3.64334174e-02 -4.56373453e-01 -4.93011415e-01 -6.81765854e-01
5.08099161e-02 -6.44632339e-01 8.66928250e-02 -1.48319781e+00
-4.97337952e-02 -6.65553510e-01 -2.82102108e-01 5.45227408e-01
-2.09633991e-01 4.74805802e-01 9.89067778e-02 4.54486877e-01
-5.51023841e-01 2.68586546e-01 6.23633742e-01 -2.78781801e-01
-3.71447980e-01 -3.60897362e-01 -1.09841871e+00 7.18353152e-01
1.10755253e+00 -3.44887197e-01 -3.53760242e-01 -8.52695704e-01
1.35517851e-01 -8.49668860e-01 6.09272838e-01 -1.37441897e+00
4.28939879e-01 -4.73586857e-01 4.36621457e-01 -1.72542006e-01
-4.78410646e-02 -7.07850516e-01 1.12082608e-01 5.33308797e-02
1.17504105e-01 1.30051598e-01 6.28020525e-01 5.93346298e-01
3.82775217e-02 -1.82895795e-01 9.88478184e-01 -2.77625173e-01
-1.77766633e+00 2.43409723e-01 -2.31243610e-01 3.91513675e-01
9.37698245e-01 -6.03886366e-01 -7.02666044e-01 -1.11938134e-01
-4.55586672e-01 4.83211316e-02 1.18580294e+00 7.44639099e-01
3.32615525e-01 -1.15940428e+00 -8.53235662e-01 8.93761292e-02
8.22908878e-01 -1.00468360e-01 5.89810193e-01 5.39217174e-01
-7.31808722e-01 3.26887071e-01 -5.44915259e-01 -5.70263684e-01
-9.67684627e-01 4.62792158e-01 3.99010539e-01 1.36387706e-01
-4.44466591e-01 1.03983140e+00 4.36932445e-01 -8.65324259e-01
-1.22458793e-01 1.27881154e-01 -1.68660775e-01 2.25828178e-02
3.14886391e-01 2.81083077e-01 1.29239127e-01 -9.00834918e-01
-3.96674871e-01 8.05146873e-01 -1.61400735e-01 -1.39921373e-02
1.44564009e+00 -3.50874722e-01 2.59498805e-01 1.43946648e-01
7.13556886e-01 -1.63845807e-01 -1.48276806e+00 -3.72464925e-01
6.86729178e-02 -6.36256754e-01 -2.88250118e-01 -1.09566772e+00
-6.83609962e-01 4.85339195e-01 8.29101562e-01 -1.16036780e-01
1.04637682e+00 6.00321703e-02 4.48697180e-01 4.37998742e-01
4.15817261e-01 -1.67440033e+00 -1.09389007e-01 8.71238530e-01
8.94223452e-01 -1.35964549e+00 -3.70435044e-02 -8.27850699e-02
-8.64983201e-01 6.32847667e-01 9.03731346e-01 -1.59978084e-02
5.29668987e-01 -3.52816373e-01 3.77576292e-01 -2.37585694e-01
-1.48515135e-01 -4.96946126e-01 3.17751735e-01 1.33680582e+00
-5.30613363e-02 8.27215537e-02 2.39090562e-01 1.27576694e-01
-3.56195569e-01 -2.86361575e-01 6.53383970e-01 9.21053231e-01
-3.32480192e-01 -6.88250840e-01 -6.56207860e-01 1.49043381e-01
-3.26485895e-02 -1.30732879e-01 -8.29611719e-01 9.42799866e-01
2.68181533e-01 8.37851584e-01 4.86979671e-02 -3.76981050e-01
5.98368406e-01 3.66201922e-02 3.59213591e-01 -6.71363115e-01
-4.35653657e-01 -4.67253745e-01 3.29628773e-02 -2.86748469e-01
-1.42974362e-01 -7.40268052e-01 -7.35881031e-01 -4.15693820e-01
-8.69422220e-03 -2.31290981e-01 8.28263521e-01 6.48862839e-01
6.78225517e-01 2.35031828e-01 3.58856440e-01 -8.27348232e-01
-3.65100175e-01 -9.69105065e-01 -5.03527045e-01 6.63314462e-01
1.89689070e-01 -7.33181119e-01 -3.74394029e-01 3.05606127e-01] | [9.829038619995117, 1.4682080745697021] |
58047c54-56bb-4d65-8d9f-300ae337b082 | one-for-all-unified-workload-prediction-for | 2306.01507 | null | https://arxiv.org/abs/2306.01507v1 | https://arxiv.org/pdf/2306.01507v1.pdf | One for All: Unified Workload Prediction for Dynamic Multi-tenant Edge Cloud Platforms | Workload prediction in multi-tenant edge cloud platforms (MT-ECP) is vital for efficient application deployment and resource provisioning. However, the heterogeneous application patterns, variable infrastructure performance, and frequent deployments in MT-ECP pose significant challenges for accurate and efficient workload prediction. Clustering-based methods for dynamic MT-ECP modeling often incur excessive costs due to the need to maintain numerous data clusters and models, which leads to excessive costs. Existing end-to-end time series prediction methods are challenging to provide consistent prediction performance in dynamic MT-ECP. In this paper, we propose an end-to-end framework with global pooling and static content awareness, DynEformer, to provide a unified workload prediction scheme for dynamic MT-ECP. Meticulously designed global pooling and information merging mechanisms can effectively identify and utilize global application patterns to drive local workload predictions. The integration of static content-aware mechanisms enhances model robustness in real-world scenarios. Through experiments on five real-world datasets, DynEformer achieved state-of-the-art in the dynamic scene of MT-ECP and provided a unified end-to-end prediction scheme for MT-ECP. | ['Wenyu Wang', 'Cheng Zhang', 'Xiaofei Wang', 'Heng Zhang', 'Zheng Wang', 'Shaoyuan Huang'] | 2023-06-02 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [-4.10276949e-01 -9.67442393e-01 -4.19593185e-01 -1.48148999e-01
-4.18704301e-01 -2.03395799e-01 -1.74509704e-01 -2.38089282e-02
3.85012329e-01 3.74826819e-01 3.70530821e-02 -2.83082604e-01
-3.60591799e-01 -5.18840134e-01 -1.56556040e-01 -5.56450844e-01
-3.87201577e-01 5.63435972e-01 5.17518520e-01 -2.21883599e-02
1.62977085e-01 5.59612036e-01 -1.67268789e+00 7.27019489e-01
8.48603249e-01 1.63346779e+00 7.01002359e-01 9.61518586e-01
-1.80248052e-01 1.29420567e+00 -5.36258519e-01 1.75758228e-02
2.99086958e-01 9.53122675e-02 -5.68327546e-01 1.51710898e-01
-2.99452335e-01 -4.03178036e-02 -3.62042576e-01 3.35550696e-01
8.01932573e-01 -1.52672127e-01 1.09420165e-01 -1.99973798e+00
-2.02498198e-01 3.89486283e-01 -9.07992244e-01 6.87858820e-01
-1.92045808e-01 3.30532730e-01 7.08288729e-01 -7.98907399e-01
4.24471945e-01 4.25888509e-01 7.07094967e-01 -5.76337380e-03
-1.10678267e+00 -5.25035501e-01 2.59706378e-01 8.23250711e-01
-1.45691454e+00 -6.67846382e-01 7.11108685e-01 -4.42613125e-01
1.18212414e+00 7.55073071e-01 3.83951426e-01 3.97211164e-01
3.94751817e-01 6.41680062e-01 7.22411215e-01 -4.87486832e-02
3.33815426e-01 6.06620386e-02 -2.79894948e-01 2.91321110e-02
-2.83223212e-01 -1.91554114e-01 -1.07303476e+00 -2.62642413e-01
3.11748534e-01 1.03140570e-01 -2.28168190e-01 -1.22044630e-01
-1.12187266e+00 -6.03192486e-02 1.16399266e-01 -5.92480972e-02
-1.00414813e+00 -1.06000565e-01 9.84026611e-01 7.71484971e-02
8.14635336e-01 8.90617892e-02 -9.89192545e-01 -6.65875912e-01
-1.47709262e+00 -1.07043199e-01 8.16905737e-01 1.47296548e+00
5.40485382e-01 7.17225224e-02 -6.00796700e-01 9.04661536e-01
-3.03446174e-01 5.02482116e-01 2.91632503e-01 -1.31866491e+00
5.71007788e-01 4.74121541e-01 2.56468635e-03 -1.17496192e+00
-4.30213332e-01 -7.97851443e-01 -9.27504957e-01 -3.20758492e-01
-4.40512329e-01 -2.50367045e-01 -2.81197190e-01 1.49686825e+00
3.10165882e-01 9.15088058e-01 -3.85219216e-01 7.77313173e-01
1.08433604e-01 8.81342769e-01 1.66969821e-01 -7.32502580e-01
1.11916745e+00 -1.29768753e+00 -3.21077168e-01 6.55796602e-02
5.90003550e-01 -1.07545829e+00 1.16703701e+00 2.04864815e-01
-1.18984020e+00 -6.17034793e-01 -5.39333820e-01 5.07333994e-01
2.51016673e-02 1.37261152e-01 2.33574241e-01 1.97662815e-01
-1.06838000e+00 2.70015061e-01 -8.16688597e-01 -3.06938201e-01
4.42044944e-01 2.02563062e-01 2.88207173e-01 -9.22268555e-02
-4.79163855e-01 3.44722629e-01 1.24458306e-01 -2.26111531e-01
-7.93643117e-01 -1.61727536e+00 -3.05741690e-02 5.66933215e-01
5.63833356e-01 -6.73704445e-01 1.27802765e+00 -4.32443589e-01
-1.32888818e+00 3.72471750e-01 -3.80755842e-01 -1.49928689e-01
2.77472168e-01 1.00359552e-01 -9.10770357e-01 2.06970766e-01
6.34818897e-02 -6.23504110e-02 3.68188798e-01 -1.22692168e+00
-8.47993851e-01 -3.27984780e-01 -4.60651577e-01 1.72336712e-01
-6.65318608e-01 3.27067673e-01 -6.58804417e-01 -4.76906359e-01
-1.64570123e-01 -5.98542809e-01 -2.69126177e-01 -4.03808773e-01
-2.96896636e-01 1.67392999e-01 1.42851484e+00 -4.01238620e-01
1.94174623e+00 -2.06447864e+00 -2.79897630e-01 3.23382802e-02
5.40496886e-01 1.11987002e-01 -2.40829661e-01 5.57967722e-01
1.66072652e-01 -3.32953595e-02 3.83066297e-01 -5.53694844e-01
1.80194117e-02 1.76619336e-01 -4.86877769e-01 -1.86550859e-02
-1.66766942e-01 6.71247303e-01 -7.20678866e-01 -7.74182379e-01
3.78688425e-01 2.49484047e-01 -5.69564342e-01 5.81493139e-01
-3.63803238e-01 2.84016043e-01 -4.34857123e-02 8.80327165e-01
9.43174124e-01 -7.49823093e-01 1.25391409e-01 -3.09365481e-01
-2.31094196e-01 -1.76298857e-01 -9.01094139e-01 1.20850945e+00
-1.13107157e+00 7.86633372e-01 1.74546808e-01 -8.65670919e-01
7.39581704e-01 3.31672758e-01 1.25268507e+00 -7.85701156e-01
-6.60155937e-02 2.37770930e-01 -1.99378192e-01 -5.70507526e-01
7.36152351e-01 2.42289588e-01 1.25628337e-01 5.45729756e-01
-4.09946948e-01 4.45101172e-01 1.87447131e-01 1.63372010e-01
1.20198143e+00 -1.12870693e-01 -6.42212182e-02 -3.47913086e-01
3.07151109e-01 7.99553618e-02 1.07006145e+00 1.93758279e-01
-5.19284546e-01 4.97495174e-01 3.21429580e-01 -5.58871806e-01
-1.13598585e+00 -9.56343830e-01 8.49053562e-02 1.57732224e+00
1.03849664e-01 -6.02495492e-01 -4.85120773e-01 -8.61001164e-02
-2.47333854e-01 5.16307294e-01 -2.37884164e-01 6.16482534e-02
-4.37413782e-01 -9.85110104e-01 -4.49477173e-02 4.16200399e-01
5.93929946e-01 -8.24372232e-01 -8.35501313e-01 7.12129414e-01
-5.74702203e-01 -1.75614345e+00 -8.62441778e-01 -1.34766728e-01
-6.70349836e-01 -9.36686933e-01 -2.39935100e-01 -3.42507243e-01
2.30418116e-01 8.58909845e-01 1.70261967e+00 1.78501636e-01
-4.77885872e-01 1.70902431e-01 -2.72242635e-01 -3.40795577e-01
-7.73642287e-02 2.61007667e-01 4.17889953e-02 2.03371316e-01
5.50230801e-01 -1.07449555e+00 -1.09641182e+00 8.90887380e-01
-4.94414270e-01 3.28604102e-01 2.26588368e-01 2.91578501e-01
1.09421170e+00 5.58961689e-01 9.00601983e-01 -8.87595296e-01
5.53991258e-01 -1.04676831e+00 -3.19053829e-01 5.53162873e-01
-1.06366801e+00 -7.95892119e-01 1.10806417e+00 -3.87297213e-01
-9.87257361e-01 -4.00750697e-01 3.40644032e-01 -9.57525134e-01
1.97782684e-02 4.79004413e-01 -9.06026065e-02 2.64433801e-01
5.84731102e-01 5.23072124e-01 -3.31805378e-01 -3.77013534e-01
-9.96295065e-02 9.95032310e-01 5.52105665e-01 -8.10878158e-01
1.53328657e-01 2.70013303e-01 1.07439961e-02 -5.38609147e-01
-5.91355205e-01 -7.56691933e-01 -3.06483597e-01 -5.80360532e-01
1.80888638e-01 -1.27661312e+00 -9.18455541e-01 4.63831723e-01
-7.46267200e-01 -7.74160624e-01 -1.55342266e-01 -1.49047701e-02
-8.20488393e-01 1.35157436e-01 -7.90364385e-01 -7.40522444e-01
-7.51837313e-01 -1.13533068e+00 9.07549441e-01 9.09629539e-02
4.01441082e-02 -7.50655115e-01 -1.60780013e-01 2.52956092e-01
1.15994823e+00 1.85199976e-01 6.92900479e-01 -2.31136121e-02
-6.36321664e-01 -1.06564194e-01 -7.37604797e-01 2.98853554e-02
1.26382470e-01 5.39554298e-01 -8.75101209e-01 -3.53086799e-01
-1.16433047e-01 2.13723928e-01 1.69520691e-01 4.13365811e-01
2.13425303e+00 -3.47352445e-01 -4.22833562e-01 7.73392081e-01
1.75785172e+00 -3.01880352e-02 5.41240573e-01 -2.08234787e-02
6.18499696e-01 3.96608293e-01 7.93117940e-01 1.22825778e+00
7.87975609e-01 8.47293496e-01 4.08743888e-01 3.33617516e-02
-2.43431747e-01 1.81075543e-01 1.56785354e-01 1.61609745e+00
-3.34962904e-02 -3.89600396e-01 -1.15056956e+00 7.49463379e-01
-2.04105926e+00 -1.00037873e+00 -1.98145524e-01 1.96269000e+00
5.93842804e-01 2.85811245e-01 5.18248737e-01 1.49804384e-01
5.59699476e-01 2.46837921e-02 -8.12518239e-01 -3.72566551e-01
6.35358766e-02 -1.55467406e-01 4.07903403e-01 -2.71799296e-01
-4.68215555e-01 5.43175220e-01 6.12722826e+00 1.05618894e+00
-1.32948887e+00 5.34670174e-01 1.10086775e+00 -5.50104082e-01
1.18053414e-01 -2.62177438e-01 -4.05200124e-01 1.29351902e+00
1.53794527e+00 -7.62921154e-01 5.81913888e-01 1.25318122e+00
7.96670020e-01 -7.85393417e-02 -7.15824604e-01 1.25118256e+00
-3.98078740e-01 -1.66325092e+00 -5.49004197e-01 6.77397475e-02
9.78727221e-01 6.24600410e-01 -6.66025206e-02 3.12779933e-01
-6.71311989e-02 -5.08014739e-01 4.83590931e-01 4.76694494e-01
1.03755105e+00 -7.25850821e-01 7.20502317e-01 3.97478551e-01
-1.81423056e+00 -3.97597283e-01 -4.90955979e-01 -1.29624158e-01
4.10898715e-01 1.20751727e+00 -4.53645706e-01 6.04646444e-01
9.11190569e-01 5.54444611e-01 -2.98025966e-01 1.25801718e+00
8.30159068e-01 7.59054959e-01 -1.90692499e-01 4.07213300e-01
-3.75015140e-01 2.99215287e-01 2.94420183e-01 1.34078074e+00
6.39393389e-01 5.80836013e-02 5.87736309e-01 4.60170418e-01
-2.07362652e-01 2.92019248e-01 1.74273565e-01 1.55119151e-01
1.05420578e+00 1.46400511e+00 -7.78572261e-01 -5.04835188e-01
-5.41044772e-01 6.28953576e-01 1.40391782e-01 3.63491684e-01
-1.10897684e+00 -7.39061311e-02 1.31128335e+00 5.11057973e-01
1.01976462e-01 -2.01452717e-01 -7.60450900e-01 -1.09049737e+00
1.09624162e-01 -6.43967628e-01 2.77345568e-01 -6.50527656e-01
-1.53698206e+00 8.84724438e-01 -2.96611130e-01 -1.57111990e+00
1.90413147e-01 -2.45857194e-01 -1.10774374e+00 8.26107800e-01
-1.41175675e+00 -9.40880895e-01 -1.09014618e+00 6.17161632e-01
8.50942612e-01 -1.20926104e-01 5.78892291e-01 8.02674294e-01
-1.05747759e+00 4.75242674e-01 -1.32861240e-02 -6.73648417e-01
5.45770943e-01 -9.94352102e-01 1.81698859e-01 9.79039490e-01
-3.51824403e-01 -2.01710746e-01 7.62225091e-01 -3.10489088e-01
-1.36071014e+00 -1.55406511e+00 4.25832152e-01 -3.09367061e-01
5.82749486e-01 -1.59843862e-01 -9.33512032e-01 1.54868752e-01
1.48449048e-01 6.07044160e-01 1.04343009e+00 1.10482395e-01
-9.92961153e-02 -7.38994598e-01 -8.17137003e-01 4.35688853e-01
1.01089048e+00 -5.64441800e-01 6.57105982e-01 5.54471076e-01
1.04884744e+00 -2.96155751e-01 -1.29667783e+00 5.43624640e-01
2.49189645e-01 -1.03484249e+00 7.18213737e-01 -4.70887244e-01
4.50512707e-01 -3.78805041e-01 -5.41610420e-01 -1.17845547e+00
-6.84577823e-01 -7.76649117e-01 -7.44487166e-01 1.19632208e+00
1.90672457e-01 -2.02434808e-01 1.04863656e+00 7.73644984e-01
-4.27259028e-01 -1.40842628e+00 -5.98563015e-01 -9.45707798e-01
-4.70622212e-01 -7.41588950e-01 1.20275044e+00 8.14218700e-01
7.54049271e-02 -4.48025167e-02 -5.75162649e-01 2.82689959e-01
5.92693508e-01 6.72703147e-01 9.52535629e-01 -8.60996127e-01
-3.83579969e-01 -6.45772636e-01 -2.60074139e-01 -7.78154373e-01
-1.94640115e-01 -5.82797050e-01 -1.02917694e-01 -1.33140111e+00
4.65381563e-01 -8.62892807e-01 -6.97049201e-01 1.96786568e-01
-2.51086056e-01 2.65725344e-01 5.61861873e-01 6.29731715e-01
-1.14030766e+00 4.78922576e-01 9.31652665e-01 4.37272310e-01
-3.48655909e-01 2.68562049e-01 -4.32218283e-01 1.11389339e-01
1.14814079e+00 -2.41922915e-01 -7.21412182e-01 -4.70825344e-01
1.62285104e-01 2.50320762e-01 -8.49668533e-02 -1.37245572e+00
6.12869620e-01 -6.97969019e-01 1.50608927e-01 -6.36693060e-01
1.01828478e-01 -9.26284015e-01 7.42694736e-01 4.24966067e-02
3.57565790e-01 4.87474620e-01 4.27665949e-01 3.69098455e-01
-3.14078629e-01 7.15964913e-01 6.21656775e-01 2.86569834e-01
-1.00345337e+00 1.07346368e+00 -1.42505720e-01 2.45744362e-01
1.33642149e+00 -4.42706458e-02 -6.31728828e-01 -2.41535440e-01
-5.36621988e-01 7.33529031e-01 5.82839549e-01 2.76258171e-01
4.56607968e-01 -1.35260177e+00 -7.51575053e-01 -9.72163528e-02
3.12142193e-01 -2.56154686e-01 1.02337992e+00 1.35615444e+00
-7.40452290e-01 1.02570817e-01 -2.87802219e-01 -7.91252911e-01
-1.23630011e+00 1.02292097e+00 2.27561727e-01 -4.80183363e-01
-2.71636188e-01 3.80977362e-01 -7.52189085e-02 -1.59065515e-01
-4.43102457e-02 2.69468606e-01 4.94065613e-01 -2.28310615e-01
6.06845081e-01 8.69937956e-01 3.35083246e-01 -6.26255751e-01
-3.49335849e-01 1.75318018e-01 2.84167230e-01 6.67833984e-01
1.39549375e+00 -8.06639850e-01 -3.42998356e-01 2.19569519e-01
1.18495405e+00 -1.91210315e-01 -1.29885924e+00 -5.27178407e-01
-3.11963670e-02 -9.68652844e-01 2.57711470e-01 -8.93438399e-01
-1.53217781e+00 7.90460110e-01 3.75590146e-01 4.98542160e-01
1.99947023e+00 -2.24503189e-01 1.34723639e+00 -4.16797489e-01
7.22920001e-01 -1.24101353e+00 3.12557518e-02 4.89423946e-02
5.27130306e-01 -1.11409581e+00 -1.68697625e-01 -7.39325762e-01
-6.59934223e-01 8.67844999e-01 1.11698401e+00 5.73522389e-01
1.12522137e+00 6.88404977e-01 5.64266779e-02 -6.95580691e-02
-1.70062387e+00 1.15404882e-01 -2.30141096e-02 3.92050356e-01
1.02477044e-01 6.44183099e-01 2.58052140e-01 6.80723965e-01
2.11582403e-03 2.89344758e-01 5.43157518e-01 8.69520545e-01
-1.63026348e-01 -9.98707891e-01 2.26719454e-02 8.54956686e-01
-4.14375037e-01 5.80558218e-02 6.26171410e-01 3.69625241e-02
1.37153640e-01 1.07377779e+00 3.02039891e-01 -8.70589375e-01
2.37731755e-01 -2.22068191e-01 -1.63344070e-01 -3.26564640e-01
-4.90930885e-01 1.06360637e-01 -2.55143344e-01 -1.02006567e+00
-6.10233247e-02 -4.61288512e-01 -8.24911237e-01 -1.22107577e+00
-2.33840663e-02 -8.32175389e-02 6.98228598e-01 3.91650945e-01
1.41518462e+00 1.03716075e+00 1.48771274e+00 -7.63590395e-01
-1.69731721e-01 -8.20077658e-01 -6.90347612e-01 3.22694957e-01
-8.45315233e-02 -4.01616424e-01 -5.51981889e-02 1.12642862e-01] | [6.983275890350342, 2.775489330291748] |
7c779bb3-7fea-4533-8bf5-4c3ed2059ddd | on-the-generalization-of-basicvsr-to-video | 2204.05308 | null | https://arxiv.org/abs/2204.05308v2 | https://arxiv.org/pdf/2204.05308v2.pdf | On the Generalization of BasicVSR++ to Video Deblurring and Denoising | The exploitation of long-term information has been a long-standing problem in video restoration. The recent BasicVSR and BasicVSR++ have shown remarkable performance in video super-resolution through long-term propagation and effective alignment. Their success has led to a question of whether they can be transferred to different video restoration tasks. In this work, we extend BasicVSR++ to a generic framework for video restoration tasks. In tasks where inputs and outputs possess identical spatial size, the input resolution is reduced by strided convolutions to maintain efficiency. With only minimal changes from BasicVSR++, the proposed framework achieves compelling performance with great efficiency in various video restoration tasks including video deblurring and denoising. Notably, BasicVSR++ achieves comparable performance to Transformer-based approaches with up to 79% of parameter reduction and 44x speedup. The promising results demonstrate the importance of propagation and alignment in video restoration tasks beyond just video super-resolution. Code and models are available at https://github.com/ckkelvinchan/BasicVSR_PlusPlus. | ['Chen Change Loy', 'Xiangyu Xu', 'Shangchen Zhou', 'Kelvin C. K. Chan'] | 2022-04-11 | null | null | null | null | ['video-super-resolution', 'video-restoration'] | ['computer-vision', 'computer-vision'] | [ 3.43498856e-01 -4.52887714e-01 1.24892429e-03 -1.57423496e-01
-9.09720719e-01 -2.19358310e-01 4.31301862e-01 -5.23963273e-01
-2.82777399e-01 7.87761211e-01 6.95470452e-01 -1.80325061e-01
-4.03772183e-02 -4.29729909e-01 -8.04764330e-01 -6.04154587e-01
-1.09795099e-02 -3.34882557e-01 4.51782674e-01 -4.29402679e-01
3.74891341e-01 4.35363203e-01 -1.47086406e+00 5.29606521e-01
9.49567378e-01 7.08674312e-01 7.23290563e-01 8.61890852e-01
7.62505159e-02 9.87927973e-01 -2.67386675e-01 -2.23740712e-01
3.56473863e-01 -1.91131383e-01 -8.76193345e-01 -1.21811675e-02
7.67269194e-01 -6.31649852e-01 -1.01202321e+00 1.16180599e+00
7.23368287e-01 2.50059128e-01 7.25724697e-02 -7.81298637e-01
-9.92367148e-01 5.57780445e-01 -8.10357511e-01 8.32867324e-01
4.63314742e-01 2.12000027e-01 6.87337220e-01 -1.20716095e+00
6.96175873e-01 1.28710747e+00 8.18490326e-01 6.58818901e-01
-1.34152687e+00 -6.00524902e-01 1.73002124e-01 6.23495221e-01
-1.39736998e+00 -9.20759618e-01 3.41566414e-01 -2.83037603e-01
1.10529721e+00 3.47208679e-01 2.31446072e-01 9.34669018e-01
-3.84328291e-02 5.81172109e-01 1.19354033e+00 -1.40047371e-01
-1.27121493e-01 -2.54998207e-01 1.98153749e-01 3.10688466e-01
-9.72969010e-02 1.39937952e-01 -8.47230732e-01 1.47105485e-01
1.28067887e+00 -3.49073932e-02 -9.31672990e-01 3.50813359e-01
-1.21053886e+00 4.83413607e-01 5.84126413e-01 3.48831028e-01
-3.69896978e-01 3.24897081e-01 3.75632912e-01 4.85545784e-01
6.38797343e-01 1.71125993e-01 -2.44853064e-01 -2.13242080e-02
-1.16494071e+00 2.69490242e-01 1.36231393e-01 9.44170415e-01
5.01862705e-01 4.07988220e-01 -3.75988811e-01 1.14071453e+00
-2.97293086e-02 4.25733030e-01 4.25552994e-01 -1.58091831e+00
4.41836208e-01 -1.54644489e-01 2.05559731e-01 -8.69038582e-01
-5.84521554e-02 -6.30975842e-01 -1.20222545e+00 3.88837546e-01
1.83429837e-01 2.54745275e-01 -9.19894099e-01 1.60563767e+00
6.82981759e-02 5.42798221e-01 9.36703905e-02 1.21650147e+00
1.03766048e+00 9.34650302e-01 -2.13310942e-01 -3.44735295e-01
1.22053444e+00 -1.05144298e+00 -8.98416102e-01 -1.55579537e-01
8.15222040e-02 -1.00007474e+00 9.31539416e-01 3.17202330e-01
-1.47835600e+00 -6.72639430e-01 -9.52620983e-01 -4.81513828e-01
2.47989610e-01 -2.09460735e-01 4.72153455e-01 2.11741611e-01
-1.80830979e+00 1.00920129e+00 -6.51034772e-01 -3.44497889e-01
6.11666143e-01 3.27568263e-01 -5.06144166e-01 -5.86222112e-01
-1.02993560e+00 7.13940203e-01 -6.80402592e-02 3.30210149e-01
-9.16782320e-01 -9.25476193e-01 -6.75683618e-01 -3.13526951e-02
2.62119740e-01 -9.07003045e-01 1.00549543e+00 -6.78421557e-01
-1.31402969e+00 6.10867560e-01 -4.83676434e-01 -5.04098892e-01
5.84522605e-01 -3.86585385e-01 -4.64774817e-01 1.63680002e-01
4.19116616e-02 4.27531183e-01 1.01701987e+00 -1.25545776e+00
-5.29593766e-01 -2.75729567e-01 8.86618942e-02 3.90198767e-01
-3.23955387e-01 4.69313830e-01 -7.65000761e-01 -9.54398274e-01
3.41562420e-01 -5.52092910e-01 -3.64622265e-01 1.54777989e-01
-9.37465951e-02 2.59919465e-01 7.05477834e-01 -1.25266767e+00
1.14178550e+00 -2.09370637e+00 3.81975025e-01 -4.19207156e-01
6.18859172e-01 4.87930030e-01 -3.90217304e-01 2.25641727e-01
-1.38055667e-01 5.34544773e-02 -3.11286390e-01 -3.53380531e-01
-4.21078354e-01 1.56755019e-02 -5.02843261e-01 5.64502597e-01
-2.66956948e-02 7.93793619e-01 -6.18589044e-01 -3.37372273e-01
4.19473797e-01 1.11398554e+00 -5.94101310e-01 1.00271478e-01
2.60485619e-01 6.80778921e-01 -9.53479484e-02 5.92251897e-01
9.82004702e-01 -4.67224687e-01 -1.56726185e-02 -5.80749810e-01
-2.78376788e-01 1.82342395e-01 -1.16118896e+00 1.70050871e+00
-5.62798679e-01 1.00142372e+00 5.23063958e-01 -8.52653801e-01
5.45409441e-01 3.75618875e-01 3.81774187e-01 -1.05585003e+00
-3.17514598e-01 1.55600473e-01 -4.95777875e-01 -4.99198169e-01
8.79468501e-01 -2.25694776e-02 5.20813823e-01 3.95356752e-02
1.06581531e-01 1.93280250e-01 2.64446378e-01 5.22094548e-01
1.23080921e+00 2.77355820e-01 2.50457786e-02 -3.22284162e-01
5.91051102e-01 -1.37516722e-01 6.44898832e-01 7.80432999e-01
-1.59143791e-01 1.09033632e+00 -1.05106696e-01 -1.93373173e-01
-1.15260208e+00 -1.12049520e+00 -2.23233119e-01 1.02765489e+00
3.40536445e-01 -4.67504412e-01 -5.88854909e-01 2.29376019e-03
-3.85112703e-01 2.90483892e-01 -3.65528524e-01 2.65790403e-01
-1.02019024e+00 -9.10368919e-01 3.35445881e-01 5.19807994e-01
6.91056848e-01 -8.64080608e-01 -1.01943187e-01 1.61120787e-01
-6.81732476e-01 -1.44861305e+00 -6.58489704e-01 -3.68352860e-01
-1.28716981e+00 -7.68861175e-01 -1.20977294e+00 -8.16586614e-01
4.47941512e-01 9.11046565e-01 1.12822855e+00 2.47915655e-01
-2.64061928e-01 2.65568614e-01 -3.41206163e-01 6.32135570e-01
-2.27656558e-01 -2.79220879e-01 6.65356666e-02 -1.14053622e-01
-3.56324255e-01 -1.10744107e+00 -9.48543966e-01 3.91398698e-01
-8.78766239e-01 2.31016234e-01 3.23494405e-01 7.10395932e-01
5.51586449e-01 1.49995356e-03 4.53258455e-01 -5.71503818e-01
4.69952911e-01 -2.57762402e-01 -4.25620019e-01 6.46234956e-03
-5.94837129e-01 -2.35541612e-01 5.52208602e-01 -2.71466523e-01
-1.13117409e+00 -4.68742669e-01 -3.36601734e-01 -6.49925590e-01
1.01163127e-01 1.92661062e-01 2.95195915e-02 -2.25851879e-01
5.40796459e-01 5.03793597e-01 -2.26806719e-02 -8.48637879e-01
2.81533957e-01 5.01263022e-01 8.47902179e-01 -2.16336042e-01
9.44757581e-01 6.60318911e-01 -2.78901786e-01 -9.71090198e-01
-4.30426061e-01 -4.57913458e-01 -3.87200683e-01 -2.58141488e-01
7.63634562e-01 -1.36922252e+00 -5.35898209e-01 7.93511093e-01
-1.07579017e+00 -4.98909175e-01 -2.83685893e-01 3.51061314e-01
-3.88694108e-01 6.51292920e-01 -1.13454282e+00 -4.37793702e-01
-4.79544371e-01 -1.27259421e+00 8.07359397e-01 2.75367349e-01
3.28840911e-01 -8.11021924e-01 -3.28078538e-01 7.22580791e-01
1.05840969e+00 -2.79884070e-01 3.86460394e-01 2.03206226e-01
-1.03681767e+00 2.64125109e-01 -6.72120988e-01 4.20602262e-01
4.06595655e-02 -4.51057106e-01 -1.02151167e+00 -6.95309579e-01
5.55807576e-02 2.58446168e-02 1.28061545e+00 6.63676918e-01
1.05064905e+00 -2.95217246e-01 -1.06814742e-01 1.21798348e+00
1.60320246e+00 -7.62275457e-02 1.09420848e+00 3.82735103e-01
7.94974923e-01 2.36844048e-01 3.44543606e-01 2.31234416e-01
3.14350277e-01 9.47473526e-01 4.39563334e-01 -2.11740896e-01
-8.69859338e-01 1.28230780e-01 6.23039842e-01 9.81233418e-01
-5.51465094e-01 -1.58102363e-01 -4.73116070e-01 4.79390472e-01
-1.73173082e+00 -1.24569309e+00 -3.47353607e-01 2.11612177e+00
6.86921239e-01 -3.09600860e-01 -2.69475669e-01 3.84744518e-02
9.58825946e-01 5.90241253e-01 -4.72310513e-01 -4.66376878e-02
-5.48368812e-01 1.70719862e-01 5.02100527e-01 8.85935903e-01
-9.88342106e-01 9.68427658e-01 6.44530773e+00 1.06796241e+00
-9.99759972e-01 4.46395338e-01 5.18748641e-01 -2.35859945e-01
-2.04518691e-01 -1.06784560e-01 -6.59101427e-01 2.79092312e-01
7.18100190e-01 -1.44990847e-01 8.24519157e-01 1.58012360e-01
5.31443775e-01 2.84100417e-02 -6.16676629e-01 1.16645217e+00
7.51785785e-02 -1.65747201e+00 8.91871285e-03 -2.78423410e-02
7.42294729e-01 3.19071263e-01 1.63354799e-01 -8.79610553e-02
1.95775107e-01 -1.22286022e+00 4.21578258e-01 4.57828254e-01
1.12302887e+00 -4.42382187e-01 4.83154297e-01 -1.17192268e-01
-1.35546231e+00 -1.24792226e-01 -5.31042516e-01 -2.65089009e-04
6.82410598e-01 6.49512649e-01 -1.63802788e-01 7.12081373e-01
1.16228127e+00 1.20618618e+00 -4.47322696e-01 1.04113972e+00
-5.78599656e-03 5.15949368e-01 8.21112841e-02 9.48535085e-01
-2.20048741e-01 -2.87977666e-01 8.68252337e-01 1.26940835e+00
5.36555171e-01 2.35715896e-01 -8.38741586e-02 5.83069146e-01
-5.26806936e-02 -7.85108209e-02 -2.28336737e-01 3.08044910e-01
2.71972358e-01 1.09573090e+00 -3.78497452e-01 -3.21813524e-01
-5.27724683e-01 1.34907722e+00 1.08328216e-01 6.04481578e-01
-8.92297149e-01 3.02462336e-02 1.02061665e+00 4.61841911e-01
4.14333880e-01 -2.51727134e-01 -1.29015535e-01 -1.46514952e+00
1.23988725e-01 -9.24193382e-01 2.17052951e-01 -1.03551733e+00
-1.06170535e+00 9.55516338e-01 -3.05057138e-01 -1.20132673e+00
2.52532810e-01 -3.18300992e-01 -2.51778930e-01 1.08741510e+00
-1.83254063e+00 -1.08697605e+00 -5.22409558e-01 9.40544546e-01
8.57253611e-01 3.22236791e-02 4.42049652e-01 7.50544250e-01
-5.64705908e-01 4.67339694e-01 3.45461696e-01 -2.38504796e-03
7.84269512e-01 -8.98093641e-01 6.62373662e-01 1.35195541e+00
-2.46276304e-01 5.40374100e-01 1.05262756e+00 -5.82820952e-01
-1.54471028e+00 -1.06975448e+00 5.98602295e-01 -1.12373605e-01
5.01768291e-01 -7.90136680e-02 -1.18498588e+00 6.80519700e-01
4.14119095e-01 3.54437023e-01 1.44719675e-01 -2.11534575e-01
-6.58604205e-01 -1.91423908e-01 -1.15014136e+00 6.09295070e-01
1.57033753e+00 -5.73849976e-01 -1.77416563e-01 3.22173804e-01
9.01971817e-01 -4.35727030e-01 -9.19555008e-01 4.21974719e-01
3.29541296e-01 -1.23354697e+00 1.54559827e+00 -2.31800959e-01
7.85214007e-01 -5.21792054e-01 -5.11256993e-01 -1.04031932e+00
-6.23863757e-01 -7.93322861e-01 -2.32612953e-01 1.13402438e+00
6.80472180e-02 -7.15590358e-01 3.06622684e-01 2.99564749e-01
-3.40454102e-01 -3.81318182e-01 -9.35491025e-01 -6.43777192e-01
-7.41467327e-02 -2.93348402e-01 1.78670570e-01 9.74189222e-01
-4.34069008e-01 6.56737685e-02 -9.03896868e-01 4.54799831e-01
9.75981414e-01 9.84398574e-02 4.32229787e-01 -8.04278076e-01
-4.25939560e-01 -3.39212954e-01 -3.44205201e-01 -1.44502902e+00
-1.82596073e-01 -9.21741605e-01 -2.03855351e-01 -1.81987023e+00
4.86066014e-01 -2.77936727e-01 -3.09597343e-01 1.90958053e-01
-1.60546541e-01 9.36537087e-01 3.71725947e-01 4.28825349e-01
-5.55926859e-01 4.06088114e-01 1.39329278e+00 5.11164069e-02
-1.11753516e-01 -2.75024235e-01 -8.06076407e-01 5.39836764e-01
7.03684151e-01 -8.55999812e-02 -1.86855108e-01 -9.56165791e-01
8.18642750e-02 5.62510669e-01 7.03348696e-01 -9.11756635e-01
2.79768676e-01 7.81229232e-03 2.71159440e-01 -2.24766359e-01
7.27442443e-01 -3.49490941e-01 3.33072275e-01 2.35171393e-01
-1.88839272e-01 2.75162727e-01 1.79618850e-01 5.16205788e-01
-3.02722782e-01 2.42539376e-01 1.03028321e+00 -1.77895412e-01
-1.00635982e+00 4.24155384e-01 -4.05098826e-01 4.15566824e-02
4.62641478e-01 -4.09307480e-01 -6.10400617e-01 -5.35210192e-01
-9.82595146e-01 4.69986461e-02 6.20011747e-01 4.65671122e-01
1.02105892e+00 -1.09941661e+00 -1.13570821e+00 -1.09612472e-01
-4.58595276e-01 -2.71351963e-01 1.04890478e+00 1.00735664e+00
-5.14119685e-01 1.66653290e-01 -2.74131358e-01 -5.96143603e-01
-1.67092359e+00 5.91368198e-01 3.90132844e-01 -3.86338420e-02
-1.35069466e+00 9.42927957e-01 4.36053962e-01 2.12912094e-02
2.58442648e-02 1.99066147e-01 -2.52083540e-01 -2.45712683e-01
1.15997684e+00 7.30868459e-01 -8.49969983e-02 -8.43190193e-01
-2.70188004e-01 6.72134578e-01 -1.91389248e-01 -6.36906028e-02
1.72592294e+00 -7.64035404e-01 -2.99383253e-01 -1.80015396e-02
1.04988730e+00 3.18698399e-02 -1.49205577e+00 -4.59921986e-01
-3.51106435e-01 -8.98538053e-01 3.83727372e-01 -7.34279931e-01
-1.66400516e+00 7.21962988e-01 8.06161523e-01 -2.29759231e-01
1.51873863e+00 -4.29600887e-02 8.12942088e-01 -2.33566254e-01
5.72270691e-01 -5.61946929e-01 -4.79341224e-02 4.25920039e-01
1.19023550e+00 -1.18035567e+00 1.03642263e-01 -6.54605091e-01
-4.52839702e-01 8.42990458e-01 4.14956301e-01 -2.95689583e-01
4.32239980e-01 5.62931895e-01 2.83949245e-02 1.41435489e-01
-7.04588592e-01 -1.32089257e-01 9.80606861e-03 7.39290118e-01
5.60122669e-01 -2.46351615e-01 -2.18729958e-01 1.71227187e-01
8.30369741e-02 1.09383069e-01 9.42329764e-01 5.30486524e-01
-3.69153500e-01 -9.23196018e-01 -4.94157463e-01 2.23902002e-01
-6.93095386e-01 -5.34905493e-01 4.10200179e-01 3.68668288e-01
-3.79417598e-01 1.14102197e+00 -2.06417471e-01 -2.91185439e-01
8.77379850e-02 -3.91884625e-01 6.16313398e-01 -2.08761185e-01
-4.87175047e-01 1.20037332e-01 1.35697471e-02 -9.99839485e-01
-5.54146707e-01 -5.85852981e-01 -9.68963504e-01 -6.48770213e-01
-1.08799957e-01 -4.97899875e-02 3.80122185e-01 6.28677487e-01
5.57562828e-01 7.69239962e-01 4.32818919e-01 -1.11713088e+00
-3.84588927e-01 -9.29200292e-01 -4.46100891e-01 2.34532163e-01
6.70400083e-01 -2.69846380e-01 -3.80080223e-01 2.58415163e-01] | [11.063648223876953, -2.010988473892212] |
e4cc465c-a612-4a51-a5dc-c45924f29375 | statistical-beamformer-exploiting-non | 2306.07562 | null | https://arxiv.org/abs/2306.07562v1 | https://arxiv.org/pdf/2306.07562v1.pdf | Statistical Beamformer Exploiting Non-stationarity and Sparsity with Spatially Constrained ICA for Robust Speech Recognition | In this paper, we present a statistical beamforming algorithm as a pre-processing step for robust automatic speech recognition (ASR). By modeling the target speech as a non-stationary Laplacian distribution, a mask-based statistical beamforming algorithm is proposed to exploit both its output and masked input variance for robust estimation of the beamformer. In addition, we also present a method for steering vector estimation (SVE) based on a noise power ratio obtained from the target and noise outputs in independent component analysis (ICA). To update the beamformer in the same ICA framework, we derive ICA with distortionless and null constraints on target speech, which yields beamformed speech at the target output and noises at the other outputs, respectively. The demixing weights for the target output result in a statistical beamformer with the weighted spatial covariance matrix (wSCM) using a weighting function characterized by a source model. To enhance the SVE, the strict null constraints imposed by the Lagrange multiplier methods are relaxed by generalized penalties with weight parameters, while the strict distortionless constraints are maintained. Furthermore, we derive an online algorithm based on an optimization technique of recursive least squares (RLS) for practical applications. Experimental results on various environments using CHiME-4 and LibriCSS datasets demonstrate the effectiveness of the presented algorithm compared to conventional beamforming and blind source extraction (BSE) based on ICA on both batch and online processing. | ['Hyung-Min Park', 'Ui-Hyeop Shin'] | 2023-06-13 | null | null | null | null | ['robust-speech-recognition', 'automatic-speech-recognition'] | ['speech', 'speech'] | [ 4.21593308e-01 -4.73201096e-01 3.49351645e-01 -2.38748401e-01
-8.39763641e-01 -5.89587510e-01 3.27594757e-01 -6.06979012e-01
-4.18915749e-01 5.61606586e-01 5.92790127e-01 -2.20126495e-01
-4.55861837e-01 -1.78259775e-01 -3.70196402e-01 -1.28003442e+00
2.81605981e-02 -2.53445238e-01 -1.42591566e-01 9.12074894e-02
1.16729222e-01 6.04700625e-01 -1.28637302e+00 -5.16329333e-02
9.88845229e-01 1.07048893e+00 4.92703378e-01 6.55091286e-01
7.55723864e-02 3.41520607e-01 -7.34606206e-01 -1.37807308e-02
4.62840199e-01 -4.72992092e-01 2.49107718e-01 4.29577500e-01
7.97696263e-02 1.55870859e-02 -4.73374009e-01 1.36224413e+00
8.53694856e-01 4.87429082e-01 5.30527294e-01 -1.03900766e+00
-4.87794429e-01 5.96853971e-01 -6.64783061e-01 2.48992592e-01
2.15606600e-01 -1.35068372e-01 6.45142734e-01 -1.56290138e+00
-3.46444435e-02 1.13962972e+00 3.53307068e-01 1.85626373e-01
-1.05619931e+00 -8.23827744e-01 1.02288522e-01 3.37977380e-01
-1.87470818e+00 -1.20754397e+00 1.03223372e+00 -2.48716265e-01
5.01321971e-01 6.66160643e-01 2.33220726e-01 9.04024482e-01
-2.20356420e-01 6.92909181e-01 1.12450671e+00 -4.82366055e-01
3.68565410e-01 -9.21728648e-03 -9.23340768e-02 4.92636859e-01
2.44346559e-01 1.02087393e-01 -8.61180067e-01 -9.81400758e-02
6.08366847e-01 -3.78739148e-01 -7.11436510e-01 -1.14413500e-01
-1.28370690e+00 4.67970461e-01 1.96672156e-01 6.23427987e-01
-5.84043324e-01 -1.42747745e-01 -2.75713474e-01 4.63246455e-04
5.39406717e-01 -4.99783680e-02 -2.74380654e-01 3.27498972e-01
-1.34053481e+00 -2.37676159e-01 6.19359016e-01 8.73046815e-01
5.44247627e-01 7.53678799e-01 -4.29450452e-01 1.13001359e+00
1.02619863e+00 1.25591540e+00 5.83234668e-01 -6.80154622e-01
7.02611625e-01 -1.43115986e-02 1.99706540e-01 -1.12533498e+00
-2.61782467e-01 -1.06729853e+00 -1.26013732e+00 7.05967546e-02
3.11247796e-01 -4.20124024e-01 -1.05641282e+00 1.68159819e+00
1.95859104e-01 5.93269289e-01 2.11550608e-01 1.06483591e+00
6.09506845e-01 8.81960273e-01 -4.01374996e-01 -8.66860569e-01
9.42808747e-01 -7.27995515e-01 -1.39564085e+00 -4.42652345e-01
-1.46183297e-01 -1.05923927e+00 3.94393533e-01 5.42202592e-01
-1.12078238e+00 -5.04295886e-01 -1.05694640e+00 4.14720356e-01
-8.15220475e-02 5.43438077e-01 -5.28070293e-02 1.08441913e+00
-1.05311310e+00 -1.74814790e-01 -7.38306284e-01 9.91418585e-02
-6.83575049e-02 2.26180092e-01 -2.42992714e-01 1.08397059e-01
-7.87885427e-01 6.87816799e-01 2.11916417e-02 8.13237965e-01
-6.23183727e-01 -5.57231486e-01 -8.58443737e-01 1.92596808e-01
2.75971234e-01 -2.78700769e-01 7.66661227e-01 -9.64338958e-01
-1.91919184e+00 7.55884945e-02 -7.32920885e-01 -1.85096353e-01
1.45414695e-01 -1.05800949e-01 -8.48408163e-01 8.14103708e-03
-9.68272910e-02 -1.56677887e-01 1.42441738e+00 -1.44127667e+00
-4.02389884e-01 -3.78827125e-01 -8.54936361e-01 4.58378613e-01
-4.41256225e-01 2.73742825e-01 -7.35265136e-01 -1.27493858e+00
8.94203126e-01 -5.79442620e-01 -2.04515949e-01 -4.52924907e-01
-4.40156847e-01 3.94979805e-01 5.95853865e-01 -1.18867624e+00
1.55823767e+00 -2.70341325e+00 3.81765634e-01 8.81836891e-01
-1.12689324e-01 1.67681828e-01 -3.46122384e-01 1.70050204e-01
-2.93317765e-01 -4.47209388e-01 -5.48832119e-01 -4.28535223e-01
-1.27003267e-01 -6.51393831e-02 -3.01022261e-01 8.14552307e-01
-1.37597695e-01 3.75506848e-01 -7.25809813e-01 -2.09899843e-01
3.52080017e-01 5.59340060e-01 -3.93683821e-01 4.24689293e-01
6.11889422e-01 6.00355625e-01 -8.79817754e-02 6.00296140e-01
9.57422495e-01 2.57087290e-01 4.81662005e-02 -4.46729511e-01
-3.62911105e-01 -1.05868936e-01 -1.99605143e+00 1.51022387e+00
-4.89163190e-01 4.89480168e-01 1.03714478e+00 -8.46614540e-01
1.02026188e+00 5.22637665e-01 3.28715891e-01 -3.37996513e-01
9.69184339e-02 3.52310151e-01 2.62636632e-01 -3.24359566e-01
1.24095447e-01 -3.96313928e-02 3.51803362e-01 1.30806908e-01
1.89686149e-01 7.48813078e-02 8.60664062e-03 1.53698042e-01
7.92887688e-01 -3.32451075e-01 2.45093733e-01 -2.19728768e-01
1.08979321e+00 -7.30135798e-01 5.60965717e-01 7.42829323e-01
-3.84432636e-02 7.05511570e-01 -3.39625806e-01 4.11477298e-01
-4.70049918e-01 -1.11620331e+00 -1.28803626e-01 9.45075393e-01
-6.46002144e-02 -1.47175146e-02 -6.09509826e-01 -8.73651877e-02
-3.28329414e-01 9.41321790e-01 -1.23182818e-01 2.17320159e-01
-6.17753506e-01 -1.10817397e+00 2.19245896e-01 2.45822564e-01
4.75352108e-01 -4.90195870e-01 2.23173238e-02 3.95470351e-01
-2.46231854e-01 -1.10680330e+00 -8.52949619e-01 2.36824527e-01
-3.90199482e-01 -4.22992885e-01 -9.15272176e-01 -5.99985719e-01
9.97083604e-01 7.09348023e-01 1.29940137e-01 -3.27991754e-01
2.29424119e-01 5.42182088e-01 -3.58084530e-01 -2.62578756e-01
-2.45780736e-01 -6.57154322e-01 5.27688742e-01 1.08922899e+00
-2.64026731e-01 -6.15827441e-01 -3.50077748e-01 3.76181275e-01
-8.27616453e-01 -9.38798040e-02 8.04208696e-01 7.79307961e-01
3.54795218e-01 3.92643571e-01 4.97130692e-01 -8.30988437e-02
7.41539419e-01 -4.12412375e-01 -8.22312415e-01 2.67452240e-01
-3.95759016e-01 -1.23534970e-01 5.06584585e-01 -5.26245892e-01
-1.69008958e+00 3.00771296e-01 -1.28393590e-01 -5.28719723e-01
2.92439945e-02 6.39750481e-01 -7.29875624e-01 -2.18788177e-01
4.22395587e-01 6.29322350e-01 -9.28529799e-02 -6.47461355e-01
6.96798801e-01 1.02543938e+00 7.77837217e-01 -2.18591839e-02
1.48046267e+00 3.25938284e-01 -4.85763550e-02 -1.13400531e+00
-4.00607049e-01 -8.04448009e-01 -5.53700566e-01 -2.81429023e-01
5.83340287e-01 -1.00215995e+00 -3.53318930e-01 6.81653917e-01
-1.00718856e+00 1.56685218e-01 1.30102903e-01 9.73598540e-01
-9.75059494e-02 5.42751849e-01 1.88442424e-03 -1.36617124e+00
-3.08597416e-01 -1.13041461e+00 6.95731223e-01 9.62378979e-02
1.07838690e-01 -7.33726501e-01 -1.78268224e-01 1.87129393e-01
6.67391479e-01 -4.79641050e-01 3.31279844e-01 -6.92236066e-01
-4.94929671e-01 -2.14466527e-01 1.91309154e-02 5.45268476e-01
3.77134115e-01 -2.97273129e-01 -1.10166466e+00 -4.25668567e-01
4.52118397e-01 5.92638433e-01 6.21953189e-01 6.59587741e-01
8.73625636e-01 -4.96907979e-01 -1.09127283e-01 8.36349666e-01
1.31151116e+00 4.74173188e-01 3.37442249e-01 -6.62866160e-02
5.50944507e-01 3.14779758e-01 1.92914307e-01 6.01321399e-01
-2.05488905e-01 7.48815119e-01 2.95816865e-02 -5.78792430e-02
-3.81426334e-01 3.24709639e-02 5.81263781e-01 1.40983665e+00
1.33527160e-01 -4.45623606e-01 -5.89345634e-01 3.57851177e-01
-1.62742615e+00 -9.71235633e-01 -2.85504550e-01 2.60352755e+00
6.96747303e-01 -3.12174022e-01 -4.46746111e-01 3.94189000e-01
7.95269132e-01 2.86849767e-01 -3.46998006e-01 1.17182299e-01
-5.07051408e-01 2.76235849e-01 7.42441773e-01 9.68755364e-01
-1.01901042e+00 5.53506792e-01 5.82350349e+00 1.07384217e+00
-9.53682482e-01 2.73598015e-01 8.51948485e-02 -7.08274469e-02
-1.69533074e-01 -3.03046733e-01 -5.29688179e-01 4.40030068e-01
6.70419395e-01 -1.86779752e-01 9.57926512e-01 3.04416239e-01
7.85823464e-01 -4.69776951e-02 -4.79927033e-01 1.30536044e+00
2.49067008e-01 -6.72706544e-01 -3.16730291e-01 -1.83697492e-01
5.92773616e-01 -2.00356588e-01 2.87359059e-01 -1.54714659e-01
4.48830128e-02 -6.60961032e-01 1.00397098e+00 5.65935016e-01
6.50297105e-01 -4.69982237e-01 5.67475796e-01 3.50302964e-01
-1.32497573e+00 -3.75853032e-01 -7.89629668e-02 1.84134915e-01
3.34880114e-01 8.81939590e-01 -7.40192533e-01 7.12734342e-01
3.98761064e-01 3.46884310e-01 -1.86349392e-01 1.42008626e+00
-6.21439695e-01 8.63938510e-01 -5.07503092e-01 2.10072711e-01
-1.44996718e-01 -6.00234509e-01 1.20765579e+00 1.32915962e+00
6.37364686e-01 4.59860384e-01 -2.00420707e-01 5.93194067e-01
1.14848942e-01 2.89824247e-01 -1.27327234e-01 1.47124499e-01
8.17632735e-01 1.36513758e+00 -5.92589021e-01 -1.20822057e-01
-4.89700854e-01 8.92782152e-01 -3.77363235e-01 1.07379603e+00
-6.46824777e-01 -3.79555464e-01 4.49072808e-01 -4.37969446e-01
3.83433521e-01 -5.89317203e-01 -4.35590863e-01 -1.17951524e+00
6.70696869e-02 -9.59501863e-01 -2.36464329e-02 -7.19493210e-01
-1.17227852e+00 5.42632163e-01 -8.43211189e-02 -1.32581043e+00
-6.06231168e-02 -4.66501296e-01 -6.95968091e-01 1.31274068e+00
-1.47024095e+00 -8.19148600e-01 1.20464565e-05 8.59494328e-01
5.91676831e-01 -4.80855018e-01 5.67375302e-01 4.56164747e-01
-1.02237010e+00 6.02530301e-01 4.92017537e-01 -3.71391773e-02
5.40399432e-01 -9.10233915e-01 -9.01857093e-02 1.65344560e+00
4.01210099e-01 7.82571554e-01 9.61279750e-01 -5.85726559e-01
-1.68643701e+00 -9.52968121e-01 7.37651467e-01 1.13646410e-01
7.78213561e-01 -4.42268580e-01 -7.98866749e-01 5.42645633e-01
3.28326881e-01 -1.10829353e-01 7.75671899e-01 -3.36567312e-01
-1.14918940e-01 -2.43285030e-01 -9.41351891e-01 4.76583540e-01
7.58661866e-01 -2.31448114e-01 -6.49353027e-01 3.34065735e-01
4.07754034e-01 -3.97274584e-01 -2.82976776e-01 4.26442027e-01
3.37491274e-01 -5.27912140e-01 1.12119842e+00 -3.79055850e-02
-4.97026086e-01 -9.35733676e-01 -5.07698834e-01 -1.50843108e+00
-6.51380360e-01 -1.00719726e+00 -1.23181321e-01 1.41601849e+00
5.82471430e-01 -6.27552271e-01 2.15652451e-01 3.61763269e-01
-3.25702459e-01 -7.90942460e-02 -1.28912461e+00 -7.98586488e-01
-7.56041408e-01 -7.16936171e-01 1.97788313e-01 4.91892636e-01
-2.57668436e-01 1.61730990e-01 -6.06433809e-01 9.97841358e-01
9.54642713e-01 -3.14877123e-01 4.22104776e-01 -6.66654766e-01
-3.09203863e-01 -3.41170281e-01 4.48682811e-03 -1.21251619e+00
1.58306256e-01 -6.96041346e-01 5.32799602e-01 -1.37236142e+00
-2.62940943e-01 -5.67628965e-02 -5.36254585e-01 2.47610599e-01
-2.96760321e-01 6.00130521e-02 2.79840112e-01 1.34306923e-01
-1.35415820e-02 7.19791472e-01 6.67422235e-01 -1.30212545e-01
-4.18390185e-01 4.45342302e-01 -4.72794592e-01 8.13041866e-01
4.21317667e-01 -3.71801525e-01 -3.00764799e-01 -5.32002926e-01
-2.08884031e-01 7.50690922e-02 5.42966276e-02 -1.03440702e+00
7.17947602e-01 1.16526656e-01 4.41901743e-01 -5.95211923e-01
5.82904935e-01 -1.13135326e+00 3.03374529e-01 2.13728219e-01
-1.09038390e-01 -4.25140262e-01 6.01039939e-02 7.23101377e-01
-2.62709022e-01 -3.34905446e-01 7.86224902e-01 3.99911553e-01
-4.23720568e-01 -3.35348099e-02 -7.85498619e-01 -5.42393446e-01
5.97126544e-01 -6.65088445e-02 2.65863895e-01 -6.36605859e-01
-8.03433955e-01 -2.90781427e-02 -4.51457649e-01 1.73130646e-01
9.74008620e-01 -1.26823401e+00 -9.47675407e-01 6.22567177e-01
-4.07043874e-01 -4.93388683e-01 2.78848022e-01 9.43114817e-01
7.41170421e-02 3.59273106e-01 2.94667184e-01 -2.65121132e-01
-1.30394650e+00 4.78197992e-01 1.91417947e-01 2.70863175e-01
-1.06454641e-01 1.13211346e+00 1.93779945e-01 -8.92630294e-02
5.25692165e-01 3.88458520e-02 -4.02527273e-01 9.24210995e-02
7.40088344e-01 4.90091920e-01 2.61910617e-01 -1.03916669e+00
-4.46251690e-01 5.12076557e-01 6.76190138e-01 -7.30421841e-01
1.22657919e+00 -6.72881663e-01 -3.35470259e-01 8.17627013e-02
1.06585538e+00 9.39336240e-01 -8.28071177e-01 -4.71247792e-01
-9.12739486e-02 -6.54570222e-01 6.25442028e-01 -9.11955178e-01
-1.10523307e+00 5.68628550e-01 7.96247959e-01 5.90480911e-03
1.59662187e+00 -4.08673763e-01 2.19573840e-01 2.30536014e-01
2.06096560e-01 -7.38205016e-01 -2.59652466e-01 3.83451849e-01
1.27994752e+00 -7.14216530e-01 -1.57687932e-01 -4.54035461e-01
-2.39011407e-01 1.10458434e+00 9.19983089e-02 2.87557095e-01
8.18818569e-01 3.99569333e-01 2.25656524e-01 3.70920628e-01
-1.25752211e-01 -4.09407318e-01 6.61805630e-01 4.23973978e-01
8.96787196e-02 1.22731822e-02 -3.72419983e-01 8.77081096e-01
-9.41390917e-03 -4.59458858e-01 2.80573785e-01 4.65026438e-01
-6.29933357e-01 -9.23322618e-01 -1.27399302e+00 3.63324098e-02
-2.70062774e-01 -4.80926305e-01 -8.23567528e-03 1.91639066e-02
-1.30980909e-01 1.76722038e+00 -2.38959104e-01 -3.26870173e-01
3.57298940e-01 4.95297648e-02 1.41624346e-01 -4.43873048e-01
-2.27582201e-01 9.92487907e-01 -1.30765095e-01 -2.56413221e-01
-3.54865879e-01 -6.17957354e-01 -1.01144707e+00 3.80811721e-01
-7.09066331e-01 2.55744249e-01 1.07025254e+00 9.35755670e-01
2.57954091e-01 5.38600028e-01 1.08559740e+00 -9.22003746e-01
-6.03209734e-01 -1.10495269e+00 -8.68888140e-01 -7.45842233e-02
3.13576639e-01 -4.65881944e-01 -8.91389251e-01 2.40848228e-01] | [15.030355453491211, 5.817958831787109] |
5af3a171-8142-444e-918c-003a9567b5b9 | an-exploratory-study-on-temporally-evolving | null | null | https://aclanthology.org/2021.nlp4dh-1.21 | https://aclanthology.org/2021.nlp4dh-1.21.pdf | An Exploratory Study on Temporally Evolving Discussion around Covid-19 using Diachronic Word Embeddings | Covid 19 has seen the world go into a lock down and unconventional social situations throughout. During this time, the world saw a surge in information sharing around the pandemic and the topics shared in the time were diverse. People’s sentiments have changed during this period. Given the wide spread usage of Online Social Networks (OSN) and support groups, the user sentiment is well reflected in online discussions. In this work, we aim to show the topics under discussion, evolution of discussions, change in user sentiment during the pandemic. Alongside which, we also demonstrate the possibility of exploratory analysis to find pressing topics, change in perception towards the topics and ways to use the knowledge extracted from online discussions. For our work we employ Diachronic Word embeddings which capture the change in word usage over time. With the help of analysis from temporal word usages, we show the change in people’s option on covid-19 from being a conspiracy, to the post-covid topics that surround vaccination. | ['Arun Balaji Buduru', 'Ponnurangam Kumaraguru', 'Asanobu Kitamoto', 'Avinash Tulasi'] | null | null | null | null | nlp4dh-icon-2021-12 | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [-2.94016331e-01 2.45695025e-01 -2.09755555e-01 -2.35267013e-01
8.34515169e-02 -8.71336401e-01 1.05279732e+00 9.39942420e-01
-6.40956044e-01 6.02546155e-01 1.32200027e+00 -4.41115767e-01
-1.55215174e-01 -8.01764667e-01 -1.96693167e-02 -5.39426923e-01
-8.83827880e-02 2.65291065e-01 4.25494015e-02 -9.93568182e-01
3.56127053e-01 2.80452184e-02 -1.05083811e+00 2.55076408e-01
3.19714665e-01 3.41241121e-01 2.41231658e-02 5.10117412e-01
-7.14290440e-01 3.49814713e-01 -6.62246168e-01 -4.16353822e-01
-1.07590090e-02 9.22758505e-02 -6.45316422e-01 -1.26817480e-01
-1.19724348e-01 4.03480195e-02 -1.56098321e-01 6.93751991e-01
5.13440251e-01 1.28399000e-01 4.50584829e-01 -1.10090399e+00
-4.80210155e-01 5.62248707e-01 -6.21081233e-01 7.49089420e-01
6.77906513e-01 -9.62184295e-02 8.61678779e-01 -4.37513202e-01
1.28640771e+00 1.41548371e+00 9.94620562e-01 2.05323890e-01
-9.12540019e-01 -5.12332320e-01 2.86773324e-01 -2.05976471e-01
-9.59573030e-01 7.12364092e-02 4.51802462e-01 -1.15664148e+00
9.55572724e-01 1.97770506e-01 1.12015808e+00 1.39549923e+00
3.90434951e-01 4.19719070e-02 1.11331594e+00 -3.11213434e-01
1.53824054e-02 7.48810649e-01 4.56973463e-01 1.16582185e-01
3.65415484e-01 -3.97160769e-01 -6.61307931e-01 -6.59618020e-01
-1.01805635e-01 5.67209423e-01 -2.02999637e-01 1.88729763e-01
-1.18320191e+00 1.33106792e+00 3.03879976e-01 1.01227307e+00
-5.34838855e-01 -4.25450832e-01 6.77408457e-01 3.50316286e-01
1.10520172e+00 4.35583800e-01 -6.01640642e-01 -7.14865446e-01
-5.30776918e-01 1.73255399e-01 1.08749497e+00 -7.33161867e-02
5.60111344e-01 -6.85732782e-01 -1.01202361e-01 7.50449717e-01
4.08013463e-01 4.27770823e-01 5.51802456e-01 -9.95701030e-02
2.28316322e-01 9.31242406e-01 1.54185012e-01 -1.83513403e+00
-7.71091342e-01 -3.12245429e-01 -3.15584898e-01 -3.69638443e-01
3.70006382e-01 -8.24835300e-01 -5.30457914e-01 1.81413198e+00
5.23585498e-01 -2.20845655e-01 -1.98356077e-01 3.73386919e-01
4.11668479e-01 8.60240996e-01 1.77932918e-01 -2.92800397e-01
1.52781022e+00 -2.19670907e-01 -1.17079437e+00 1.70812503e-01
7.19497621e-01 -8.70867729e-01 5.71000218e-01 1.94544598e-01
-4.93654191e-01 -1.89379558e-01 -8.36957335e-01 2.01388195e-01
-1.14916348e+00 -9.29016173e-01 4.43784475e-01 8.79965663e-01
-1.22510839e+00 4.55730975e-01 -5.05551159e-01 -1.35179126e+00
2.17039272e-01 -5.84919974e-02 -4.34046209e-01 4.31227803e-01
-1.46668518e+00 1.11741471e+00 2.21210584e-01 -3.19246709e-01
-6.18098211e-03 -7.36488402e-01 -5.41471362e-01 -4.95071054e-01
1.05432205e-01 -3.70481938e-01 8.72191429e-01 -9.59642768e-01
-1.04577088e+00 8.59385669e-01 -7.18003064e-02 -2.76047170e-01
4.08648372e-01 -1.94474772e-01 -8.58293891e-01 -2.47376055e-01
4.28876877e-02 4.98613656e-01 3.31238389e-01 -8.16658139e-01
-6.15818679e-01 -4.03363794e-01 1.73453584e-01 -1.40688315e-01
-6.84087038e-01 3.92506838e-01 8.20990875e-02 -5.64315200e-01
-4.25178170e-01 -8.53127658e-01 -2.06530064e-01 -3.47954452e-01
-1.70792967e-01 -6.06649518e-01 1.01341343e+00 -7.11595893e-01
1.60116851e+00 -2.21247196e+00 4.28886749e-02 1.85082883e-01
3.88185889e-01 3.99597362e-02 3.86555552e-01 1.46676326e+00
3.43437642e-01 5.25977075e-01 1.22562610e-01 -1.10736350e-02
9.88241062e-02 3.50854337e-01 -3.69204611e-01 5.94462752e-01
-2.12623700e-01 5.35419941e-01 -1.21555328e+00 -5.03761545e-02
-6.79337755e-02 7.48955190e-01 -6.01510108e-01 -3.86886224e-02
-1.28973871e-01 4.32272643e-01 -6.51434481e-01 8.73120129e-02
5.62662899e-01 -3.99943233e-01 3.03816468e-01 5.66682592e-02
-8.13105464e-01 3.47062737e-01 -3.56027037e-01 1.32741416e+00
-3.53903145e-01 1.00314867e+00 7.35441744e-02 -4.43062097e-01
6.71813965e-01 4.12909001e-01 4.47501928e-01 -4.49715436e-01
5.38671911e-01 -1.68824539e-01 1.67919710e-01 -6.73885942e-01
7.31202126e-01 -9.61459950e-02 -2.33258918e-01 9.24392700e-01
-2.23789528e-01 1.64419606e-01 -9.82314572e-02 3.39377075e-01
7.39374578e-01 -3.59863579e-01 6.80542707e-01 -8.05846453e-01
3.12991999e-02 2.29884565e-01 1.36516795e-01 3.25513989e-01
-3.57498676e-01 1.41459197e-01 5.98729908e-01 -5.85062504e-01
-6.12865806e-01 -8.09914768e-01 -2.80584902e-01 1.26714897e+00
-4.63645943e-02 -5.46688855e-01 -5.78999341e-01 -1.46906897e-01
9.46747214e-02 6.62474573e-01 -1.13386607e+00 4.36049670e-01
-3.86006445e-01 -9.53670502e-01 -1.32384613e-01 -3.15261424e-01
2.13974684e-01 -9.56655025e-01 -7.89755464e-01 4.39665198e-01
3.59476395e-02 -6.28575563e-01 -3.45114052e-01 -2.06169888e-01
-4.57952946e-01 -8.74665320e-01 -7.67795384e-01 -3.73621911e-01
5.21339297e-01 1.68804184e-01 1.00617135e+00 -1.30358249e-01
-3.68393391e-01 5.64763665e-01 -5.67587793e-01 -9.60634232e-01
-2.67636091e-01 1.56205252e-01 2.03502938e-01 -6.41527250e-02
5.77577949e-01 -4.37829822e-01 -8.44144225e-01 -1.51152685e-01
-1.15194535e+00 -1.24263078e-01 -2.84574777e-01 2.36148193e-01
-4.53940541e-01 -2.75277764e-01 5.77463031e-01 -1.29333985e+00
8.56637537e-01 -1.58106291e+00 1.37886122e-01 -1.46563813e-01
-5.02924800e-01 -5.45869172e-01 -2.26969644e-03 -2.53182799e-01
-8.67826164e-01 -1.10027397e+00 -2.76005622e-02 6.13122404e-01
1.59760281e-01 7.00280011e-01 7.70359039e-01 6.15980923e-01
6.88900113e-01 -3.76772255e-01 3.07592034e-01 -4.39635009e-01
3.53897750e-01 1.28294218e+00 -4.49084729e-01 1.00787766e-01
7.07999289e-01 8.95023644e-01 -1.02847552e+00 -1.29317474e+00
-9.62602317e-01 -7.94825137e-01 -1.20563768e-01 -4.76593137e-01
1.35604417e+00 -7.47741580e-01 -6.64108276e-01 5.64687908e-01
-1.22548616e+00 -2.23966584e-01 1.08660050e-01 2.62065470e-01
4.09879088e-01 1.42399207e-01 -3.73210222e-01 -8.75024021e-01
-2.52348125e-01 -5.77273428e-01 3.34906846e-01 3.50220323e-01
-1.12323105e+00 -1.78873694e+00 9.11425591e-01 2.88245499e-01
1.06311131e+00 6.88927829e-01 8.22063506e-01 -1.10929918e+00
1.95824832e-01 2.17387587e-01 -1.82390586e-01 -2.24428207e-01
1.15201807e+00 -6.42690584e-02 -7.95674503e-01 -5.73117435e-01
-1.15764253e-01 6.35404810e-02 2.58683413e-01 1.19174451e-01
3.69650841e-01 -6.37709439e-01 -6.55288458e-01 -9.09584537e-02
1.44320047e+00 3.23900014e-01 2.70927072e-01 6.00796998e-01
2.10494861e-01 1.05859363e+00 6.21870011e-02 8.29392850e-01
3.73854339e-01 5.91762722e-01 6.94617778e-02 -1.82585530e-02
4.34509784e-01 -2.20658973e-01 5.22131205e-01 1.23554504e+00
2.13361368e-01 -4.51681077e-01 -1.14345574e+00 8.71824682e-01
-1.43706477e+00 -7.49680161e-01 7.59080052e-03 1.75189769e+00
7.04515815e-01 1.98300630e-01 1.95917696e-01 -4.35689300e-01
5.34553766e-01 6.35270715e-01 -1.75368860e-02 -9.72194374e-01
9.16248187e-02 -2.07357407e-02 4.86116111e-01 8.04779947e-01
-5.83657503e-01 3.86053741e-01 6.41329193e+00 3.35512638e-01
-1.28904986e+00 5.12062013e-01 6.66476786e-01 1.50572851e-01
-8.10024559e-01 -2.87698984e-01 -6.25025570e-01 7.40741372e-01
1.00094497e+00 -3.40943158e-01 1.24113776e-01 2.30013236e-01
5.60891509e-01 -1.85237855e-01 -3.56250405e-01 5.56964397e-01
2.99878061e-01 -1.46212757e+00 -1.57547489e-01 4.65371609e-02
9.26807284e-01 4.19755280e-01 1.14950739e-01 1.16684809e-01
2.67740428e-01 -7.29343116e-01 2.36601129e-01 3.70977968e-01
4.10084277e-01 -3.36346388e-01 6.64748609e-01 1.44898355e-01
-5.64493656e-01 1.62570819e-01 9.23318341e-02 -4.21703368e-01
5.87561727e-01 6.66806698e-01 -1.16900551e+00 2.02535361e-01
7.17491567e-01 8.05809438e-01 -2.55757928e-01 4.17950422e-01
3.01909745e-01 6.12851262e-01 -1.39777511e-01 -5.42388916e-01
5.90573430e-01 -4.59793597e-01 7.48099506e-01 1.52113557e+00
2.79144466e-01 -1.50195342e-02 -1.70646369e-01 1.50904134e-01
2.45523155e-01 3.14432889e-01 -9.29463625e-01 -7.43575811e-01
1.57657057e-01 1.16821218e+00 -9.47382390e-01 -1.79933518e-01
-4.46612954e-01 7.22122133e-01 7.53198843e-03 3.04818153e-01
-5.33248365e-01 -1.85191184e-01 1.05359769e+00 5.70990503e-01
1.31590948e-01 -2.59976745e-01 1.86583400e-01 -7.39950001e-01
-4.22280461e-01 -6.58291161e-01 3.70839447e-01 -3.77351195e-01
-1.41354191e+00 6.29044473e-01 5.93942553e-02 -4.90579516e-01
1.29563799e-02 -3.72212797e-01 -8.47650528e-01 6.41600430e-01
-1.08755481e+00 -5.39897561e-01 3.16810273e-02 6.05165698e-02
5.77855885e-01 -6.00664644e-03 7.46125162e-01 1.35655969e-01
-2.13449225e-01 1.93691283e-01 2.15567112e-01 -2.45236978e-01
6.18989348e-01 -9.25346792e-01 3.47520322e-01 9.47811678e-02
-3.43198359e-01 9.39451337e-01 1.16404331e+00 -8.81142020e-01
-1.04555798e+00 -5.39209545e-01 1.45779407e+00 -5.74589550e-01
1.54525745e+00 -5.14962256e-01 -5.56487381e-01 6.65671289e-01
1.06176078e+00 -1.03194964e+00 1.24438965e+00 4.63780314e-01
-1.72465652e-01 3.61968011e-01 -1.26878440e+00 8.90088499e-01
8.89877081e-01 -5.62557697e-01 -9.55910206e-01 7.34039307e-01
1.12283349e+00 6.25723228e-02 -6.01924539e-01 -9.94428098e-02
5.79541326e-01 -8.66531909e-01 6.02607906e-01 -7.61895835e-01
9.75148305e-02 1.81628227e-01 -1.20269336e-01 -1.55158675e+00
-8.93567652e-02 -1.25625384e+00 5.68463027e-01 1.25844443e+00
5.20721793e-01 -1.29612219e+00 3.60853195e-01 3.72726470e-01
2.10178062e-01 -6.81611598e-01 -9.30338383e-01 1.31818280e-01
1.82044879e-02 -3.55229639e-02 2.55019069e-01 1.41142559e+00
4.08466935e-01 7.87244439e-02 -1.74949437e-01 -2.98512995e-01
-2.62319624e-01 -1.76344663e-01 6.55061781e-01 -1.21210384e+00
-1.97242737e-01 -2.37911865e-01 -4.21496004e-01 -5.08561969e-01
-5.84119916e-01 -6.47701383e-01 -5.67477226e-01 -1.57133508e+00
2.85563350e-01 -3.25958312e-01 -2.25342140e-01 -7.59370476e-02
1.24669284e-01 -5.71964756e-02 3.21240127e-01 -6.83364943e-02
-3.22734922e-01 -6.16493486e-02 9.88217950e-01 -7.80201852e-02
-4.73378807e-01 -3.49575758e-01 -1.05182087e+00 6.07276261e-01
8.94605994e-01 -3.39795530e-01 -3.91985327e-01 -1.80107150e-02
1.39193797e+00 -4.06368643e-01 -3.14968377e-01 -4.89563942e-01
-2.21654758e-01 -4.12599854e-02 -9.33123231e-02 -5.92500389e-01
2.52233505e-01 -8.30217004e-01 4.53162014e-01 8.49564135e-01
-1.46355227e-01 6.05297983e-01 4.23538059e-01 5.51122606e-01
1.68747883e-02 1.48197055e-01 2.36257210e-01 1.12364799e-01
-2.14005515e-01 9.86873433e-02 -9.81095970e-01 3.00028414e-01
9.98895884e-01 -3.35803568e-01 -4.75744069e-01 -4.67654020e-01
-9.11076784e-01 2.77723491e-01 4.96859342e-01 8.23090613e-01
-7.61319464e-03 -8.81994069e-01 -6.55005217e-01 -1.90317631e-01
-8.53628665e-02 -4.62574422e-01 6.50798559e-01 9.92409110e-01
-5.75627804e-01 3.63111675e-01 -2.70976812e-01 -3.08241397e-01
-1.16023302e+00 2.59642750e-01 -5.63210510e-02 -1.85429186e-01
-5.33764660e-01 6.84467554e-01 1.30336165e-01 -5.07898331e-01
6.50289059e-02 -4.08997685e-01 -8.16694498e-01 1.25115407e+00
6.77340448e-01 5.88684440e-01 -3.87221187e-01 -6.48597836e-01
-4.86054868e-01 4.09442782e-01 -1.71637893e-01 -7.80920684e-02
1.65913939e+00 -3.37199897e-01 -4.43074435e-01 1.17952681e+00
1.57161522e+00 5.21723866e-01 -6.34536445e-01 3.43158059e-02
1.49883479e-01 -5.01038015e-01 -3.37749898e-01 -7.66135633e-01
-5.42020202e-01 4.71372008e-01 6.97364569e-01 1.09561086e+00
3.97457063e-01 1.35522142e-01 1.01044202e+00 9.11570638e-02
1.09028459e-01 -9.72341776e-01 -1.47787005e-01 5.14585853e-01
8.76267135e-01 -1.04178917e+00 -8.05432200e-02 -1.04441509e-01
-6.42213583e-01 9.66423869e-01 -1.56396434e-01 1.27028435e-01
1.42024827e+00 -7.65098631e-02 4.74483848e-01 -7.86376119e-01
-6.42669797e-01 2.21816719e-01 -2.03585118e-01 4.63031858e-01
6.66815579e-01 3.47326100e-01 -7.18921602e-01 1.67356301e-02
-3.36525291e-01 -2.44501382e-01 8.76253188e-01 7.01690555e-01
-6.35223091e-01 -1.09212983e+00 -1.12608075e-01 3.84052724e-01
-9.52366352e-01 1.96760654e-01 -6.09898746e-01 9.08389151e-01
3.85114044e-01 1.00625241e+00 4.22956914e-01 -2.33967319e-01
-2.55534276e-02 2.46003971e-01 -1.91898838e-01 -6.95747316e-01
-1.02838933e+00 -4.23175305e-01 3.63614202e-01 -2.01789394e-01
-6.63937211e-01 -8.77567768e-01 -9.52761948e-01 -6.84647143e-01
-7.28816539e-02 4.29582030e-01 9.83278036e-01 8.82823408e-01
4.19636905e-01 3.65628451e-01 6.05155051e-01 -4.31270838e-01
5.05389757e-02 -1.14631486e+00 -2.30654731e-01 4.65367436e-01
5.56620061e-01 -5.81481576e-01 -5.73624671e-01 -3.65658462e-01] | [8.516326904296875, 9.802871704101562] |
f02a478e-0338-4bd7-b4b2-36de21ecff3c | weakly-supervised-named-entity-tagging-with | 2107.02282 | null | https://arxiv.org/abs/2107.02282v1 | https://arxiv.org/pdf/2107.02282v1.pdf | Weakly Supervised Named Entity Tagging with Learnable Logical Rules | We study the problem of building entity tagging systems by using a few rules as weak supervision. Previous methods mostly focus on disambiguation entity types based on contexts and expert-provided rules, while assuming entity spans are given. In this work, we propose a novel method TALLOR that bootstraps high-quality logical rules to train a neural tagger in a fully automated manner. Specifically, we introduce compound rules that are composed from simple rules to increase the precision of boundary detection and generate more diverse pseudo labels. We further design a dynamic label selection strategy to ensure pseudo label quality and therefore avoid overfitting the neural tagger. Experiments on three datasets demonstrate that our method outperforms other weakly supervised methods and even rivals a state-of-the-art distantly supervised tagger with a lexicon of over 2,000 terms when starting from only 20 simple rules. Our method can serve as a tool for rapidly building taggers in emerging domains and tasks. Case studies show that learned rules can potentially explain the predicted entities. | ['Zhe Feng', 'Julian McAuley', 'Jingbo Shang', 'Haibo Ding', 'Jiacheng Li'] | 2021-07-05 | null | https://aclanthology.org/2021.acl-long.352 | https://aclanthology.org/2021.acl-long.352.pdf | acl-2021-5 | ['boundary-detection'] | ['computer-vision'] | [ 7.59082437e-02 7.08025396e-01 -4.93134826e-01 -5.64247072e-01
-6.63885176e-01 -8.82303536e-01 5.43410361e-01 4.41007227e-01
-5.22227705e-01 1.10027611e+00 -2.91110110e-02 -3.40030611e-01
-3.20956893e-02 -9.46280420e-01 -8.89486313e-01 -3.04912090e-01
-8.48139226e-02 9.85105872e-01 5.34575760e-01 -2.75245577e-01
-1.98726565e-01 2.99463533e-02 -1.54873145e+00 3.89048070e-01
1.23852789e+00 7.25372612e-01 -4.13210690e-03 3.33703905e-02
-3.68593186e-01 8.50092232e-01 -3.99396390e-01 -7.87373006e-01
3.29609960e-02 6.17146157e-02 -1.10326731e+00 -3.18707854e-01
4.62420940e-01 1.34141326e-01 2.64726073e-01 1.08450937e+00
1.76692232e-01 8.76739547e-02 6.72336996e-01 -8.86198163e-01
-6.08539999e-01 1.36760688e+00 -1.34035116e-02 -1.81546107e-01
1.58717528e-01 -5.93036115e-01 1.48509598e+00 -8.42016220e-01
8.35327327e-01 8.59948456e-01 1.11405051e+00 7.57348418e-01
-1.27422023e+00 -8.02603662e-01 4.00392890e-01 -4.38604727e-02
-1.48960912e+00 -4.03237730e-01 3.80587995e-01 -4.19115037e-01
1.24043608e+00 5.10382988e-02 1.21847391e-01 7.24372625e-01
-3.04284900e-01 5.45089126e-01 9.07544672e-01 -7.94585288e-01
2.89627403e-01 4.47533756e-01 4.93509948e-01 9.84264791e-01
8.08624446e-01 -7.74714723e-02 -4.42595840e-01 -4.08409089e-01
4.39277232e-01 -4.59708840e-01 -1.11591265e-01 -4.68971640e-01
-1.15678465e+00 7.56863773e-01 2.88573474e-01 6.57323658e-01
-2.73902446e-01 -2.21324369e-01 2.15302229e-01 -1.96551196e-02
6.62280381e-01 1.02724707e+00 -1.10548162e+00 2.64136374e-01
-9.73324358e-01 1.88065425e-01 1.13043773e+00 1.17550039e+00
1.05074573e+00 -3.14405054e-01 -7.88979381e-02 1.04070854e+00
2.36651242e-01 3.65744084e-01 3.76754880e-01 -7.81529486e-01
2.64148623e-01 7.73410439e-01 3.89053941e-01 -5.98155618e-01
-6.63029552e-01 -5.29411495e-01 -3.83109182e-01 -1.33878112e-01
4.48635757e-01 -3.33096296e-01 -1.06669903e+00 1.79049563e+00
3.73540014e-01 2.59519428e-01 9.81672853e-02 4.16252464e-01
8.50693524e-01 2.18694866e-01 4.79699492e-01 -5.69545515e-02
1.51279497e+00 -8.43292534e-01 -6.57178164e-01 -4.72168267e-01
1.17184114e+00 -3.52651328e-01 7.83326685e-01 3.12452197e-01
-8.34896266e-01 -4.55544949e-01 -9.60228741e-01 1.81577817e-01
-7.87253916e-01 3.11635584e-01 9.39437807e-01 5.96382797e-01
-9.73785222e-01 7.00411320e-01 -7.64291823e-01 -5.99196732e-01
1.58578396e-01 5.52174449e-01 -4.68885452e-01 2.04683006e-01
-1.71105218e+00 1.13177526e+00 1.06824279e+00 -3.36426735e-01
-2.66502291e-01 -5.65560877e-01 -1.01250744e+00 -1.60823967e-02
6.32885873e-01 -7.31198311e-01 1.39017320e+00 -7.29809403e-01
-1.12361848e+00 1.19519973e+00 -9.06617045e-02 -6.21644437e-01
-9.20697749e-02 -3.34020168e-01 -5.93086660e-01 -1.68515742e-01
4.73022193e-01 7.63553858e-01 1.73996478e-01 -1.40117967e+00
-1.09273160e+00 -4.19770181e-02 1.53236151e-01 4.85287644e-02
-3.79721433e-01 -1.33597031e-02 -3.79191667e-01 -4.96421486e-01
1.36271026e-02 -8.81480277e-01 -2.72797406e-01 -7.13270605e-01
-3.81382942e-01 -5.95291376e-01 3.20259817e-02 -3.98540854e-01
1.51121759e+00 -1.62963355e+00 -3.36776733e-01 3.14697891e-01
2.95476437e-01 4.56565738e-01 2.09256709e-01 5.09985462e-02
3.27755287e-02 3.80084008e-01 -2.16388538e-01 -7.23076984e-02
1.47568241e-01 4.12311018e-01 -3.44511598e-01 -1.78982660e-01
2.11371869e-01 9.89153087e-01 -1.16475022e+00 -7.10323453e-01
-3.58457625e-01 9.34788957e-02 -4.83577818e-01 1.09972395e-01
-4.51140761e-01 -9.61718261e-02 -5.30506849e-01 5.78235030e-01
2.56040037e-01 -6.15712225e-01 7.36325204e-01 -7.35520795e-02
8.50687176e-02 7.09429741e-01 -1.15719247e+00 1.57669640e+00
-5.19287527e-01 1.27256483e-01 -7.89546520e-02 -9.06907558e-01
1.09184217e+00 3.63581866e-01 2.14558095e-01 -2.27337658e-01
5.30578569e-02 5.02698839e-01 -2.65352994e-01 -4.09603745e-01
5.87078452e-01 -2.14231893e-01 -3.41901541e-01 2.79654175e-01
3.38371456e-01 3.03897262e-01 3.92723262e-01 1.39196083e-01
1.06850457e+00 4.81075257e-01 6.08075202e-01 -2.53898114e-01
3.24160129e-01 4.00165290e-01 9.27501082e-01 9.64896142e-01
5.93000231e-03 1.08400956e-01 4.96328883e-02 -5.27662396e-01
-8.81422520e-01 -7.59547770e-01 -3.21675152e-01 1.53201282e+00
1.04970291e-01 -5.81234157e-01 -8.04448009e-01 -1.32872379e+00
5.65803051e-02 8.46840382e-01 -5.69278657e-01 1.04008324e-01
-6.05659366e-01 -7.44233549e-01 8.98241639e-01 6.39349937e-01
3.48690480e-01 -1.13934314e+00 -1.88516811e-01 4.40318018e-01
-3.89393657e-01 -1.28601706e+00 2.08941437e-02 8.16705346e-01
-6.37745619e-01 -9.22664404e-01 -1.17365740e-01 -1.20425355e+00
7.12925553e-01 -2.05875769e-01 1.55660486e+00 1.95140883e-01
4.18993652e-01 -2.25630984e-01 -4.82615232e-01 -1.93562701e-01
-5.22147596e-01 7.23412097e-01 3.07550728e-01 -4.54690486e-01
1.02629650e+00 -5.00188887e-01 2.21439023e-02 3.40232372e-01
-6.85118496e-01 1.04752414e-01 5.93154311e-01 9.74453509e-01
4.63806063e-01 1.40818954e-01 7.21851230e-01 -1.68476903e+00
3.71256649e-01 -5.73266387e-01 -5.16484916e-01 6.85662329e-01
-1.19018781e+00 5.83339214e-01 5.97331405e-01 -3.97682577e-01
-1.23576272e+00 3.10871363e-01 9.97564867e-02 1.96585029e-01
-5.45757532e-01 6.46231830e-01 -2.16731355e-01 1.08146966e-01
1.12093866e+00 -2.32573986e-01 -7.18286216e-01 -6.54755414e-01
6.66521549e-01 6.86617255e-01 6.44714952e-01 -9.65794146e-01
8.46189976e-01 1.71363175e-01 -3.74346763e-01 1.00021221e-01
-1.40785921e+00 -5.45092821e-01 -8.89788747e-01 1.21924393e-01
5.39926410e-01 -9.71064687e-01 -2.84931242e-01 -2.87622511e-02
-1.07813871e+00 -3.49898458e-01 -1.18223369e-01 3.62525225e-01
-2.31256515e-01 4.72770073e-02 -7.19320238e-01 -3.79026055e-01
-3.81461620e-01 -3.44320089e-01 1.08400178e+00 2.65039414e-01
-5.71382761e-01 -1.14926362e+00 3.48354250e-01 3.23465586e-01
8.20045918e-02 -8.18279106e-03 8.01483333e-01 -1.46030474e+00
-2.74414212e-01 -2.07258519e-02 -1.14033684e-01 1.75636373e-02
1.31993905e-01 -4.05768812e-01 -9.80841100e-01 9.66290087e-02
-7.36011922e-01 -4.32755411e-01 8.35010469e-01 -7.49228969e-02
5.88047266e-01 -4.49346602e-01 -9.23258722e-01 4.52195466e-01
1.22658873e+00 2.16576070e-01 2.45109603e-01 7.77088940e-01
7.13153839e-01 6.51604593e-01 8.85666668e-01 3.73931834e-03
5.76960921e-01 6.79966211e-01 -1.21044382e-01 -7.42426813e-02
1.73381791e-01 -5.16119301e-01 3.23062651e-02 6.75774336e-01
-1.27390385e-01 -4.54884619e-02 -1.28284097e+00 7.74816513e-01
-2.09421396e+00 -8.37294161e-01 -1.39092859e-02 1.97343206e+00
1.60388815e+00 4.21910167e-01 -9.29426923e-02 -1.10905856e-01
9.11395073e-01 -3.40623826e-01 -2.63380706e-01 -3.03261671e-02
-1.86136775e-02 3.50380987e-01 6.71856344e-01 6.60143077e-01
-1.45482612e+00 1.68852854e+00 6.45854616e+00 7.87386596e-01
-7.17845738e-01 1.12751268e-01 2.78205007e-01 3.41070265e-01
-3.69548112e-01 2.41092876e-01 -1.48439550e+00 2.36041933e-01
8.19788575e-01 3.02049257e-02 1.59696475e-01 1.02550876e+00
-3.46952379e-01 1.91211537e-01 -1.04545748e+00 5.04898310e-01
-5.74557409e-02 -1.26026595e+00 -2.96246354e-02 -2.86185499e-02
9.02217448e-01 1.11180857e-01 -5.03246367e-01 6.48730338e-01
1.19055390e+00 -8.38050604e-01 6.24602199e-01 4.30327535e-01
8.26477885e-01 -4.62643117e-01 9.12230968e-01 3.79252404e-01
-1.14972079e+00 1.44698143e-01 -2.08416983e-01 -3.23491134e-02
1.05049096e-01 8.36297095e-01 -1.31846189e+00 5.81842601e-01
6.63816094e-01 6.01320028e-01 -5.84683478e-01 8.71602416e-01
-6.36027157e-01 9.30094540e-01 -4.87731636e-01 -6.88994750e-02
6.83987960e-02 4.38345313e-01 1.97445840e-01 1.53237092e+00
1.21975616e-01 2.00304657e-01 3.96502316e-01 7.17100501e-01
-2.82681704e-01 2.39543110e-01 -5.95851719e-01 1.65699318e-01
9.41382051e-01 1.39311600e+00 -9.46536183e-01 -6.11231804e-01
-3.61275136e-01 6.45157456e-01 8.08557451e-01 2.55025506e-01
-7.11789608e-01 -5.16990960e-01 2.58539706e-01 4.28512469e-02
4.60650593e-01 4.99575846e-02 -4.22322839e-01 -1.38512671e+00
-1.30801685e-02 -6.52119756e-01 6.38427734e-01 -5.97529948e-01
-1.29990411e+00 8.40300024e-01 6.82522133e-02 -1.01675868e+00
-5.68572819e-01 -6.58891857e-01 -1.75668463e-01 4.37197536e-01
-1.75913477e+00 -1.40321136e+00 -1.06812388e-01 1.89563110e-01
9.32443514e-03 -7.04302564e-02 1.09934866e+00 3.83972436e-01
-3.42565477e-01 6.56416059e-01 -1.10269122e-01 4.76438135e-01
9.76730645e-01 -1.52461791e+00 4.35807109e-01 7.84646034e-01
4.94058549e-01 9.71717715e-01 5.98767459e-01 -1.02984583e+00
-1.74310267e-01 -1.25364530e+00 1.65279329e+00 -6.74325109e-01
6.93243444e-01 -4.35754418e-01 -1.08576369e+00 8.95574808e-01
2.85971463e-02 4.30378281e-02 8.91606688e-01 8.36146474e-01
-6.93820417e-01 -8.81249667e-04 -9.34861124e-01 2.24473715e-01
1.33779955e+00 -3.15059453e-01 -1.05748951e+00 2.17225134e-01
6.93218827e-01 -3.17999780e-01 -8.34558189e-01 8.05500031e-01
4.87656593e-01 -4.53381032e-01 7.13758051e-01 -8.30352485e-01
-6.73338119e-03 -5.84518731e-01 1.33274600e-01 -1.20409536e+00
-6.74836576e-01 -5.42445958e-01 -2.35914707e-01 1.57966113e+00
1.09666550e+00 -5.19577324e-01 7.67904222e-01 7.20840514e-01
4.16646665e-03 -4.99277562e-01 -5.50112963e-01 -8.84520054e-01
-5.83251417e-02 -2.72164255e-01 7.62482464e-01 1.50885224e+00
4.89611477e-01 5.57331622e-01 -1.54720396e-01 3.36919248e-01
3.89613271e-01 5.78530021e-02 3.56911182e-01 -1.69178295e+00
-2.23244086e-01 -2.01202318e-01 -1.46794870e-01 -9.31515098e-01
5.65826893e-01 -1.14887857e+00 4.40764219e-01 -1.47221351e+00
1.84655651e-01 -1.06878102e+00 -5.89740217e-01 1.28946447e+00
-5.32601714e-01 2.27797687e-01 -3.62407446e-01 2.10461244e-01
-9.98501182e-01 -7.87241682e-02 4.94315684e-01 2.55031697e-02
-2.25238204e-01 -2.07594767e-01 -9.86544907e-01 8.96748960e-01
6.76241100e-01 -8.27601552e-01 -1.66495830e-01 -2.97020227e-01
7.65734017e-01 -6.41965687e-01 -3.07341907e-02 -8.54874015e-01
3.78435493e-01 -9.02473554e-02 7.35669211e-02 -1.66362345e-01
-3.16111535e-01 -7.86664307e-01 9.72039327e-02 -1.08206244e-02
-5.39886355e-01 -3.74363273e-01 1.22325718e-01 2.97670484e-01
-1.59628525e-01 -3.25608730e-01 4.64153796e-01 -1.58624172e-01
-1.00097942e+00 -6.32671043e-02 -5.11230379e-02 2.92170435e-01
8.60430419e-01 8.63184780e-02 -3.78541142e-01 6.29253611e-02
-8.62320423e-01 3.50770921e-01 5.32572806e-01 4.94020134e-01
-5.57083152e-02 -1.44650662e+00 -5.14041364e-01 -5.39283566e-02
4.58495200e-01 9.75772887e-02 -4.81347471e-01 2.40931392e-01
-2.67563403e-01 6.67092919e-01 7.41972253e-02 -3.19223344e-01
-1.15852916e+00 4.65877801e-01 3.15523088e-01 -8.23879957e-01
-3.12391520e-01 9.27793622e-01 -4.45129983e-02 -7.93586016e-01
2.81537861e-01 -3.92667532e-01 -6.50284290e-01 5.70751801e-02
3.40550601e-01 -1.97053522e-01 2.82176971e-01 -5.20603001e-01
-5.04727423e-01 4.01684910e-01 -2.08303064e-01 2.10102014e-02
1.41854382e+00 2.94189639e-02 -3.29647996e-02 2.82277972e-01
4.36610937e-01 2.73204744e-01 -7.74647534e-01 -6.92165136e-01
6.68113053e-01 1.99002877e-01 -4.42297123e-02 -1.21866131e+00
-6.35985255e-01 2.89952993e-01 1.05367504e-01 4.32170540e-01
8.45145166e-01 3.54203463e-01 7.46129096e-01 7.46375024e-01
6.55629277e-01 -1.21289134e+00 -5.23881316e-01 7.53210068e-01
1.02888532e-01 -1.17886317e+00 -2.46336311e-01 -6.04324222e-01
-6.26956046e-01 8.71068358e-01 8.20224226e-01 2.46105477e-01
5.45447886e-01 5.18834889e-01 2.33781621e-01 -2.23745450e-01
-8.07109416e-01 -7.62457907e-01 5.07530391e-01 6.05704308e-01
7.68017352e-01 1.39591366e-01 -3.70588452e-01 1.09973443e+00
-2.35521659e-01 1.38469726e-01 2.61911610e-03 9.14264977e-01
-7.44367599e-01 -1.64853036e+00 -1.92097262e-01 3.69278342e-01
-6.99132919e-01 -5.65005839e-01 -5.42076170e-01 6.16192162e-01
7.44829535e-01 8.46567035e-01 -2.87998736e-01 -4.61373329e-01
2.23222435e-01 6.64109349e-01 1.90218940e-01 -1.09344399e+00
-6.52279854e-01 -2.89535910e-01 7.85418272e-01 -2.47689575e-01
-7.75841475e-01 -5.17400563e-01 -1.48260951e+00 1.83634683e-01
-8.80962968e-01 6.95694447e-01 2.74029136e-01 1.23941219e+00
5.15366793e-01 1.65951043e-01 2.14447141e-01 -2.06002653e-01
-2.64503807e-01 -9.91267145e-01 -3.98525685e-01 5.40191472e-01
-6.51970878e-03 -9.32847500e-01 -7.69369006e-02 4.48834717e-01] | [9.558032989501953, 8.843116760253906] |
d4c4e7f9-9fb7-48cb-bb1a-92aeae60851a | quantization-guided-jpeg-artifact-correction | 2004.09320 | null | https://arxiv.org/abs/2004.09320v2 | https://arxiv.org/pdf/2004.09320v2.pdf | Quantization Guided JPEG Artifact Correction | The JPEG image compression algorithm is the most popular method of image compression because of its ability for large compression ratios. However, to achieve such high compression, information is lost. For aggressive quantization settings, this leads to a noticeable reduction in image quality. Artifact correction has been studied in the context of deep neural networks for some time, but the current state-of-the-art methods require a different model to be trained for each quality setting, greatly limiting their practical application. We solve this problem by creating a novel architecture which is parameterized by the JPEG files quantization matrix. This allows our single model to achieve state-of-the-art performance over models trained for specific quality settings. | ['Ser-Nam Lim', 'Max Ehrlich', 'Abhinav Shrivastava', 'Larry Davis'] | 2020-04-17 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/570_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123530290.pdf | eccv-2020-8 | ['jpeg-artifact-correction'] | ['computer-vision'] | [ 5.72870135e-01 -3.05792123e-01 -2.89196074e-01 -3.30078334e-01
-8.88913274e-01 -6.80735111e-02 3.58363718e-01 1.19638667e-01
-3.97171646e-01 2.82996356e-01 2.62162298e-01 -1.54786453e-01
2.39993334e-01 -7.48037636e-01 -8.22693586e-01 -4.99343544e-01
-9.74554494e-02 1.18299611e-01 3.32171440e-01 -9.34397727e-02
4.10360664e-01 1.65068626e-01 -1.30412316e+00 4.92045134e-01
6.45843089e-01 1.10610688e+00 2.73354024e-01 8.78161967e-01
4.89471573e-03 9.22263205e-01 -6.81767762e-01 -3.84275734e-01
4.44746524e-01 -5.81827462e-01 -5.01908660e-01 -9.76999253e-02
8.38198364e-01 -1.03249013e+00 -8.53235424e-01 1.22799027e+00
5.55956423e-01 -7.32878372e-02 4.98346180e-01 -9.85554576e-01
-6.51724458e-01 6.73004806e-01 -5.53059757e-01 2.79430181e-01
-2.44730269e-03 2.00076759e-01 9.19069409e-01 -5.58795571e-01
3.79262298e-01 1.10487747e+00 6.08855844e-01 4.87922400e-01
-1.40890992e+00 -7.62049675e-01 -3.34196985e-01 4.29625690e-01
-1.25434542e+00 -6.40385926e-01 6.52847886e-01 -1.09978996e-01
1.17017567e+00 -1.72273852e-02 7.67915010e-01 8.61062765e-01
4.82314855e-01 5.65609992e-01 6.79601669e-01 -3.12315464e-01
3.15002024e-01 -3.52994174e-01 -3.89102936e-01 2.09423691e-01
1.53198913e-01 4.73908661e-03 -6.56054497e-01 2.13542134e-01
1.01959467e+00 2.05973089e-02 -5.41355312e-01 -3.80278826e-01
-7.93941319e-01 8.34287226e-01 6.54892921e-01 2.57534325e-01
-2.28825167e-01 8.48537266e-01 4.74241614e-01 4.99329269e-01
2.54805654e-01 4.93593305e-01 -2.00261503e-01 -4.25183088e-01
-1.62656415e+00 4.21068370e-01 4.55728829e-01 5.83826542e-01
3.42303336e-01 2.24606708e-01 -7.31020495e-02 9.34279263e-01
2.98931062e-01 1.44476488e-01 6.75238252e-01 -1.39929533e+00
5.90864301e-01 2.20064834e-01 -7.43666515e-02 -9.88760769e-01
-1.89434633e-01 -6.34962440e-01 -1.06173074e+00 5.36283016e-01
2.91107744e-01 4.35596257e-01 -1.02348948e+00 1.81070471e+00
-4.35802579e-01 1.40311912e-01 -1.17434151e-02 8.35758924e-01
2.60889232e-01 6.61043167e-01 -1.54430762e-01 -3.67920985e-03
1.02504134e+00 -9.52596486e-01 -8.22353065e-01 -5.30900955e-01
3.05808961e-01 -6.89745486e-01 1.12716138e+00 7.39271820e-01
-1.44779861e+00 -4.97822911e-01 -1.61201620e+00 -4.03050959e-01
1.37318959e-02 -3.49617004e-01 4.06967014e-01 6.43336833e-01
-1.19263268e+00 1.02009010e+00 -1.12688541e+00 -1.05794758e-01
7.13228405e-01 4.70688432e-01 -2.26267725e-01 -3.39281559e-01
-1.05183351e+00 9.13602769e-01 2.52451897e-01 -2.33891085e-01
-9.31334436e-01 -8.01916301e-01 -6.05625629e-01 5.83211660e-01
1.19453594e-01 -4.43655998e-01 1.48524308e+00 -8.63911986e-01
-1.28257275e+00 6.66471481e-01 3.95188592e-02 -9.40405607e-01
6.41870201e-01 -4.40875173e-01 -2.58949786e-01 4.61109012e-01
-1.89711288e-01 8.28463674e-01 1.12182832e+00 -9.79209125e-01
-3.03775519e-01 -7.77442902e-02 -6.60290569e-03 8.72952715e-02
-6.78488135e-01 2.89651323e-02 -8.28218102e-01 -8.21075976e-01
1.36658162e-01 -7.63769090e-01 -1.53497532e-01 4.67409819e-01
6.36256784e-02 2.86648452e-01 7.09273279e-01 -8.15729916e-01
1.45018888e+00 -2.29862213e+00 -5.89223951e-02 -1.62910715e-01
3.80916506e-01 3.46377820e-01 -1.60544887e-01 2.52594143e-01
3.05840075e-02 4.49206114e-01 -3.95010948e-01 -5.79739690e-01
-1.77638903e-01 2.06389949e-02 -6.05316721e-02 3.10030460e-01
1.46965116e-01 5.46701252e-01 -5.98308146e-01 -3.48270953e-01
3.01092565e-01 8.52424622e-01 -1.04681098e+00 2.70180792e-01
-4.71315570e-02 -5.53359985e-02 7.25949109e-02 2.25613087e-01
8.95310640e-01 -4.95489478e-01 3.04536730e-01 -5.28855562e-01
2.30703890e-01 4.65386927e-01 -9.66606796e-01 1.81247497e+00
-4.76483673e-01 9.24558997e-01 -1.18012708e-02 -8.24691832e-01
3.91426891e-01 4.16098326e-01 6.31941974e-01 -9.23719227e-01
2.41545469e-01 2.98663288e-01 2.78509200e-01 -3.65550280e-01
6.16113484e-01 -1.60709098e-01 3.65794301e-01 3.22572738e-01
-1.98233724e-02 -3.07661444e-01 2.64648944e-01 2.82929391e-01
1.23761892e+00 -1.81795403e-01 8.08181763e-02 1.47261307e-01
-1.57529667e-01 -4.97395068e-01 3.63807887e-01 8.44679654e-01
-3.51630598e-01 1.17178380e+00 2.32710987e-01 -3.22432905e-01
-1.50451863e+00 -9.33633804e-01 -2.01387018e-01 7.19236732e-01
1.32478654e-01 -6.34592354e-01 -9.28030729e-01 -1.95336014e-01
-1.84492677e-01 5.72296262e-01 -3.32877010e-01 -3.88147712e-01
-7.61357009e-01 -6.56888485e-01 6.45824790e-01 5.16486466e-01
9.33601022e-01 -8.61782551e-01 -1.04467702e+00 5.05127311e-01
-3.36848438e-01 -1.26310658e+00 -4.98552859e-01 2.67670512e-01
-1.04922533e+00 -8.24188352e-01 -7.38779485e-01 -3.59889150e-01
1.98502854e-01 3.33149076e-01 1.35838318e+00 3.96834612e-01
-8.13689679e-02 -5.33401454e-03 -2.60808080e-01 -1.19987018e-01
-7.40990758e-01 -4.39467728e-02 9.13186371e-03 -3.41999680e-01
1.88078061e-01 -7.66413033e-01 -9.98187959e-01 -4.44298014e-02
-1.43668401e+00 1.16387628e-01 7.30861902e-01 8.17183614e-01
5.05615234e-01 3.85160625e-01 1.39342993e-01 -5.18913507e-01
6.57394946e-01 -1.15912803e-01 -4.43579137e-01 -7.24835321e-02
-8.65906537e-01 1.71360657e-01 7.40592182e-01 -3.40938300e-01
-5.69629490e-01 -2.77113736e-01 -3.30061376e-01 -5.58321595e-01
5.13796024e-02 5.73135734e-01 -6.01966679e-03 -3.61526161e-01
7.10293889e-01 1.17100604e-01 5.49104325e-02 -5.59904218e-01
1.48991615e-01 5.99247277e-01 8.51332486e-01 -6.99238330e-02
5.07919550e-01 1.69830799e-01 6.68226704e-02 -6.43383503e-01
-5.26664019e-01 -1.10179204e-02 -3.25920820e-01 -2.48808973e-02
7.01723933e-01 -1.07657468e+00 -2.75634587e-01 6.71653032e-01
-9.86331582e-01 -4.05221045e-01 -1.54554799e-01 3.79665643e-01
-6.82740867e-01 5.65894425e-01 -8.46659064e-01 -4.06214178e-01
-3.74218166e-01 -1.45314574e+00 6.35665715e-01 1.86740272e-02
-1.25671238e-01 -6.34132028e-01 -1.45943850e-01 1.64398089e-01
1.02041721e+00 -3.04542352e-02 9.47261333e-01 -1.07636310e-01
-7.00851977e-01 -2.04574451e-01 -4.97824520e-01 6.94877446e-01
-1.88000035e-02 -2.83160955e-01 -9.57443476e-01 -5.48261225e-01
2.92031080e-01 -4.41678047e-01 1.19072914e+00 5.35440683e-01
1.55542374e+00 -2.91986167e-01 -3.02150529e-02 9.30473089e-01
1.74930358e+00 1.77552830e-02 1.03761578e+00 4.94207621e-01
3.77271026e-01 -1.98630512e-01 3.95914279e-02 2.95571595e-01
1.12306699e-01 8.58265936e-01 6.77620471e-01 -1.62433550e-01
-5.88270187e-01 -2.28181452e-01 1.49375081e-01 5.66887438e-01
8.17342252e-02 -4.61161613e-01 -8.60059857e-01 4.17431116e-01
-1.52548349e+00 -9.59850907e-01 3.16994160e-01 2.29667425e+00
1.01147079e+00 5.18487692e-01 -2.00607732e-01 5.60171366e-01
2.36198843e-01 3.24157953e-01 -5.98373830e-01 -2.75206178e-01
-3.01337335e-03 2.71855474e-01 7.13466704e-01 3.29979300e-01
-1.00041938e+00 5.90963304e-01 7.48779583e+00 8.38532805e-01
-1.37189126e+00 4.48227003e-02 6.78712726e-01 -4.72440094e-01
-6.10235194e-03 -1.44614533e-01 -3.80547523e-01 6.82774484e-01
1.26604867e+00 7.88954422e-02 7.40440190e-01 6.68633103e-01
3.43540192e-01 -1.69010058e-01 -1.12737095e+00 1.36432636e+00
1.72046900e-01 -1.37865603e+00 1.99497357e-01 1.63482636e-01
5.80677390e-01 3.76156092e-01 3.38086188e-01 2.25194339e-02
-3.51860188e-02 -1.16779959e+00 7.28782773e-01 -1.46347203e-03
1.18319166e+00 -4.82742965e-01 5.69009781e-01 -9.26191360e-03
-8.28582466e-01 -2.19994828e-01 -5.83744526e-01 6.29673749e-02
3.47851306e-01 6.74060404e-01 -2.97758788e-01 -9.05741006e-03
8.83824527e-01 6.01285398e-01 -6.47918165e-01 1.42710137e+00
8.08306560e-02 7.85265744e-01 -3.95331621e-01 4.96772200e-01
9.16662738e-02 1.59701526e-01 1.88248813e-01 1.03720772e+00
6.49268806e-01 -2.07775727e-01 -5.35450131e-02 7.85447001e-01
-5.34635246e-01 -2.94673592e-01 -3.11675966e-01 -1.42627698e-03
4.82112378e-01 6.36602223e-01 -4.87708390e-01 -4.25575972e-01
-5.41761220e-01 1.13931680e+00 2.71683007e-01 2.75338650e-01
-6.85515165e-01 -2.82258958e-01 7.17860222e-01 4.27934110e-01
3.54047865e-01 -3.41299236e-01 -3.98830414e-01 -1.17575407e+00
2.43537575e-02 -1.09081185e+00 -2.00912841e-02 -6.87221825e-01
-8.13765645e-01 3.99009109e-01 -1.29268572e-01 -1.32678330e+00
-2.81215757e-01 -5.56984305e-01 -2.60014236e-01 6.51276827e-01
-1.77060163e+00 -7.24543929e-01 -4.06969905e-01 3.45346481e-01
6.76154613e-01 -7.98418224e-02 8.10231388e-01 6.30164862e-01
-2.78603435e-01 8.25371265e-01 3.77855331e-01 -3.08279619e-02
8.52680147e-01 -1.04499316e+00 5.95520794e-01 1.05403495e+00
-1.77617697e-03 4.43914562e-01 1.06528044e+00 -4.10101086e-01
-1.41759133e+00 -8.82870197e-01 6.93107843e-01 -1.47162423e-01
2.87624478e-01 -1.03922702e-01 -1.09992969e+00 4.17899281e-01
4.06343997e-01 5.60840741e-02 5.16608536e-01 -2.57654935e-01
-7.08269000e-01 -2.36381039e-01 -1.29744530e+00 5.90689421e-01
7.39071369e-01 -6.15567625e-01 1.63382273e-02 1.34292245e-01
7.07982838e-01 -5.10953546e-01 -9.20889795e-01 2.57504612e-01
7.06977844e-01 -1.35741198e+00 1.08903635e+00 -3.93181108e-02
1.06690586e+00 -1.40828222e-01 -4.05723065e-01 -1.30200887e+00
-4.77888077e-01 -3.89232367e-01 -4.52388108e-01 7.54022121e-01
2.01169521e-01 -1.22294068e-01 8.35817337e-01 6.85280979e-01
2.81758830e-02 -6.38904691e-01 -8.66421580e-01 -8.18455338e-01
1.53479859e-01 -5.61759591e-01 5.99389374e-01 4.85789657e-01
-1.08913891e-01 1.46360502e-01 -6.76462293e-01 -1.08042374e-01
6.93004310e-01 -3.00388575e-01 4.57151383e-01 -8.26696098e-01
-6.75796866e-01 -7.21781552e-01 -5.66941798e-01 -1.33437848e+00
-4.22469586e-01 -6.33397281e-01 1.59705251e-01 -1.49583983e+00
3.59228373e-01 -4.03793126e-01 -4.35585022e-01 2.76581258e-01
-1.36467561e-01 8.29888046e-01 5.28199553e-01 3.95136565e-01
-3.51780206e-01 4.20831084e-01 8.67502570e-01 -3.49387825e-01
1.27763283e-02 -4.54531640e-01 -7.33609080e-01 4.18304741e-01
8.46350729e-01 -4.79823560e-01 -4.40607250e-01 -1.03269386e+00
2.56161124e-01 -8.08658153e-02 3.20846736e-01 -1.66940904e+00
3.60812768e-02 1.81752384e-01 6.09582663e-01 -3.56562585e-01
5.18559098e-01 -9.21779454e-01 2.03021720e-01 6.39107108e-01
-5.94481587e-01 1.31510824e-01 1.63042873e-01 4.61861759e-01
-4.54399675e-01 -2.24669099e-01 1.29494250e+00 -1.38279155e-01
-3.55935037e-01 2.82777816e-01 -3.37470561e-01 -2.56211441e-02
2.97885001e-01 -2.80537367e-01 -1.54068083e-01 -7.56563067e-01
-2.54222304e-01 -1.80895239e-01 8.16017687e-01 3.83466870e-01
7.48472810e-01 -1.27679670e+00 -7.11089730e-01 3.52705181e-01
-2.02076614e-01 -1.41695186e-01 2.47229367e-01 4.89439756e-01
-7.76933730e-01 1.12362020e-01 -3.85681868e-01 -5.78591824e-01
-1.08937109e+00 5.57809830e-01 4.78545427e-01 -4.72677976e-01
-8.03753436e-01 6.83528721e-01 -2.73395795e-03 3.65927368e-01
3.26297283e-01 -2.56814897e-01 -4.17557471e-02 -5.34723401e-01
8.35009813e-01 2.26807401e-01 3.02780598e-01 -3.91860783e-01
1.09130919e-01 5.54562628e-01 -2.39073932e-01 -1.76504746e-01
1.21881807e+00 -1.22591697e-01 1.54396147e-01 -1.78641342e-02
1.35710192e+00 -3.05250734e-01 -1.37032080e+00 -1.12442719e-02
-3.99012238e-01 -8.78252864e-01 5.38758516e-01 -6.82151258e-01
-1.36115646e+00 1.23053336e+00 1.05872130e+00 1.42455578e-01
1.43240130e+00 -6.10953927e-01 1.05172694e+00 1.43444151e-01
2.30572909e-01 -9.81218636e-01 1.58784971e-01 3.80555183e-01
8.73205423e-01 -1.37929857e+00 1.52091518e-01 -4.20712940e-02
-1.89388812e-01 1.13420641e+00 3.51716846e-01 -2.14972168e-01
6.26898527e-01 4.70567614e-01 7.98995122e-02 9.47317854e-02
-5.89922249e-01 4.78167534e-01 -5.54321818e-02 4.82850105e-01
7.14222312e-01 -9.11709219e-02 -2.45708302e-01 2.08833978e-01
-3.23103040e-01 3.55030268e-01 7.10453510e-01 9.44425046e-01
-5.16905129e-01 -1.14155447e+00 -2.07548916e-01 8.37622702e-01
-8.41994345e-01 -3.93149614e-01 1.55811369e-01 3.81891310e-01
-3.99220698e-02 9.74369228e-01 2.12029919e-01 -3.81027877e-01
3.81462350e-02 -1.26716077e-01 7.07010806e-01 -2.44150743e-01
-5.63105047e-01 2.49194235e-01 -2.66280532e-01 -8.87821436e-01
-2.38997698e-01 -3.63278180e-01 -1.06263244e+00 -5.87212741e-01
-1.59724921e-01 -2.57625401e-01 7.39655554e-01 6.35941684e-01
3.54893446e-01 7.31491864e-01 2.67837167e-01 -1.03645551e+00
-6.30581737e-01 -9.88112748e-01 -5.88711500e-01 7.35232353e-01
5.85696340e-01 -3.90665531e-01 -3.15004736e-01 1.61135167e-01] | [11.386543273925781, -1.6547811031341553] |
46922323-69a4-4015-9aa4-9ea1b8e4a601 | towards-open-temporal-graph-neural-networks | 2303.15015 | null | https://arxiv.org/abs/2303.15015v2 | https://arxiv.org/pdf/2303.15015v2.pdf | Towards Open Temporal Graph Neural Networks | Graph neural networks (GNNs) for temporal graphs have recently attracted increasing attentions, where a common assumption is that the class set for nodes is closed. However, in real-world scenarios, it often faces the open set problem with the dynamically increased class set as the time passes by. This will bring two big challenges to the existing dynamic GNN methods: (i) How to dynamically propagate appropriate information in an open temporal graph, where new class nodes are often linked to old class nodes. This case will lead to a sharp contradiction. This is because typical GNNs are prone to make the embeddings of connected nodes become similar, while we expect the embeddings of these two interactive nodes to be distinguishable since they belong to different classes. (ii) How to avoid catastrophic knowledge forgetting over old classes when learning new classes occurred in temporal graphs. In this paper, we propose a general and principled learning approach for open temporal graphs, called OTGNet, with the goal of addressing the above two challenges. We assume the knowledge of a node can be disentangled into class-relevant and class-agnostic one, and thus explore a new message passing mechanism by extending the information bottleneck principle to only propagate class-agnostic knowledge between nodes of different classes, avoiding aggregating conflictive information. Moreover, we devise a strategy to select both important and diverse triad sub-graph structures for effective class-incremental learning. Extensive experiments on three real-world datasets of different domains demonstrate the superiority of our method, compared to the baselines. | ['Jun Zhou', 'Xiaolu Zhang', 'Changsheng Li', 'Kaituo Feng'] | 2023-03-27 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [ 2.44189397e-01 3.06551278e-01 -2.98543662e-01 5.12034446e-02
-1.24228569e-02 -8.23295474e-01 4.21084493e-01 2.84191310e-01
-1.48918748e-01 7.47179329e-01 -9.79947001e-02 -2.82998025e-01
-6.06875062e-01 -1.34751725e+00 -6.66550219e-01 -9.39019024e-01
-6.35841608e-01 6.44757032e-01 6.70423865e-01 -3.98241282e-01
-1.77653104e-01 2.70864487e-01 -1.40518773e+00 -1.26746073e-01
8.74937713e-01 6.47201419e-01 -1.11796938e-01 4.35778111e-01
-2.01279745e-01 6.90498710e-01 -4.06628489e-01 -6.46686792e-01
4.93953437e-01 -3.43904793e-01 -1.03039622e+00 4.01361063e-02
2.62507796e-01 2.14658082e-02 -6.66807532e-01 1.22592247e+00
2.82123923e-01 1.64305478e-01 4.74248618e-01 -1.58448792e+00
-6.93817973e-01 1.20044458e+00 -7.11695969e-01 3.88687611e-01
-2.86081228e-02 1.67587176e-01 1.22207725e+00 -2.97169179e-01
7.33350277e-01 1.24371314e+00 5.25520742e-01 5.76275945e-01
-1.32534444e+00 -8.19473624e-01 8.60960364e-01 5.11000037e-01
-1.43459427e+00 -5.10564037e-02 1.01837134e+00 -3.54708940e-01
5.15701652e-01 2.24845096e-01 9.33804810e-01 1.27371442e+00
1.80723697e-01 5.63240886e-01 8.27882528e-01 -2.25469675e-02
3.22527617e-01 3.19174752e-02 4.16586220e-01 7.28988469e-01
6.52658165e-01 1.57248676e-01 -5.21234751e-01 -6.78861514e-02
5.15956938e-01 1.98870718e-01 -3.85261506e-01 -8.98393154e-01
-1.02167773e+00 9.10466909e-01 7.38450527e-01 4.15730417e-01
-2.95114089e-02 4.21649544e-03 4.03961837e-01 7.86533535e-01
4.14264143e-01 1.62592575e-01 -7.29624152e-01 3.20185214e-01
-2.46239513e-01 -1.61647558e-01 1.08911836e+00 8.84004533e-01
1.00164366e+00 -1.77719012e-01 7.60902613e-02 5.83434284e-01
2.53138810e-01 2.06862956e-01 3.30268234e-01 -3.74855697e-01
5.47953844e-01 9.97226238e-01 -4.72191334e-01 -1.39773858e+00
-4.12850827e-01 -7.08891809e-01 -1.27306318e+00 -9.71110165e-02
4.54467684e-01 -1.35404304e-01 -1.08883715e+00 2.15616059e+00
6.11203969e-01 4.00373250e-01 4.57987422e-03 4.57721978e-01
7.40154922e-01 5.09672463e-01 -1.94414169e-01 -4.50591266e-01
1.12038958e+00 -6.89598083e-01 -5.30959070e-01 -2.97964513e-01
5.61163962e-01 -5.36825359e-02 7.20636308e-01 1.32483631e-01
-5.68586946e-01 -3.38468045e-01 -1.19565201e+00 2.42807761e-01
-6.31392062e-01 -9.27781284e-01 8.08316052e-01 5.97645581e-01
-1.08189952e+00 8.11499715e-01 -7.15743482e-01 -4.06422615e-01
4.89610493e-01 5.25282860e-01 -3.20965528e-01 -2.12790612e-02
-1.60401249e+00 4.72461432e-01 8.82805347e-01 1.90913349e-01
-9.89177525e-01 -6.14463866e-01 -8.09838474e-01 1.68311551e-01
1.04808438e+00 -7.47251451e-01 8.43129933e-01 -1.09139884e+00
-1.09407127e+00 4.50557142e-01 2.93302894e-01 -2.71379977e-01
4.33584720e-01 3.19253653e-01 -6.36831105e-01 4.38283309e-02
4.00050133e-02 2.94266641e-01 1.01604140e+00 -1.25202262e+00
-5.79965591e-01 -4.92270470e-01 4.14164066e-01 1.58397958e-01
-7.27045119e-01 -6.85336709e-01 -3.66344422e-01 -5.13367474e-01
3.44373852e-01 -1.00204372e+00 -2.50233948e-01 -1.73818227e-02
-4.33000118e-01 -3.55043650e-01 9.44299996e-01 -6.98394030e-02
1.45432472e+00 -1.97506666e+00 4.54140514e-01 4.61442173e-01
1.00246477e+00 1.96563840e-01 -2.97367036e-01 5.18543899e-01
-1.61019236e-01 2.79074013e-01 -2.65981883e-01 1.22827500e-01
-3.10417444e-01 6.07676148e-01 -2.20905706e-01 2.67207086e-01
-5.10312878e-02 7.54148841e-01 -1.44528949e+00 -4.54075962e-01
-2.29388222e-01 2.32828200e-01 -3.96004498e-01 -2.38251030e-01
-2.95130014e-01 2.87883461e-01 -5.10454059e-01 4.16023284e-01
7.61164367e-01 -6.06106162e-01 5.92240095e-01 -3.51635106e-02
4.09013450e-01 7.16687068e-02 -1.44618130e+00 1.22567582e+00
6.72133081e-03 2.76740760e-01 -1.32918447e-01 -1.04222834e+00
7.67616093e-01 1.08028188e-01 3.54853541e-01 -4.58462089e-01
5.35298325e-02 -3.55551876e-02 3.83666694e-01 -3.31869185e-01
2.09930614e-01 -7.05243275e-02 -2.99647381e-03 4.48283017e-01
1.56651556e-01 2.21898094e-01 4.17140275e-01 6.51634157e-01
1.31008565e+00 -3.34741622e-01 1.02689847e-01 -2.75444090e-01
3.80314142e-01 -3.54065627e-01 9.09840763e-01 9.07967985e-01
-3.85134190e-01 1.66160047e-01 9.98306930e-01 -5.92667341e-01
-5.09622812e-01 -1.19352579e+00 3.15664440e-01 1.05299008e+00
4.94183421e-01 -4.72332239e-01 -2.85571486e-01 -1.27399254e+00
-5.21092378e-02 2.16270566e-01 -8.78026664e-01 -6.50980055e-01
-5.60258150e-01 -7.50346780e-01 1.39839277e-01 3.41817558e-01
3.40706915e-01 -8.54843199e-01 -8.27541277e-02 1.55154333e-01
-1.77439317e-01 -7.89290249e-01 -5.86407900e-01 3.42482477e-01
-9.31306779e-01 -1.46710849e+00 -3.93310934e-01 -7.81283379e-01
8.02026391e-01 4.27071184e-01 1.20433688e+00 2.56272107e-01
9.19872820e-02 4.14096951e-01 -4.79156882e-01 -1.67175829e-01
-4.22713101e-01 4.76475447e-01 8.00988898e-02 4.14566576e-01
2.16431051e-01 -1.04204047e+00 -6.15533292e-01 3.60491335e-01
-1.03857279e+00 -7.23712742e-02 4.71317261e-01 7.30849802e-01
2.57543564e-01 5.37346184e-01 7.24248469e-01 -1.19924438e+00
6.00065053e-01 -7.36308098e-01 -6.65010393e-01 4.68042701e-01
-9.19633210e-01 2.52315998e-01 8.13349307e-01 -8.68455291e-01
-7.37935126e-01 -2.52009779e-01 4.83177751e-01 -3.72949451e-01
4.03188974e-01 7.71844506e-01 -1.93367869e-01 -1.51362643e-01
6.42996132e-01 1.63379475e-01 -6.88026771e-02 -6.32485747e-02
4.58331645e-01 4.24636230e-02 1.77307144e-01 -4.91499126e-01
1.27885997e+00 5.23717403e-01 1.89349741e-01 -6.04506135e-01
-8.46181393e-01 -2.55201191e-01 -7.85767376e-01 -2.77215719e-01
4.21842426e-01 -6.48845732e-01 -6.93855286e-01 5.27181327e-01
-9.03485179e-01 -2.85401970e-01 -4.78195935e-01 6.24099188e-02
-3.41687277e-02 5.95940411e-01 -5.19362032e-01 -3.94506752e-01
-6.95085526e-02 -8.33323658e-01 4.38596368e-01 4.14750278e-01
6.96153268e-02 -1.27564454e+00 2.25629598e-01 -7.86793306e-02
2.02254012e-01 2.58633614e-01 1.15125060e+00 -9.44505334e-01
-8.32722068e-01 -4.67060842e-02 -1.20039351e-01 -2.20650272e-03
2.94533432e-01 -5.14607579e-02 -6.32939458e-01 -6.95040107e-01
-3.19458812e-01 -2.45675847e-01 1.08602273e+00 -5.36685325e-02
8.64613533e-01 -5.57542562e-01 -6.62582517e-01 4.24553573e-01
1.30189514e+00 2.78983340e-02 2.74595976e-01 -2.39770431e-02
9.67074096e-01 5.54321527e-01 1.51608914e-01 2.66965479e-01
5.87583721e-01 3.03136736e-01 5.29513359e-01 2.49078393e-01
-1.47509262e-01 -4.47165638e-01 2.82912374e-01 1.20409358e+00
9.24684033e-02 -6.16749883e-01 -8.36649895e-01 6.27952754e-01
-1.99572122e+00 -7.89995193e-01 -1.09103778e-02 2.27460194e+00
8.94312143e-01 4.01283592e-01 6.98265135e-02 1.53659850e-01
8.48904431e-01 4.65950519e-01 -6.99666798e-01 6.12980835e-02
-5.49558476e-02 -1.02477044e-01 2.97737628e-01 3.86725754e-01
-9.23301876e-01 7.43081272e-01 5.51226044e+00 7.62017846e-01
-9.87160265e-01 1.30580589e-01 4.82010245e-01 1.37887880e-01
-5.18515825e-01 4.42893714e-01 -6.80447280e-01 3.57054204e-01
7.59912550e-01 -4.87709016e-01 5.42713344e-01 4.54786092e-01
-3.92206460e-01 1.26763001e-01 -1.29452074e+00 6.67577326e-01
-4.88721728e-02 -1.10262215e+00 3.30843478e-01 2.27905251e-02
8.36700022e-01 6.00364432e-02 -1.27694622e-01 6.18536353e-01
7.68271863e-01 -5.83493829e-01 3.69243383e-01 3.28912556e-01
5.49820006e-01 -7.06885815e-01 3.72882128e-01 4.09039646e-01
-1.53266490e+00 -3.09130937e-01 -1.89695701e-01 2.36219112e-02
-1.86667025e-01 6.15601301e-01 -6.38147652e-01 8.66609752e-01
7.66798377e-01 9.14530277e-01 -7.91115940e-01 9.30527091e-01
-3.40388596e-01 4.99293864e-01 -2.36082613e-01 -1.08643562e-01
5.74335866e-02 -2.05202788e-01 7.22667217e-01 5.73006272e-01
-1.37848079e-01 3.74395475e-02 3.05360883e-01 7.05650330e-01
-2.85909861e-01 -2.24954888e-01 -8.43902111e-01 -1.98785007e-01
6.05325401e-01 1.23305488e+00 -1.17903507e+00 -2.77326375e-01
-4.23987687e-01 8.39901030e-01 6.60595596e-01 5.37045121e-01
-7.42981732e-01 -5.29544592e-01 3.65913063e-01 1.61949635e-01
4.06110764e-01 -7.58164674e-02 2.60062009e-01 -1.31066656e+00
4.20292728e-02 -7.07878470e-01 1.13066649e+00 -1.92038476e-01
-1.62200475e+00 5.15341997e-01 6.14761598e-02 -1.16739643e+00
2.23171130e-01 -3.57809454e-01 -5.12936831e-01 2.48372212e-01
-1.40089941e+00 -1.00543487e+00 -2.17219993e-01 7.98124313e-01
2.86871105e-01 1.10137992e-01 4.86613184e-01 3.77959728e-01
-5.77008724e-01 7.04219043e-01 7.01112077e-02 3.57565492e-01
4.60293978e-01 -1.19390166e+00 2.62007236e-01 1.04937685e+00
2.95864671e-01 5.67150235e-01 4.99639332e-01 -8.69031787e-01
-1.43961096e+00 -1.26320660e+00 7.38252878e-01 -4.88495469e-01
1.05896318e+00 -7.07704365e-01 -1.19688272e+00 7.99172819e-01
-9.26666111e-02 3.23933214e-01 3.82896721e-01 5.40262163e-01
-7.19708502e-01 -4.22185779e-01 -7.81899214e-01 7.64190674e-01
1.53173733e+00 -4.46071118e-01 -3.57617438e-01 4.08085763e-01
1.17462635e+00 -1.01356559e-01 -7.30782866e-01 6.38436139e-01
2.60029346e-01 -7.37736344e-01 8.30420196e-01 -7.67925322e-01
7.12981299e-02 -3.63444865e-01 3.62797976e-01 -1.53309298e+00
-5.55930138e-01 -9.35630083e-01 -6.13161802e-01 1.20179200e+00
4.00188982e-01 -1.09511983e+00 7.18793094e-01 1.23925447e-01
2.21929073e-01 -6.07976675e-01 -1.08054006e+00 -1.02132308e+00
1.80382133e-02 -7.05645010e-02 5.61419487e-01 1.26086569e+00
-1.67929754e-02 7.87206709e-01 -3.37824643e-01 3.35976779e-01
6.44004583e-01 2.27767840e-01 5.14708877e-01 -1.78926873e+00
-3.88755172e-01 -5.84276915e-01 -4.80259061e-01 -8.00899565e-01
4.15449627e-02 -1.14517665e+00 -1.69399574e-01 -1.35940528e+00
4.81488585e-01 -6.63319349e-01 -6.12101197e-01 5.98300040e-01
-2.81435251e-01 -2.69476265e-01 -3.18255834e-03 5.35297468e-02
-9.06836092e-01 7.15866983e-01 1.43221140e+00 -4.85315055e-01
-4.05499518e-01 1.08957790e-01 -9.15546179e-01 5.34467697e-01
5.66389382e-01 -7.13992476e-01 -1.00302982e+00 -1.93004012e-01
7.73980916e-01 1.62987541e-02 3.40584844e-01 -6.77154422e-01
4.68197018e-01 -2.04800099e-01 -6.06917478e-02 -5.03667116e-01
-4.94894460e-02 -1.03384483e+00 2.76482433e-01 6.64557695e-01
-1.76668555e-01 -9.62060764e-02 1.35929687e-02 1.30739605e+00
1.29619420e-01 1.45507902e-01 6.07381761e-01 -1.60836414e-01
-8.55012476e-01 9.79798913e-01 -2.32757963e-02 4.47229624e-01
1.12772393e+00 -3.34222883e-01 -5.52087128e-01 -4.12263721e-01
-8.57451141e-01 6.91639185e-01 1.91630989e-01 8.03085566e-01
4.82925624e-01 -1.23756301e+00 -5.86003006e-01 2.34363869e-01
3.77180129e-01 2.36971378e-01 3.76546472e-01 8.23168397e-01
2.20494922e-02 -1.55419752e-01 9.13876817e-02 -6.39346480e-01
-1.08875275e+00 1.03556490e+00 4.19208437e-01 -5.29066205e-01
-6.01606429e-01 8.29042375e-01 3.76973927e-01 -6.57096267e-01
3.35958153e-01 -5.39986454e-02 -3.52805465e-01 3.20840925e-01
1.85453087e-01 2.78881758e-01 -6.86086863e-02 -3.06562334e-01
-2.76176006e-01 4.13164496e-01 -4.73957568e-01 3.64869535e-01
1.32790220e+00 -3.20524812e-01 -3.18895549e-01 6.29047096e-01
1.02930319e+00 -2.13438660e-01 -1.00103390e+00 -6.83176756e-01
7.57273585e-02 -2.88581222e-01 -3.24279398e-01 -5.00298619e-01
-1.35285878e+00 5.57539940e-01 4.97411460e-01 7.84281254e-01
1.16395462e+00 2.95594692e-01 6.00145936e-01 4.91269916e-01
5.21990657e-01 -9.62044060e-01 2.74446189e-01 6.34962142e-01
5.33222854e-01 -1.12887979e+00 -1.64098721e-02 -6.38782144e-01
-1.69912890e-01 1.08482409e+00 9.04183447e-01 1.74491987e-01
9.66223598e-01 -1.43336102e-01 -3.02921891e-01 -5.21860301e-01
-9.78737175e-01 -1.75319582e-01 1.02883734e-01 6.48910463e-01
-1.68246046e-01 1.40495151e-01 -1.28201798e-01 2.78159052e-01
1.06952801e-01 -3.37439537e-01 7.08564103e-01 8.37426066e-01
-2.19052330e-01 -1.14093816e+00 1.32906735e-02 5.79585016e-01
-1.61363259e-02 -3.58179882e-02 -4.93430972e-01 8.85241032e-01
1.14494018e-01 7.09342539e-01 -1.93917319e-01 -5.72267532e-01
5.57864010e-02 -1.60714731e-01 4.73527879e-01 -6.26214027e-01
-2.77505159e-01 -2.74910599e-01 -8.66706669e-02 -3.24353814e-01
-3.20183694e-01 -4.47846532e-01 -1.01403248e+00 -3.66628021e-01
-6.19322538e-01 1.38883457e-01 -1.57881081e-01 7.96502709e-01
4.34243202e-01 7.41365314e-01 8.10285270e-01 -2.82285929e-01
-6.43680334e-01 -5.24874628e-01 -6.70428634e-01 3.90984863e-01
5.55704415e-01 -7.99391210e-01 -6.51590168e-01 -3.14740688e-01] | [7.298735618591309, 6.080990791320801] |
35863e8e-c7a1-4991-81bf-eedb181e82c6 | a-health-telemonitoring-platform-based-on | 2207.13913 | null | https://arxiv.org/abs/2207.13913v2 | https://arxiv.org/pdf/2207.13913v2.pdf | A health telemonitoring platform based on data integration from different sources | The management of people with long-term or chronic illness is one of the biggest challenges for national health systems. In fact, these diseases are among the leading causes of hospitalization, especially for the elderly, and huge amount of resources required to monitor them leads to problems with sustainability of the healthcare systems. The increasing diffusion of portable devices and new connectivity technologies allows the implementation of telemonitoring system capable of providing support to health care providers and lighten the burden on hospitals and clinics. In this paper, we present the implementation of a telemonitoring platform for healthcare, designed to capture several types of physiological health parameters from different consumer mobile and custom devices. Consumer medical devices can be integrated into the platform via the Google Fit ecosystem that supports hundreds of devices, while custom devices can directly interact with the platform with standard communication protocols. The platform is designed to process the acquired data using machine learning algorithms, and to provide patients and physicians the physiological health parameters with a user-friendly, comprehensive, and easy to understand dashboard which monitors the parameters through time. Preliminary usability tests show a good user satisfaction in terms of functionality and usefulness. | ['Raimondo Schettini', 'Matteo Romanato', 'Paolo Napoletano', 'Gianluigi Ciocca'] | 2022-07-28 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-4.12536830e-01 -2.42613316e-01 -2.35183701e-01 -2.42347613e-01
-2.53899217e-01 -2.69242913e-01 -4.72281635e-01 6.15120351e-01
-3.06295216e-01 7.11992085e-01 1.49536848e-01 -6.03266299e-01
-1.80880815e-01 -8.01113367e-01 1.04235798e-01 -4.23515856e-01
-1.75519153e-01 4.54837650e-01 -2.04924587e-02 -2.16725186e-01
-2.46899873e-01 1.53436810e-01 -1.60278261e+00 3.63028169e-01
9.02729094e-01 1.31433189e+00 6.96536079e-02 5.16255498e-01
3.02342296e-01 5.83716780e-02 -5.79988480e-01 2.80964792e-01
-7.28015304e-02 -2.27103308e-01 -2.40232021e-01 -3.77779961e-01
-3.65867436e-01 -6.53154969e-01 2.07510158e-01 6.41124845e-01
1.16228163e+00 -4.58476126e-01 -4.04781587e-02 -1.08066571e+00
-6.96503818e-02 3.70966941e-02 1.81054503e-01 1.03766412e-01
8.67426693e-01 2.82845706e-01 7.42489249e-02 -2.60772556e-01
7.51410872e-02 7.46472776e-01 1.11506319e+00 4.22408283e-01
-9.28487897e-01 -3.81523550e-01 -4.11855161e-01 1.39167085e-01
-1.27841341e+00 -5.85987329e-01 4.67590094e-02 -4.54503596e-01
1.10614288e+00 6.87509418e-01 1.29571688e+00 1.03151989e+00
7.14801013e-01 -2.15476975e-01 9.00854528e-01 -2.30741769e-01
4.51216638e-01 7.06794798e-01 4.17063236e-01 2.50819951e-01
8.56096625e-01 -2.61295974e-01 -1.39143616e-01 -6.73191905e-01
6.27667010e-01 7.08916247e-01 -6.09906852e-01 -2.94691771e-02
-1.21112585e+00 2.03395694e-01 -6.51748152e-03 5.40946901e-01
-8.40015233e-01 -2.58605778e-01 4.50372815e-01 3.03221136e-01
5.94086833e-02 1.69893786e-01 -9.68076408e-01 -6.11275136e-01
-2.74117470e-01 -1.47564903e-01 1.02585578e+00 8.77689779e-01
5.46626672e-02 -5.26326418e-01 -1.81896567e-01 4.97998148e-01
5.57571292e-01 8.74858499e-01 9.39284265e-01 -6.40181482e-01
-2.55048946e-02 1.02423322e+00 4.69202220e-01 -7.20391631e-01
-1.07579350e+00 -3.26613426e-01 -9.58388269e-01 -1.40722439e-01
2.26794910e-02 -6.68231070e-01 -3.49842370e-01 1.09388471e+00
5.19415975e-01 -2.42992282e-01 -8.09365138e-02 6.11213803e-01
8.12172949e-01 2.70505726e-01 5.07787228e-01 -6.40629530e-01
1.63815653e+00 -1.24052845e-01 -9.69509363e-01 1.59081638e-01
7.26991296e-01 -2.91520953e-01 1.32271481e+00 4.00312334e-01
-1.06400990e+00 -4.60085332e-01 -8.78822625e-01 2.09557593e-01
-5.29140115e-01 5.68271205e-02 4.57217157e-01 1.10633492e+00
-9.99521136e-01 6.62107348e-01 -1.10645187e+00 -1.01289189e+00
2.24325851e-01 5.24547398e-01 -1.58089474e-01 2.17099160e-01
-1.07377493e+00 8.62643957e-01 1.97199345e-01 -2.15259925e-01
-4.72170673e-03 -8.82827401e-01 -5.20128489e-01 8.79525468e-02
-4.91410702e-01 -1.20883763e+00 1.00161386e+00 -4.45978522e-01
-1.50434697e+00 5.03523231e-01 1.19103573e-01 6.80124611e-02
4.76724327e-01 -4.29038405e-01 -7.27693796e-01 1.86240911e-01
1.76197633e-01 -4.79412777e-03 2.32773032e-02 -2.34281227e-01
-4.88802165e-01 -6.92230999e-01 -4.42658365e-01 6.86139464e-02
-6.53914392e-01 -3.45825516e-02 -2.24098176e-01 8.15778598e-02
9.69199389e-02 -7.63045073e-01 -1.25466555e-01 4.35927175e-02
-1.21636063e-01 3.77956510e-01 7.04859853e-01 -7.67626703e-01
1.40688479e+00 -1.99986184e+00 -4.92454261e-01 2.99414128e-01
6.75279275e-02 2.33513713e-01 6.18778229e-01 4.25007701e-01
5.16791316e-03 1.29804179e-01 2.30311856e-01 1.95123747e-01
-1.58766061e-01 1.56629294e-01 6.19725943e-01 3.90359432e-01
-6.11789882e-01 7.62769878e-01 -5.59455216e-01 -1.62048578e-01
4.65753943e-01 8.89930308e-01 -2.86896437e-01 2.52322644e-01
1.36756405e-01 9.68999743e-01 -5.68576992e-01 8.37709963e-01
3.17869872e-01 -5.61593592e-01 4.90824193e-01 -1.51308537e-01
-1.90410122e-01 3.22489798e-01 -9.42023814e-01 1.45018029e+00
-3.59629273e-01 -2.36605868e-01 4.47464846e-02 -6.97657168e-01
5.81102610e-01 9.38885033e-01 1.00533152e+00 -7.62299538e-01
5.35908639e-01 1.90345600e-01 -3.29361916e-01 -1.44743836e+00
-3.48255783e-01 1.25806823e-01 -1.59071088e-02 3.63954306e-01
-4.86870468e-01 6.39890671e-01 -3.05465311e-01 -3.13051343e-01
1.08780813e+00 -1.52045712e-01 8.57172370e-01 -2.62184769e-01
3.29078674e-01 -1.90359160e-01 2.83250630e-01 2.38958076e-01
-2.85222858e-01 3.09089702e-02 -1.68776914e-01 -6.08703315e-01
-5.17239094e-01 -7.33689368e-01 -2.68862396e-01 7.54691660e-01
-1.58833966e-01 -4.85180914e-01 -5.73776305e-01 -3.28430049e-02
1.17456339e-01 9.26809087e-02 -1.02211699e-01 -5.78025095e-02
6.71027750e-02 -7.97195315e-01 2.48561829e-01 4.62276489e-01
7.12530792e-01 -9.07945216e-01 -1.41586125e+00 4.88713205e-01
-4.70920146e-01 -7.50401378e-01 -1.39686018e-01 -8.28415379e-02
-1.41977406e+00 -9.89541888e-01 -5.67818761e-01 -6.76942289e-01
3.33519757e-01 1.64669886e-01 7.32462764e-01 3.73342574e-01
-5.93598187e-01 7.04777479e-01 -1.97303355e-01 -6.98287606e-01
1.25567149e-03 1.43862322e-01 2.03009188e-01 -2.91567713e-01
5.05877495e-01 -8.83665979e-01 -1.08716607e+00 2.89217889e-01
-3.28921825e-01 -2.83356130e-01 2.34318584e-01 2.31514320e-01
5.61857164e-01 -6.57203421e-02 7.87968814e-01 -8.07719111e-01
8.43511760e-01 -8.08145821e-01 -2.45598421e-01 8.49087983e-02
-1.07872415e+00 -4.27801847e-01 2.69915879e-01 -2.46667489e-01
-4.50856268e-01 1.21374175e-01 -1.42613396e-01 3.17381829e-01
-5.01005828e-01 5.99971831e-01 -4.95858133e-01 6.93172961e-02
6.95998967e-01 -1.40477896e-01 1.63461149e-01 -6.04243279e-01
-2.02575132e-01 1.32953060e+00 4.30223346e-01 7.88638890e-02
-1.74902171e-01 1.96619719e-01 -1.72409251e-01 -8.19817483e-01
-1.84340432e-01 -7.17455506e-01 -3.18505317e-01 -3.85244727e-01
8.56976509e-01 -9.91621912e-01 -1.28676760e+00 3.56789768e-01
-7.14651883e-01 -1.12578750e-01 6.42015338e-02 6.34429514e-01
-1.28889591e-01 -9.12854224e-02 -4.76277679e-01 -7.38310099e-01
-9.59373295e-01 -7.50890076e-01 6.61905885e-01 2.89693594e-01
-7.64398396e-01 -9.39771712e-01 2.37035546e-02 1.75047070e-01
1.05648184e+00 6.78460896e-01 8.72449279e-01 -2.15380833e-01
-4.74321954e-02 -7.52554774e-01 1.93455443e-01 -5.21512069e-02
6.74973428e-01 -4.44574058e-01 -7.90737092e-01 -3.86136860e-01
4.43303734e-01 1.57668695e-01 -2.19332799e-01 7.17928052e-01
9.17717159e-01 -3.67477387e-01 -7.80967832e-01 4.69892085e-01
1.29013026e+00 3.96284252e-01 9.82384384e-01 4.03571784e-01
2.12060630e-01 2.48802036e-01 3.05363536e-01 7.22195804e-01
6.13330662e-01 4.65578437e-01 2.61897057e-01 -2.32061103e-01
5.12027979e-01 2.70184636e-01 2.85069823e-01 7.25777090e-01
-2.86613941e-01 9.76006016e-02 -7.75119781e-01 -3.06408387e-02
-1.85636544e+00 -7.52203286e-01 -4.12547648e-01 2.61191845e+00
6.55074060e-01 -3.01221430e-01 4.02734786e-01 4.11566973e-01
6.78474009e-01 -9.70411777e-01 -6.21152282e-01 -1.45649210e-01
5.65525830e-01 1.84698001e-01 4.36709136e-01 7.71148726e-02
-7.34092295e-01 -3.21097188e-02 6.92561960e+00 -5.29646039e-01
-1.42399216e+00 4.76151854e-01 5.07423341e-01 -2.11283922e-01
4.30270761e-01 -6.36308491e-01 -2.95651793e-01 7.39330292e-01
1.63811421e+00 9.95576680e-02 4.82408404e-02 8.64558458e-01
9.65555489e-01 -3.01993489e-01 -9.25384641e-01 1.24735200e+00
-4.28417951e-01 -9.44575489e-01 -4.90626752e-01 1.96527131e-02
2.24651337e-01 -4.58887815e-02 -2.59188026e-01 -9.26387608e-02
-6.67396605e-01 -6.13775849e-01 1.20845184e-01 8.83219898e-01
9.80928719e-01 -4.11194175e-01 1.03802502e+00 2.78603673e-01
-9.18577790e-01 -2.15078965e-01 8.11120123e-02 -4.87824380e-01
2.97505379e-01 6.16135120e-01 -5.25667548e-01 1.50242478e-01
1.24897110e+00 3.60113353e-01 -3.13776165e-01 1.39240980e+00
2.71547675e-01 3.21041256e-01 -3.57950807e-01 -7.11429641e-02
-8.22383583e-01 -6.89801127e-02 -7.78748691e-02 9.17950571e-01
7.36555696e-01 3.89703453e-01 1.72605574e-01 1.63651049e-01
4.61259305e-01 4.60087121e-01 -4.16768074e-01 1.36921838e-01
4.45761085e-01 1.20246816e+00 -6.20390058e-01 -5.68141282e-01
-4.01455820e-01 9.03610408e-01 -5.21923006e-01 -1.26772095e-02
-7.25272298e-01 -3.56155872e-01 6.12946451e-01 7.91226804e-01
-3.11540484e-01 -4.72107390e-03 -6.15737319e-01 -7.61290431e-01
1.44817099e-01 -1.04036605e+00 4.10983384e-01 -6.31926000e-01
-9.62918520e-01 3.52857709e-01 -3.22627723e-01 -1.39502800e+00
1.45959267e-02 -3.48122418e-01 -3.92787784e-01 8.97758603e-01
-1.09152389e+00 -4.86549556e-01 -1.00910175e+00 1.06852329e+00
-1.36430085e-01 1.86874300e-01 1.71381795e+00 8.60488713e-01
-5.28702080e-01 3.05045784e-01 -2.90923305e-02 -5.37023902e-01
7.19710827e-01 -7.79112816e-01 -1.86332896e-01 -2.59084180e-02
-1.01079714e+00 9.04738367e-01 4.12128001e-01 -8.09916198e-01
-1.48271775e+00 -9.70504403e-01 9.83861566e-01 -3.51864010e-01
-8.41515362e-02 -6.95860460e-02 -6.24657452e-01 3.41805160e-01
3.01281810e-02 -6.12598881e-02 1.26382160e+00 2.82530747e-02
3.22935581e-01 -3.65783483e-01 -1.57730734e+00 2.06546023e-01
5.89254022e-01 -2.54415721e-01 -8.29235613e-02 4.73838121e-01
3.22929204e-01 -3.09742242e-01 -1.20042086e+00 2.52666682e-01
1.01238990e+00 -1.00597787e+00 7.34950066e-01 -3.87089223e-01
-3.15238506e-01 -1.29445955e-01 2.16649249e-01 -1.01557839e+00
-2.19717547e-01 -7.63069987e-01 -8.90541729e-03 8.55650902e-01
4.60326612e-01 -1.29936326e+00 3.06696117e-01 1.25191903e+00
-1.54427032e-03 -5.54357946e-01 -6.18269444e-01 -3.65853012e-01
-7.49489009e-01 -2.48733625e-01 7.54719257e-01 6.36703134e-01
1.08927059e+00 3.26990455e-01 -2.11338758e-01 3.10317844e-01
1.10405549e-01 -3.47388774e-01 5.42734027e-01 -1.70674956e+00
-3.86979461e-01 1.91645205e-01 -5.29238641e-01 -4.29565102e-01
-1.12183094e+00 -4.18231428e-01 -2.97624290e-01 -1.98270869e+00
2.32313797e-02 -4.74714965e-01 -3.15405279e-01 6.29010856e-01
3.10689006e-02 2.23387316e-01 -5.41104376e-01 3.25940669e-01
-4.02494699e-01 -1.87476024e-01 8.01484227e-01 1.54404163e-01
-8.01800370e-01 3.67773563e-01 -8.44823718e-01 5.49151361e-01
1.19300067e+00 -3.84233892e-01 -4.87178981e-01 -2.31581002e-01
3.00395131e-01 2.72182971e-01 2.91700900e-01 -1.49473488e+00
1.40877306e-01 2.40863293e-01 7.62054980e-01 -4.72366177e-02
8.30756724e-02 -1.32640254e+00 8.75832438e-01 1.02614462e+00
3.09264958e-01 4.09497797e-01 2.17221275e-01 1.60121113e-01
3.47129077e-01 4.97839123e-01 5.64187825e-01 -1.03800751e-01
1.69822320e-01 8.78161788e-02 -7.40679383e-01 -4.37331736e-01
1.18174982e+00 -2.34232262e-01 -4.10247177e-01 -3.48738998e-01
-1.06609654e+00 1.47176296e-01 3.50830108e-01 2.27625251e-01
3.06034029e-01 -1.09748960e+00 4.50008214e-02 4.90753233e-01
1.36650801e-01 -3.64444941e-01 3.68691117e-01 1.34617949e+00
-7.02958822e-01 5.27984858e-01 -5.60251534e-01 -5.61557949e-01
-1.40003455e+00 5.81471920e-01 5.12485325e-01 3.56303841e-01
-9.96833563e-01 -3.90473083e-02 -6.50857568e-01 7.08329231e-02
5.75401962e-01 -8.63451362e-01 -3.40494365e-01 -4.25263420e-02
1.04170609e+00 7.35523164e-01 4.32070583e-01 -1.15452297e-01
-5.90728939e-01 4.61625338e-01 4.57692266e-01 1.08438872e-01
1.32765162e+00 -5.10752141e-01 -1.52187929e-01 5.27524412e-01
7.80184925e-01 -2.62084961e-01 -4.01825100e-01 3.97149384e-01
-2.15119839e-01 -9.14200209e-03 -5.91836236e-02 -1.15242434e+00
-8.59449446e-01 4.92835790e-01 1.56211233e+00 4.79642719e-01
1.29236472e+00 -2.34417006e-01 8.82669091e-01 3.07810158e-01
6.25777841e-01 -9.05597746e-01 -3.86962652e-01 -2.23211899e-01
4.22278434e-01 -1.08461165e+00 -1.56083435e-01 -4.22604501e-01
-3.79156768e-01 1.08359206e+00 9.08481330e-02 3.99691403e-01
1.19031894e+00 2.72232771e-01 5.07228315e-01 -2.88827866e-01
-4.18262750e-01 1.48783088e-01 -3.32823694e-01 8.42024028e-01
6.78014517e-01 3.10683459e-01 -6.77756011e-01 1.00745797e+00
2.72253174e-02 8.24129105e-01 2.12289304e-01 1.20429778e+00
-7.15339839e-01 -1.19510508e+00 -4.85864401e-01 6.89233184e-01
-7.01376319e-01 8.80814902e-03 9.69252512e-02 2.55188435e-01
4.49706942e-01 1.32845020e+00 -3.21413040e-01 -1.29011720e-01
7.17218995e-01 4.27732378e-01 1.54957280e-01 -5.53990364e-01
-7.75199831e-01 1.15230203e-01 -3.78050692e-02 -7.11836874e-01
-1.96960777e-01 -4.33031082e-01 -1.38012826e+00 -1.77801326e-01
1.24772400e-01 1.03920750e-01 1.00318706e+00 7.43674159e-01
9.66253877e-01 6.84338748e-01 2.72573233e-01 -5.59999347e-01
-2.09925562e-01 -1.15865898e+00 -6.21127844e-01 1.94486514e-01
2.47442350e-01 -3.79057497e-01 6.91380575e-02 9.43001583e-02] | [13.733424186706543, 3.170393943786621] |
01992804-c6b3-4d65-9df7-59ef421df084 | inferring-attention-shift-ranks-of-objects | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Siris_Inferring_Attention_Shift_Ranks_of_Objects_for_Image_Saliency_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Siris_Inferring_Attention_Shift_Ranks_of_Objects_for_Image_Saliency_CVPR_2020_paper.pdf | Inferring Attention Shift Ranks of Objects for Image Saliency | Psychology studies and behavioural observation show that humans shift their attention from one location to another when viewing an image of a complex scene. This is due to the limited capacity of the human visual system in simultaneously processing multiple visual inputs. The sequential shifting of attention on objects in a non-task oriented viewing can be seen as a form of saliency ranking. Although there are methods proposed for predicting saliency rank, they are not able to model this human attention shift well, as they are primarily based on ranking saliency values from binary prediction. Following psychological studies, in this paper, we propose to predict the saliency rank by inferring human attention shift. Due to the lack of such data, we first construct a large-scale salient object ranking dataset. The saliency rank of objects is defined by the order that an observer attends to these objects based on attention shift. The final saliency rank is an average across the saliency ranks of multiple observers. We then propose a learning-based CNN to leverage both bottom-up and top-down attention mechanisms to predict the saliency rank. Experimental results show that the proposed network achieves state-of-the-art performances on salient object rank prediction. Code and dataset are available at https://github.com/SirisAvishek/Attention_Shift_Ranks
| [' Rynson W.H. Lau', ' Xianghua Xie', ' Gary K.L. Tam', ' Jianbo Jiao', 'Avishek Siris'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['saliency-ranking'] | ['computer-vision'] | [ 4.65109289e-01 -1.51807740e-01 -2.65314281e-01 -5.08723915e-01
-4.35334086e-01 -1.46768779e-01 4.23294872e-01 3.61903220e-01
-2.22322434e-01 1.95648000e-01 5.12790263e-01 1.15801811e-01
-1.11094065e-01 -4.72942710e-01 -7.90488362e-01 -3.08030576e-01
2.32042577e-02 1.71002090e-01 7.30709493e-01 -1.48460209e-01
9.03592587e-01 3.56964730e-02 -1.99778903e+00 5.71081936e-01
8.95362020e-01 1.18576837e+00 7.34839797e-01 6.01841867e-01
3.74539137e-01 9.08264577e-01 -2.11457744e-01 -7.53770024e-02
3.53336424e-01 -5.47955990e-01 -9.23796892e-01 -5.03583625e-02
8.06588411e-01 -5.33165991e-01 -3.54627639e-01 1.16992867e+00
4.71548706e-01 2.83437103e-01 4.52574849e-01 -1.24992061e+00
-1.11366248e+00 5.73973894e-01 -6.93667948e-01 8.58304381e-01
4.11745667e-01 -1.87433898e-01 1.56830776e+00 -1.09287739e+00
3.30342710e-01 1.14654696e+00 -2.48803142e-02 3.64495754e-01
-1.02154183e+00 -4.95168507e-01 3.85573298e-01 8.44979942e-01
-9.62189317e-01 -2.49626160e-01 9.43923295e-01 -5.80007195e-01
5.78919709e-01 4.58856642e-01 8.32736731e-01 7.41038918e-01
1.43385023e-01 1.11590219e+00 1.27459967e+00 -2.66080499e-01
7.56823570e-02 -6.41684681e-02 3.58065158e-01 3.70359808e-01
1.95421845e-01 3.92980315e-02 -9.99731123e-01 1.12034395e-01
6.36991858e-01 3.87429327e-01 -1.93631262e-01 -4.25586671e-01
-1.41611314e+00 5.41290522e-01 1.25196123e+00 -7.16657862e-02
-4.80288595e-01 7.04887286e-02 1.51762560e-01 2.42920984e-02
4.02952820e-01 5.16353428e-01 -2.92524010e-01 2.66379327e-01
-9.19608116e-01 3.12408388e-01 7.14140609e-02 7.63384461e-01
8.76158357e-01 -2.16766581e-01 -5.58840454e-01 8.38682353e-01
3.08020115e-01 4.00868088e-01 6.93517506e-01 -9.27544534e-01
1.92614093e-01 7.15388119e-01 1.37059748e-01 -1.33773160e+00
-3.74090105e-01 -4.10721242e-01 -4.03911620e-01 1.81126073e-01
2.63282508e-01 5.08300662e-01 -8.90650332e-01 1.68793261e+00
7.68954605e-02 1.14960864e-01 -4.45247233e-01 1.57872725e+00
1.03870285e+00 3.70213062e-01 2.55293667e-01 1.12408601e-01
1.52484465e+00 -1.14887488e+00 -3.17648947e-01 -7.18305528e-01
-6.57066144e-03 -5.09722412e-01 1.26525331e+00 2.66700208e-01
-1.25344586e+00 -8.41573000e-01 -8.59818578e-01 -3.15912545e-01
-2.70041525e-01 -3.83808613e-02 6.04593933e-01 -1.61039531e-01
-1.07862008e+00 4.56039965e-01 -4.64369714e-01 -4.20863271e-01
7.32702971e-01 2.54067987e-01 1.09305177e-02 2.00366423e-01
-1.23890507e+00 1.02778304e+00 3.04761320e-01 5.77394515e-02
-1.07672250e+00 -5.92184007e-01 -7.13169992e-01 3.49685609e-01
3.42526227e-01 -6.03960693e-01 1.26654661e+00 -1.36370265e+00
-8.61365795e-01 1.09616590e+00 -5.61665654e-01 -3.93759072e-01
1.53948918e-01 -4.46541280e-01 -1.50250807e-01 2.80549794e-01
4.94441330e-01 1.05789709e+00 9.96959329e-01 -1.39508247e+00
-1.05394542e+00 -3.49010050e-01 2.38077059e-01 5.31154096e-01
-3.07384491e-01 2.50788152e-01 -1.66766077e-01 -4.97703642e-01
2.77764529e-01 -8.56723011e-01 -1.12124592e-01 -1.73330382e-01
-4.02471781e-01 -3.41749191e-01 5.13894320e-01 -5.27540207e-01
1.23958278e+00 -1.98440802e+00 1.77205682e-01 -1.62094817e-01
6.27351224e-01 -2.03112559e-03 -7.91295245e-02 1.55954197e-01
-1.83708981e-01 7.28205368e-02 7.80921355e-02 -6.53318018e-02
1.26561066e-02 -4.03833210e-01 -5.78020036e-01 3.83917332e-01
2.80749172e-01 1.06958485e+00 -1.21761179e+00 -5.13687849e-01
1.42582521e-01 7.44045675e-02 -6.71534717e-01 1.26696840e-01
8.43036994e-02 4.53616023e-01 -4.13644344e-01 7.06878424e-01
5.93524396e-01 -5.89582682e-01 -2.30010599e-01 -2.32946500e-01
-5.63452505e-02 5.37333846e-01 -6.85721993e-01 1.38944995e+00
1.17448486e-01 9.15947974e-01 -6.09152555e-01 -7.55443573e-01
6.72888219e-01 -2.36933514e-01 2.02960730e-01 -1.01872718e+00
1.01622254e-01 1.01454541e-01 3.77262145e-01 -2.01791272e-01
8.21127772e-01 1.37363359e-01 -8.17743838e-02 3.94848198e-01
-9.79584232e-02 3.00303876e-01 2.63024986e-01 3.43228340e-01
8.39174330e-01 -1.22822255e-01 4.58911926e-01 -3.57185543e-01
3.02915424e-01 -6.95271119e-02 5.79938233e-01 8.80191982e-01
-6.25774741e-01 8.40109706e-01 4.06802952e-01 -5.19675434e-01
-8.87977958e-01 -1.13436890e+00 -9.23845991e-02 2.17206430e+00
7.42322266e-01 -7.02829808e-02 -5.45659363e-01 -3.60929400e-01
4.57807770e-03 5.59404373e-01 -9.02083755e-01 -4.32479054e-01
-4.01113391e-01 -3.11136663e-01 -2.66477078e-01 5.76049626e-01
4.27773088e-01 -1.78330398e+00 -1.16255701e+00 -2.15116188e-01
-3.39389920e-01 -7.65954733e-01 -6.85971797e-01 1.25292554e-01
-6.45080924e-01 -1.04247856e+00 -6.93404138e-01 -9.66854751e-01
8.24376523e-01 8.67187262e-01 1.22368264e+00 2.47150496e-01
-1.19969256e-01 2.69029826e-01 -4.03105408e-01 -6.35595202e-01
2.03023523e-01 1.06426403e-01 5.49382195e-02 2.11450472e-01
4.81404126e-01 -1.79707274e-01 -1.19459701e+00 4.10628855e-01
-7.27399290e-01 3.76012385e-01 7.27358222e-01 5.31299949e-01
6.01666927e-01 -4.16249871e-01 7.21230924e-01 -6.43315494e-01
4.02711034e-01 -6.44572377e-01 -3.65299344e-01 3.66871059e-02
-3.85143667e-01 -1.51793405e-01 1.26880690e-01 -5.21765530e-01
-7.36562669e-01 -6.70130923e-02 5.14207721e-01 -3.43505323e-01
-1.58353478e-01 4.31632847e-01 2.05133274e-01 1.79581001e-01
4.81317699e-01 1.53488293e-01 -4.30061311e-01 -1.05609111e-01
1.15255192e-01 4.04148579e-01 3.98570001e-01 -1.15715645e-01
5.72182715e-01 5.02812505e-01 -1.61000729e-01 -3.56644958e-01
-1.63033235e+00 -7.37982810e-01 -6.86329722e-01 -5.57460010e-01
8.88018310e-01 -9.49918866e-01 -8.01969767e-01 3.88922513e-01
-9.31330502e-01 -2.98627645e-01 -1.40806004e-01 2.82218695e-01
-5.56736767e-01 1.35700881e-01 -3.44805986e-01 -4.72799778e-01
-2.45760173e-01 -9.71168518e-01 1.20692432e+00 5.45881212e-01
-3.68085504e-01 -6.80822790e-01 -1.96967870e-01 4.89214569e-01
3.85494679e-01 9.11429301e-02 7.00696170e-01 -6.32415533e-01
-8.20784807e-01 -9.69490260e-02 -5.93641698e-01 -5.74107468e-02
1.69352684e-02 -3.60481888e-01 -1.01706803e+00 -2.25998402e-01
-1.29380360e-01 -4.13312018e-01 1.19668055e+00 6.69171214e-01
1.44424236e+00 -2.12155640e-01 -2.35005677e-01 2.43812338e-01
1.16277790e+00 -1.61818638e-01 5.55111825e-01 5.20665646e-01
8.73111308e-01 7.74791658e-01 8.74354005e-01 4.06238496e-01
7.02154160e-01 5.92301488e-01 8.07396352e-01 -2.16252476e-01
-7.61557072e-02 -2.00556979e-01 2.40273625e-01 5.82103968e-01
-1.16683990e-01 8.13212022e-02 -9.01189506e-01 8.40972900e-01
-1.90122998e+00 -1.24739611e+00 -2.91640580e-01 2.38627338e+00
7.36905694e-01 2.85579711e-01 1.28825307e-01 -1.12494685e-01
1.08314252e+00 2.77447850e-01 -8.15835476e-01 -2.80250758e-01
8.95893481e-03 -2.43261889e-01 3.21962863e-01 2.57679433e-01
-1.17510474e+00 1.02941144e+00 6.20302773e+00 4.52004224e-01
-1.23141658e+00 5.58385737e-02 6.92786992e-01 -3.44792515e-01
-1.03035383e-01 -3.98380347e-02 -6.41014874e-01 6.11953735e-01
5.63813269e-01 -4.48138386e-01 4.38436121e-01 8.27114880e-01
1.10372700e-01 -3.96867186e-01 -1.13530302e+00 8.81078064e-01
4.20456499e-01 -7.13290632e-01 1.64435312e-01 1.12082157e-02
7.54800200e-01 7.72753656e-02 5.92912376e-01 1.92776546e-01
-4.94331159e-02 -1.00260282e+00 1.15267253e+00 7.15629756e-01
3.19224536e-01 -3.30086976e-01 4.68189150e-01 2.32194409e-01
-1.06222975e+00 -3.19265753e-01 -6.30733311e-01 -4.62777466e-01
-7.23924115e-02 4.04486507e-01 -6.51679397e-01 -2.28683740e-01
1.18204975e+00 9.24137175e-01 -1.06600583e+00 1.37545490e+00
-2.62326896e-01 3.92401814e-01 2.67587662e-01 -5.80977350e-02
-6.53949305e-02 2.83463478e-01 4.83806014e-01 9.22673881e-01
1.71380222e-01 1.22148499e-01 2.03792408e-01 8.13584507e-01
-1.04238398e-01 -6.95891976e-02 -3.44016820e-01 3.25207263e-01
3.03427368e-01 1.38409901e+00 -7.45211601e-01 -4.11209941e-01
-4.08558667e-01 9.53753531e-01 5.77934861e-01 2.94146091e-01
-9.54679310e-01 -1.03370227e-01 5.22881508e-01 3.82200390e-01
5.15025139e-01 1.11900382e-01 -3.86786312e-01 -9.76169288e-01
1.53657002e-02 -5.21097541e-01 3.56633544e-01 -1.15796328e+00
-1.28675294e+00 4.05747831e-01 -6.13329597e-02 -1.31105900e+00
1.87806696e-01 -5.21487772e-01 -5.40947735e-01 7.84183860e-01
-1.80358100e+00 -1.07643652e+00 -6.32006705e-01 3.04399818e-01
8.53702605e-01 8.64918083e-02 2.40111575e-01 -7.00878724e-02
-1.71370104e-01 2.95390278e-01 -4.56694573e-01 -3.62163745e-02
9.21424985e-01 -1.46505535e+00 1.64390758e-01 7.82595158e-01
-6.58851815e-03 6.70070410e-01 6.72464907e-01 -5.99921882e-01
-8.66655529e-01 -8.86148095e-01 1.04168844e+00 -7.86838710e-01
6.97052658e-01 -3.24277788e-01 -1.12357032e+00 4.25036281e-01
2.84919590e-01 1.74579307e-01 6.62858307e-01 2.20677733e-01
-3.32735538e-01 -9.01449174e-02 -6.66026771e-01 7.16538966e-01
1.20372450e+00 -5.24322391e-01 -9.19055700e-01 9.54069495e-02
6.07688427e-01 -2.59619623e-01 -5.18343568e-01 4.49996978e-01
6.69543266e-01 -1.24306989e+00 1.06902790e+00 -5.81000626e-01
9.77816522e-01 -5.00208974e-01 -1.22501940e-01 -1.17580640e+00
-1.07103181e+00 8.33138078e-02 -1.47937402e-01 8.62950385e-01
2.37819344e-01 -3.98056537e-01 5.62593818e-01 4.48905766e-01
-2.80488878e-01 -6.84890449e-01 -6.20711446e-01 -2.23973930e-01
-3.70727122e-01 -3.86856571e-02 3.91315997e-01 7.36250997e-01
-3.96464206e-02 5.28342187e-01 -3.67039561e-01 3.16932917e-01
5.90788841e-01 5.46667576e-01 4.23968047e-01 -1.49641466e+00
2.45323703e-02 -8.15304220e-01 -5.27531803e-01 -9.84484315e-01
8.86712149e-02 -9.16891575e-01 3.58561754e-01 -1.63506460e+00
9.71666038e-01 4.17826790e-03 -1.10166907e+00 5.26436269e-01
-7.75108755e-01 5.30872345e-01 4.11316633e-01 5.64763367e-01
-1.39186609e+00 3.99561644e-01 1.46319020e+00 -1.08769804e-01
-6.66248500e-02 -1.64275214e-01 -1.21151543e+00 8.24343085e-01
9.03013051e-01 -3.42344582e-01 -2.84999073e-01 -4.43383813e-01
4.58584458e-01 -2.27881670e-01 7.70780861e-01 -1.13631272e+00
2.63234675e-01 -3.95809174e-01 6.89322114e-01 -6.59279943e-01
1.37572974e-01 -6.18581414e-01 -4.48925406e-01 3.36771727e-01
-7.53276587e-01 3.23160142e-02 2.92876177e-02 5.04771471e-01
-2.18921781e-01 -1.00022890e-01 8.26343060e-01 -7.67899901e-02
-1.27276158e+00 2.94620514e-01 -1.03735477e-01 8.71031657e-02
8.00689757e-01 -3.00077140e-01 -6.35883808e-01 -3.89641017e-01
-5.90220392e-01 2.76364714e-01 4.83739972e-01 8.15248668e-01
8.72956753e-01 -1.53629696e+00 -6.87976837e-01 8.48590303e-03
4.45469797e-01 -2.49558136e-01 3.61688077e-01 7.76133657e-01
-6.76664114e-02 3.81390035e-01 -5.96938491e-01 -7.07217395e-01
-1.24472904e+00 7.52830029e-01 9.40587074e-02 3.93932424e-02
-1.71672463e-01 7.93558538e-01 9.65809762e-01 1.42785450e-02
-8.83548036e-02 -3.10214013e-01 -5.95299244e-01 8.79022777e-02
6.43817604e-01 4.11189139e-01 -2.10894585e-01 -9.51452315e-01
-4.48907495e-01 6.55990958e-01 -1.95516348e-01 1.50362343e-01
1.18293834e+00 -3.71446401e-01 -2.30551556e-01 8.04520547e-01
8.30412865e-01 -1.71960220e-01 -1.44921958e+00 -2.83797175e-01
-5.91555163e-02 -7.94010222e-01 -9.34544113e-03 -7.73117840e-01
-9.19249356e-01 8.51329684e-01 6.27445459e-01 3.77149105e-01
1.24508476e+00 4.68420148e-01 6.09644294e-01 1.44613788e-01
2.40230933e-01 -1.11956000e+00 6.40775740e-01 5.69692492e-01
1.14760697e+00 -1.73387432e+00 1.09524488e-01 -3.06212902e-01
-1.07136559e+00 4.44473684e-01 9.85514641e-01 -4.28519636e-01
4.19025600e-01 -6.48827016e-01 -1.21383801e-01 -3.43204290e-01
-8.53678703e-01 -6.78300381e-01 9.32989419e-01 3.59218180e-01
4.60581154e-01 2.56761342e-01 -4.86187674e-02 6.10250235e-01
-2.83960909e-01 -3.15273643e-01 4.61652517e-01 7.18918800e-01
-1.05814040e+00 -2.99465775e-01 -3.33295614e-01 9.07681942e-01
-4.12035525e-01 -2.96964169e-01 -4.59894657e-01 1.59607381e-01
7.69197494e-02 1.00642681e+00 2.77211487e-01 -3.15077692e-01
2.49071598e-01 -2.66316831e-01 2.25566998e-01 -8.69281530e-01
-4.59677368e-01 -1.99971214e-01 -4.80467796e-01 -5.73950827e-01
-5.22202015e-01 -7.48786390e-01 -1.19722855e+00 -4.40218411e-02
-1.57349795e-01 -2.52276570e-01 1.62720367e-01 6.29108250e-01
3.82066518e-01 6.53958678e-01 6.23759151e-01 -1.13425076e+00
-9.33797732e-02 -9.71308291e-01 -5.69090128e-01 9.09892082e-01
4.52152193e-01 -8.66261125e-01 -3.88798773e-01 1.67685330e-01] | [9.957599639892578, 0.3373660147190094] |
d73e85a8-7e67-4ee9-b7d1-1832290118f0 | learning-articulated-shape-with-keypoint | 2304.14396 | null | https://arxiv.org/abs/2304.14396v1 | https://arxiv.org/pdf/2304.14396v1.pdf | Learning Articulated Shape with Keypoint Pseudo-labels from Web Images | This paper shows that it is possible to learn models for monocular 3D reconstruction of articulated objects (e.g., horses, cows, sheep), using as few as 50-150 images labeled with 2D keypoints. Our proposed approach involves training category-specific keypoint estimators, generating 2D keypoint pseudo-labels on unlabeled web images, and using both the labeled and self-labeled sets to train 3D reconstruction models. It is based on two key insights: (1) 2D keypoint estimation networks trained on as few as 50-150 images of a given object category generalize well and generate reliable pseudo-labels; (2) a data selection mechanism can automatically create a "curated" subset of the unlabeled web images that can be used for training -- we evaluate four data selection methods. Coupling these two insights enables us to train models that effectively utilize web images, resulting in improved 3D reconstruction performance for several articulated object categories beyond the fully-supervised baseline. Our approach can quickly bootstrap a model and requires only a few images labeled with 2D keypoints. This requirement can be easily satisfied for any new object category. To showcase the practicality of our approach for predicting the 3D shape of arbitrary object categories, we annotate 2D keypoints on giraffe and bear images from COCO -- the annotation process takes less than 1 minute per image. | ['Dimitris Metaxas', 'Ligong Han', 'Georgios Pavlakos', 'Anastasis Stathopoulos'] | 2023-04-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Stathopoulos_Learning_Articulated_Shape_With_Keypoint_Pseudo-Labels_From_Web_Images_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Stathopoulos_Learning_Articulated_Shape_With_Keypoint_Pseudo-Labels_From_Web_Images_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-reconstruction'] | ['computer-vision'] | [-1.86125994e-01 3.48101348e-01 -3.87703329e-01 -4.68041569e-01
-1.19588220e+00 -8.15512657e-01 5.52648008e-01 -2.34089822e-01
-3.14845383e-01 5.20452201e-01 1.83093362e-02 -8.60539228e-02
1.74450636e-01 -3.58041316e-01 -1.02176428e+00 -4.52522576e-01
2.46792706e-03 9.76848900e-01 3.30213398e-01 2.72974700e-01
2.68144548e-01 1.02046812e+00 -1.72396100e+00 -3.75471786e-02
4.76963818e-01 1.18112504e+00 5.64128458e-01 7.37526476e-01
-3.21925282e-01 3.74679476e-01 -3.95407528e-01 -2.12946400e-01
6.61054194e-01 -2.77970523e-01 -8.05017591e-01 6.91566706e-01
8.47783446e-01 -6.37449145e-01 -6.00508563e-02 8.62249076e-01
3.71432155e-01 -1.75301835e-01 9.69821870e-01 -1.20746911e+00
-3.81259531e-01 2.10365012e-01 -5.20748138e-01 -1.26380980e-01
3.27414513e-01 -3.76880430e-02 9.03235972e-01 -1.34193110e+00
9.94628251e-01 1.26595628e+00 6.88884676e-01 8.17084432e-01
-1.25858712e+00 -5.55285335e-01 6.28916197e-04 -4.77466583e-02
-1.36943805e+00 -4.85777467e-01 9.18496847e-01 -4.59122717e-01
5.58709800e-01 5.54496497e-02 8.73764515e-01 9.30615306e-01
-6.98430762e-02 1.06888199e+00 1.15645099e+00 -4.18569416e-01
2.78372079e-01 1.74856305e-01 6.09255070e-03 8.82022381e-01
8.97208154e-02 9.88628566e-02 -4.37165588e-01 -1.86781839e-01
1.15204620e+00 -6.68616816e-02 6.27530217e-02 -1.06172335e+00
-1.33691907e+00 8.01432133e-01 3.97969306e-01 -2.27067634e-01
-3.27173799e-01 3.18502784e-01 2.01152325e-01 2.27430850e-01
5.77725947e-01 2.85846174e-01 -6.89054549e-01 7.33247250e-02
-6.73149586e-01 4.02296990e-01 6.64158046e-01 1.33281612e+00
1.10228157e+00 -8.04368109e-02 4.85797346e-01 1.04756796e+00
4.01427299e-01 8.62982988e-01 3.24082136e-01 -1.31637394e+00
2.61896849e-01 5.29459953e-01 2.90094882e-01 -5.74366987e-01
-3.45315814e-01 -9.36945230e-02 -4.05207902e-01 4.39202905e-01
4.54690874e-01 1.27480373e-01 -1.31669891e+00 1.59748840e+00
7.86940038e-01 -2.58602083e-01 -9.21772271e-02 9.82299924e-01
8.43349040e-01 2.74033725e-01 -2.33482391e-01 -3.95271406e-02
1.15475821e+00 -7.40899384e-01 -5.01191542e-02 -3.11031491e-01
2.79208988e-01 -7.10342646e-01 1.07928252e+00 2.12819800e-01
-1.09041619e+00 -7.02657402e-01 -7.15945601e-01 9.20005217e-02
-1.42153546e-01 1.54081523e-01 6.29185140e-01 2.89583892e-01
-9.87363577e-01 3.64250273e-01 -8.18506360e-01 -5.99254191e-01
3.54720950e-01 3.32912326e-01 -7.33617365e-01 -2.60062944e-02
-4.19588476e-01 9.69989896e-01 3.35724175e-01 -1.74685970e-01
-1.34901071e+00 -4.96543735e-01 -9.93043959e-01 -4.28632051e-01
3.21152449e-01 -5.60595214e-01 1.40178323e+00 -9.20286655e-01
-1.32896316e+00 1.41408217e+00 -1.70494765e-01 -2.50220269e-01
6.65260732e-01 -3.84093523e-01 1.03210256e-01 5.86438596e-01
4.49177563e-01 1.30502713e+00 1.04219186e+00 -1.85389841e+00
-8.24709535e-01 -4.68314439e-01 -5.33156730e-02 3.67683649e-01
3.01697314e-01 -3.02282482e-01 -5.76545954e-01 -6.07777357e-01
8.97757888e-01 -1.11672807e+00 -8.07746798e-02 3.41965199e-01
-4.66165721e-01 -2.65387803e-01 7.80401945e-01 -4.34675366e-01
2.38255560e-01 -1.89497125e+00 1.50061056e-01 1.83079541e-01
5.17298840e-02 -1.90639988e-01 -1.97222814e-01 1.14642061e-01
3.26703452e-02 -3.04204151e-02 -1.11873271e-02 -2.17028290e-01
-3.41709517e-02 3.81436467e-01 -3.48942250e-01 5.24927437e-01
1.15586542e-01 8.74865294e-01 -7.82076061e-01 -6.49242163e-01
5.17369986e-01 8.80204737e-02 -5.55755615e-01 3.99026006e-01
-3.67514968e-01 4.88717288e-01 -3.54894876e-01 9.40244734e-01
5.60563505e-01 -3.86763692e-01 -5.77161796e-02 -3.19882929e-01
1.53013259e-01 1.18764222e-01 -1.36576295e+00 1.72024596e+00
-4.30057764e-01 3.93094659e-01 5.97467311e-02 -8.27693999e-01
1.06414628e+00 1.79963127e-01 6.72750652e-01 -1.30660698e-01
-5.91447391e-02 4.25004840e-01 -5.31679094e-01 -4.58535939e-01
2.16942623e-01 -2.72919118e-01 -1.07355930e-01 5.89341819e-01
4.52454418e-01 -7.14118242e-01 -6.91629350e-02 6.47916123e-02
7.14250743e-01 5.24379849e-01 2.12874770e-01 -1.87990323e-01
1.12155378e-01 2.38903895e-01 3.80121619e-01 7.01083601e-01
-7.58973137e-02 8.50805044e-01 7.97206983e-02 -7.16654181e-01
-1.51510596e+00 -1.29460895e+00 -2.46335894e-01 8.51153255e-01
3.47463012e-01 8.36136844e-03 -5.03792226e-01 -9.34273243e-01
3.23025405e-01 3.41239840e-01 -6.43959582e-01 2.06718221e-01
-5.64972878e-01 -1.74451396e-01 2.42147818e-01 6.49739861e-01
5.40138364e-01 -9.31056798e-01 -7.43704379e-01 -3.66521217e-02
-2.19239533e-01 -9.40674067e-01 -3.62902850e-01 4.37227458e-01
-1.07082117e+00 -1.32147431e+00 -9.28255200e-01 -1.03006709e+00
1.03015423e+00 5.35847902e-01 1.16455579e+00 9.29430220e-03
-1.68088019e-01 8.90155435e-01 -4.02301103e-01 -2.62158841e-01
-4.85935718e-01 8.26115012e-02 1.98749349e-01 -4.12486762e-01
2.93944031e-01 -4.54774886e-01 -4.36148882e-01 6.18224561e-01
-5.19830823e-01 1.23281799e-01 7.27273047e-01 9.44964826e-01
8.82212281e-01 -4.95018631e-01 5.28609514e-01 -7.70335138e-01
-7.87734762e-02 -1.68356016e-01 -6.25548124e-01 1.87043279e-01
-3.59453827e-01 2.00415656e-01 3.16147715e-01 -6.72025740e-01
-8.94788682e-01 5.61070085e-01 -1.44166097e-01 -7.33677268e-01
-5.09323537e-01 1.97444662e-01 3.02677900e-02 -1.25986338e-01
7.40501344e-01 1.48628131e-01 1.89052671e-01 -7.33563960e-01
6.77769542e-01 7.73480594e-01 6.65212035e-01 -8.11713517e-01
1.06678414e+00 5.94740629e-01 -6.93165734e-02 -9.31756973e-01
-9.17868435e-01 -5.49501419e-01 -1.00678957e+00 -2.74867445e-01
7.58871555e-01 -1.19093966e+00 -5.49286842e-01 2.75572240e-01
-9.12329376e-01 -4.06644613e-01 -5.45221090e-01 6.58041358e-01
-1.03434408e+00 3.63023371e-01 -5.47370493e-01 -7.13428557e-01
-2.02192351e-01 -8.94081891e-01 1.56711340e+00 5.72563410e-02
-2.46050343e-01 -6.99430406e-01 -8.90446976e-02 3.56324852e-01
-1.46851525e-01 1.42820939e-01 8.74568760e-01 -8.06786656e-01
-7.90949762e-01 -2.44404882e-01 -8.36310163e-02 3.37690026e-01
3.02851778e-02 -2.24725962e-01 -9.10852432e-01 -3.44936758e-01
-1.31905228e-01 -8.34908843e-01 5.31743348e-01 3.81061375e-01
9.81762588e-01 -1.26679927e-01 -3.96230847e-01 5.02808809e-01
1.15327215e+00 -2.01306231e-02 1.48410484e-01 1.51587367e-01
6.20397329e-01 5.03342330e-01 8.98783803e-01 3.25236589e-01
5.18001974e-01 5.58689058e-01 5.02949834e-01 4.11106348e-02
-2.42747352e-01 -6.14972889e-01 5.82849011e-02 6.47594810e-01
1.06364982e-02 1.44716635e-01 -8.46169770e-01 8.02371800e-01
-1.55663848e+00 -7.52615750e-01 2.38412321e-01 2.20700264e+00
8.46032262e-01 7.99138397e-02 2.57748544e-01 -1.45353779e-01
6.90725148e-01 -8.07118937e-02 -8.92239332e-01 2.13524014e-01
1.46464258e-01 3.36838700e-02 5.48582137e-01 3.87688339e-01
-1.12794936e+00 9.75060105e-01 7.21542978e+00 5.26511252e-01
-9.41116929e-01 -6.97247013e-02 5.29446363e-01 1.90358058e-01
-3.35330397e-01 3.48764509e-02 -9.23881412e-01 1.29728749e-01
3.19413304e-01 2.21145287e-01 2.36503676e-01 1.50354910e+00
-1.13563292e-01 -9.19753835e-02 -1.20491076e+00 1.18425834e+00
2.65823573e-01 -1.24206853e+00 2.14781538e-01 1.92264795e-01
6.92391634e-01 5.67166433e-02 -2.61393934e-01 8.19848925e-02
6.08111680e-01 -6.45729363e-01 1.05362380e+00 2.85803616e-01
9.85356867e-01 -4.78017211e-01 4.61816698e-01 4.84170586e-01
-1.03013933e+00 2.04965621e-02 -5.99303246e-01 3.57644796e-01
5.25448583e-02 2.57072538e-01 -1.14107680e+00 1.90353125e-01
8.48635018e-01 5.61930418e-01 -5.50355434e-01 9.55791056e-01
-4.22700882e-01 2.98285872e-01 -6.11434817e-01 5.72262816e-02
5.07106399e-03 -2.90528834e-02 4.71045762e-01 6.82674766e-01
4.10930723e-01 1.82267293e-01 3.09008926e-01 6.27300739e-01
-1.20267227e-01 -2.93220282e-02 -7.42456377e-01 2.59301007e-01
7.48534620e-01 1.18326640e+00 -9.13738549e-01 -4.64043111e-01
-3.02038938e-01 8.52860153e-01 3.19352001e-01 1.54264510e-01
-4.95746881e-01 -4.09135595e-02 3.82192820e-01 2.35389635e-01
5.44304073e-01 -3.90814185e-01 -2.11486444e-01 -1.23254156e+00
-1.35026097e-01 -7.77998626e-01 2.50322849e-01 -1.20060182e+00
-1.28817773e+00 3.02983254e-01 3.69513571e-01 -1.28346252e+00
-6.33650720e-01 -7.14002669e-01 -6.94546849e-02 4.79690075e-01
-1.08796072e+00 -1.49675465e+00 -4.64230150e-01 4.76535380e-01
7.20895112e-01 -1.91259265e-01 7.47186244e-01 1.48847364e-02
2.42291585e-01 2.15952173e-01 -1.29896812e-02 1.06080666e-01
5.70489526e-01 -1.49857867e+00 5.47222972e-01 4.76411700e-01
5.22100449e-01 3.93512875e-01 6.12136066e-01 -6.53368652e-01
-1.41267312e+00 -8.25721025e-01 6.06733382e-01 -8.10123324e-01
2.24678010e-01 -3.50502610e-01 -6.01779997e-01 8.99419010e-01
-3.98667008e-01 7.40855858e-02 4.85486627e-01 5.41855171e-02
-3.68014395e-01 -1.87883992e-02 -1.18698585e+00 4.86398876e-01
1.33092725e+00 -4.97817785e-01 -8.41662228e-01 4.78509843e-01
4.74607587e-01 -5.76859236e-01 -8.56297374e-01 4.41371173e-01
7.06636429e-01 -6.92246616e-01 1.21960282e+00 -5.81067204e-01
2.76748121e-01 -2.89209276e-01 -5.01772046e-01 -1.22483933e+00
7.36035854e-02 -2.01011032e-01 -1.43334731e-01 8.54065776e-01
4.42004986e-02 -2.39340365e-01 1.06750464e+00 4.46991444e-01
-6.44543543e-02 -5.26062846e-01 -8.92467916e-01 -7.44358718e-01
-1.70352757e-02 -2.83452213e-01 4.83128101e-01 5.39182007e-01
-3.83117735e-01 4.49548781e-01 -4.26577777e-01 5.40086105e-02
9.42855775e-01 5.75095534e-01 1.45399511e+00 -1.35783529e+00
-1.63346939e-02 1.38286009e-01 -4.94942009e-01 -1.69210637e+00
2.83478856e-01 -8.22896183e-01 2.87758559e-01 -1.32509625e+00
3.28502744e-01 -8.26957881e-01 2.65224069e-01 6.12576663e-01
1.73373781e-02 5.82958698e-01 1.44317567e-01 4.98403609e-01
-4.44056094e-01 3.95393640e-01 1.29093933e+00 3.75413597e-02
-4.81802374e-02 1.18247770e-01 -3.89480859e-01 9.35083747e-01
4.46930408e-01 -2.75155663e-01 -3.15336823e-01 -3.49253953e-01
-6.31850362e-02 1.91441327e-01 6.70011997e-01 -7.45392025e-01
-1.63022399e-01 -2.54686117e-01 7.31916904e-01 -1.13703990e+00
4.53452170e-01 -1.04967618e+00 2.47520447e-01 3.06088895e-01
-2.37533078e-01 -8.67921859e-02 -1.32838637e-01 6.62045240e-01
1.01354294e-01 -5.02402425e-01 9.00759339e-01 -6.27017379e-01
-9.53729987e-01 3.07388365e-01 -1.53510839e-01 8.94654170e-03
1.00068688e+00 -4.20076102e-01 1.41469389e-01 -5.04397869e-01
-9.59852159e-01 -6.43585948e-03 9.95931327e-01 3.38032335e-01
7.80101717e-01 -1.47500145e+00 -4.93312180e-01 4.02956784e-01
2.31444478e-01 4.00542110e-01 -7.67688677e-02 1.85944512e-01
-7.13710189e-01 1.92818388e-01 -1.92115888e-01 -1.23435092e+00
-1.22132754e+00 5.24801910e-01 2.78470099e-01 2.70290643e-01
-6.95400476e-01 6.90214813e-01 1.18491374e-01 -9.19247270e-01
1.79981232e-01 -3.72146606e-01 2.03429028e-01 -1.19126633e-01
1.37880504e-01 9.69167203e-02 -1.59316480e-01 -7.76978016e-01
-2.52243459e-01 8.94340515e-01 2.26632673e-02 -2.96626806e-01
1.42018580e+00 -2.44239345e-01 2.79234916e-01 6.55017316e-01
1.25610459e+00 -1.16969161e-01 -1.65038764e+00 -2.89048553e-01
-8.25039670e-02 -6.64688826e-01 -4.16976035e-01 -6.68767512e-01
-7.64238954e-01 7.88404167e-01 8.25034678e-01 -6.87129125e-02
6.86700642e-01 6.26844525e-01 5.39229453e-01 5.23054063e-01
8.37719381e-01 -9.73773003e-01 4.58837837e-01 2.87217468e-01
8.08650136e-01 -1.45876479e+00 8.19651335e-02 -3.01423132e-01
-6.64738953e-01 1.10376406e+00 4.88566786e-01 -2.17302322e-01
6.03709757e-01 -3.82223725e-02 3.78926754e-01 -3.54123563e-01
-6.01907790e-01 -2.30137959e-01 2.77213514e-01 7.20079601e-01
-3.27529572e-02 -1.28499726e-02 7.81717226e-02 8.75442326e-02
-2.67544806e-01 -1.68823496e-01 2.73153931e-01 9.87827182e-01
-7.85701632e-01 -9.66093123e-01 -4.74341452e-01 5.51782727e-01
-1.10455006e-01 3.07392120e-01 -3.19517106e-01 1.08794451e+00
-9.39675570e-02 5.08810818e-01 1.57539606e-01 -3.36985439e-01
2.54114121e-01 1.74275428e-01 6.70339584e-01 -7.76063740e-01
-8.08875076e-03 3.42456430e-01 -4.56987508e-03 -3.28348488e-01
-6.93634868e-01 -7.14015663e-01 -1.14034021e+00 -8.21736604e-02
-5.93235910e-01 1.93392053e-01 8.39576304e-01 9.00738776e-01
2.33033091e-01 -3.32719296e-01 6.87627852e-01 -1.32718873e+00
-8.49624038e-01 -9.92960274e-01 -5.88110387e-01 6.13704801e-01
2.17700571e-01 -1.03869975e+00 -4.77409869e-01 4.54680741e-01] | [8.085453033447266, -2.960296392440796] |
761ed6c2-e289-45d8-b358-841cccc6a4bf | zero-shot-composed-image-retrieval-with | 2303.15247 | null | https://arxiv.org/abs/2303.15247v1 | https://arxiv.org/pdf/2303.15247v1.pdf | Zero-Shot Composed Image Retrieval with Textual Inversion | Composed Image Retrieval (CIR) aims to retrieve a target image based on a query composed of a reference image and a relative caption that describes the difference between the two images. The high effort and cost required for labeling datasets for CIR hamper the widespread usage of existing methods, as they rely on supervised learning. In this work, we propose a new task, Zero-Shot CIR (ZS-CIR), that aims to address CIR without requiring a labeled training dataset. Our approach, named zero-Shot composEd imAge Retrieval with textuaL invErsion (SEARLE), maps the visual features of the reference image into a pseudo-word token in CLIP token embedding space and integrates it with the relative caption. To support research on ZS-CIR, we introduce an open-domain benchmarking dataset named Composed Image Retrieval on Common Objects in context (CIRCO), which is the first dataset for CIR containing multiple ground truths for each query. The experiments show that SEARLE exhibits better performance than the baselines on the two main datasets for CIR tasks, FashionIQ and CIRR, and on the proposed CIRCO. The dataset, the code and the model are publicly available at https://github.com/miccunifi/SEARLE . | ['Alberto del Bimbo', 'Marco Bertini', 'Lorenzo Agnolucci', 'Alberto Baldrati'] | 2023-03-27 | null | null | null | null | ['zero-shot-composed-image-retrieval-zs-cir', 'composed-image-retrieval'] | ['computer-vision', 'computer-vision'] | [ 3.02055359e-01 -4.59811032e-01 -3.68666083e-01 -3.06523263e-01
-1.73106980e+00 -6.61728442e-01 8.34726155e-01 -2.46237978e-01
-3.70845228e-01 4.10023272e-01 9.13772881e-02 -9.16010737e-02
-4.28883955e-02 -3.47294688e-01 -7.41745114e-01 -6.63658559e-01
5.03090024e-01 4.98336434e-01 2.57264338e-02 -2.31096998e-01
2.91452944e-01 1.30203981e-02 -1.57669854e+00 4.25297379e-01
6.06097460e-01 1.23265171e+00 5.39984465e-01 4.91357923e-01
6.95283040e-02 8.33503902e-01 -5.21148384e-01 -4.68978107e-01
2.47685254e-01 -2.89460599e-01 -9.06098425e-01 -8.55656266e-02
7.14263558e-01 -3.78086150e-01 -3.69050711e-01 9.18201685e-01
6.21255636e-01 9.31598991e-02 1.00814652e+00 -1.61390400e+00
-1.31397712e+00 2.91756392e-01 -5.51531434e-01 1.52615570e-02
3.52813244e-01 -4.68872078e-02 1.29392040e+00 -1.44223499e+00
9.98814344e-01 1.24236465e+00 3.75264525e-01 4.97386575e-01
-9.88763332e-01 -7.32153773e-01 -1.16982251e-01 4.31250989e-01
-1.86885583e+00 -4.00035501e-01 4.09929097e-01 -4.41325456e-01
6.56380415e-01 3.48507017e-01 2.26860028e-02 1.18830299e+00
-1.10943392e-01 9.47287083e-01 9.41946507e-01 -4.98273313e-01
-3.08316126e-02 2.18738511e-01 3.01773340e-01 3.75113696e-01
1.20073810e-01 1.40175641e-01 -3.03446352e-01 1.22000903e-01
3.01541805e-01 -3.90945971e-02 -3.53130043e-01 -3.83310556e-01
-1.36101151e+00 6.93851352e-01 6.91102684e-01 2.04262391e-01
-6.27623647e-02 4.33471799e-01 4.75319982e-01 3.19593608e-01
5.47401190e-01 3.71685475e-01 -3.76530826e-01 2.11786419e-01
-8.51917386e-01 2.25125045e-01 4.74999249e-01 1.35229397e+00
8.50845575e-01 -3.71756375e-01 -5.03920496e-01 9.96908665e-01
2.58878112e-01 1.00481462e+00 4.83674109e-01 -9.68848825e-01
3.98395717e-01 1.91371381e-01 3.07526827e-01 -8.28300893e-01
3.56955230e-01 -1.61807299e-01 -6.32430851e-01 -3.07620943e-01
-2.64179111e-02 4.69401121e-01 -1.11821890e+00 1.62705636e+00
3.24135423e-02 1.12799808e-01 3.51084411e-01 1.16161168e+00
1.39147377e+00 8.29218447e-01 1.65821314e-02 3.69570792e-01
1.54793274e+00 -1.35173726e+00 -7.52142072e-01 -4.32003558e-01
7.86253691e-01 -1.08242333e+00 1.31386673e+00 5.84838055e-02
-5.86254060e-01 -6.00111604e-01 -1.04962921e+00 -5.64454079e-01
-9.18318868e-01 5.70337832e-01 1.85999304e-01 1.75197423e-01
-1.11681581e+00 -4.89072502e-02 -1.32288963e-01 -3.68124545e-01
1.60968930e-01 -1.37427449e-01 -6.07300997e-01 -7.86214352e-01
-1.37897635e+00 8.83878231e-01 5.47681570e-01 2.24871099e-01
-1.07572055e+00 -6.20360196e-01 -1.13112187e+00 5.34352921e-02
5.43201208e-01 -5.11454761e-01 1.36506689e+00 -1.16521716e+00
-9.01964545e-01 1.27815866e+00 -7.63605013e-02 -1.42549574e-01
2.32859284e-01 -3.21276039e-01 -3.22099149e-01 2.76688039e-01
6.66323185e-01 1.16457677e+00 1.08424330e+00 -1.55178583e+00
-1.73687130e-01 -1.05451062e-01 1.35312080e-01 1.52539372e-01
7.85780400e-02 1.71915814e-01 -1.19022965e+00 -8.08989525e-01
-3.60364318e-01 -1.15291834e+00 2.40377098e-01 1.04001164e-02
-4.11941051e-01 -2.94580489e-01 8.89195144e-01 -5.86243868e-01
9.70887542e-01 -2.55680346e+00 -2.15370171e-02 -2.03278273e-01
-9.57800001e-02 3.79072428e-01 -7.70881534e-01 7.36411214e-01
-1.84386253e-01 1.24924324e-01 -3.14452112e-01 -4.93837595e-01
1.53685763e-01 1.80091888e-01 -6.65243745e-01 2.62208730e-01
2.76137590e-01 1.38722408e+00 -1.22713840e+00 -7.99577892e-01
2.32695147e-01 5.41662037e-01 -1.43700927e-01 3.31970394e-01
-2.90633500e-01 2.91328162e-01 -3.60095650e-01 8.01533222e-01
7.56478727e-01 -5.50258040e-01 -1.89365506e-01 -5.17464399e-01
5.92806712e-02 -2.05756485e-01 -7.31834173e-01 1.93798184e+00
-5.96981525e-01 8.31131756e-01 -3.97999942e-01 -8.18856716e-01
8.15128922e-01 2.35935748e-01 2.56450117e-01 -1.16721761e+00
2.04941332e-02 4.99990642e-01 -5.87638259e-01 -6.16398096e-01
8.34705293e-01 2.60857284e-01 -5.12284756e-01 4.40570831e-01
1.97189286e-01 -2.96792030e-01 4.08579916e-01 6.30721331e-01
7.71564305e-01 4.24104393e-01 9.01468694e-02 2.01058132e-03
6.06138527e-01 3.00385773e-01 2.09705353e-01 9.68422055e-01
-7.04222023e-02 1.16180968e+00 1.62168950e-01 -1.00191481e-01
-1.07848740e+00 -1.20664597e+00 -1.46684408e-01 1.08174658e+00
7.24331021e-01 -4.52096164e-01 -4.16547090e-01 -4.77893054e-01
-8.84970278e-02 5.73233008e-01 -7.49911129e-01 -1.74847484e-01
-2.51969725e-01 -3.56440008e-01 6.61182046e-01 2.86933124e-01
6.71478868e-01 -1.14785743e+00 -2.79927880e-01 -3.09219241e-01
-7.43342757e-01 -1.37366819e+00 -9.08600450e-01 -2.15089172e-01
-1.80629611e-01 -1.15596545e+00 -1.00370955e+00 -1.05311060e+00
6.13614202e-01 9.06814754e-01 1.38268554e+00 1.49181217e-01
-6.59905076e-01 8.16662610e-01 -8.12702775e-01 -2.72115320e-01
-3.45128551e-02 1.48053756e-02 -4.71335292e-01 3.19864079e-02
2.80228198e-01 2.09420085e-01 -7.41120934e-01 3.28289628e-01
-1.40836859e+00 1.41424373e-01 8.13805819e-01 9.75258887e-01
7.79495716e-01 -6.42347813e-01 6.64281011e-01 -7.95982480e-01
4.27581131e-01 -6.32724524e-01 -6.30856276e-01 6.78032994e-01
-5.31914294e-01 1.25920147e-01 1.68978661e-01 -2.65744984e-01
-1.05439520e+00 -1.04212046e-01 2.28225961e-01 -8.03473592e-01
2.06128031e-01 6.53395414e-01 1.10383339e-01 8.33374411e-02
4.19174254e-01 3.74189019e-01 -2.12653786e-01 -4.09237504e-01
9.84379530e-01 1.03256381e+00 8.29732597e-01 -5.36547303e-01
7.90484190e-01 4.13567424e-01 -3.81771266e-01 -6.62539721e-01
-1.30587065e+00 -1.05701315e+00 -4.95930821e-01 -2.59365439e-01
9.09492135e-01 -1.38297558e+00 -1.90377191e-01 2.48202458e-01
-1.27911997e+00 -1.95835292e-01 -6.43563643e-02 3.04384351e-01
-5.66418767e-01 2.89019138e-01 -2.37812281e-01 -5.14698923e-01
-6.60135865e-01 -1.24015331e+00 1.82979202e+00 -9.24871787e-02
1.66040882e-01 -6.78368926e-01 2.19058707e-01 6.72523677e-01
4.65138018e-01 1.15265837e-02 6.39624238e-01 -4.49322909e-01
-9.72004056e-01 -3.92221093e-01 -8.81752670e-01 5.91504276e-01
2.39656065e-02 -4.05427441e-02 -1.08729088e+00 -4.93941247e-01
-5.30105948e-01 -9.73280907e-01 1.05756950e+00 1.35749191e-01
1.25181556e+00 -5.23368269e-02 -1.56931907e-01 4.91694033e-01
1.96339238e+00 -7.96626508e-02 9.15657580e-01 4.20650959e-01
5.26990414e-01 4.89962697e-01 1.35938978e+00 5.52385710e-02
4.99171495e-01 8.46273601e-01 4.41272318e-01 -1.14202835e-01
-5.14302969e-01 -3.15497071e-01 4.14855301e-01 9.88772750e-01
3.47084999e-01 -5.31654000e-01 -7.88937330e-01 9.13233399e-01
-1.89847517e+00 -7.53548682e-01 1.22377932e-01 2.06841469e+00
7.15949357e-01 -5.40031075e-01 -5.98577619e-01 -3.23670924e-01
5.82161129e-01 3.74101222e-01 -3.75442564e-01 -6.60412088e-02
-1.94529936e-01 1.97213829e-01 5.20495176e-01 4.16229665e-01
-1.15094328e+00 1.21306717e+00 5.30937243e+00 1.15217268e+00
-1.09412920e+00 4.02252614e-01 4.31739032e-01 2.78276324e-01
-1.20352969e-01 1.89078629e-01 -6.91639483e-01 2.80334622e-01
8.37315798e-01 -2.79571563e-01 2.61043191e-01 7.15886474e-01
-7.87224621e-02 -2.36183137e-01 -1.01179338e+00 1.31568599e+00
6.89281881e-01 -1.10156095e+00 3.72891337e-01 -3.78075182e-01
8.58225107e-01 3.80949855e-01 4.00449932e-01 5.58645487e-01
-1.80568025e-02 -1.00132656e+00 8.19238007e-01 5.35015821e-01
1.35662842e+00 -1.49716675e-01 9.56327617e-01 -6.75388873e-02
-1.23513520e+00 1.99283361e-01 -4.60204184e-01 5.32528520e-01
-5.08434102e-02 7.75928497e-02 -6.34104013e-01 9.78137016e-01
7.61276007e-01 9.88780975e-01 -8.90825510e-01 1.04731178e+00
-2.49259904e-01 3.07194710e-01 6.77649081e-02 2.33997986e-01
3.84833097e-01 -2.34355390e-01 3.36589485e-01 1.13310909e+00
3.24451149e-01 -2.52784133e-01 9.93034840e-02 9.16912675e-01
-2.88470626e-01 2.17901915e-01 -8.26883435e-01 -1.47747129e-01
5.31318247e-01 1.33381069e+00 -5.13868213e-01 -5.40888309e-01
-3.59353721e-01 1.20828938e+00 -1.36996331e-02 6.41543925e-01
-9.39179242e-01 -6.25855684e-01 3.13790679e-01 -3.20218384e-01
4.23102677e-01 1.53438240e-01 3.15092504e-01 -1.37047243e+00
1.42053872e-01 -8.68879735e-01 3.93001288e-01 -1.64836180e+00
-1.33677769e+00 7.51278400e-01 3.46400410e-01 -1.68385124e+00
-2.35203817e-01 -3.84340942e-01 -1.87392026e-01 7.81043589e-01
-2.14899540e+00 -1.61029601e+00 -8.03514361e-01 6.90276980e-01
8.64512086e-01 5.28906509e-02 8.54287148e-01 7.72443295e-01
-4.89496291e-01 4.89053696e-01 3.84151101e-01 2.14279830e-01
1.27112126e+00 -9.87113655e-01 3.12530935e-01 5.81010878e-01
4.60423619e-01 4.94223356e-01 5.43449819e-01 -4.06324267e-01
-1.46650326e+00 -1.47825289e+00 8.99486482e-01 -5.15516520e-01
5.68333030e-01 -4.74350870e-01 -7.12360859e-01 8.27486634e-01
3.46053600e-01 4.24150527e-01 4.49013859e-01 -4.51801836e-01
-8.81177187e-01 -1.33431464e-01 -7.23150849e-01 4.66977984e-01
7.04556942e-01 -8.66804779e-01 -5.84655881e-01 6.10903740e-01
1.06696486e+00 -1.51434273e-01 -6.83666587e-01 2.28093192e-01
5.92880428e-01 -4.21742409e-01 1.21146262e+00 -2.86260664e-01
6.13684714e-01 -4.80714351e-01 -5.20346999e-01 -1.08926618e+00
1.04163602e-01 -4.35670763e-02 5.29622018e-01 1.20446038e+00
3.10454249e-01 -4.92034882e-01 1.17424689e-01 1.21833831e-01
-7.05291927e-02 -3.35576475e-01 -7.54433990e-01 -1.07072091e+00
-9.45788547e-02 -4.12343204e-01 4.38617885e-01 7.67474651e-01
-7.50874400e-01 4.80410039e-01 -4.35537934e-01 1.23677604e-01
6.68383181e-01 3.91200393e-01 9.07271981e-01 -8.59712541e-01
1.14129037e-01 1.07457727e-01 -3.69862199e-01 -9.70919549e-01
2.69008577e-01 -1.23668003e+00 3.70898396e-01 -1.60919368e+00
4.95188177e-01 -5.15336394e-01 -4.37254995e-01 4.58845168e-01
-1.57613993e-01 9.39832985e-01 3.55429649e-01 5.64989209e-01
-1.24480438e+00 7.87946463e-01 1.06860042e+00 -6.03554964e-01
3.11323702e-01 -6.13818228e-01 -5.84683061e-01 2.58027762e-01
6.05099797e-01 -6.77042067e-01 -5.27225971e-01 -7.09733188e-01
1.66470230e-01 -3.79502848e-02 6.58739686e-01 -5.56665897e-01
1.17941201e-01 1.61903933e-01 -1.17532328e-01 -9.25061464e-01
5.96328557e-01 -7.14624882e-01 1.71296403e-01 1.67857170e-01
-4.94933695e-01 1.75763160e-01 -5.16272895e-02 5.40015101e-01
-7.09330440e-01 -4.44020212e-01 5.18341303e-01 -4.69225720e-02
-1.13976455e+00 3.59784603e-01 5.73497042e-02 2.92020857e-01
9.28978205e-01 1.20140694e-01 -5.77321768e-01 -4.59277034e-01
-2.19312966e-01 4.93920207e-01 3.82541597e-01 8.23193550e-01
9.14703488e-01 -1.55910206e+00 -8.84982049e-01 -1.12427682e-01
1.04674554e+00 -1.03488259e-01 2.17109546e-01 7.17233717e-01
-6.65958405e-01 9.04494941e-01 1.27779189e-02 -6.19763613e-01
-1.39383745e+00 8.36397469e-01 1.79470450e-01 -1.67416781e-01
-5.19087613e-01 6.28425121e-01 4.63152409e-01 -5.58486819e-01
6.86511099e-02 -1.12488255e-01 -1.07499979e-01 6.41890094e-02
6.04641676e-01 -6.49446100e-02 -2.44444311e-02 -9.56768036e-01
-2.09412813e-01 7.83750474e-01 -1.44387782e-01 -2.16664121e-01
1.08452439e+00 -3.83721590e-01 -4.09375429e-01 5.17095506e-01
1.75665367e+00 -3.66943896e-01 -7.36981690e-01 -5.40536880e-01
1.21630087e-01 -8.12050283e-01 1.73161566e-01 -8.39286923e-01
-1.09024537e+00 7.36029267e-01 7.43278205e-01 -3.16830397e-01
1.13374066e+00 2.64295220e-01 7.93064296e-01 5.96961558e-01
3.81825954e-01 -7.93441415e-01 3.70343983e-01 4.38420564e-01
1.25123668e+00 -1.61979496e+00 -7.00944588e-02 -3.77153397e-01
-7.77545750e-01 7.28991330e-01 4.03449774e-01 -1.61851674e-01
4.33068573e-01 -4.35113370e-01 3.34101796e-01 -3.48781407e-01
-8.42717767e-01 -6.23012364e-01 4.84883070e-01 5.45507789e-01
1.78242996e-01 -1.18611172e-01 -2.22262502e-01 1.94843158e-01
1.85295448e-01 2.10595667e-01 3.09152424e-01 8.14172745e-01
-2.15449631e-02 -1.08578551e+00 -2.87890673e-01 2.43710488e-01
-2.96443045e-01 -5.13394773e-01 -4.73053902e-01 8.45232844e-01
-7.81517997e-02 9.08230722e-01 -2.65005957e-02 -1.05181411e-01
2.42856488e-01 -1.24642558e-01 3.85018021e-01 -6.62069678e-01
-1.74024150e-01 -1.64081246e-01 -6.03673235e-02 -6.31693661e-01
-6.57646656e-01 -1.72268212e-01 -7.01174736e-01 2.21141517e-01
-3.79784346e-01 2.01706767e-01 7.53205538e-01 6.74820364e-01
4.03986603e-01 2.61530638e-01 6.34426415e-01 -7.53600955e-01
-4.61420685e-01 -9.58917797e-01 -5.00819027e-01 7.43786454e-01
3.77258390e-01 -7.33538985e-01 -4.62960899e-01 2.32152849e-01] | [10.8616943359375, 1.1513326168060303] |
ff576dd6-9c77-470c-913f-4eaf60d675ce | neuroevolution-for-rts-micro | 1803.10288 | null | http://arxiv.org/abs/1803.10288v1 | http://arxiv.org/pdf/1803.10288v1.pdf | Neuroevolution for RTS Micro | This paper uses neuroevolution of augmenting topologies to evolve control
tactics for groups of units in real-time strategy games. In such games, players
build economies to generate armies composed of multiple types of units with
different attack and movement characteristics to combat each other. This paper
evolves neural networks to control movement and attack commands, also called
micro, for a group of ranged units skirmishing with a group of melee units. Our
results show that neuroevolution of augmenting topologies can effectively
generate neural networks capable of good micro for our ranged units against a
group of hand-coded melee units. The evolved neural networks lead to kiting
behavior for the ranged units which is a common tactic used by professional
players in ranged versus melee skirmishes in popular real-time strategy games
like Starcraft. The evolved neural networks also generalized well to other
starting positions and numbers of units. We believe these results indicate the
potential of neuroevolution for generating effective micro in real-time
strategy games. | ['Sushil J. Louis', 'Siming Liu', 'Aavaas Gajurel', 'Daniel J Mendez'] | 2018-03-27 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [-6.09190837e-02 9.37678888e-02 4.94963139e-01 3.16028684e-01
3.32323194e-01 -8.65349889e-01 4.26920503e-01 -7.41311193e-01
-8.80439162e-01 9.36270297e-01 -4.68740761e-01 -2.64971912e-01
-2.61952907e-01 -1.14474499e+00 -3.00704390e-01 -7.23816693e-01
-4.96369630e-01 8.09886932e-01 3.45639199e-01 -1.58844316e+00
5.41206539e-01 5.56985497e-01 -1.63429141e+00 -4.38079005e-03
6.26471579e-01 2.65859097e-01 2.63992697e-01 1.36965406e+00
2.39748806e-01 1.00740016e+00 -1.56334317e+00 -4.07451928e-01
6.12542927e-01 -1.07273495e+00 -2.25373536e-01 -3.08534294e-01
-1.01398699e-01 1.35168836e-01 -2.10260242e-01 8.09829235e-01
8.53623509e-01 5.61758101e-01 4.06497657e-01 -1.19569290e+00
-1.70069993e-01 1.36616015e+00 -3.51715922e-01 4.44722205e-01
1.08425595e-01 3.40703040e-01 5.77257216e-01 1.30928874e-01
5.19824564e-01 1.28928614e+00 9.94299173e-01 1.12811387e+00
-8.86875629e-01 -9.64787424e-01 -5.06059468e-01 -3.14206243e-01
-1.28255308e+00 -5.24792820e-02 4.27044362e-01 3.12900990e-02
1.28136718e+00 6.58290386e-01 1.73670864e+00 9.81877446e-01
6.78722858e-01 3.69066387e-01 9.86874163e-01 -3.13604355e-01
3.92870545e-01 -4.13051158e-01 -8.45482230e-01 6.22940898e-01
3.52860063e-01 7.31281459e-01 -3.03028375e-01 -1.44864231e-01
1.52205753e+00 -5.65213859e-01 2.67080814e-02 1.61370710e-01
-9.19586241e-01 1.23560715e+00 2.99434096e-01 5.32504022e-01
-6.06370389e-01 5.77279031e-01 2.09019661e-01 7.70281911e-01
-3.64082068e-01 2.05015492e+00 -5.27920842e-01 -9.31610584e-01
-5.43760717e-01 5.89252293e-01 1.20890033e+00 5.84039807e-01
1.24335140e-01 1.08998179e+00 2.62219518e-01 7.49640822e-01
-4.32461172e-01 2.64644325e-01 8.95311832e-01 -1.29838681e+00
-1.15576498e-01 5.50793767e-01 -2.44194254e-01 -7.85631239e-01
-3.97211313e-01 -5.54678977e-01 -4.05834287e-01 1.02403033e+00
2.06873849e-01 -9.79722798e-01 -7.00614929e-01 1.80553377e+00
8.20486844e-02 7.09905252e-02 2.32786134e-01 4.74637806e-01
3.98511201e-01 5.67218781e-01 -4.65355098e-01 2.48925462e-02
8.39344561e-01 -9.10576642e-01 -2.83572584e-01 -9.52182561e-02
4.38261390e-01 -2.92755812e-01 9.46928859e-01 5.23201525e-01
-1.71146333e+00 -3.55553746e-01 -1.18692505e+00 1.19622242e+00
-2.69213587e-01 -6.92941904e-01 8.83832991e-01 1.13941860e+00
-1.27779591e+00 8.12756598e-01 -7.14087844e-01 -4.49862301e-01
-7.21925944e-02 9.09897327e-01 3.34614292e-02 9.68116939e-01
-1.38678813e+00 1.18361497e+00 7.31170654e-01 -1.98067129e-01
-1.10110211e+00 -3.12857270e-01 -5.35612226e-01 1.03468873e-01
5.30502200e-01 -5.30001163e-01 1.32390535e+00 -1.23212111e+00
-1.99555540e+00 5.79258800e-01 1.26193237e+00 -5.90548754e-01
4.23227400e-01 5.79201639e-01 -2.14174792e-01 -9.35846046e-02
-1.92064285e-01 6.95741117e-01 7.25422978e-01 -7.97042251e-01
-8.73008668e-01 2.41360486e-01 6.81736887e-01 5.41591406e-01
-3.07847530e-01 2.57890910e-01 3.38957608e-01 -1.11737752e+00
-5.39376855e-01 -1.09594500e+00 -3.47676784e-01 -7.45352030e-01
9.09033492e-02 1.95635799e-02 3.88557971e-01 3.46320234e-02
1.13387442e+00 -1.50057125e+00 8.95300210e-01 2.43198812e-01
3.65377516e-01 4.11530137e-01 -3.29896390e-01 6.18886590e-01
-2.92557664e-02 3.93306285e-01 1.17850818e-01 5.28198838e-01
-1.95109770e-01 6.55838013e-01 3.84181142e-01 -4.22046557e-02
-1.14164829e-01 1.08238590e+00 -8.84589076e-01 -1.11909539e-01
-3.02774817e-01 -6.30566701e-02 -8.15222144e-01 3.26500654e-01
-4.40169722e-02 2.33572692e-01 -4.84667748e-01 6.58840358e-01
-5.71542643e-02 5.94191730e-01 2.70699002e-02 6.98568702e-01
-4.74652737e-01 -3.95316154e-01 -7.08569050e-01 1.14899898e+00
-1.74466670e-01 5.74038565e-01 3.89520556e-01 -7.33039260e-01
1.04466486e+00 -2.72208229e-02 3.51095289e-01 -5.24118781e-01
8.99341881e-01 2.60981858e-01 1.27077317e+00 -1.13129899e-01
7.16538668e-01 -5.55295110e-01 -6.53503597e-01 6.09988332e-01
-2.03849748e-02 -1.20711601e+00 5.96621513e-01 -1.77496031e-01
1.52770603e+00 -9.73069668e-02 2.17562258e-01 -2.54880190e-01
1.38992921e-01 3.33558857e-01 7.44110286e-01 1.15363586e+00
-2.32840762e-01 2.27788270e-01 6.41970694e-01 -4.33184534e-01
-1.38290679e+00 -7.00340629e-01 6.71637118e-01 1.47213244e+00
-1.50417522e-01 -2.22522929e-01 -8.91477346e-01 -6.25912473e-02
-1.34644896e-01 8.80155504e-01 -1.06776929e+00 -6.45251334e-01
-1.08834743e+00 -6.43206000e-01 1.62975001e+00 4.01584864e-01
8.02210689e-01 -1.56827390e+00 -1.45534062e+00 4.41261739e-01
4.67305899e-01 -1.00102141e-01 -3.72618914e-01 5.16839325e-01
-4.85026360e-01 -9.22864676e-01 -7.67167330e-01 -7.83750772e-01
1.83070749e-01 -1.54620223e-02 1.07264972e+00 6.51069880e-01
-3.55653375e-01 1.78213254e-01 -6.05587065e-01 -7.86374509e-01
-9.96081471e-01 1.31224722e-01 2.77457416e-01 -9.52465653e-01
-3.68822545e-01 -6.93721294e-01 -8.73207152e-02 8.00205410e-01
-9.85790849e-01 1.19449519e-01 6.07036233e-01 1.27583647e+00
-1.99946940e-01 6.39321089e-01 3.30054700e-01 -3.75482053e-01
1.47306025e+00 -2.04578578e-01 -5.76153159e-01 1.85776651e-01
-1.18586212e-01 1.15025602e-02 6.94093108e-01 -1.19477451e+00
-8.70912194e-01 -6.05966210e-01 2.99699921e-02 -2.46876463e-01
4.51997310e-01 4.37818974e-01 2.71447003e-01 -7.44963706e-01
1.05396008e+00 1.25992015e-01 4.18452531e-01 1.63856149e-01
8.06785971e-02 3.24757546e-01 4.99950945e-01 -9.41981912e-01
7.91832268e-01 -2.49228403e-01 -1.47331938e-01 -6.52620554e-01
5.30292332e-01 4.89332289e-01 -1.79098740e-01 -5.11477947e-01
8.90341997e-01 -2.17531115e-01 -8.49486470e-01 9.36993241e-01
-4.65087444e-01 -7.85358131e-01 -8.88657093e-01 1.73319459e-01
-9.25724745e-01 -3.97881329e-01 -7.27164328e-01 -5.39641142e-01
-2.36129463e-01 -1.15000010e+00 2.36638457e-01 8.69771123e-01
-4.95321810e-01 -7.99394965e-01 5.07012129e-01 -6.24789521e-02
8.52896512e-01 6.27602518e-01 7.88927138e-01 -8.03216219e-01
2.54771948e-01 -3.27960372e-01 7.93147027e-01 1.03161544e-01
1.38212427e-01 5.05903959e-01 2.76066959e-01 -4.48913872e-01
5.34257144e-02 -5.50613463e-01 1.42932385e-01 1.73670352e-01
2.66132623e-01 -9.08501744e-02 3.15524428e-03 6.62306547e-01
7.62637019e-01 1.29749298e+00 7.39002466e-01 9.65506434e-01
2.22036660e-01 5.57365119e-01 5.73601365e-01 5.42197824e-01
-3.16332608e-01 3.85743707e-01 6.13198459e-01 5.73208481e-02
1.87829092e-01 3.72045189e-02 4.68862623e-01 8.12174439e-01
-1.06635141e+00 -4.42250907e-01 -8.31135929e-01 1.45546973e-01
-1.52167726e+00 -1.47912073e+00 4.89007413e-01 1.74148417e+00
9.06063318e-01 2.24053085e-01 8.09681714e-01 6.68828785e-02
1.03729701e+00 -1.91282764e-01 -6.83881938e-01 -1.20980191e+00
-5.38584590e-01 9.21909809e-01 6.45950615e-01 1.95895150e-01
-4.11218643e-01 1.29391623e+00 7.21006775e+00 8.80887270e-01
-1.08334398e+00 -2.79044986e-01 3.31981540e-01 -8.04886758e-01
-1.84210047e-01 -1.74905106e-01 -2.10269570e-01 4.50561285e-01
1.01343465e+00 -7.29630828e-01 9.18376446e-01 6.47667110e-01
2.00574994e-02 1.16498500e-01 -3.62973899e-01 5.45863271e-01
6.39741123e-02 -1.66967595e+00 -2.07403257e-01 2.59856969e-01
9.37635958e-01 -4.72078353e-01 1.87799528e-01 8.67344379e-01
1.44865596e+00 -1.49112940e+00 7.91572869e-01 5.10372641e-03
6.56566441e-01 -1.32855833e+00 6.17919326e-01 3.61870646e-01
-1.02809346e+00 -5.51963091e-01 -4.47726130e-01 -4.87656713e-01
1.16093948e-01 -9.78386998e-01 -7.57718027e-01 3.39109562e-02
6.55828774e-01 -1.35860354e-01 -4.67433244e-01 9.72359240e-01
-2.00303122e-01 3.68819356e-01 -3.90138179e-01 -8.30204308e-01
7.19686210e-01 -2.84453958e-01 9.26533341e-01 5.76367319e-01
5.58201730e-01 4.30843949e-01 -2.61870891e-01 6.40823364e-01
5.11691928e-01 -3.74019086e-01 -6.35423601e-01 -6.45392299e-01
5.87085247e-01 1.09485185e+00 -9.58535314e-01 -1.68430462e-01
5.61887562e-01 8.44174623e-01 6.67864382e-02 -2.20088556e-01
-1.21405327e+00 -8.50776672e-01 9.29026663e-01 3.80832143e-02
2.44275630e-01 -3.11789632e-01 5.22097796e-02 -7.50724256e-01
-1.11244500e+00 -1.67283559e+00 2.89327800e-01 -8.05643618e-01
-8.58635128e-01 1.12033892e+00 3.11444163e-01 -8.11110139e-01
-9.43850040e-01 -5.22346258e-01 -1.28569829e+00 4.83082950e-01
1.40220389e-01 -9.29102421e-01 -1.04020126e-01 5.80009460e-01
6.77541196e-01 -1.20461571e+00 5.82484722e-01 -5.14788449e-01
-8.55237722e-01 7.58114934e-01 -4.12767380e-02 1.92437526e-02
8.25157091e-02 -1.27823734e+00 7.82862067e-01 5.99282563e-01
-2.41782531e-01 6.08815253e-01 9.75350976e-01 -8.67695928e-01
-1.40966189e+00 -1.74702227e-01 -3.63934696e-01 1.97405107e-02
8.35302055e-01 -1.34418264e-01 -1.29463673e-01 2.91219980e-01
6.21890843e-01 -8.45111370e-01 6.33561611e-01 -2.88943857e-01
4.32509422e-01 4.17787671e-01 -1.35812521e+00 1.29140508e+00
1.29630315e+00 3.07586700e-01 -7.48976111e-01 -2.73379415e-01
6.07971549e-01 -6.50282979e-01 -5.95568001e-01 2.03885168e-01
8.11204851e-01 -1.07094395e+00 8.85931849e-01 -1.06480396e+00
3.07295740e-01 -7.52377650e-03 1.51606262e-01 -1.91894376e+00
-6.02032661e-01 -1.26543999e+00 5.09743333e-01 7.42804289e-01
2.35952035e-01 -9.88539279e-01 1.04198968e+00 4.44328904e-01
-2.60256529e-01 -4.89074349e-01 -9.13745522e-01 -9.68852758e-01
4.76704180e-01 2.82512635e-01 6.85785592e-01 9.16905284e-01
1.46443937e-02 2.63096746e-02 -4.77861762e-01 -6.20316446e-01
2.71913797e-01 -5.10453522e-01 8.04431438e-01 -1.14088058e+00
-1.02523327e+00 -1.03888655e+00 -7.89933562e-01 -2.09300995e-01
-3.48154176e-03 -4.13114250e-01 -1.16089858e-01 -7.45190680e-01
-5.86453557e-01 -5.61524510e-01 7.08469376e-02 5.36484301e-01
5.35382554e-02 1.62214488e-01 6.05319262e-01 -1.79344825e-02
1.22318886e-01 2.20767319e-01 1.36315858e+00 2.65415553e-02
-5.70777953e-01 3.84378582e-02 -1.09547901e+00 8.07609916e-01
1.06639898e+00 -4.38117474e-01 -7.30885029e-01 -1.75165296e-01
5.17199039e-01 2.17322707e-01 -3.83741349e-01 -1.29327059e+00
9.56607088e-02 -7.19213486e-01 2.52299756e-01 2.01149985e-01
3.19012552e-01 -3.64750564e-01 7.19301522e-01 1.24641562e+00
-1.53042719e-01 1.03342366e+00 4.51276392e-01 -5.44427382e-03
7.74387494e-02 -3.83490533e-01 7.26087809e-01 -7.66179025e-01
-8.50249052e-01 -2.20449030e-01 -1.27276385e+00 9.22117457e-02
1.70409727e+00 -9.97454524e-01 -3.57328176e-01 -7.40811169e-01
-6.32794321e-01 2.40801632e-01 7.22029805e-01 2.37286612e-01
5.50749540e-01 -1.03110993e+00 -9.60916996e-01 1.57423064e-01
-5.32567501e-01 -4.17200774e-01 -2.43191525e-01 2.36434981e-01
-1.65580535e+00 -3.46424937e-01 -1.57821894e+00 9.60253328e-02
-1.55168796e+00 5.66806123e-02 1.07398856e+00 -1.04289085e-01
-2.19538361e-01 1.60834610e+00 -3.68838906e-02 -4.98029470e-01
-2.23722443e-01 8.01155269e-02 -3.93944353e-01 8.84554163e-03
2.70441890e-01 5.79984844e-01 -4.93094206e-01 -4.22954947e-01
-1.70915484e-01 2.04566330e-01 1.40339077e-01 -5.58578134e-01
1.39947402e+00 6.74530923e-01 -2.32060179e-01 -3.27510089e-02
2.25327715e-01 8.75209868e-02 -9.32517827e-01 8.28577757e-01
-4.57106948e-01 -4.33647960e-01 -4.23687220e-01 -7.88395107e-01
-1.23310137e+00 8.86128023e-02 1.20891854e-01 3.24579597e-01
1.24028897e+00 -5.71086645e-01 4.17275071e-01 7.31826007e-01
7.69849777e-01 -1.13418603e+00 5.77149093e-01 8.17404687e-01
1.07403982e+00 -1.66766226e-01 -3.82912368e-01 3.43271166e-01
-1.01142216e+00 1.15449321e+00 1.21090221e+00 -7.20651984e-01
-7.72338882e-02 9.50398088e-01 1.23823069e-01 -1.83182865e-01
-1.00317276e+00 -2.06462368e-01 -4.91597861e-01 1.12047160e+00
5.22618070e-02 -3.56588475e-02 -6.63160145e-01 2.36578271e-01
-1.14830613e+00 -5.60635746e-01 9.71745551e-01 1.53499818e+00
-6.51323140e-01 -1.33627737e+00 -7.03307390e-01 2.60753751e-01
-2.71451175e-01 -5.29522263e-02 -1.17462718e+00 1.50056183e+00
4.01169747e-01 7.49994993e-01 2.95595348e-01 -1.06896091e+00
3.96944284e-01 -5.05430222e-01 7.57393062e-01 -5.36549985e-01
-1.78695118e+00 -1.87536478e-02 4.94720966e-01 -1.21871814e-01
1.29915595e-01 -4.14278477e-01 -1.14315665e+00 -1.09111154e+00
-2.16393113e-01 6.35859728e-01 2.90279746e-01 1.63451523e-01
-1.24234632e-01 7.52690554e-01 4.05612528e-01 -1.18879354e+00
-6.97632432e-01 -9.19384241e-01 -9.09555912e-01 9.71970260e-02
-5.46521068e-01 -6.94725037e-01 -1.85550824e-01 -6.74318671e-01] | [3.573101043701172, 1.6348488330841064] |
de438c0c-7fdd-4fc7-86e8-a9fd1a854d0d | mahtm-a-multi-agent-framework-for | 2303.08447 | null | https://arxiv.org/abs/2303.08447v1 | https://arxiv.org/pdf/2303.08447v1.pdf | MAHTM: A Multi-Agent Framework for Hierarchical Transactive Microgrids | Integrating variable renewable energy into the grid has posed challenges to system operators in achieving optimal trade-offs among energy availability, cost affordability, and pollution controllability. This paper proposes a multi-agent reinforcement learning framework for managing energy transactions in microgrids. The framework addresses the challenges above: it seeks to optimize the usage of available resources by minimizing the carbon footprint while benefiting all stakeholders. The proposed architecture consists of three layers of agents, each pursuing different objectives. The first layer, comprised of prosumers and consumers, minimizes the total energy cost. The other two layers control the energy price to decrease the carbon impact while balancing the consumption and production of both renewable and conventional energy. This framework also takes into account fluctuations in energy demand and supply. | ['Martin Takac', 'Yongli Zhu', 'Roberto Gutierrez', 'Nicolas Cuadrado'] | 2023-03-15 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-5.16234040e-01 -1.11944601e-01 -3.41084659e-01 2.58728832e-01
-3.73710752e-01 -6.64157450e-01 2.84434378e-01 2.66668290e-01
-6.52212277e-02 1.27655673e+00 -1.06877409e-01 -1.24321890e-03
-2.64298320e-01 -1.28588533e+00 -1.55445812e-02 -1.19664133e+00
1.63089223e-02 2.99335390e-01 -3.96612495e-01 -9.26052928e-02
1.41638368e-01 3.84085715e-01 -1.22896600e+00 -6.25604093e-01
1.18230057e+00 1.20527720e+00 2.61592448e-01 3.03810000e-01
4.76678878e-01 3.02264929e-01 -6.93017721e-01 4.25427407e-01
3.75814468e-01 -4.79582191e-01 -1.27549633e-01 -1.96517885e-01
-8.29403758e-01 -5.10543108e-01 4.53723580e-01 1.38844407e+00
6.25850201e-01 2.99209982e-01 5.95707834e-01 -1.81797457e+00
-7.58330941e-01 5.99214315e-01 -5.35936654e-01 8.67416039e-02
-6.18861467e-02 3.61916274e-01 1.17846715e+00 -1.11482047e-01
-9.29283649e-02 8.37586820e-01 -1.58068776e-01 5.02069294e-01
-1.06407952e+00 -5.18642783e-01 1.77558690e-01 5.37833512e-01
-1.14901030e+00 4.93332818e-02 5.21717310e-01 -1.66327238e-01
1.56775296e+00 5.79421878e-01 1.30526674e+00 1.96529791e-01
5.26315928e-01 9.50023308e-02 1.27205718e+00 -5.89504182e-01
8.15030992e-01 2.41353307e-02 -1.66382402e-01 -2.73520648e-01
7.54348874e-01 2.85497308e-01 -2.16506347e-02 -1.55859450e-02
1.69552505e-01 -2.62082487e-01 -2.55432904e-01 -2.56586343e-01
-6.48122668e-01 7.61428893e-01 3.21986794e-01 4.12302107e-01
-7.62865126e-01 2.08722487e-01 1.87302753e-02 1.30982891e-01
1.17962457e-01 4.12338167e-01 -3.80895913e-01 1.74153790e-01
-4.41672921e-01 -1.89026475e-01 6.89052463e-01 8.18406045e-01
5.08010089e-01 7.23149836e-01 2.43463084e-01 3.58858109e-01
7.02347577e-01 7.61298954e-01 4.94753122e-01 -1.19107759e+00
-4.80597690e-02 5.78138173e-01 8.61340761e-01 -3.86671364e-01
-3.72953713e-01 -2.72519529e-01 -6.03121400e-01 7.71993697e-01
-1.60076886e-01 -7.72312164e-01 -3.42764348e-01 1.59986424e+00
3.41555893e-01 -7.51034796e-01 1.25865102e-01 6.98987901e-01
-1.32615998e-01 9.43047941e-01 3.47938001e-01 -1.00168943e+00
1.25665522e+00 -8.61942589e-01 -1.15145099e+00 1.77741051e-01
-1.72251329e-01 -3.79961222e-01 3.56654495e-01 4.80540097e-02
-1.50639307e+00 -9.27867927e-03 -1.32990658e+00 4.46533740e-01
-7.87568808e-01 -4.09574956e-02 -1.88765824e-01 6.17322862e-01
-1.06009734e+00 2.75112242e-01 -6.93791032e-01 -1.93608582e-01
-2.51688987e-01 5.61784863e-01 7.41652727e-01 8.05140674e-01
-1.27629352e+00 1.46731901e+00 5.97453535e-01 1.06856592e-01
-6.05220258e-01 -5.84437907e-01 -6.57021642e-01 4.16969359e-01
4.73110437e-01 -8.44419599e-01 1.36650181e+00 -8.57232809e-01
-1.80164337e+00 -7.73053169e-02 3.64370346e-01 -4.87975180e-01
5.96840739e-01 1.11065820e-01 -6.39755130e-01 -1.54673353e-01
8.37450102e-03 -1.58806406e-02 2.95997113e-01 -1.06166053e+00
-1.14041567e+00 -5.17664969e-01 -1.23942532e-01 8.55828166e-01
-3.83260012e-01 -4.55804050e-01 5.38893938e-01 -2.31626377e-01
-8.49759281e-01 -8.99197817e-01 -3.64555240e-01 -7.02894568e-01
-1.99855313e-01 -6.17646515e-01 1.03154516e+00 -5.12960792e-01
9.32687998e-01 -1.50005651e+00 2.79923946e-01 2.82180876e-01
-3.46839815e-01 -4.34891209e-02 1.98894948e-01 8.96974742e-01
1.77883193e-01 1.78314239e-01 4.09904607e-02 2.32021183e-01
4.94068414e-01 3.12563926e-01 2.49488160e-01 1.54282168e-01
-5.04835322e-02 6.45029306e-01 -9.49391067e-01 1.96210518e-01
7.11904228e-01 2.21679345e-01 1.61874697e-01 4.04537827e-01
-6.74036860e-01 1.65445045e-01 -5.29837728e-01 6.28719270e-01
4.77304995e-01 9.05095264e-02 6.58702731e-01 -9.88331810e-02
-6.59873307e-01 -5.46366572e-02 -1.55178118e+00 9.30030167e-01
-7.36841619e-01 6.42621815e-02 6.45457029e-01 -9.52034771e-01
4.71667618e-01 2.42145732e-01 9.13865030e-01 -1.19731200e+00
-6.28604814e-02 4.61357474e-01 1.58672389e-02 -3.54867160e-01
5.43957472e-01 -2.39612505e-01 1.59059525e-01 6.84414506e-01
-1.59563377e-01 -1.28799558e-01 5.63804924e-01 -2.80025244e-01
3.41090173e-01 1.99365422e-01 8.84751379e-01 -8.23932350e-01
5.96090794e-01 -1.64130419e-01 7.99433470e-01 -7.30847865e-02
-3.92910540e-01 -7.41204798e-01 5.68571016e-02 -8.19882303e-02
-8.44608665e-01 -9.65202808e-01 3.31183076e-01 8.46207142e-01
5.44469893e-01 2.75829136e-01 -6.50491953e-01 -1.46181807e-01
1.55663028e-01 1.54618025e+00 -2.74850994e-01 8.39130804e-02
-4.32190627e-01 -1.16052163e+00 -5.20586014e-01 2.43487433e-01
3.71396482e-01 -7.20903456e-01 -1.57211661e+00 1.78568944e-01
-7.98666030e-02 -5.85385501e-01 -5.57369947e-01 4.04256701e-01
-1.22461133e-01 -1.16912138e+00 -4.76403415e-01 -2.78457522e-01
6.02527261e-01 1.23355515e-01 1.03171790e+00 -7.83816352e-02
-2.27416694e-01 4.46343124e-01 6.53195530e-02 -1.01594234e+00
-2.31262103e-01 -8.09184313e-02 3.25318038e-01 -2.25496784e-01
2.55971346e-02 -4.45252448e-01 -9.65755999e-01 1.84785500e-01
-6.63190246e-01 -7.43637159e-02 1.78626239e-01 4.97370690e-01
8.33431900e-01 8.36612761e-01 1.35398567e+00 -5.17601483e-02
9.63352978e-01 -5.79071522e-01 -1.24543703e+00 5.60216188e-01
-1.32739723e+00 -4.93920408e-02 9.89032209e-01 1.32108539e-01
-1.14277005e+00 -2.33090341e-01 2.99034894e-01 1.54113278e-01
1.77662641e-01 5.15096961e-03 -6.36631668e-01 6.49984479e-02
-5.00589788e-01 2.46287927e-01 3.03200800e-02 -1.76924422e-01
4.20950741e-01 4.17836666e-01 1.82843387e-01 -2.86378413e-01
7.63175726e-01 -5.34380488e-02 2.24065438e-01 -4.51460153e-01
-9.57905278e-02 -1.82708487e-01 -2.06368372e-01 -5.85012913e-01
7.95050502e-01 -8.44104826e-01 -1.40412438e+00 4.02616978e-01
-6.09039545e-01 -2.55553365e-01 -6.31307244e-01 2.73836106e-01
-6.50980830e-01 3.65263671e-02 1.95713751e-02 -1.21205592e+00
-8.49074543e-01 -1.05121148e+00 2.35892348e-02 9.03717816e-01
8.55376646e-02 -1.04956806e+00 2.42396653e-01 -5.92228621e-02
7.97944188e-01 7.38308787e-01 1.14733016e+00 -1.19169749e-01
-8.37323546e-01 4.31833148e-01 2.27596253e-01 5.58612287e-01
7.65719473e-01 3.41725722e-02 -4.38402712e-01 -8.92811477e-01
5.78110032e-02 -1.47169352e-01 -1.09168433e-01 4.23871100e-01
6.25525355e-01 -7.68514216e-01 -1.49283158e-02 -1.12708844e-01
2.45906210e+00 1.00966430e+00 1.83020383e-01 6.41736209e-01
8.70299339e-02 5.34174979e-01 6.97668016e-01 7.13770270e-01
7.11168349e-01 4.47787613e-01 1.06737900e+00 2.11175799e-01
2.88474917e-01 3.11286926e-01 3.10047328e-01 7.06086218e-01
-1.44084662e-01 -6.20254517e-01 -4.56453025e-01 7.10613251e-01
-1.97546625e+00 -8.80621195e-01 2.84216583e-01 2.13235545e+00
4.45922852e-01 -2.45572731e-01 4.54824716e-01 3.81964296e-02
6.09030902e-01 1.10668309e-01 -1.15578938e+00 -7.89885283e-01
-7.29324371e-02 -2.93954998e-01 6.82455420e-01 4.92030561e-01
-3.27970952e-01 -2.08008531e-02 5.87644958e+00 3.52047086e-01
-1.02298641e+00 2.58340128e-02 6.19311929e-01 -5.19011378e-01
-4.67338413e-01 -2.04316452e-01 -3.84269744e-01 9.33552265e-01
1.30924714e+00 -1.16830230e+00 1.06783283e+00 4.59889442e-01
1.04489768e+00 -7.00664222e-01 -8.81941438e-01 2.15392664e-01
-4.40814286e-01 -1.27336156e+00 -2.38902479e-01 2.16853991e-01
1.03560865e+00 1.35725811e-01 -4.33447301e-01 -3.75027359e-02
6.96047723e-01 -6.07699394e-01 1.03844786e+00 7.71011829e-01
5.71509637e-02 -1.52317405e+00 6.19533062e-01 5.90797603e-01
-1.46988416e+00 -5.91608465e-01 1.89280629e-01 4.67881709e-02
4.86713797e-01 4.39755261e-01 -2.22131923e-01 9.87980664e-01
7.85372257e-01 2.77425852e-02 1.52111396e-01 6.95201218e-01
-1.30413741e-01 -1.17592007e-01 -3.53367239e-01 -3.63152266e-01
1.96787506e-01 -9.76895511e-01 3.56103212e-01 7.66523600e-01
3.00884008e-01 3.18885595e-01 3.61260176e-01 8.73543739e-01
-1.09212890e-01 3.35884206e-02 -2.91164786e-01 6.69183433e-02
9.38468754e-01 1.53937387e+00 -5.96323788e-01 -2.22995758e-01
-2.02458918e-01 3.89976740e-01 -1.80759698e-01 3.97270828e-01
-6.49962962e-01 -4.95818794e-01 7.14445412e-01 -3.23444873e-01
1.07285045e-01 -4.13634777e-02 -5.68538964e-01 -6.94619417e-01
1.92041602e-02 -5.70098341e-01 4.59492207e-01 -4.00481492e-01
-1.35209751e+00 -2.26615947e-02 -2.88387365e-03 -9.13568139e-01
-5.54174125e-01 -2.51318127e-01 -9.01099682e-01 1.25532365e+00
-2.29715323e+00 -7.15709329e-01 -4.63364907e-02 3.76320332e-01
7.12559342e-01 -7.42957145e-02 8.74464631e-01 8.55975673e-02
-1.15334582e+00 -2.30474278e-01 7.73766696e-01 -4.61928934e-01
-1.45841360e-01 -1.74945486e+00 -4.44235116e-01 9.83455896e-01
-6.95455492e-01 6.41669147e-03 6.44553065e-01 -4.01323229e-01
-1.75158727e+00 -9.75587726e-01 3.51635993e-01 6.43139005e-01
8.62142920e-01 2.75565892e-01 -3.51668566e-01 1.36348590e-01
1.45858109e+00 -5.81831217e-01 8.00232947e-01 -7.62504041e-01
1.68676570e-01 -3.63957137e-01 -1.67311084e+00 4.98823583e-01
2.84246984e-03 -7.90460855e-02 -9.22120512e-02 7.62132406e-02
3.30328882e-01 1.34924307e-01 -1.20289516e+00 1.85247362e-01
5.31550169e-01 -4.24422711e-01 5.08721769e-01 -2.02570438e-01
-2.99248070e-01 -3.87273401e-01 -2.94489384e-01 -2.17949891e+00
-1.99846655e-01 -7.93816805e-01 -6.52621031e-01 1.35885799e+00
3.39611053e-01 -9.75452900e-01 2.27873608e-01 8.41032326e-01
3.86366062e-02 -6.89924955e-01 -1.25821161e+00 -8.52161586e-01
1.77633867e-01 4.60752547e-01 8.99090707e-01 7.87946224e-01
3.53306979e-01 3.45409721e-01 -1.26470238e-01 5.17198324e-01
1.30137968e+00 3.19979787e-01 -1.08304627e-01 -7.26571858e-01
1.93762854e-01 -6.54511333e-01 5.10314107e-01 1.28351703e-01
-1.67630941e-01 -3.68969291e-01 -4.92417067e-02 -2.21383739e+00
-6.59885556e-02 7.24493293e-03 -7.59398520e-01 5.39135754e-01
1.65130228e-01 -7.73062706e-02 7.59795725e-01 2.72317361e-02
-1.72888637e-01 9.02290881e-01 9.22131717e-01 -2.03877777e-01
-2.99262643e-01 -4.61157933e-02 -6.09639108e-01 6.77577376e-01
1.36284876e+00 -8.76583531e-02 -7.10769176e-01 -4.01516825e-01
1.26661286e-02 3.48160177e-01 -2.37389565e-01 -7.56780505e-01
2.04364926e-01 -1.13336480e+00 5.43448515e-02 -7.07138956e-01
-1.11152597e-01 -1.43547904e+00 7.77800322e-01 1.02640355e+00
-3.29446681e-02 4.92398441e-01 -2.05561817e-01 6.36945844e-01
1.87301219e-01 -1.07291892e-01 1.27987778e+00 -2.82660097e-01
-6.82904899e-01 -1.47284418e-01 -7.74977684e-01 -4.78113264e-01
1.94710004e+00 1.00558080e-01 -7.16425061e-01 -5.63705340e-02
-6.58358455e-01 1.23667037e+00 4.14110512e-01 5.72452009e-01
4.38542068e-02 -1.15493166e+00 -4.11013365e-01 -4.37382668e-01
-4.53613728e-01 -4.70664829e-01 2.15991586e-02 1.88928649e-01
-3.40304494e-01 3.86089504e-01 -6.93978131e-01 5.89393489e-02
-8.88404906e-01 3.68460894e-01 1.02103615e+00 -6.14511609e-01
8.47438872e-02 -1.09431207e-01 -5.12023032e-01 4.25466932e-02
5.88092804e-02 -1.86003149e-02 -2.42375538e-01 6.60738468e-01
3.92230570e-01 1.24260712e+00 1.05122663e-01 -2.81813383e-01
-4.64096069e-01 3.68863940e-01 6.31022334e-01 -2.60028802e-03
1.44965875e+00 -7.61448801e-01 -3.50422025e-01 4.53710020e-01
6.29945755e-01 -2.33388811e-01 -1.05036318e+00 3.60776573e-01
3.16088319e-01 -1.10061243e-01 3.80657643e-01 -1.54125118e+00
-1.19148946e+00 7.36917332e-02 7.65326321e-01 1.09051430e+00
1.61869860e+00 -8.32404137e-01 5.68182170e-01 2.21080184e-02
6.13032520e-01 -1.87646580e+00 -1.01052113e-01 4.52972725e-02
7.16080070e-01 -7.95712411e-01 1.56387061e-01 9.24143642e-02
-3.95004183e-01 1.12539995e+00 7.46393740e-01 -1.90759003e-01
5.96207440e-01 3.32366049e-01 -2.82522917e-01 2.60669649e-01
-8.93370271e-01 -1.80854902e-01 -3.68321270e-01 5.66097856e-01
1.66037504e-03 8.75152111e-01 -8.67271185e-01 2.30644226e-01
2.97121376e-01 -2.01739490e-01 8.28693867e-01 7.95900345e-01
-6.05583668e-01 -1.12023342e+00 -4.92285877e-01 3.79254699e-01
-3.91311496e-01 4.22709256e-01 9.72778536e-03 7.88985729e-01
5.95183253e-01 1.20918787e+00 1.20180793e-01 4.23853546e-01
6.58045530e-01 -1.70109510e-01 1.37167368e-02 -2.22920731e-01
-5.37667453e-01 3.33495706e-01 1.14992127e-01 -7.96414912e-02
-6.48942828e-01 -7.32635975e-01 -1.47694206e+00 -2.42977262e-01
-3.59261870e-01 7.20690429e-01 1.12894857e+00 8.40051174e-01
2.41372824e-01 8.00594628e-01 1.43653333e+00 -7.16081798e-01
-1.05813372e+00 -6.22447073e-01 -1.00929308e+00 -4.35320348e-01
4.47729409e-01 -3.03410351e-01 -4.69574958e-01 -3.18072170e-01] | [5.59123420715332, 2.589179039001465] |
8e7129dd-43d0-4c3f-b382-7fe705255fbc | visual-appearance-based-person-retrieval-in | 1910.14565 | null | https://arxiv.org/abs/1910.14565v1 | https://arxiv.org/pdf/1910.14565v1.pdf | Visual Appearance Based Person Retrieval in Unconstrained Environment Videos | Visual appearance-based person retrieval is a challenging problem in surveillance. It uses attributes like height, cloth color, cloth type and gender to describe a human. Such attributes are known as soft biometrics. This paper proposes person retrieval from surveillance video using height, torso cloth type, torso cloth color and gender. The approach introduces an adaptive torso patch extraction and bounding box regression to improve the retrieval. The algorithm uses fine-tuned Mask R-CNN and DenseNet-169 for person detection and attribute classification respectively. The performance is analyzed on AVSS 2018 challenge II dataset and it achieves 11.35% improvement over state-of-the-art based on average Intersection over Union measure. | ['Mehul S. Raval', 'Shivansh Dave', 'Hiren Galiyawala'] | 2019-10-31 | null | null | null | null | ['person-retrieval'] | ['computer-vision'] | [-2.65134007e-01 -2.83048034e-01 1.70802921e-01 -5.48964858e-01
-5.16448796e-01 -6.07190132e-01 5.21833003e-01 2.72974133e-01
-5.00557244e-01 5.82329631e-01 5.43201119e-02 4.21843886e-01
2.10322171e-01 -7.25589812e-01 -5.04806042e-01 -6.22304082e-01
-6.00128174e-02 5.44234753e-01 -2.40038875e-02 -2.10893750e-01
-2.53819823e-01 4.89906460e-01 -1.48671484e+00 3.73513028e-02
2.27128044e-01 1.46301699e+00 -8.74727964e-01 9.19991136e-01
4.88966376e-01 9.81509462e-02 -8.98253500e-01 -8.31468821e-01
8.19835901e-01 2.49190986e-01 -3.01237673e-01 1.97479483e-02
1.19923329e+00 -7.36094892e-01 -5.68225741e-01 8.52220297e-01
8.37850571e-01 -9.33686942e-02 1.01777124e+00 -1.22856045e+00
-9.39670980e-01 3.66222896e-02 -1.10104442e+00 2.42636368e-01
7.67924070e-01 1.56338960e-01 4.45042163e-01 -4.63039637e-01
4.57113177e-01 1.57068706e+00 1.08706343e+00 7.74730086e-01
-8.59552503e-01 -7.19672799e-01 -2.42073163e-02 1.45287663e-01
-1.89062321e+00 -2.37304956e-01 4.51019257e-01 -4.18041289e-01
8.08005691e-01 5.65267205e-01 1.15544534e+00 1.09249556e+00
6.89732954e-02 8.19346309e-01 1.23929930e+00 -3.16446900e-01
-3.39599729e-01 3.79119724e-01 4.62987006e-01 1.30446482e+00
6.71224356e-01 1.98607072e-02 -1.29233226e-01 -4.11006719e-01
7.17921197e-01 2.41613403e-01 2.44771317e-02 1.16221618e-03
-7.16832697e-01 5.00613868e-01 4.61551100e-01 -3.68122607e-01
-2.57384777e-01 3.37303042e-01 5.58765888e-01 1.08228184e-01
4.84539390e-01 -2.78373510e-02 -2.41549984e-01 3.84791642e-01
-9.86388266e-01 5.61592758e-01 6.32839203e-01 1.28459346e+00
2.00383916e-01 2.90728454e-02 -8.50557685e-01 9.87605512e-01
4.92544234e-01 1.41884232e+00 -1.29486203e-01 -4.66532230e-01
-1.41580431e-02 9.81492817e-01 2.31135055e-01 -1.17200041e+00
-4.13681686e-01 -1.26606315e-01 -8.20977807e-01 -1.48705944e-01
4.94892538e-01 -1.66387051e-01 -1.67043340e+00 1.46922755e+00
6.62811518e-01 -1.82987198e-01 -4.22343433e-01 1.23044753e+00
1.47831762e+00 4.91639256e-01 1.82347283e-01 1.24249667e-01
1.94408298e+00 -7.24826515e-01 -5.98900318e-01 1.40490457e-02
-2.51056343e-01 -5.83988070e-01 3.95965368e-01 2.02068925e-01
-1.00387335e+00 -6.65968955e-01 -9.25983727e-01 4.83455136e-03
-8.60018492e-01 6.67396247e-01 4.65671986e-01 1.26030099e+00
-9.29718256e-01 -4.13089432e-03 -2.75534719e-01 -8.21558952e-01
6.52798593e-01 7.50354528e-01 -4.31943238e-01 1.57594740e-01
-1.07043266e+00 6.28140390e-01 -1.54610708e-01 1.76346853e-01
-1.03420675e+00 -3.96047622e-01 -8.34042370e-01 -3.43005031e-01
9.63228345e-02 -1.13885260e+00 6.50428534e-01 -7.11532652e-01
-1.27254725e+00 1.59021568e+00 8.04849118e-02 -4.06119585e-01
6.86074972e-01 -2.20988512e-01 -4.83760685e-01 3.80071908e-01
-1.28666818e-01 5.97345531e-01 1.36663938e+00 -1.07183683e+00
-4.64842230e-01 -1.02899694e+00 -6.68742135e-02 8.31445903e-02
-4.01355684e-01 3.14847648e-01 -7.60061681e-01 -5.97341776e-01
-1.45700231e-01 -1.02439821e+00 1.43290848e-01 2.68878192e-01
-5.20246387e-01 -4.01337475e-01 7.98480868e-01 -1.17855859e+00
8.65702510e-01 -1.53324389e+00 -7.16057271e-02 5.07277369e-01
4.96414512e-01 3.47100079e-01 2.82293499e-01 -1.77092671e-01
4.56450075e-01 -1.57956973e-01 1.62243992e-01 -4.66933876e-01
1.95621073e-01 -3.99216324e-01 3.05583805e-01 1.14818192e+00
-5.43644056e-02 9.31634367e-01 -3.50949556e-01 -9.27781880e-01
3.72553945e-01 8.87161255e-01 6.96508139e-02 5.03155112e-01
4.34439719e-01 6.11133836e-02 -4.03604209e-01 1.48656619e+00
9.46841180e-01 2.67559677e-01 -2.44320139e-01 -6.07465625e-01
2.38524035e-01 -8.23822439e-01 -1.00002766e+00 1.24439240e+00
1.55251652e-01 7.54675150e-01 2.97796249e-01 -6.17516935e-01
1.17535686e+00 2.86960095e-01 3.68316203e-01 -3.86020511e-01
7.67646194e-01 -2.14196131e-01 -3.87630999e-01 -6.40145898e-01
4.31477666e-01 3.58662605e-01 -1.97058544e-01 -3.13653558e-01
-7.74204880e-02 3.84219676e-01 -3.42074595e-02 4.24871780e-02
7.65000999e-01 1.95706949e-01 3.40793639e-01 -4.44903493e-01
7.23677099e-01 -1.34481952e-01 3.75390232e-01 8.99800777e-01
-8.00634623e-01 7.41473496e-01 -2.17997096e-02 -1.11053967e+00
-1.35589468e+00 -1.12819707e+00 -2.28584051e-01 1.19075871e+00
3.02513301e-01 1.49801029e-02 -1.23003483e+00 -6.78441048e-01
4.58906174e-01 -4.11835134e-01 -1.34732401e+00 -5.93575370e-03
-6.19688988e-01 -7.80268788e-01 1.01282859e+00 6.30887747e-01
9.50198770e-01 -7.84592986e-01 -6.43672764e-01 -4.17029619e-01
8.14364925e-02 -1.21553195e+00 -5.96271038e-01 -6.75680041e-01
-2.56778330e-01 -1.21901882e+00 -1.32730985e+00 -5.40727437e-01
8.83622944e-01 5.78494370e-02 1.07986557e+00 2.94466764e-01
-1.16995180e+00 9.93852437e-01 -2.36639738e-01 -7.97218382e-01
3.91509831e-01 -2.09165793e-02 3.60670805e-01 3.94019753e-01
7.15836287e-01 2.38802105e-01 -1.26408803e+00 1.08628988e-01
-3.24670136e-01 -4.04561788e-01 4.74633723e-01 5.76393485e-01
4.60148811e-01 -5.56882620e-01 4.36325558e-02 -6.34910285e-01
3.65730345e-01 -1.75080359e-01 -5.99545419e-01 7.55566776e-01
-3.58670950e-01 -5.48748493e-01 1.92795753e-01 -6.02040529e-01
-7.23280489e-01 8.15591887e-02 3.03148568e-01 -5.23353279e-01
-3.47291082e-01 -5.19832313e-01 -1.32772192e-01 -3.13517988e-01
4.77940381e-01 1.52265579e-01 -1.74554318e-01 -2.26545915e-01
-1.29268467e-02 8.41413379e-01 5.28309405e-01 -5.25190949e-01
8.86027992e-01 7.44254053e-01 5.19401133e-02 -9.49250162e-01
-7.61501253e-01 -7.14943111e-01 -6.32346153e-01 -6.90723121e-01
1.28273141e+00 -1.01042080e+00 -1.37691224e+00 6.93044722e-01
-1.02189481e+00 1.45494238e-01 2.48827755e-01 2.73096003e-02
-5.50201871e-02 1.25514507e-01 -5.88030398e-01 -1.43046987e+00
-1.24919701e+00 -7.79432893e-01 1.64563549e+00 6.54716015e-01
2.16022655e-01 -5.42211771e-01 -2.92153746e-01 6.81632578e-01
4.14662302e-01 9.88932312e-01 1.56020578e-02 -5.24270892e-01
-4.14064676e-01 -9.39647913e-01 -6.68769538e-01 -1.24735542e-01
-2.17563361e-01 2.89236028e-02 -1.09020424e+00 -6.63783729e-01
-7.29592741e-01 -2.05529690e-01 1.07033372e+00 6.94790781e-01
1.19084203e+00 -3.82278234e-01 -5.68194091e-01 9.23176646e-01
1.45938551e+00 -7.29533285e-02 4.12425965e-01 1.60885617e-01
1.02534449e+00 6.15523279e-01 3.65715623e-01 6.37086868e-01
5.10857642e-01 8.48441601e-01 1.12282872e-01 -2.50981539e-01
-3.72898966e-01 2.16528922e-01 6.32565618e-02 -2.06261352e-02
-6.56166017e-01 -1.89168453e-01 -8.71596992e-01 4.06692833e-01
-1.70464790e+00 -8.83476973e-01 -7.38709569e-02 2.18301606e+00
3.90927047e-01 -1.61741942e-01 8.09657931e-01 -2.82355785e-01
9.88374233e-01 -8.66523907e-02 -3.30937833e-01 -1.13489814e-01
-3.20167154e-01 -2.12124479e-03 1.09541297e+00 1.87567964e-01
-1.76589346e+00 8.75860929e-01 6.03121233e+00 5.57475328e-01
-6.61333978e-01 1.82070449e-01 8.38911474e-01 -6.18017204e-02
7.04082191e-01 -7.50183344e-01 -1.23209834e+00 2.01300919e-01
3.65350872e-01 3.85600775e-01 4.79650408e-01 6.16656363e-01
-4.24369097e-01 -4.22958583e-02 -9.61394846e-01 1.48298120e+00
7.71861672e-01 -6.66618645e-01 -7.72453472e-02 -1.45723388e-01
5.47550619e-01 -5.19006610e-01 2.65228957e-01 3.05046111e-01
-1.46726444e-01 -1.17934406e+00 7.31771886e-01 9.66688752e-01
1.08280885e+00 -7.18393922e-01 9.99160469e-01 -3.63879025e-01
-1.67306793e+00 -1.18381470e-01 -5.31012237e-01 4.55888242e-01
-2.18359500e-01 -1.75810814e-01 -6.39164388e-01 3.39460284e-01
1.31704330e+00 2.86482215e-01 -9.59390521e-01 1.21003866e+00
4.40719455e-01 3.37820917e-01 -4.90354836e-01 -2.63902664e-01
-1.63418084e-01 9.44114551e-02 7.04694927e-01 1.82589912e+00
-7.16087967e-02 3.77536833e-01 5.76099932e-01 6.22384608e-01
-8.45717341e-02 1.35373905e-01 -7.02189386e-01 4.82193321e-01
3.17178160e-01 1.58880317e+00 -7.10232735e-01 -6.70215368e-01
-2.13563070e-01 8.80387664e-01 -1.35150477e-01 3.87610257e-01
-1.12824094e+00 -1.91757426e-01 5.18087506e-01 4.47663486e-01
2.87674982e-02 7.42896274e-02 -2.97761094e-02 -8.16606104e-01
-6.81314617e-02 -8.42451513e-01 6.09443724e-01 -4.17464286e-01
-1.44926393e+00 6.92212462e-01 3.33013505e-01 -8.93458366e-01
4.22131211e-01 -8.78501773e-01 -2.02945262e-01 7.27260530e-01
-9.43581760e-01 -2.18672395e+00 -1.07723129e+00 6.15577638e-01
4.93030787e-01 -7.63971508e-01 7.95623064e-01 4.10777271e-01
-8.09419692e-01 1.24158537e+00 -2.88297564e-01 7.38520920e-01
8.55755806e-01 -1.36591363e+00 1.09954134e-01 3.98907721e-01
-4.43388849e-01 8.31351697e-01 7.41140246e-01 -9.54162717e-01
-1.74833965e+00 -1.18109894e+00 2.93850861e-02 -8.75397027e-01
-1.51497424e-01 -6.05679929e-01 -2.96016335e-01 3.94764692e-01
3.34884256e-01 4.24110472e-01 6.26493514e-01 -6.92727640e-02
-7.16317594e-01 -5.89452684e-01 -1.73056507e+00 2.43132874e-01
7.35880613e-01 -2.51050353e-01 -1.93570018e-01 3.94182116e-01
1.89087674e-01 -5.70176601e-01 -9.42911029e-01 4.96620029e-01
1.27106643e+00 -5.84593832e-01 1.57680213e+00 -5.88351071e-01
-3.33451271e-01 -2.29919001e-01 -1.93399355e-01 -4.98469353e-01
-3.71901572e-01 -4.61524844e-01 -3.93176258e-01 1.19833970e+00
9.13443863e-02 -2.29299024e-01 8.39879870e-01 9.15552676e-01
3.99959415e-01 -6.88851237e-01 -5.42880774e-01 -7.92930067e-01
-3.67208570e-01 3.59715551e-01 7.22009838e-01 5.13763309e-01
-7.63315201e-01 2.98687946e-02 -7.98483968e-01 2.84614474e-01
1.32227361e+00 -1.76950380e-01 1.00004661e+00 -1.45723236e+00
7.34068826e-02 -2.60518432e-01 -7.95177162e-01 -3.46332580e-01
-1.69542730e-01 -3.99665594e-01 -1.58373177e-01 -1.47317660e+00
8.27581823e-01 -1.25596687e-01 -3.44657153e-01 4.12554860e-01
-2.09239930e-01 9.46519792e-01 2.53501743e-01 -1.20224893e-01
-1.06778693e+00 1.31489828e-01 8.19128871e-01 -7.47007847e-01
-2.37032510e-02 -2.50554481e-03 -1.17405966e-01 5.90927958e-01
8.62078428e-01 -1.87826261e-01 2.67290622e-01 -2.59535074e-01
-2.30093449e-01 -5.07717654e-02 9.21063125e-01 -1.14910495e+00
2.56319642e-01 1.09842852e-01 1.50153744e+00 -7.71721780e-01
9.78692234e-01 -9.06641841e-01 -1.01902954e-01 5.14412224e-01
-9.24810693e-02 2.64720678e-01 3.77171606e-01 6.23288691e-01
4.89991963e-01 2.11185515e-01 8.82212937e-01 -2.46629402e-01
-3.73818457e-01 7.70181894e-01 4.39725854e-02 -2.04115361e-01
1.14116585e+00 -5.56054235e-01 -4.81057554e-01 -2.59961516e-01
-6.68628097e-01 1.11235872e-01 4.05562133e-01 4.67914760e-01
8.63881230e-01 -1.36537933e+00 -1.10669339e+00 1.25421152e-01
3.75214100e-01 -6.72608733e-01 1.47977382e-01 6.20423317e-01
-6.86988592e-01 2.58931249e-01 -4.23245788e-01 -6.09210253e-01
-2.05884457e+00 4.34784740e-01 5.29940128e-01 1.15523815e-01
-4.09845233e-01 1.02444983e+00 7.41581693e-02 -1.67489901e-01
5.48762262e-01 2.44966716e-01 -4.74737942e-01 -1.12997524e-01
6.94325864e-01 7.10241437e-01 -4.21758085e-01 -1.05331981e+00
-8.14534843e-01 1.00604379e+00 3.44647616e-02 1.35903284e-01
9.56504881e-01 -1.33537501e-01 -2.03057438e-01 -1.31117910e-01
1.07853627e+00 -2.83378005e-01 -9.55306888e-01 -1.87606476e-02
-4.04463947e-01 -6.69267952e-01 -2.12681234e-01 -9.97231781e-01
-1.17427683e+00 6.26331270e-01 1.56516194e+00 2.81204283e-01
9.74938035e-01 -3.61764804e-02 9.12584126e-01 3.90418053e-01
2.05803022e-01 -1.52017510e+00 -6.38533309e-02 3.06936838e-02
9.93524671e-01 -1.71206725e+00 4.75061715e-01 -3.93025815e-01
-3.25363815e-01 9.01528955e-01 9.96109664e-01 -4.28792715e-01
5.66770077e-01 -2.93292315e-03 -4.20123339e-03 -4.06518221e-01
2.07531452e-03 -3.62281471e-01 1.09612978e+00 8.76536012e-01
4.52653557e-01 5.29524148e-01 -1.84371203e-01 5.02506971e-01
1.17621925e-02 -4.54847395e-01 -2.84308791e-01 5.82129538e-01
-2.93746948e-01 -4.80891347e-01 -1.00816941e+00 7.31712162e-01
-9.14207280e-01 1.64561331e-01 -8.62159669e-01 6.16375983e-01
5.13650060e-01 9.49670553e-01 -5.96686974e-02 -3.42231393e-01
3.50328714e-01 -1.55981615e-01 9.63566780e-01 -2.00405657e-01
-1.09033263e+00 3.75366397e-02 9.85730588e-02 -3.57878327e-01
-4.42975372e-01 -4.34315860e-01 -5.97198486e-01 -5.92238486e-01
-2.13700503e-01 -3.21956933e-01 6.74386859e-01 6.56231403e-01
-1.70831792e-02 2.91367233e-01 2.78382421e-01 -6.73231542e-01
-3.08955461e-01 -1.11871076e+00 -3.36526543e-01 7.06665337e-01
4.32688296e-01 -7.83948660e-01 3.13540787e-01 2.16079906e-01] | [14.578531265258789, 0.8842045664787292] |
ba3f7656-a413-44d0-8338-07c8bede20c3 | one-adapter-for-all-programming-languages | 2303.15822 | null | https://arxiv.org/abs/2303.15822v1 | https://arxiv.org/pdf/2303.15822v1.pdf | One Adapter for All Programming Languages? Adapter Tuning for Code Search and Summarization | As pre-trained models automate many code intelligence tasks, a widely used paradigm is to fine-tune a model on the task dataset for each programming language. A recent study reported that multilingual fine-tuning benefits a range of tasks and models. However, we find that multilingual fine-tuning leads to performance degradation on recent models UniXcoder and CodeT5. To alleviate the potentially catastrophic forgetting issue in multilingual models, we fix all pre-trained model parameters, insert the parameter-efficient structure adapter, and fine-tune it. Updating only 0.6\% of the overall parameters compared to full-model fine-tuning for each programming language, adapter tuning yields consistent improvements on code search and summarization tasks, achieving state-of-the-art results. In addition, we experimentally show its effectiveness in cross-lingual and low-resource scenarios. Multilingual fine-tuning with 200 samples per programming language approaches the results fine-tuned with the entire dataset on code summarization. Our experiments on three probing tasks show that adapter tuning significantly outperforms full-model fine-tuning and effectively overcomes catastrophic forgetting. | ['Xiangke Liao', 'Wei Dong', 'Shaoliang Peng', 'Wei Luo', 'Shanshan Li', 'Boxing Chen', 'Deze Wang'] | 2023-03-28 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-3.57541502e-01 -1.91240206e-01 -4.72958177e-01 -2.58376777e-01
-1.27531159e+00 -6.79015994e-01 3.60082269e-01 1.28939062e-01
-4.67343777e-01 3.80218923e-01 4.17254269e-01 -4.45075244e-01
1.50352940e-01 -2.09834963e-01 -9.83326554e-01 -1.37888983e-01
2.17295930e-01 3.91404837e-01 7.51896128e-02 -2.28364527e-01
4.71960545e-01 -1.93393588e-01 -1.32967365e+00 5.96046507e-01
1.44673693e+00 -3.60701419e-02 6.58238292e-01 6.66179717e-01
-5.23125827e-01 6.81698561e-01 -6.55235231e-01 -6.82052016e-01
-2.96897236e-02 9.33344141e-02 -1.00307226e+00 -3.00489604e-01
8.42444062e-01 -9.62792858e-02 1.20349322e-02 9.39823031e-01
5.16793430e-01 -2.12702826e-01 3.41853350e-01 -6.52179003e-01
-9.27001238e-01 1.29406476e+00 -8.69743109e-01 3.15086097e-01
9.67243165e-02 2.07530722e-01 7.70165384e-01 -9.41913307e-01
6.06213987e-01 1.19836771e+00 1.16264653e+00 5.63344955e-01
-1.59027445e+00 -7.89750397e-01 2.69570798e-01 -1.39830843e-01
-1.48371458e+00 -6.53263032e-01 2.45392308e-01 -8.09420586e-01
1.88238502e+00 -3.51258628e-02 1.35116145e-01 1.02399492e+00
6.59103513e-01 6.21579766e-01 5.21293223e-01 -4.62253094e-01
-2.72566587e-01 1.99642077e-01 2.60877550e-01 9.01501358e-01
3.42533827e-01 -3.96887451e-01 -5.38768411e-01 -4.70600277e-01
1.02311730e-01 -1.06936514e-01 -8.37613866e-02 -3.70072991e-01
-1.31173635e+00 8.21111202e-01 7.40727484e-02 4.91073430e-01
-9.06128362e-02 4.49447513e-01 9.72606361e-01 4.77885932e-01
5.77970326e-01 9.02521968e-01 -1.09387851e+00 -5.98349988e-01
-1.15671313e+00 1.57101423e-01 7.69659221e-01 1.37037373e+00
9.29868042e-01 1.09988175e-01 -3.54631364e-01 1.32654226e+00
-1.15756402e-02 5.16962945e-01 1.14545941e+00 -1.02786207e+00
9.22710717e-01 6.92221701e-01 -3.42023522e-01 -3.26223969e-01
-3.93723518e-01 -8.53567660e-01 -3.91837507e-01 -3.20377290e-01
9.37285945e-02 -2.13801965e-01 -6.52813911e-01 1.88265729e+00
-1.68910503e-01 -4.06418979e-01 -9.46732908e-02 1.29110217e-01
5.28840899e-01 3.94991159e-01 2.82740116e-01 4.29336987e-02
1.16024613e+00 -1.61806893e+00 -3.81658345e-01 -6.66485429e-01
1.22905684e+00 -1.23347914e+00 1.79606378e+00 2.85007328e-01
-1.08214986e+00 -7.13537335e-01 -1.03907156e+00 -3.88509810e-01
-1.00994796e-01 4.04946536e-01 7.07319260e-01 6.54006124e-01
-1.22979319e+00 2.71769047e-01 -1.04556358e+00 -6.56860471e-01
6.87440187e-02 5.64432926e-02 -2.54618347e-01 -2.54501462e-01
-6.13433480e-01 9.93288636e-01 3.26221228e-01 -6.67802930e-01
-1.02945936e+00 -1.34880137e+00 -7.69831002e-01 2.07079738e-01
1.95070758e-01 -1.04762316e+00 1.69576037e+00 -5.27911007e-01
-1.10562205e+00 8.97293806e-01 -4.54223424e-01 -4.66605395e-01
2.95302421e-01 -5.08334041e-01 -2.04116166e-01 -6.75272703e-01
3.81148338e-01 5.31003296e-01 7.89760053e-01 -9.00880396e-01
-4.67361718e-01 5.58051988e-02 1.30071118e-01 -3.00369710e-02
-6.75831735e-01 1.00849710e-01 -8.82816732e-01 -8.59266758e-01
-4.15530086e-01 -1.01358986e+00 -2.06541829e-02 -7.48218894e-01
-2.53650188e-01 -1.62749499e-01 4.14080858e-01 -7.39275634e-01
1.83646905e+00 -2.21551442e+00 2.64854014e-01 -6.24480069e-01
1.50215387e-01 2.10562184e-01 -7.18585789e-01 4.23348308e-01
-5.31082647e-03 3.36741716e-01 -3.79055738e-01 -6.76270604e-01
5.90923913e-02 1.78578392e-01 -2.39232481e-01 7.46287555e-02
-7.88049251e-02 1.02718866e+00 -6.97429776e-01 -3.71833146e-01
-9.07484964e-02 3.27791691e-01 -1.40864718e+00 3.32743600e-02
-4.36539173e-01 6.89256042e-02 -2.82467335e-01 5.42265713e-01
4.26874638e-01 -3.91289711e-01 6.62974417e-02 9.53770280e-02
-1.89977974e-01 5.37447989e-01 -5.65515935e-01 2.51110053e+00
-1.25181067e+00 5.16016185e-01 -9.65468660e-02 -3.33811969e-01
6.22664928e-01 -2.74106441e-03 5.52617498e-02 -8.43722522e-01
-5.29540539e-01 4.14852649e-01 -6.85220584e-02 -8.01210165e-01
7.27210641e-01 2.96883672e-01 -6.33102238e-01 7.51921833e-01
2.32123166e-01 -2.95606464e-01 5.43000937e-01 4.71280158e-01
1.42328644e+00 1.54641777e-01 3.38021904e-01 -7.70805240e-01
4.86059159e-01 2.29601085e-01 3.23094755e-01 9.57543850e-01
2.16729537e-01 2.75443941e-01 2.00851545e-01 -4.91771758e-01
-1.03234029e+00 -7.38327503e-01 -1.37364089e-01 1.93998516e+00
-5.55626631e-01 -1.10260618e+00 -1.18239117e+00 -8.09723377e-01
2.18338966e-01 1.03393030e+00 -6.30800307e-01 -3.19775134e-01
-7.66860604e-01 -9.80980337e-01 8.54334712e-01 2.53621161e-01
3.11938733e-01 -7.20420122e-01 -4.19596851e-01 3.50341201e-01
-5.36979854e-01 -6.80757880e-01 -1.23365128e+00 3.67860228e-01
-8.78803492e-01 -8.58091295e-01 -4.20334429e-01 -6.34635031e-01
4.55786616e-01 3.90115410e-01 1.87746572e+00 3.14356327e-01
-2.03296483e-01 2.69335598e-01 -1.77260995e-01 4.87310026e-04
-8.60468209e-01 1.08331919e+00 -1.55381233e-01 -7.95077384e-01
5.30005574e-01 -5.45296729e-01 -1.35692418e-01 1.56238582e-03
-6.91536307e-01 -7.08834752e-02 8.10471773e-01 8.46017838e-01
2.70733684e-01 -3.06461096e-01 6.17806017e-01 -1.29953039e+00
9.30415690e-01 -5.19851506e-01 -4.11731392e-01 5.64924240e-01
-1.05549467e+00 6.88859820e-01 5.42122602e-01 -5.18196821e-01
-1.23161995e+00 -2.19467491e-01 -2.60159373e-02 -1.73338577e-01
3.08532238e-01 8.24256122e-01 2.08708808e-01 -6.06256798e-02
1.09560418e+00 2.66763002e-01 -4.77173448e-01 -8.84164274e-01
5.93328714e-01 5.30399382e-01 5.61823070e-01 -1.09955740e+00
5.52294493e-01 -1.06597766e-01 -9.42020178e-01 -3.87394935e-01
-9.21174288e-01 -3.49492460e-01 -5.46868920e-01 4.03429240e-01
5.58104753e-01 -1.29272223e+00 -1.34222582e-01 2.80973881e-01
-1.28432500e+00 -6.97701216e-01 -1.13856327e-03 1.28379196e-01
-4.42307442e-01 2.58069068e-01 -7.63750196e-01 -2.91491542e-02
-6.16400421e-01 -1.51866460e+00 1.19885409e+00 -2.02209830e-01
-5.59943318e-01 -1.08809507e+00 4.65760559e-01 4.81812268e-01
9.47976112e-01 -3.85665864e-01 1.49777186e+00 -3.74372154e-01
-4.22427297e-01 2.09795564e-01 -1.49944723e-01 1.93677157e-01
4.01619412e-02 5.34965061e-02 -9.73569274e-01 -6.79484427e-01
-9.38696265e-02 -4.70043033e-01 1.05858791e+00 2.23446399e-01
1.08811677e+00 -4.08845454e-01 -3.24324250e-01 9.16627824e-01
1.44921768e+00 -2.05311432e-01 5.55231869e-02 4.54598695e-01
8.28000844e-01 1.43808708e-01 3.94184053e-01 5.69667399e-01
8.37442219e-01 6.90361917e-01 5.06269485e-02 4.45960313e-01
-5.44907629e-01 -3.18661124e-01 6.71787262e-01 1.59687710e+00
4.46460485e-01 1.73427299e-01 -1.22205937e+00 8.47937405e-01
-1.77091432e+00 -5.29611409e-01 -8.25680345e-02 2.07967091e+00
1.46546018e+00 9.87584516e-02 -2.05631822e-01 -7.26198733e-01
2.77117848e-01 -2.54893750e-02 -7.74337947e-01 -6.90255642e-01
-3.75046059e-02 1.25187799e-01 5.66314697e-01 6.59510374e-01
-7.27855623e-01 1.29461253e+00 7.01013470e+00 1.03266001e+00
-1.08150959e+00 7.36804545e-01 1.84680849e-01 -3.87120843e-01
-6.10693514e-01 6.03465252e-02 -1.08552122e+00 4.91582394e-01
1.17365038e+00 -6.77236557e-01 9.43471014e-01 1.09796500e+00
-9.68529806e-02 2.15651141e-03 -1.36498129e+00 8.50758076e-01
2.72850066e-01 -1.28382647e+00 2.28588998e-01 -2.09120348e-01
1.27764189e+00 8.45056057e-01 1.37622384e-02 1.18902624e+00
6.74188912e-01 -7.52230167e-01 9.65860605e-01 3.66292238e-01
9.75280941e-01 -6.19896591e-01 4.90227461e-01 5.48713446e-01
-1.10716212e+00 -1.39537856e-01 -4.06646937e-01 -5.36865788e-03
-1.93826616e-01 6.58019960e-01 -8.13471079e-01 1.51037797e-01
7.61190176e-01 6.60894990e-01 -1.43567038e+00 8.60655963e-01
3.26539576e-02 5.30119479e-01 -2.30604354e-02 3.39874208e-01
2.66917571e-02 4.08524871e-01 2.63622075e-01 1.81129885e+00
4.72719669e-01 -7.63493419e-01 1.46677762e-01 9.85385001e-01
-3.25485706e-01 6.94412887e-02 -4.22450632e-01 -4.86759543e-02
6.67584121e-01 9.81721938e-01 2.40642205e-02 -5.40971875e-01
-5.53816319e-01 8.66907835e-01 8.67749095e-01 2.27363691e-01
-8.02476764e-01 -4.84640777e-01 7.66293883e-01 1.55180665e-02
2.47405306e-01 -2.29116321e-01 -4.10485655e-01 -1.42993283e+00
9.57411081e-02 -1.33458173e+00 4.15907204e-01 -6.39655590e-01
-1.01127934e+00 7.52253175e-01 2.48878030e-03 -6.74859703e-01
-5.91496706e-01 -2.72192191e-02 -2.97034353e-01 8.92567158e-01
-1.37078333e+00 -1.09542572e+00 -2.38089621e-01 3.30213845e-01
1.17589164e+00 -2.39309192e-01 8.92532706e-01 7.20037341e-01
-6.42525434e-01 1.11509204e+00 3.15009177e-01 -3.06314141e-01
1.23797679e+00 -1.26132166e+00 9.64904308e-01 9.81362343e-01
-2.14625541e-02 1.33001745e+00 7.27157533e-01 -6.88894272e-01
-1.44746196e+00 -1.27620184e+00 1.25619590e+00 -9.44868863e-01
8.78404379e-01 -4.08301294e-01 -1.10437536e+00 9.32390749e-01
5.04899681e-01 -3.49550158e-01 3.83409500e-01 5.15766025e-01
-8.79285038e-01 -1.09296544e-02 -7.30181813e-01 4.85966533e-01
1.08201313e+00 -8.92326534e-01 -6.10000253e-01 4.05751824e-01
1.13248253e+00 -5.19882202e-01 -1.07675612e+00 1.57135203e-01
4.62304980e-01 -8.85757565e-01 7.39761233e-01 -5.27754843e-01
4.62049335e-01 -1.13964446e-01 -1.48886010e-01 -1.72867811e+00
-5.18907130e-01 -6.43871069e-01 -1.25656545e-01 1.34255910e+00
7.19035506e-01 -4.85377580e-01 1.83512270e-01 3.70933264e-01
-4.98029411e-01 -3.72160673e-01 -5.45071006e-01 -8.22793067e-01
6.17523074e-01 -4.62868333e-01 6.02632046e-01 1.03383934e+00
5.44312969e-02 5.18184841e-01 -1.42011061e-01 -2.24049658e-01
6.41091883e-01 1.13294698e-01 1.00934339e+00 -8.24016929e-01
-6.66639030e-01 -6.58116341e-01 5.58442235e-01 -1.07005167e+00
5.10119796e-01 -1.21007180e+00 -1.07673714e-02 -1.21049392e+00
8.33241343e-01 -5.53595543e-01 -6.28989860e-02 7.89917648e-01
-3.81907403e-01 -8.45455900e-02 8.28149393e-02 4.42583382e-01
-7.74944484e-01 3.09921563e-01 7.95304179e-01 -4.68624383e-01
-1.36382625e-01 -2.95948654e-01 -1.02336013e+00 5.27938724e-01
6.42139673e-01 -7.50919998e-01 -4.65462297e-01 -1.33729899e+00
4.99697655e-01 -1.08364135e-01 -3.40745360e-01 -1.01373172e+00
2.60904819e-01 7.70620555e-02 -2.08450958e-01 -2.27296278e-01
-2.58235455e-01 -3.43626350e-01 2.86716998e-01 6.65878296e-01
-5.87649584e-01 5.03870010e-01 7.77575135e-01 2.60203630e-01
2.32537594e-02 -4.86398071e-01 8.70705366e-01 -1.64622068e-01
-6.10631466e-01 -8.78656283e-02 -3.73289436e-01 5.87815166e-01
5.52145958e-01 3.06254387e-01 -8.38325918e-01 2.42216691e-01
-2.50969797e-01 1.87212005e-01 7.41112411e-01 1.00749958e+00
-7.48175159e-02 -1.14218736e+00 -7.56321073e-01 3.62133443e-01
7.14055598e-01 -3.19301158e-01 3.23044688e-01 8.27858388e-01
-4.03510839e-01 8.53809297e-01 -5.55758551e-02 -6.30857289e-01
-1.06909049e+00 7.43928850e-01 2.58007526e-01 -6.08819187e-01
-2.35355958e-01 9.04276192e-01 3.73577237e-01 -9.35706079e-01
4.73852605e-02 -6.40776575e-01 2.59688318e-01 -2.03536823e-01
4.59852844e-01 1.96801767e-01 5.88733137e-01 -2.59131581e-01
-3.20122421e-01 6.27719760e-01 -6.32033587e-01 2.11322665e-01
1.23598099e+00 -2.60115057e-01 -4.51198429e-01 5.24247408e-01
1.31636488e+00 4.91565645e-01 -1.15991497e+00 -5.57305336e-01
2.73148417e-01 -2.41445661e-01 -2.09294390e-02 -1.11894774e+00
-9.00891721e-01 6.76080287e-01 2.66783267e-01 -2.23880023e-01
8.41127992e-01 4.88702618e-02 6.27208233e-01 8.00433099e-01
7.97791481e-01 -8.93015683e-01 2.24946719e-02 9.61443663e-01
9.39811766e-01 -1.03043818e+00 -6.96395412e-02 1.48928508e-01
-4.36633378e-01 7.60098338e-01 9.84264553e-01 2.36989439e-01
3.75461608e-01 5.84135532e-01 -5.96074294e-03 -1.13934558e-02
-1.37884820e+00 4.18445468e-01 1.92156821e-01 2.84650207e-01
1.05340719e+00 2.99942959e-02 -1.08459704e-01 6.41142309e-01
-3.99203837e-01 -8.16779286e-02 5.14288485e-01 8.95971060e-01
-3.81575823e-01 -1.16953611e+00 -1.14257768e-01 6.43198490e-01
-5.61085820e-01 -7.71873713e-01 -4.27028760e-02 6.92692995e-01
4.24943771e-03 7.48980463e-01 -1.42683074e-01 -4.37143773e-01
4.53406990e-01 3.03371280e-01 5.15819252e-01 -1.04160488e+00
-1.27772105e+00 -1.02922559e-01 1.08319444e-04 -7.34426379e-01
1.30171239e-01 -5.59030354e-01 -9.26002920e-01 -6.90459788e-01
-2.18341783e-01 2.22238377e-01 6.76067948e-01 7.09935129e-01
9.61071193e-01 8.57039094e-01 3.52588519e-02 -6.35484040e-01
-7.97527492e-01 -1.40786231e+00 5.24252094e-02 1.51886508e-01
3.55294436e-01 -4.11723405e-01 -3.28933150e-01 3.82287353e-01] | [7.702566623687744, 7.906900405883789] |
9cc67a4c-d3a3-4c08-9683-c5c2cd34e9e8 | end-to-end-abstractive-summarization-for | 2004.02016 | null | https://arxiv.org/abs/2004.02016v4 | https://arxiv.org/pdf/2004.02016v4.pdf | A Hierarchical Network for Abstractive Meeting Summarization with Cross-Domain Pretraining | With the abundance of automatic meeting transcripts, meeting summarization is of great interest to both participants and other parties. Traditional methods of summarizing meetings depend on complex multi-step pipelines that make joint optimization intractable. Meanwhile, there are a handful of deep neural models for text summarization and dialogue systems. However, the semantic structure and styles of meeting transcripts are quite different from articles and conversations. In this paper, we propose a novel abstractive summary network that adapts to the meeting scenario. We design a hierarchical structure to accommodate long meeting transcripts and a role vector to depict the difference among speakers. Furthermore, due to the inadequacy of meeting summary data, we pretrain the model on large-scale news summary data. Empirical results show that our model outperforms previous approaches in both automatic metrics and human evaluation. For example, on ICSI dataset, the ROUGE-1 score increases from 34.66% to 46.28%. | ['Michael Zeng', 'Chenguang Zhu', 'Xuedong Huang', 'Ruochen Xu'] | 2020-04-04 | null | https://aclanthology.org/2020.findings-emnlp.19 | https://aclanthology.org/2020.findings-emnlp.19.pdf | findings-of-the-association-for-computational | ['meeting-summarization'] | ['natural-language-processing'] | [ 5.25038466e-02 2.08782494e-01 -5.17766438e-02 -6.54638469e-01
-1.13437772e+00 -5.00771284e-01 6.12618029e-01 1.70205057e-01
-2.03786999e-01 8.84705901e-01 1.07394075e+00 2.02446710e-02
3.12648296e-01 -2.92451739e-01 -2.49772489e-01 -2.42030263e-01
2.88517118e-01 4.51652408e-01 -2.42733106e-01 -4.08818156e-01
3.65326017e-01 -1.17978945e-01 -1.10670900e+00 6.32320821e-01
1.13665819e+00 6.07566476e-01 1.48850322e-01 7.45497704e-01
-4.18849885e-01 4.30120409e-01 -1.33100283e+00 -4.27682877e-01
-1.58490837e-01 -7.43099630e-01 -7.24966407e-01 1.91571325e-01
5.81685543e-01 -4.78160381e-01 -4.15330231e-01 7.45508611e-01
7.64392912e-01 3.34260762e-01 6.34741664e-01 -7.68553257e-01
-5.18560886e-01 1.15998256e+00 -5.33823967e-01 3.78136411e-02
5.11462092e-01 -9.01360586e-02 1.29317236e+00 -6.16618514e-01
3.67303669e-01 1.22716892e+00 5.56704640e-01 7.66847253e-01
-8.69140327e-01 -6.41629994e-01 5.07394552e-01 -3.68426330e-02
-9.50669229e-01 -7.65381038e-01 7.66673148e-01 -1.80944413e-01
9.89392638e-01 6.92012429e-01 5.46204269e-01 1.29023480e+00
9.01112631e-02 1.02459049e+00 2.91359723e-01 -1.31824777e-01
7.16075897e-02 2.12456807e-02 3.51358175e-01 3.86884928e-01
3.54101062e-01 -9.06713426e-01 -9.13478613e-01 -1.74254403e-01
4.23723400e-01 -7.25639537e-02 -3.62527043e-01 1.57712579e-01
-1.29811502e+00 6.38743162e-01 -1.36877806e-03 3.50379407e-01
-4.09727544e-01 -1.74963266e-01 6.05982959e-01 7.51616135e-02
7.29989290e-01 6.87468529e-01 -1.48210928e-01 -1.65970489e-01
-1.21891582e+00 4.10634637e-01 1.28535485e+00 8.42011452e-01
2.96493649e-01 2.15752944e-01 -6.31616414e-01 1.06764185e+00
1.11174516e-01 2.85615265e-01 5.82420051e-01 -9.55117345e-01
7.94698238e-01 5.66812158e-01 2.04756036e-01 -1.20894897e+00
-4.73837525e-01 -5.67489684e-01 -1.26921022e+00 -7.65171409e-01
1.28526822e-01 -3.81203949e-01 -5.68828046e-01 1.61873722e+00
-1.39320239e-01 -1.24953680e-01 1.20763756e-01 7.23618925e-01
1.50762546e+00 8.85971725e-01 -2.15676531e-01 -4.85577255e-01
1.08045971e+00 -1.23364723e+00 -1.15333402e+00 -3.52281123e-01
4.21418935e-01 -6.92470551e-01 1.00824893e+00 2.41466120e-01
-1.42166042e+00 -4.32554632e-01 -1.08164334e+00 -1.01188414e-01
2.83634573e-01 3.43030095e-01 4.52736735e-01 3.28742534e-01
-1.11819649e+00 3.59966874e-01 -6.45971656e-01 -4.78805006e-01
3.23694721e-02 3.55212063e-01 -3.94854881e-02 2.05234706e-01
-1.10104287e+00 5.92118442e-01 2.00069204e-01 3.23845685e-01
-4.05058116e-01 -3.67073178e-01 -8.69218767e-01 4.52906489e-01
1.72156528e-01 -9.16360736e-01 1.71188080e+00 -7.74131358e-01
-1.84311152e+00 4.79539335e-01 -4.77828264e-01 -3.16125631e-01
5.92524707e-01 -3.68950784e-01 -1.53535441e-01 -4.32012640e-02
1.23474695e-01 4.47802991e-01 4.44678038e-01 -1.18953347e+00
-5.51474214e-01 -1.15310036e-01 -5.64275160e-02 7.39187241e-01
-5.10402083e-01 1.00527052e-02 -5.09561837e-01 -3.80579561e-01
-3.14633138e-02 -6.17085636e-01 -3.16674590e-01 -9.20766115e-01
-9.10129786e-01 -4.32421535e-01 3.43169451e-01 -1.01465702e+00
1.76405358e+00 -2.09426475e+00 2.93970287e-01 -2.36184537e-01
3.44917417e-01 2.11249784e-01 -1.44877717e-01 7.71260262e-01
3.71474952e-01 3.02318335e-01 -1.72432497e-01 -8.17214251e-01
2.08520859e-01 -1.42538741e-01 -2.50766903e-01 2.84957383e-02
-4.49917465e-02 8.34307313e-01 -6.95503533e-01 -5.12488484e-01
-1.61969528e-01 8.18774998e-02 -4.70723301e-01 4.76355940e-01
-1.54464483e-01 4.25844669e-01 -5.73460519e-01 3.79618645e-01
4.04592067e-01 -3.34759206e-01 4.89253521e-01 -2.79128682e-02
-2.01515362e-01 7.67130554e-01 -8.14803183e-01 1.91632473e+00
-2.80562580e-01 6.54207885e-01 2.84880430e-01 -8.54524910e-01
9.75443423e-01 4.90442157e-01 2.44680852e-01 -3.24750394e-01
1.30170748e-01 8.01269859e-02 -3.89755405e-02 -3.72205853e-01
1.14125419e+00 1.01483136e-01 -5.46272099e-01 6.40642345e-01
-1.55435294e-01 -2.30087042e-01 3.85874808e-01 4.55828786e-01
1.14962065e+00 -3.60447645e-01 4.09160644e-01 -3.74986380e-02
3.95356208e-01 -1.13164812e-01 8.67197573e-01 8.57708752e-01
-2.11788669e-01 9.47968185e-01 6.27473772e-01 -3.32362473e-01
-8.46048534e-01 -7.32469201e-01 4.55332816e-01 1.09960020e+00
-5.61063103e-02 -6.89783335e-01 -8.84875238e-01 -4.90379691e-01
-3.50249261e-01 7.55244851e-01 -2.29790404e-01 -1.73410606e-02
-8.22891533e-01 -5.81622005e-01 4.97701883e-01 4.54489768e-01
5.31339705e-01 -1.08989990e+00 -1.33401707e-01 3.99449915e-01
-8.70988309e-01 -1.05619514e+00 -9.69897568e-01 -2.47000396e-01
-7.33994246e-01 -5.38766801e-01 -6.73148155e-01 -8.56742740e-01
2.66336352e-01 4.82228518e-01 1.29958510e+00 -5.11689447e-02
2.52436996e-01 1.74439684e-01 -1.33830383e-01 -6.31029427e-01
-5.11153996e-01 7.21113980e-01 5.03210947e-02 -6.72908798e-02
2.26116642e-01 -4.45994288e-01 -5.80685139e-01 -8.22427124e-02
-7.13457406e-01 3.55191261e-01 6.65420234e-01 6.50953829e-01
-2.93441247e-02 -5.53900838e-01 1.04683900e+00 -8.28436494e-01
1.24701345e+00 -3.13991159e-01 1.72480028e-02 3.34348410e-01
-1.12911507e-01 -3.07358146e-01 4.49988306e-01 -2.98651397e-01
-1.23916113e+00 -2.61881679e-01 2.08499888e-03 2.08850905e-01
-1.02061117e-02 6.50048971e-01 -3.63942713e-01 8.41874719e-01
4.35150683e-01 1.56553313e-01 -4.09722514e-02 -3.50979447e-01
7.70605309e-03 9.55656111e-01 4.37853932e-01 -2.47128785e-01
4.27823573e-01 1.46237940e-01 -6.65309191e-01 -1.16827929e+00
-1.27780509e+00 -4.99365956e-01 -3.36798817e-01 -1.77144632e-01
7.93053031e-01 -9.89714503e-01 -6.71240926e-01 3.30672741e-01
-1.51051176e+00 -1.80978477e-01 2.20987387e-02 4.75001514e-01
-2.44038492e-01 6.78364575e-01 -8.05100322e-01 -7.74984837e-01
-8.35902035e-01 -8.78461540e-01 1.00558829e+00 4.12864327e-01
-8.72399628e-01 -8.51766527e-01 7.04750866e-02 6.42740130e-01
4.17171180e-01 1.01761006e-01 6.94403648e-01 -1.05668151e+00
-2.59311140e-01 -1.86670452e-01 -4.87054810e-02 2.98117548e-01
2.68937558e-01 5.43264709e-02 -7.64574409e-01 -7.21708387e-02
2.50987746e-02 -1.95989177e-01 9.14581954e-01 6.48079336e-01
1.01995516e+00 -6.30760789e-01 -1.30156592e-01 2.59737581e-01
5.92105389e-01 -2.48904303e-02 3.10518891e-01 1.88962862e-01
7.86470294e-01 8.18910956e-01 3.08232635e-01 6.05974495e-01
8.22017133e-01 2.98884749e-01 1.08495742e-01 2.72255521e-02
2.30164021e-01 -1.68201447e-01 4.88255709e-01 1.66088092e+00
1.61595941e-01 -6.49072468e-01 -7.57874608e-01 6.22660935e-01
-2.07396889e+00 -9.43274498e-01 -1.07340209e-01 1.83181667e+00
9.74849164e-01 8.62984061e-02 2.56026816e-02 -3.61675084e-01
8.99438262e-01 6.29827976e-01 -4.54323351e-01 -6.41818464e-01
-1.56652406e-01 -3.41278791e-01 -3.98989022e-02 3.93590868e-01
-9.09334362e-01 7.96048462e-01 6.48805523e+00 5.28869927e-01
-9.02114451e-01 -2.69183695e-01 5.87487996e-01 -3.79581928e-01
-4.62209344e-01 -3.95955801e-01 -8.40232551e-01 3.56026709e-01
1.00148273e+00 -6.54442608e-01 5.27284220e-02 3.46370220e-01
4.47625697e-01 7.45972916e-02 -1.22283077e+00 9.53232348e-01
4.55916554e-01 -1.44196868e+00 2.10720405e-01 2.48836987e-02
8.89645398e-01 -1.51127577e-01 -1.11338258e-01 6.12142920e-01
3.45250934e-01 -9.96538103e-01 3.81737649e-01 3.68889213e-01
4.18119729e-01 -7.66014040e-01 9.32631552e-01 6.54096782e-01
-8.04929376e-01 1.93278521e-01 -2.73016542e-01 -2.60729283e-01
3.58101696e-01 4.96677607e-01 -1.04265511e+00 5.94393075e-01
4.07097071e-01 7.16539323e-01 -2.32922226e-01 7.45720625e-01
-1.13677002e-01 7.35850155e-01 -1.76177129e-01 -2.98931837e-01
2.86953330e-01 -2.01913223e-01 6.45091236e-01 1.48443151e+00
3.02412689e-01 5.33112437e-02 4.67923939e-01 4.89903033e-01
-4.57238942e-01 2.73837626e-01 -5.69255292e-01 -1.84647113e-01
6.43559992e-01 1.34556699e+00 -3.15198600e-01 -4.99901921e-01
-2.37325296e-01 9.88382995e-01 3.66747171e-01 5.54421723e-01
-6.16804898e-01 -3.74814540e-01 5.37904024e-01 -1.98460922e-01
-1.13906730e-02 -3.19487303e-01 -4.63379681e-01 -1.39944768e+00
2.37489492e-01 -1.06501734e+00 1.13779493e-01 -3.04907024e-01
-1.12703907e+00 7.01708734e-01 -1.84326932e-01 -8.99530411e-01
-4.73108292e-01 1.73355147e-01 -1.07462239e+00 6.68765604e-01
-1.25829899e+00 -1.08926630e+00 -3.99608642e-01 -1.20346189e-01
1.15115452e+00 -1.74228460e-01 8.15946519e-01 4.38980535e-02
-1.00524640e+00 7.15731800e-01 2.58959740e-01 1.91329554e-01
1.01989007e+00 -1.27240932e+00 6.72906399e-01 5.85462213e-01
-1.37758076e-01 7.22983241e-01 1.03973246e+00 -5.49996257e-01
-1.13659024e+00 -9.43026900e-01 1.33837783e+00 -2.76950091e-01
5.28726935e-01 -3.73985827e-01 -1.10051847e+00 7.16220915e-01
8.26739609e-01 -9.95413184e-01 1.06707180e+00 5.22358537e-01
8.77845858e-04 -5.66454604e-02 -6.28149569e-01 9.78791118e-01
8.53879809e-01 -1.93489924e-01 -8.56367350e-01 2.72576481e-01
8.26050460e-01 -4.32283908e-01 -5.47173321e-01 2.32881457e-01
6.08733952e-01 -8.91896248e-01 4.10598963e-01 -5.09355962e-01
6.34705126e-01 9.57035869e-02 1.64201073e-02 -1.57204926e+00
-2.29344800e-01 -8.91318440e-01 -1.09788060e-01 1.50917172e+00
5.03968716e-01 -3.32244098e-01 7.53621042e-01 7.04751849e-01
-5.78770578e-01 -6.34970546e-01 -5.85983276e-01 -2.93570369e-01
-1.74950942e-01 1.47669941e-01 4.67248946e-01 8.61208737e-01
3.22065353e-01 1.11183333e+00 -5.00710845e-01 -2.07128316e-01
3.00386250e-01 2.40005732e-01 1.03257692e+00 -1.26235294e+00
-1.92782596e-01 -6.06335461e-01 2.44270399e-01 -1.40671921e+00
4.59029913e-01 -6.48793697e-01 3.33582878e-01 -2.28365660e+00
4.96786177e-01 1.63159564e-01 -1.19256943e-01 7.29075968e-02
-5.02469897e-01 -1.52973354e-01 8.88909623e-02 1.79477528e-01
-1.09104836e+00 8.05585027e-01 1.12402153e+00 -2.79427826e-01
-5.06178141e-01 1.59282207e-01 -1.13366771e+00 8.33959818e-01
1.09493065e+00 -1.64243683e-01 -1.56510979e-01 -7.58964121e-01
2.17146292e-01 3.50053042e-01 -2.25970641e-01 -6.99513257e-01
4.12018061e-01 -1.71703607e-01 1.64958224e-01 -9.59571302e-01
4.39419448e-01 -3.41712758e-02 -2.40310222e-01 4.67381701e-02
-7.04709828e-01 5.36651723e-02 3.85856740e-02 5.65095961e-01
-4.85934228e-01 -1.08665399e-01 1.70920834e-01 -2.02722028e-01
-1.67625546e-01 9.72007215e-03 -5.76010227e-01 1.74422324e-01
4.86876547e-01 -1.81334347e-01 -2.76876330e-01 -1.00478363e+00
-3.67521346e-01 6.37933254e-01 2.22614065e-01 4.13203835e-01
5.36323965e-01 -1.05737233e+00 -1.18494046e+00 -2.15081751e-01
-9.55893621e-02 4.63157117e-01 5.33933103e-01 6.10412657e-01
-2.60972649e-01 5.27925551e-01 7.04505742e-02 -4.09589618e-01
-1.43502951e+00 -2.06477433e-01 -4.48054597e-02 -5.92892408e-01
-5.42152047e-01 7.77505040e-01 5.57596982e-01 -4.36461121e-01
4.71263409e-01 -4.02524918e-01 -5.27662218e-01 3.90530705e-01
7.25131333e-01 3.88613105e-01 -3.72418463e-02 -4.17499393e-01
-6.30148724e-02 7.64360130e-02 -5.06852806e-01 -1.70208842e-01
1.39007115e+00 -3.63148898e-01 -1.52572870e-01 8.09468567e-01
8.79366219e-01 2.57469028e-01 -9.68226016e-01 -2.23457754e-01
1.45539299e-01 1.71611272e-02 -3.56653929e-01 -5.63200951e-01
-5.48304379e-01 7.65949845e-01 -3.31330210e-01 4.09370691e-01
7.56344497e-01 -1.28521606e-01 1.05067682e+00 7.22054601e-01
-2.51380563e-01 -1.25801802e+00 2.61755675e-01 9.31040525e-01
1.13940465e+00 -1.28441298e+00 1.04735881e-01 -1.58692151e-01
-9.82100070e-01 8.80065858e-01 7.38861561e-01 5.79206161e-02
-7.12779388e-02 -7.55570829e-02 2.53303200e-01 -3.99879739e-02
-1.10480320e+00 3.47675353e-01 3.12239319e-01 2.19124302e-01
9.22698677e-01 2.07679018e-01 -5.33907354e-01 1.01003885e+00
-5.65887570e-01 -4.88018870e-01 8.70513439e-01 5.93836308e-01
-7.37328768e-01 -8.59157979e-01 -1.29963845e-01 5.48519790e-01
-7.36332834e-01 -3.95244993e-02 -8.59981179e-01 5.50623417e-01
-7.60171473e-01 1.34921908e+00 1.29843414e-01 -2.68753976e-01
4.39441651e-01 1.49566829e-01 3.67333577e-03 -1.09850037e+00
-9.66293395e-01 2.05847085e-01 5.80300093e-01 -9.95033532e-02
-2.58471131e-01 -6.52319491e-01 -1.21350813e+00 -4.97555524e-01
-2.46084034e-01 5.02262592e-01 7.61328459e-01 8.79770100e-01
5.14935672e-01 9.27071393e-01 6.26497924e-01 -8.85675013e-01
-8.24462891e-01 -1.49104881e+00 -4.41584945e-01 8.58271942e-02
3.28840435e-01 4.22360934e-02 -2.48194814e-01 -8.47974271e-02] | [12.598843574523926, 9.304451942443848] |
b1527fbc-25b7-43fe-ad17-d0bfab3f9526 | on-implicit-filter-level-sparsity-in | 1811.12495 | null | http://arxiv.org/abs/1811.12495v2 | http://arxiv.org/pdf/1811.12495v2.pdf | On Implicit Filter Level Sparsity in Convolutional Neural Networks | We investigate filter level sparsity that emerges in convolutional neural
networks (CNNs) which employ Batch Normalization and ReLU activation, and are
trained with adaptive gradient descent techniques and L2 regularization or
weight decay. We conduct an extensive experimental study casting our initial
findings into hypotheses and conclusions about the mechanisms underlying the
emergent filter level sparsity. This study allows new insight into the
performance gap obeserved between adapative and non-adaptive gradient descent
methods in practice. Further, analysis of the effect of training strategies and
hyperparameters on the sparsity leads to practical suggestions in designing CNN
training strategies enabling us to explore the tradeoffs between feature
selectivity, network capacity, and generalization performance. Lastly, we show
that the implicit sparsity can be harnessed for neural network speedup at par
or better than explicit sparsification / pruning approaches, with no
modifications to the typical training pipeline required. | ['Christian Theobalt', 'Dushyant Mehta', 'Kwang In Kim'] | 2018-11-29 | on-implicit-filter-level-sparsity-in-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Mehta_On_Implicit_Filter_Level_Sparsity_in_Convolutional_Neural_Networks_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Mehta_On_Implicit_Filter_Level_Sparsity_in_Convolutional_Neural_Networks_CVPR_2019_paper.pdf | cvpr-2019-6 | ['l2-regularization'] | ['methodology'] | [ 3.18677276e-01 1.39970139e-01 -1.90233111e-01 -4.64362711e-01
1.09257422e-01 -4.70602423e-01 4.23993081e-01 -7.98839778e-02
-7.63702989e-01 4.59542423e-01 1.91230193e-01 -6.50269866e-01
-1.87462986e-01 -5.71799934e-01 -7.01650083e-01 -6.12271070e-01
-2.93792933e-01 -6.22267723e-01 1.56465903e-01 -1.51002511e-01
2.12939084e-01 8.34578693e-01 -1.52865827e+00 3.71389419e-01
3.47610235e-01 1.30260944e+00 1.21304251e-01 7.03015447e-01
7.74891376e-02 7.39020765e-01 -2.95991004e-01 -2.33364642e-01
6.43963397e-01 -3.63143265e-01 -4.69320476e-01 7.70376101e-02
9.20134068e-01 -2.65859932e-01 -5.45445979e-01 8.93947959e-01
5.16538322e-01 1.40516385e-01 3.90170008e-01 -7.26605117e-01
-5.10038257e-01 6.14066243e-01 -2.37085134e-01 1.07146561e+00
-5.35279632e-01 5.06127298e-01 8.82001817e-01 -1.04091752e+00
4.01661783e-01 9.78270054e-01 1.03801882e+00 7.09021389e-01
-1.37085819e+00 -7.91726649e-01 5.03606260e-01 -3.68208766e-01
-1.11578584e+00 -1.08123732e+00 3.00027728e-01 -4.46226031e-01
1.19476485e+00 -5.43020926e-02 7.90815234e-01 8.52250099e-01
1.45763457e-01 5.12038052e-01 7.76447415e-01 -3.42346430e-01
2.98309207e-01 3.22746247e-01 2.04180762e-01 8.57614458e-01
6.44207835e-01 1.18925795e-01 -6.95766330e-01 1.11929029e-01
1.24238348e+00 -1.49257377e-01 -2.13053301e-01 1.13962285e-01
-7.08737910e-01 6.85360014e-01 5.43800056e-01 2.68426120e-01
-3.11310023e-01 5.33814967e-01 5.73000073e-01 7.15695202e-01
4.06946003e-01 8.24604034e-01 -7.66485572e-01 1.39761075e-01
-1.20388877e+00 2.08143309e-01 7.77891815e-01 8.68616581e-01
6.94221437e-01 6.91562235e-01 2.41298787e-02 6.28641486e-01
1.06538117e-01 2.93792710e-02 4.99418646e-01 -1.12084663e+00
2.98915565e-01 3.61217469e-01 -2.79843420e-01 -7.00325251e-01
-4.30694401e-01 -9.51943934e-01 -7.32844234e-01 2.92727202e-01
2.96772569e-01 -8.10239017e-01 -7.92146564e-01 1.63870931e+00
-1.68749064e-01 1.45266622e-01 -1.72576651e-01 6.57217562e-01
7.07127631e-01 1.15407996e-01 3.49615991e-01 -7.63809159e-02
1.05703330e+00 -6.98510706e-01 -3.72380346e-01 -3.64311188e-01
9.33762908e-01 -4.49196041e-01 1.10929191e+00 1.12051643e-01
-1.24498856e+00 -5.35086274e-01 -1.28552473e+00 1.04045331e-01
-2.49054492e-01 2.22028494e-01 1.03027296e+00 1.01418018e+00
-1.21138668e+00 1.00952649e+00 -1.15228832e+00 -4.48576629e-01
8.51249814e-01 7.34351456e-01 -1.60457447e-01 2.23432288e-01
-5.76082945e-01 5.16583800e-01 5.20059586e-01 1.64680004e-01
-7.01009750e-01 -1.20501041e+00 -4.66434062e-01 3.19133997e-01
-8.15911070e-02 -6.73142731e-01 1.06414282e+00 -1.37164259e+00
-1.47258377e+00 5.81692219e-01 -5.35725839e-02 -7.69471109e-01
1.12129442e-01 -3.07113647e-01 -1.49124876e-01 -1.24072313e-01
-4.94587690e-01 9.18052018e-01 7.24883914e-01 -7.05566347e-01
-5.65927625e-01 -2.23436818e-01 1.09682046e-02 1.19551606e-01
-1.01871932e+00 3.35958693e-03 -3.95885669e-02 -7.06987143e-01
2.91564822e-01 -6.85638249e-01 -3.75804782e-01 2.57588953e-01
-8.68222304e-03 1.91979930e-01 7.07113862e-01 -1.34904474e-01
1.16732931e+00 -2.26428461e+00 -2.64896184e-01 3.84428382e-01
3.90416026e-01 3.97907197e-01 -4.41508055e-01 2.97525495e-01
-1.22536160e-01 2.95581043e-01 1.70531154e-01 -4.84363437e-02
-5.36280692e-01 1.65607892e-02 -3.15384120e-01 5.27533054e-01
5.46771586e-01 8.76523912e-01 -6.18712068e-01 1.17244542e-01
-2.32590973e-01 4.53338742e-01 -9.93515968e-01 -3.07416748e-02
2.97115315e-02 1.86584935e-01 -3.33477736e-01 5.64526141e-01
2.39405364e-01 -3.69575471e-01 4.87104356e-01 -4.50590312e-01
-4.93269354e-01 5.07026732e-01 -9.01589751e-01 1.34579206e+00
-3.54276210e-01 1.20216000e+00 3.18965167e-01 -1.09889519e+00
7.52634883e-01 1.91775605e-01 1.79031238e-01 -5.38540125e-01
4.18370664e-01 2.32057363e-01 3.47604275e-01 -3.15312564e-01
2.46830940e-01 8.85786116e-03 7.46719062e-01 2.19021171e-01
5.59386611e-01 2.11425468e-01 1.20655291e-01 -9.53354090e-02
1.16980541e+00 -1.87312216e-01 -1.41587391e-01 -9.69615579e-01
2.25460380e-02 2.90508885e-02 3.50564331e-01 1.03767848e+00
4.32498790e-02 2.66429037e-01 6.02021039e-01 -6.46401227e-01
-1.37633526e+00 -6.16146624e-01 -3.73222321e-01 1.52723241e+00
-3.84196311e-01 -4.28953916e-01 -6.09749258e-01 -2.83262819e-01
-3.66461240e-02 1.06229022e-01 -8.05761516e-01 -3.70720416e-01
-6.98990822e-01 -1.07077396e+00 9.10593331e-01 8.66325498e-01
6.77105129e-01 -7.78121889e-01 -7.85114765e-01 1.13410167e-01
7.09856927e-01 -1.18748772e+00 -3.76289994e-01 7.50244737e-01
-1.69507551e+00 -7.73290277e-01 -3.02597374e-01 -8.79887640e-01
1.00101399e+00 2.91055739e-01 8.89039874e-01 3.94998819e-01
-2.86427110e-01 4.06287581e-01 -1.14380598e-01 -3.92144799e-01
8.09301510e-02 5.18167138e-01 3.03727031e-01 -3.68402839e-01
1.96545765e-01 -9.79460239e-01 -1.02215505e+00 1.29641235e-01
-8.58677566e-01 -1.28891781e-01 8.96253288e-01 8.16590130e-01
3.33816439e-01 -1.29959345e-01 6.25752151e-01 -1.16803598e+00
9.31748033e-01 -5.04274547e-01 -5.64481497e-01 -1.52642161e-01
-9.34284329e-01 2.87986219e-01 5.57508826e-01 -6.86459661e-01
-1.00969875e+00 -5.42626418e-02 -2.07943633e-01 -2.99734324e-01
1.32164061e-01 6.25357628e-01 3.01227093e-01 -8.18488598e-01
1.19275057e+00 -3.15338671e-02 1.71573330e-02 -3.14657748e-01
3.29890639e-01 1.40510604e-01 2.10697576e-01 -6.34159863e-01
4.77353007e-01 4.16280121e-01 -1.25247151e-01 -1.06435013e+00
-8.59603643e-01 -1.86386123e-01 -4.72593784e-01 1.15396313e-01
2.71516085e-01 -1.04804397e+00 -5.30677795e-01 2.58386940e-01
-7.09026396e-01 -8.41197848e-01 -6.02915049e-01 5.02193987e-01
-2.47539982e-01 -2.62640864e-01 -8.82159114e-01 -6.10146046e-01
-4.00317371e-01 -7.32787609e-01 2.87770629e-01 3.92470062e-01
-1.34406626e-01 -1.12735569e+00 -2.32651517e-01 -1.55939803e-01
9.47402358e-01 -1.31257892e-01 8.71462464e-01 -6.95842206e-01
-4.40198004e-01 4.17496748e-02 -5.43510497e-01 5.72909594e-01
-3.61741371e-02 -1.58566926e-02 -1.13714039e+00 -5.92676520e-01
2.01805569e-02 -4.90094632e-01 1.05564213e+00 6.43824935e-01
1.25106859e+00 -5.06968617e-01 -2.68263340e-01 1.17615438e+00
1.47455204e+00 -2.49617219e-01 4.18044239e-01 2.62183845e-01
4.16821450e-01 3.44732195e-01 -2.45638713e-01 4.17896956e-01
-2.58969456e-01 -3.89322825e-02 2.14000151e-01 -1.46184683e-01
-3.88967603e-01 -7.02943280e-02 1.60179853e-01 6.46078110e-01
-3.47092777e-01 6.73868954e-02 -6.70249999e-01 4.26055998e-01
-1.56533968e+00 -7.70905793e-01 3.13781857e-01 2.02455854e+00
8.25042546e-01 3.69835287e-01 -1.30098201e-02 -1.33586109e-01
3.10551733e-01 3.74122262e-02 -8.69819880e-01 -3.43584985e-01
-1.92039162e-01 4.57057476e-01 1.09910595e+00 2.41848439e-01
-8.79976153e-01 1.01114118e+00 8.46639538e+00 4.85833049e-01
-1.48388350e+00 -5.92327677e-02 9.20773983e-01 -4.91360635e-01
-1.87553450e-01 -1.22201443e-01 -1.12087166e+00 -1.27684787e-01
1.18758798e+00 2.25993916e-01 5.44144392e-01 7.64716446e-01
2.47306466e-01 2.06181705e-01 -1.12995756e+00 6.13293290e-01
-3.06949854e-01 -1.79160547e+00 -2.42694952e-02 1.29508570e-01
9.04373229e-01 5.88633060e-01 3.96938801e-01 2.53143609e-01
3.23409498e-01 -1.08228827e+00 6.36892378e-01 1.23174503e-01
8.98133814e-01 -4.90872651e-01 3.86048436e-01 1.21665440e-01
-1.06940055e+00 -5.46463430e-01 -5.91760755e-01 -4.34885710e-01
-4.34255958e-01 4.63325799e-01 -7.17636466e-01 -3.36736947e-01
6.86118841e-01 5.55017769e-01 -5.81063926e-01 1.06563926e+00
1.02679305e-01 9.86438930e-01 -4.70309347e-01 -9.19938907e-02
3.26738715e-01 1.22025944e-01 3.35328519e-01 1.61453295e+00
2.29584292e-01 1.28154248e-01 -2.51962125e-01 9.00692225e-01
-2.97239155e-01 5.09177661e-03 -4.73358065e-01 -3.06282014e-01
5.64200580e-01 1.29499125e+00 -9.16178763e-01 -1.63038537e-01
-4.85979497e-01 5.48054099e-01 6.91022694e-01 6.92715526e-01
-2.73487628e-01 -3.37159067e-01 9.12483692e-01 5.21893501e-01
5.67181170e-01 -6.33402169e-01 -6.83883250e-01 -1.07208598e+00
-2.27803767e-01 -6.11398339e-01 2.91460361e-02 -1.42244399e-01
-1.03658652e+00 5.94502509e-01 -1.98004469e-01 -5.51529646e-01
2.21932083e-01 -1.00081742e+00 -8.01660359e-01 7.05810845e-01
-1.58448732e+00 -6.30784392e-01 -6.75977021e-02 3.95482033e-01
5.71334064e-01 -2.65717745e-01 5.31589508e-01 2.25886703e-01
-7.56606162e-01 9.26267564e-01 8.55017081e-02 1.69731006e-01
9.49150026e-02 -7.52564609e-01 5.24719656e-01 8.18644226e-01
1.17024537e-02 1.05801904e+00 6.85650945e-01 -3.72349709e-01
-1.47547007e+00 -1.13213515e+00 3.41841787e-01 -4.96629179e-02
7.57483780e-01 -3.42704862e-01 -8.39698136e-01 6.85890794e-01
2.24399805e-01 2.78882980e-01 9.33532476e-01 4.68002439e-01
-4.13650811e-01 -1.42616689e-01 -9.37383890e-01 6.45978510e-01
1.39217734e+00 -3.47110063e-01 4.65729460e-02 6.20472506e-02
6.06247544e-01 -2.41002351e-01 -8.12342525e-01 3.83835286e-01
8.22231889e-01 -8.59948933e-01 7.48661518e-01 -1.04039121e+00
2.98030972e-01 2.39664376e-01 -2.58158356e-01 -1.08934283e+00
-5.39914966e-01 -6.29971981e-01 -1.08361654e-01 8.00436258e-01
7.74193585e-01 -6.07853770e-01 1.07353020e+00 8.24941635e-01
-3.16154182e-01 -9.39802945e-01 -5.50025582e-01 -6.28698230e-01
2.32401267e-01 -3.24235290e-01 4.66859788e-02 7.48082638e-01
-1.60389662e-01 7.34664872e-02 1.64119214e-01 9.37810093e-02
2.44626135e-01 -4.64148313e-01 2.06023946e-01 -1.03921103e+00
-2.82387733e-01 -6.91014230e-01 -3.85631353e-01 -1.26609385e+00
-2.05395799e-02 -7.79911935e-01 -1.53024122e-01 -8.14793825e-01
-2.05798283e-01 -5.71411252e-01 -6.68455303e-01 5.67790091e-01
1.19371846e-01 3.26326758e-01 4.70989570e-02 1.71014592e-01
-3.59587669e-01 2.29977593e-01 1.06644118e+00 2.59820580e-01
-5.73311627e-01 -1.56545490e-01 -1.01654458e+00 7.31717527e-01
9.10588384e-01 -3.20025772e-01 -7.21431553e-01 -8.69653821e-01
4.72640425e-01 -4.99980539e-01 2.20132664e-01 -1.18751013e+00
4.23928857e-01 -2.13827882e-02 6.05885863e-01 3.14005345e-01
1.40512027e-02 -5.04439175e-01 -4.44639295e-01 4.90379542e-01
-7.48777628e-01 1.12417832e-01 7.24705040e-01 4.75564331e-01
4.57626618e-02 -2.43498147e-01 9.38696265e-01 -3.07219833e-01
-5.37163734e-01 5.36259770e-01 -7.69490540e-01 1.48273364e-01
4.33509111e-01 -4.52665508e-01 -9.58309472e-02 5.45967855e-02
-7.26189971e-01 -4.79362905e-02 1.18124709e-01 -4.90730815e-02
5.03977716e-01 -1.14258027e+00 -5.68167865e-01 5.57389736e-01
-3.58379185e-01 -2.95465529e-01 -8.35838839e-02 7.90251076e-01
-5.68688035e-01 3.31732571e-01 -3.47988665e-01 -2.65203565e-01
-9.32931304e-01 4.58753332e-02 7.11454391e-01 1.67083159e-01
-4.36105192e-01 1.33088374e+00 1.35780275e-01 1.98343629e-03
4.15226400e-01 -4.11737084e-01 5.22882342e-02 -2.57445186e-01
4.43370134e-01 2.48703942e-01 1.93230525e-01 1.15392298e-01
-7.31710494e-02 1.55816048e-01 -3.09135824e-01 1.23818696e-01
1.30931568e+00 4.66884077e-02 1.01746880e-01 1.78274021e-01
9.63903785e-01 -2.66565770e-01 -1.69553375e+00 -3.48688960e-01
-3.10861580e-02 -9.30761844e-02 4.88907576e-01 -3.53671491e-01
-1.37761414e+00 8.00417185e-01 6.50388241e-01 1.90380320e-01
1.02029896e+00 -2.48069674e-01 4.84336853e-01 8.96623552e-01
-1.26835600e-01 -1.03748381e+00 1.31232604e-01 6.43785655e-01
5.81789136e-01 -1.04250419e+00 1.69988066e-01 -2.19575360e-01
-1.63471624e-01 1.28608572e+00 8.42706144e-01 -7.31433272e-01
1.09010553e+00 6.76586449e-01 -1.52040944e-01 -2.99066663e-01
-1.08530450e+00 1.67555995e-02 1.11050591e-01 4.28090870e-01
9.28365767e-01 -4.02596861e-01 -1.15129419e-01 3.91820371e-01
-2.19969302e-01 2.13630833e-02 1.12576738e-01 9.53901529e-01
-7.99992859e-01 -8.75003278e-01 2.83464611e-01 8.44201684e-01
-6.50710702e-01 -6.26785815e-01 -1.81065165e-02 6.66392565e-01
1.86046109e-01 5.94271302e-01 3.50943208e-01 -3.70862156e-01
2.85611480e-01 6.85142353e-02 7.26191103e-01 -7.70033717e-01
-9.32131588e-01 -1.41416743e-01 1.66537128e-02 -3.63016576e-01
-1.68237850e-01 -4.02381271e-01 -1.04589581e+00 -3.01027507e-01
-6.74053550e-01 -1.05289780e-01 6.87723935e-01 9.31875229e-01
6.58259153e-01 6.39592469e-01 3.40452671e-01 -7.38175571e-01
-4.97847170e-01 -7.72255182e-01 -3.91341895e-01 -1.06618956e-01
3.75228047e-01 -2.33240768e-01 -4.44080859e-01 2.15229318e-01] | [8.591538429260254, 3.2407240867614746] |
645ad8ae-2038-4fcc-9bfa-9d73a316d3f2 | graph-based-selective-outlier-ensembles | 1804.06378 | null | http://arxiv.org/abs/1804.06378v1 | http://arxiv.org/pdf/1804.06378v1.pdf | Graph-based Selective Outlier Ensembles | An ensemble technique is characterized by the mechanism that generates the
components and by the mechanism that combines them. A common way to achieve the
consensus is to enable each component to equally participate in the aggregation
process. A problem with this approach is that poor components are likely to
negatively affect the quality of the consensus result. To address this issue,
alternatives have been explored in the literature to build selective classifier
and cluster ensembles, where only a subset of the components contributes to the
computation of the consensus. Of the family of ensemble methods, outlier
ensembles are the least studied. Only recently, the selection problem for
outlier ensembles has been discussed. In this work we define a new graph-based
class of ranking selection methods. A method in this class is characterized by
two main steps: (1) Mapping the rankings onto a graph structure; and (2) Mining
the resulting graph to identify a subset of rankings. We define a specific
instance of the graph-based ranking selection class. Specifically, we map the
problem of selecting ensemble components onto a mining problem in a graph. An
extensive evaluation was conducted on a variety of heterogeneous data and
methods. Our empirical results show that our approach outperforms
state-of-the-art selective outlier ensemble techniques. | ['Giovanni Stilo', 'Carlotta Domeniconi', 'Hamed Sarvari'] | 2018-04-17 | null | null | null | null | ['outlier-ensembles'] | ['methodology'] | [ 1.28250882e-01 -4.95360456e-02 1.55744284e-01 -2.40668714e-01
-4.71547186e-01 -4.50066626e-01 5.38449645e-01 7.73988664e-01
-6.81794435e-02 7.32350230e-01 7.41895437e-02 1.16300620e-01
-6.01930678e-01 -1.20150769e+00 -3.55865300e-01 -9.27681565e-01
-2.58877933e-01 6.34009182e-01 3.10848743e-01 -2.86728621e-01
5.36766171e-01 3.85932267e-01 -1.97581494e+00 2.77715802e-01
1.48747361e+00 9.06827986e-01 -3.80246460e-01 2.36086205e-01
-8.66112709e-02 5.99736631e-01 -6.20325327e-01 -3.07987720e-01
2.32913777e-01 -7.15969145e-01 -5.60083449e-01 7.26088658e-02
3.00982118e-01 3.67419153e-01 3.09867591e-01 1.09314740e+00
4.66651201e-01 8.41763169e-02 7.70449877e-01 -1.45932460e+00
1.54772952e-01 8.00782382e-01 -5.14816999e-01 -5.14467731e-02
4.90453511e-01 -4.31051940e-01 1.21501219e+00 -9.71368611e-01
7.74032295e-01 8.12491655e-01 4.54791516e-01 7.50304461e-02
-1.43810749e+00 -5.16737998e-01 1.94646925e-01 2.11519197e-01
-1.59611058e+00 -2.34346643e-01 7.63877034e-01 -5.17767251e-01
6.39915228e-01 6.22648001e-01 5.70080698e-01 5.72494328e-01
2.65878201e-01 3.08921427e-01 1.45887375e+00 -4.75712299e-01
6.25808060e-01 2.71188710e-02 5.79770207e-01 4.47769970e-01
9.91841972e-01 -3.95124890e-02 -6.80619478e-01 -7.41413653e-01
-1.06253080e-01 -1.54938191e-01 -2.46422067e-01 -5.99261343e-01
-7.55303860e-01 8.05503905e-01 1.73922986e-01 4.56880480e-01
-4.30200934e-01 -1.41192302e-01 3.80757511e-01 6.41889811e-01
8.08172822e-01 7.65061378e-01 -1.01008557e-01 3.17503422e-01
-9.43405747e-01 3.08261752e-01 1.18942523e+00 3.95663828e-01
8.24253559e-01 -2.09234834e-01 -6.66255457e-03 5.91163754e-01
2.39941463e-01 7.85790160e-02 5.51015958e-02 -4.06447619e-01
4.21230823e-01 1.22721541e+00 -8.96149948e-02 -1.19762206e+00
-3.12229216e-01 -6.99343562e-01 -8.92904341e-01 4.41158235e-01
3.70038092e-01 -9.65796039e-02 -7.54132688e-01 1.54946172e+00
4.12435621e-01 2.02275112e-01 -2.15946585e-01 6.23764694e-01
5.47481060e-01 2.91099787e-01 -1.62065625e-01 -4.93961751e-01
8.79035413e-01 -4.14836407e-01 -3.41909051e-01 3.82497728e-01
4.87483919e-01 -6.23672307e-01 3.73158872e-01 7.41147101e-01
-6.77849829e-01 -3.51126224e-01 -1.23009586e+00 7.84939289e-01
-3.93902093e-01 -4.92358319e-02 3.21200401e-01 7.24907756e-01
-1.15061605e+00 9.73069191e-01 -4.86411870e-01 -5.22859275e-01
-1.74831986e-01 5.70181549e-01 -3.64653289e-01 5.54442741e-02
-1.12370849e+00 1.02913725e+00 4.60005194e-01 -1.20812058e-01
-3.58844489e-01 -8.18369538e-02 -3.15611243e-01 1.19811408e-02
6.71961486e-01 -8.91874254e-01 4.48020399e-01 -1.34724522e+00
-9.32343423e-01 6.23294115e-01 4.14793342e-02 -2.71174967e-01
6.07078314e-01 -1.35172412e-01 -5.20179570e-01 -2.56166101e-01
5.66498525e-02 -2.70750940e-01 9.09940302e-01 -1.70639527e+00
-8.01762819e-01 -6.79972470e-01 -1.87987059e-01 2.37797484e-01
-2.60665864e-01 -7.06263408e-02 -2.09332645e-01 -5.16466081e-01
3.14050198e-01 -1.13051069e+00 -4.11639780e-01 -9.14177358e-01
-6.70282900e-01 -5.14963567e-01 6.97070003e-01 -2.59892255e-01
1.82833314e+00 -1.80166256e+00 5.66892743e-01 1.09318101e+00
5.44281304e-01 -1.80957034e-01 1.89963020e-02 1.06291533e+00
-1.74459428e-01 4.26248521e-01 -2.22938865e-01 -3.04615200e-01
-1.82993084e-01 -1.64628886e-02 -3.82022917e-01 4.99375015e-01
-2.93247610e-01 1.19150206e-01 -9.65929329e-01 -3.11733276e-01
2.05604024e-02 -1.06359664e-02 -3.98049027e-01 1.80389404e-01
-1.34651989e-01 3.43106121e-01 -5.25524676e-01 5.80596983e-01
4.34791893e-01 8.74210671e-02 6.41912222e-01 -1.15288153e-01
-1.82565615e-01 9.90337431e-02 -1.70836461e+00 9.61068332e-01
4.89828251e-02 2.00829394e-02 -1.70960948e-01 -8.64618540e-01
1.11354804e+00 2.25968033e-01 9.75700855e-01 1.78313591e-02
9.06449035e-02 7.26414800e-01 3.24771106e-01 -1.05476581e-01
7.04603195e-01 2.25873426e-01 -2.01819837e-01 7.35171199e-01
-1.48218676e-01 7.83627555e-02 6.86438501e-01 3.86766821e-01
1.46672785e+00 -1.30968302e-01 7.20961034e-01 -5.15622735e-01
5.24733543e-01 1.05717674e-01 5.52063107e-01 9.34811831e-01
2.60873407e-01 6.04703665e-01 6.72139347e-01 -5.44010997e-01
-7.75601029e-01 -8.94935250e-01 1.57215059e-01 6.90327406e-01
1.20982990e-01 -7.95701504e-01 -6.48893118e-01 -9.75527108e-01
2.25891843e-01 6.38432741e-01 -6.59065962e-01 -1.64436206e-01
-3.57806027e-01 -8.30211699e-01 2.18339097e-02 -3.41437161e-02
1.44972727e-01 -8.90018761e-01 -3.60972732e-01 2.43852675e-01
-2.10177079e-01 -4.91894960e-01 -5.84860668e-02 1.94254518e-01
-1.12606752e+00 -1.49805856e+00 1.20247073e-01 -2.75327384e-01
9.98282075e-01 1.62908375e-01 1.53623664e+00 5.25755584e-01
2.34196499e-01 3.01625013e-01 -8.90409172e-01 -5.04396498e-01
-4.71646607e-01 3.67712528e-01 3.73727232e-01 4.69037205e-01
3.08339775e-01 -8.04862142e-01 -3.01839113e-01 4.14815724e-01
-8.86451185e-01 -3.53884578e-01 5.57398081e-01 7.20440686e-01
6.47473872e-01 4.36501771e-01 6.20997727e-01 -1.27936292e+00
9.79123116e-01 -7.10627198e-01 -4.28332925e-01 3.65165234e-01
-9.41727281e-01 2.82560259e-01 7.46620655e-01 -3.43721546e-02
-5.79695761e-01 2.46734470e-01 3.58060509e-01 -1.22738227e-01
2.40236558e-02 9.84603941e-01 -1.28531739e-01 1.42975906e-02
6.03910744e-01 -2.11328920e-02 -1.62763506e-01 -3.52046430e-01
1.53574243e-01 5.01569390e-01 6.86299801e-02 -6.71245813e-01
8.29294264e-01 4.13056523e-01 2.04641908e-01 -7.49509931e-01
-4.37213570e-01 -5.87958157e-01 -6.37508869e-01 -5.83448112e-01
4.98088479e-01 -4.29771811e-01 -2.96197563e-01 2.59420693e-01
-8.14627051e-01 2.73383617e-01 -2.97042042e-01 3.24413866e-01
-3.07643712e-01 3.88481289e-01 2.08340865e-02 -8.98375213e-01
-4.09115702e-01 -1.05308509e+00 7.49649584e-01 -1.60749126e-02
-5.23640931e-01 -5.10034740e-01 6.18680000e-01 -4.94189411e-02
1.56594276e-01 7.89131224e-01 9.11313057e-01 -1.18045056e+00
-5.08749425e-01 -4.40966398e-01 3.67169052e-01 1.75310060e-01
1.52865440e-01 2.79741734e-01 -5.53312957e-01 -4.82075393e-01
-1.43372089e-01 1.57919928e-01 1.17121351e+00 2.77223259e-01
8.92495036e-01 4.69488576e-02 -6.10019743e-01 1.99935943e-01
1.59962285e+00 9.00218114e-02 6.21956825e-01 2.58240134e-01
5.71779072e-01 6.74677253e-01 6.42359257e-01 3.36398214e-01
1.37159094e-01 8.40389073e-01 5.74275792e-01 1.76205456e-01
3.00362647e-01 -4.38837372e-02 4.16251898e-01 1.06550574e+00
-7.53274024e-01 -4.67924744e-01 -8.28813612e-01 2.95150816e-01
-2.28881931e+00 -1.00563908e+00 -4.18193728e-01 2.62724423e+00
3.30122292e-01 1.22354172e-01 4.05651361e-01 4.95593429e-01
8.54724824e-01 7.49841779e-02 -6.80929348e-02 -3.83259803e-01
1.67298373e-02 1.73228115e-01 2.45315909e-01 3.33137721e-01
-1.07861257e+00 4.35165584e-01 5.35885572e+00 5.55644095e-01
-7.74482310e-01 -2.20719233e-01 4.12924826e-01 1.11366205e-01
-4.10410970e-01 3.57677996e-01 -5.24826884e-01 5.03315389e-01
7.24502146e-01 -1.76183075e-01 1.63923085e-01 7.61328340e-01
2.66344585e-02 -4.01225626e-01 -1.13031650e+00 6.10196233e-01
1.73252255e-01 -8.43097806e-01 2.63156816e-02 3.93962324e-01
1.01974094e+00 9.80617944e-03 -3.64464968e-01 -3.04586757e-02
6.60471499e-01 -1.01263452e+00 5.01040578e-01 8.18080306e-01
4.71856862e-01 -1.05845797e+00 9.77100611e-01 1.44019946e-01
-1.27575135e+00 -1.78187221e-01 -1.23926513e-01 -2.58922577e-04
-9.64667499e-02 1.22385597e+00 -4.72188205e-01 1.21742165e+00
6.75001502e-01 5.31210661e-01 -6.11782014e-01 1.33214629e+00
-2.05065176e-01 6.86283469e-01 -4.01812375e-01 -1.17422275e-01
-2.48877183e-01 -6.85501516e-01 1.04887104e+00 7.96772838e-01
6.35943294e-01 -5.79681695e-02 4.98708755e-01 3.77630085e-01
-6.39443398e-02 4.79421735e-01 -9.47816551e-01 3.15914303e-01
5.94787478e-01 1.23481965e+00 -1.02228940e+00 -2.24934176e-01
-1.81911469e-01 5.43760538e-01 5.26733398e-01 1.46415621e-01
-4.68016952e-01 -3.01475197e-01 4.61223960e-01 2.86363959e-01
1.12114109e-01 -2.77025372e-01 -4.18295532e-01 -1.05110312e+00
-2.33925637e-02 -1.13074124e+00 8.09197128e-01 -1.67566225e-01
-1.45607412e+00 8.52875233e-01 3.34725343e-02 -1.41781509e+00
-1.37277246e-01 -2.43685693e-01 -7.45874584e-01 6.38635159e-01
-8.37416768e-01 -6.95622027e-01 -5.98544776e-01 4.85541344e-01
9.89347100e-02 -2.37441212e-01 8.69096875e-01 1.43376559e-01
-4.84150589e-01 1.18544631e-01 3.16313326e-01 -3.29484344e-01
8.44999969e-01 -1.62017202e+00 -6.63837194e-02 1.05508411e+00
2.95324862e-01 5.82769036e-01 9.15323853e-01 -1.02415216e+00
-1.14083791e+00 -1.07582080e+00 1.05559707e+00 -6.92735434e-01
3.07558626e-01 -1.53521866e-01 -7.50589192e-01 2.84130394e-01
1.69943064e-01 -2.73434252e-01 6.53395236e-01 4.55637574e-01
-2.15483993e-01 -4.27730054e-01 -9.27769184e-01 5.13513565e-01
1.05072069e+00 6.90216050e-02 -4.40272868e-01 -1.17620438e-01
-1.05479464e-01 -1.08506605e-01 -8.48097086e-01 5.72358489e-01
4.72411603e-01 -1.43868613e+00 5.17897964e-01 -3.50977004e-01
3.35008502e-01 -8.10273945e-01 2.36299783e-02 -1.96855891e+00
-2.11259976e-01 -5.23336589e-01 -1.94103852e-01 1.15056491e+00
4.29924995e-01 -9.18893218e-01 4.94645536e-01 6.90130442e-02
-3.57568860e-02 -7.77120173e-01 -8.09301138e-01 -7.13754714e-01
-3.23076993e-01 -1.13788277e-01 7.17559457e-01 8.53117287e-01
4.06711251e-02 3.31897020e-01 -3.50888103e-01 2.84737535e-02
8.45774889e-01 4.25659060e-01 9.22584593e-01 -1.85394061e+00
-1.24597885e-01 -4.11327064e-01 -5.54058909e-01 -2.31577419e-02
-1.63887247e-01 -1.02276397e+00 -8.36312026e-02 -1.59957552e+00
2.95022398e-01 -6.68171704e-01 -6.09555960e-01 1.03131995e-01
-3.23483258e-01 8.36245939e-02 1.33522585e-01 5.74342370e-01
-7.32391357e-01 1.15527481e-01 6.36095762e-01 1.48462787e-01
-5.21122754e-01 2.31309876e-01 -7.63837695e-01 5.61229348e-01
9.37800586e-01 -8.15979481e-01 -4.03619975e-01 1.67392969e-01
6.53767884e-01 -2.70979375e-01 -4.11017053e-02 -1.18045986e+00
2.60161787e-01 -9.24671516e-02 1.10852785e-01 -5.82324982e-01
-2.10873440e-01 -8.80577326e-01 7.30321288e-01 5.22998571e-01
-1.77830290e-02 4.58395630e-01 -4.67700422e-01 7.30600059e-01
-4.12348717e-01 -4.01259586e-02 5.47184169e-01 -3.42509076e-02
-5.10868728e-01 1.91198543e-01 -1.63193583e-01 -1.22948661e-01
1.09725380e+00 -2.36628443e-01 -3.49369407e-01 -4.85208690e-01
-7.67060816e-01 2.00517952e-01 6.95915878e-01 3.87942463e-01
4.05761689e-01 -1.10626042e+00 -1.13423359e+00 -1.72348946e-01
2.61086762e-01 -2.41969407e-01 -2.80089527e-01 1.17538881e+00
-3.11352700e-01 -4.07064371e-02 -1.93120595e-02 -5.54147124e-01
-1.52307701e+00 1.33918390e-01 3.15061450e-01 -7.75766492e-01
-2.79812306e-01 3.48012835e-01 -2.72850811e-01 -1.97837234e-01
-8.93543288e-02 6.99166656e-02 -5.50161362e-01 4.28261966e-01
7.93832913e-02 7.09218383e-01 3.69742900e-01 -7.23973393e-01
-6.40950978e-01 4.73371416e-01 2.30985001e-01 8.90996605e-02
1.43116176e+00 1.37972906e-01 -8.64176869e-01 5.02507389e-01
5.52567959e-01 6.06877506e-01 -3.33556265e-01 1.01918198e-01
5.02092183e-01 -3.40329409e-01 -3.12994838e-01 -7.82441676e-01
-9.81916070e-01 9.64738056e-02 2.41774485e-01 8.38170648e-01
1.38417077e+00 -2.38164648e-01 7.27496669e-02 9.55341235e-02
6.68134153e-01 -1.26960135e+00 -1.44121006e-01 4.88491327e-01
8.67548347e-01 -1.02934659e+00 4.30103064e-01 -6.82303131e-01
-4.57101732e-01 1.23313034e+00 5.82495630e-01 -4.96448994e-01
7.89845169e-01 1.23766296e-01 -1.66647837e-01 -4.91220951e-01
-7.91326582e-01 -3.77977580e-01 6.44078791e-01 1.86570868e-01
5.50285816e-01 3.16911668e-01 -1.20078695e+00 5.22148550e-01
-9.16245505e-02 -2.04522431e-01 4.14505988e-01 7.92079508e-01
-5.29500723e-01 -1.50637186e+00 -5.48061252e-01 1.02870703e+00
-2.76822686e-01 1.28071919e-01 -1.05343068e+00 6.62698567e-01
4.35030460e-01 1.33813274e+00 -2.44957909e-01 -7.93931723e-01
5.03995597e-01 1.56981826e-01 3.17334294e-01 -7.20038235e-01
-1.01747763e+00 -2.19613224e-01 6.11046851e-01 -6.06779039e-01
-5.18146992e-01 -6.04944944e-01 -8.29655945e-01 -2.43329898e-01
-7.95820832e-01 7.00232446e-01 5.32834291e-01 8.17605674e-01
4.43314046e-01 4.33619916e-01 8.69652271e-01 -6.10083520e-01
-5.73205292e-01 -7.36468852e-01 -8.42175305e-01 6.16525948e-01
-1.92600265e-01 -7.84594536e-01 -6.15281224e-01 -5.01259744e-01] | [7.669680118560791, 4.563941955566406] |
f69d11c3-0b25-4767-99f7-cacd189459ae | on-credit-assignment-in-hierarchical | 2203.03292 | null | https://arxiv.org/abs/2203.03292v1 | https://arxiv.org/pdf/2203.03292v1.pdf | On Credit Assignment in Hierarchical Reinforcement Learning | Hierarchical Reinforcement Learning (HRL) has held longstanding promise to advance reinforcement learning. Yet, it has remained a considerable challenge to develop practical algorithms that exhibit some of these promises. To improve our fundamental understanding of HRL, we investigate hierarchical credit assignment from the perspective of conventional multistep reinforcement learning. We show how e.g., a 1-step `hierarchical backup' can be seen as a conventional multistep backup with $n$ skip connections over time connecting each subsequent state to the first independent of actions inbetween. Furthermore, we find that generalizing hierarchy to multistep return estimation methods requires us to consider how to partition the environment trace, in order to construct backup paths. We leverage these insight to develop a new hierarchical algorithm Hier$Q_k(\lambda)$, for which we demonstrate that hierarchical credit assignment alone can already boost agent performance (i.e., when eliminating generalization or exploration). Altogether, our work yields fundamental insight into the nature of hierarchical backups and distinguishes this as an additional basis for reinforcement learning research. | ['Aske Plaat', 'Thomas M. Moerland', 'Joery A. de Vries'] | 2022-03-07 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 2.41451878e-02 1.03185773e-01 -3.84897470e-01 2.50043739e-02
-6.59729004e-01 -7.91704834e-01 4.20801431e-01 2.97156692e-01
-7.28802264e-01 1.01782084e+00 -6.62249178e-02 -5.85434616e-01
-3.60810608e-01 -8.84411693e-01 -7.66388237e-01 -7.89131284e-01
-4.80342984e-01 3.85902196e-01 3.24182361e-01 -6.29169583e-01
4.72885698e-01 4.89365131e-01 -1.37065804e+00 -7.60192499e-02
7.87868083e-01 5.12907863e-01 1.77776232e-01 5.84116757e-01
2.91586339e-01 9.23588574e-01 -5.29324710e-01 -7.02011911e-03
3.04065138e-01 -7.66719759e-01 -1.09148252e+00 -4.48635481e-02
-1.71460006e-02 -6.84989691e-01 -3.30563128e-01 7.24222958e-01
3.12782228e-01 5.17662168e-01 2.85956204e-01 -1.35761154e+00
-3.37348074e-01 8.40731740e-01 -5.77618062e-01 2.75518030e-01
1.82214901e-01 4.48970765e-01 1.18789983e+00 -2.11624399e-01
4.41113919e-01 1.17329741e+00 5.82949996e-01 6.78190649e-01
-1.63445187e+00 -5.75746179e-01 5.71182489e-01 2.71675020e-01
-8.38744283e-01 -3.42266291e-01 5.55069983e-01 -3.01352769e-01
1.18438554e+00 -5.17793465e-03 9.52588797e-01 7.47387052e-01
2.62254328e-01 1.05759895e+00 1.52432024e+00 -5.29407442e-01
5.24927378e-01 -1.66315839e-01 8.87659565e-02 1.04581738e+00
1.06646746e-01 7.34254360e-01 -7.38832831e-01 -7.99301919e-03
7.83613384e-01 -2.63553947e-01 -2.03960817e-02 -7.25970745e-01
-6.94252491e-01 1.06613982e+00 5.02307415e-01 -9.53480229e-02
-1.43478453e-01 6.72008753e-01 2.56461889e-01 7.20003724e-01
-3.19110811e-01 8.18515480e-01 -3.69394630e-01 -2.70753115e-01
-5.60231566e-01 5.17547786e-01 5.91850400e-01 5.98295271e-01
1.14550662e+00 3.13187212e-01 2.48923048e-01 5.43716192e-01
1.82758227e-01 1.84383750e-01 2.86736131e-01 -1.45130312e+00
2.63808280e-01 4.17030752e-01 3.42208505e-01 -4.96217161e-01
-5.76708913e-01 -3.76199067e-01 -2.90037453e-01 6.32240057e-01
3.44945997e-01 -3.29566568e-01 -5.02941906e-01 2.30620742e+00
2.14302063e-01 3.83219123e-02 1.55436829e-01 5.08548260e-01
-8.05267468e-02 5.34625590e-01 9.66940001e-02 -2.61136353e-01
8.45356166e-01 -9.73704398e-01 -2.40821108e-01 -3.51622492e-01
9.02613580e-01 -9.10859481e-02 1.17436206e+00 5.17927170e-01
-1.24953413e+00 -2.81127214e-01 -1.19275939e+00 3.45863044e-01
-2.60054022e-01 -4.81195331e-01 1.14628339e+00 6.44968212e-01
-1.24571574e+00 1.05599856e+00 -1.10235369e+00 -2.24892959e-01
1.47243768e-01 5.88411510e-01 -1.44965380e-01 9.29505005e-02
-1.35768318e+00 1.00028050e+00 4.01372880e-01 -9.26875025e-02
-1.48400426e+00 -3.11701894e-01 -7.56229162e-01 -2.93120984e-02
8.24956357e-01 -6.53732896e-01 1.61915350e+00 -7.71087527e-01
-1.69952369e+00 2.54467875e-01 -4.71879281e-02 -6.79732442e-01
3.06782067e-01 -1.37078986e-01 2.99260855e-01 1.70303553e-01
1.46022871e-01 8.19070280e-01 6.15438223e-01 -1.20568550e+00
-9.43944752e-01 -4.39783424e-01 7.51185477e-01 5.42321026e-01
-2.90002018e-01 -2.59168476e-01 4.39565592e-02 -4.31819975e-01
-8.23992044e-02 -1.20405054e+00 -6.42712474e-01 -4.49598879e-01
-1.64793432e-01 -3.42460930e-01 1.44591540e-01 -1.29189163e-01
1.22532952e+00 -1.96370614e+00 4.57428038e-01 2.06430852e-01
1.87295735e-01 -2.20598921e-01 -3.25633973e-01 8.73751044e-01
5.10601066e-02 1.03637278e-01 -3.48897636e-01 -7.37438127e-02
2.77215570e-01 3.81747901e-01 -3.26706856e-01 3.59878182e-01
1.14428326e-02 9.43096042e-01 -1.02951348e+00 -2.01570392e-01
2.42183674e-02 -1.37728482e-01 -1.06584585e+00 4.21757698e-02
-3.42098355e-01 2.94468760e-01 -4.91085291e-01 3.31255943e-01
1.37116626e-01 -2.27716118e-02 4.21441913e-01 4.08015847e-01
-2.83411890e-01 3.89819562e-01 -1.15223527e+00 1.53749073e+00
-2.06335053e-01 2.73818612e-01 1.82237729e-01 -1.14027798e+00
3.43531847e-01 4.68804426e-02 6.24516785e-01 -8.64889383e-01
-9.84718576e-02 5.93359545e-02 1.67790756e-01 -1.66125998e-01
5.91142416e-01 -4.69869107e-01 -3.95489812e-01 7.24656641e-01
-1.13991410e-01 -1.23652928e-01 4.15560007e-01 2.68682450e-01
1.25428820e+00 4.72581476e-01 2.25497097e-01 -2.27601454e-01
1.58613488e-01 2.83003330e-01 7.23451674e-01 1.26199675e+00
-4.56474662e-01 -4.08411145e-01 7.94308603e-01 -9.92454290e-02
-8.45484197e-01 -1.15332294e+00 3.74025643e-01 1.63212943e+00
2.41470039e-01 -3.85615706e-01 -7.89411366e-01 -7.35081553e-01
2.03743830e-01 5.68013966e-01 -7.55921543e-01 -3.65464300e-01
-8.55706215e-01 -7.92291164e-01 6.22421324e-01 7.78561413e-01
4.98733878e-01 -1.30104256e+00 -1.05048716e+00 4.69052106e-01
9.10543278e-02 -4.42012042e-01 -2.24780187e-01 9.53144372e-01
-1.02664900e+00 -1.16957521e+00 -2.02051699e-01 -7.53908157e-01
3.84357631e-01 2.79040426e-01 9.78341699e-01 2.28440821e-01
-2.49489397e-03 5.64575434e-01 -2.61700720e-01 8.60634223e-02
-2.79928714e-01 1.56756535e-01 2.03695714e-01 -7.38235056e-01
-2.51726396e-02 -8.08871925e-01 -6.08480394e-01 1.49767071e-01
-8.30534160e-01 -5.03008850e-02 5.41933119e-01 1.06550360e+00
2.19862595e-01 1.37847915e-01 9.17380214e-01 -8.12700987e-01
9.54762816e-01 -3.59929562e-01 -9.12204325e-01 1.06389053e-01
-7.93515921e-01 4.34791684e-01 7.52750933e-01 -1.24344960e-01
-9.14420366e-01 -1.25459418e-01 -2.18492240e-01 1.28999457e-01
-3.99244167e-02 6.30269587e-01 1.30545452e-01 -1.40081599e-01
6.23289168e-01 4.48787630e-01 4.32207957e-02 -1.97406918e-01
5.81762850e-01 3.21967639e-02 2.05444187e-01 -1.10813332e+00
5.73291600e-01 1.59411803e-01 1.11299329e-01 -5.78986406e-01
-7.08171368e-01 -2.08387785e-02 -3.83356839e-01 -6.10408746e-02
5.67074120e-01 -6.52182341e-01 -1.44131744e+00 1.96299270e-01
-4.58131045e-01 -1.23219144e+00 -4.48275030e-01 1.69782981e-01
-1.16257823e+00 3.10049027e-01 -1.03632987e+00 -7.94991195e-01
2.26242244e-01 -1.39617240e+00 3.29952568e-01 2.06811696e-01
-1.53504699e-01 -8.95917952e-01 3.94753009e-01 1.64391771e-01
1.81086525e-01 -2.01600477e-01 1.25330508e+00 -2.76470065e-01
-8.06455970e-01 5.44835806e-01 2.45488286e-01 2.75187194e-02
-2.29025465e-02 -3.38148803e-01 -5.58650136e-01 -7.11424947e-01
-1.13299094e-01 -9.11762655e-01 9.68147039e-01 3.11718583e-01
9.82651889e-01 -2.62688369e-01 -2.17197999e-01 3.74269992e-01
1.47773063e+00 6.61657453e-01 3.26563179e-01 9.11856771e-01
2.83456445e-01 4.19759721e-01 6.62815273e-01 6.14693880e-01
5.78323483e-01 5.28921902e-01 4.39680576e-01 2.21616730e-01
2.76278198e-01 -4.06536639e-01 6.63923264e-01 5.15318811e-01
4.08224650e-02 1.10704325e-01 -6.71042204e-01 4.83158946e-01
-1.83869660e+00 -1.09614396e+00 5.02967656e-01 2.26064157e+00
1.08083153e+00 4.04610097e-01 4.39212412e-01 1.06589004e-01
3.41743678e-01 1.65620968e-01 -9.90565360e-01 -6.62580848e-01
1.74869388e-01 4.23359782e-01 4.44485933e-01 9.36618209e-01
-8.31316650e-01 1.34774244e+00 7.23807812e+00 5.86936533e-01
-7.73047626e-01 -1.72723070e-01 3.16192567e-01 -7.10904226e-03
-4.24339950e-01 2.62067676e-01 -9.46245968e-01 5.36001585e-02
9.71502423e-01 -1.08408518e-01 1.01736546e+00 7.65392244e-01
1.72395289e-01 -4.78039205e-01 -1.39810610e+00 3.68375331e-01
-3.14412296e-01 -1.04398704e+00 -9.91459787e-02 1.26760200e-01
7.93635011e-01 -1.46143541e-01 1.16347037e-01 8.24461162e-01
1.29501712e+00 -9.12352920e-01 6.22442782e-01 -1.87372416e-01
3.05170149e-01 -1.18921864e+00 4.14207578e-02 5.73219717e-01
-1.25402927e+00 -5.29029548e-01 -5.13091572e-02 -3.78205597e-01
-6.13303147e-02 -3.31279367e-01 -7.07613051e-01 3.98376703e-01
4.36116725e-01 4.04555678e-01 -4.16977644e-01 8.05012643e-01
-4.42117780e-01 4.65947419e-01 -2.17455968e-01 -4.77580586e-03
6.65075719e-01 -1.89197600e-01 1.86800942e-01 8.09374988e-01
-9.01395157e-02 2.21719280e-01 6.30052030e-01 6.10739589e-01
1.19561985e-01 -3.22463334e-01 -5.44337094e-01 -8.31539109e-02
6.52595103e-01 8.22666645e-01 -7.39075005e-01 -1.51625499e-01
-2.35824227e-01 8.25847328e-01 8.50824237e-01 3.98022920e-01
-8.57749820e-01 -2.93136388e-01 7.51813471e-01 -2.39949703e-01
3.17362994e-01 -5.08798420e-01 -4.94459271e-02 -9.60111976e-01
-3.70278746e-01 -9.71060336e-01 5.27575612e-01 -2.97974914e-01
-8.78975987e-01 1.76318496e-01 -4.10910994e-02 -7.65004814e-01
-4.64214563e-01 -3.46205741e-01 -2.85868287e-01 4.92688060e-01
-1.71452725e+00 -6.19824350e-01 2.09651455e-01 6.79824471e-01
6.27162516e-01 1.79923758e-01 8.92069221e-01 -1.04029737e-01
-8.41618896e-01 6.66961730e-01 2.21507445e-01 -1.26383886e-01
4.28118765e-01 -1.54128730e+00 2.60264754e-01 7.48440921e-01
-3.89969535e-02 8.15265119e-01 6.36348844e-01 -6.45738661e-01
-1.59751558e+00 -6.89307094e-01 2.32220694e-01 -1.26487434e-01
8.32274020e-01 -1.79922339e-02 -6.49224699e-01 1.00426364e+00
2.69296825e-01 -5.48240960e-01 6.80482626e-01 4.01926845e-01
-2.11803049e-01 -1.84439540e-01 -7.62794256e-01 9.73257840e-01
1.07439959e+00 -4.44783777e-01 -7.62461305e-01 -1.19051166e-01
6.23182356e-01 -1.31923854e-01 -7.79414892e-01 1.69255659e-01
5.00224650e-01 -1.08278739e+00 9.21362042e-01 -9.15843368e-01
4.21424091e-01 -3.18489313e-01 -1.20974727e-01 -1.66420436e+00
-4.65602130e-01 -8.70712996e-01 -1.33934259e-01 6.53131068e-01
3.70031148e-01 -8.06690991e-01 1.07179928e+00 4.88176137e-01
-3.73322070e-01 -9.71403122e-01 -6.94150865e-01 -9.86946464e-01
6.16897702e-01 -1.90427050e-01 3.25160593e-01 7.10534096e-01
5.52165270e-01 1.92645058e-01 -3.47935677e-01 6.78464845e-02
6.81005716e-01 3.69473219e-01 6.52689993e-01 -8.79515171e-01
-7.74747193e-01 -6.33874476e-01 2.74912387e-01 -1.52109838e+00
2.26537794e-01 -7.15051234e-01 3.48232448e-01 -1.40930092e+00
2.37174958e-01 -9.23856854e-01 -7.12180078e-01 7.44021237e-01
-8.28258991e-02 -5.11235278e-03 3.83972019e-01 7.40437880e-02
-8.11828613e-01 7.66039431e-01 1.25376701e+00 1.23249106e-01
-4.89243209e-01 -1.42343834e-01 -9.46244955e-01 6.77169919e-01
1.13018239e+00 -4.38444614e-01 -6.74446881e-01 -3.63672465e-01
4.00171340e-01 5.24101317e-01 2.23946691e-01 -8.58432651e-01
2.87566185e-01 -7.69720078e-01 1.33956105e-01 -2.61481732e-01
3.49259764e-01 -4.73203987e-01 -3.38746458e-01 8.33234489e-01
-7.32092142e-01 4.51140583e-01 2.46804401e-01 5.89946091e-01
1.63399547e-01 -4.09660995e-01 8.18039358e-01 -5.15309393e-01
-7.72648871e-01 1.18550986e-01 -9.25579309e-01 2.84448236e-01
1.05314660e+00 -2.01865509e-01 -3.39497715e-01 -2.14285478e-01
-9.22308147e-01 5.25747299e-01 4.81319606e-01 -7.02794045e-02
5.63796401e-01 -1.08648384e+00 -2.01993927e-01 2.17769239e-02
-1.66590363e-01 -2.55165458e-01 1.75115228e-01 7.18032479e-01
-2.12825805e-01 4.35046107e-01 -5.24013281e-01 -1.54217795e-01
-9.97390747e-01 5.73828697e-01 4.01144475e-01 -5.11475980e-01
-5.19036472e-01 8.72605205e-01 1.54967383e-01 -2.51937211e-01
3.75427186e-01 -1.59576133e-01 -7.01281205e-02 7.92567879e-02
3.10140669e-01 3.80240113e-01 -2.62807161e-01 8.65143687e-02
-2.67579406e-01 3.40545684e-01 -4.72282648e-01 -6.79713309e-01
1.47355866e+00 -2.04599828e-01 8.64810422e-02 3.40866446e-01
6.27029538e-01 -1.81178778e-01 -1.75090039e+00 1.48726739e-02
1.93775237e-01 -2.53041893e-01 -1.61398143e-01 -7.44836688e-01
-7.86622584e-01 7.87340105e-01 3.25117767e-01 3.97302717e-01
1.02074754e+00 -1.51608422e-01 4.63923305e-01 1.01828468e+00
7.92471051e-01 -1.31701303e+00 4.09050345e-01 7.28153288e-01
5.10221601e-01 -9.14574325e-01 1.02872448e-02 1.36878133e-01
-5.25196016e-01 9.12140608e-01 9.30509627e-01 -3.15530717e-01
2.37601772e-01 1.59679085e-01 -4.68457669e-01 -4.21254896e-02
-1.04692495e+00 -4.22326267e-01 -6.62995100e-01 5.70733607e-01
1.30683318e-01 1.37602746e-01 -1.27255097e-01 1.98744267e-01
-3.07324499e-01 -9.55324620e-02 6.30358040e-01 1.45703661e+00
-1.02669239e+00 -1.56254268e+00 -1.21595412e-02 2.29404032e-01
-2.11285606e-01 -8.94425288e-02 -4.00051251e-02 8.49148989e-01
-9.32367966e-02 8.46742153e-01 -2.66422302e-01 -2.29419857e-01
2.37311482e-01 5.12068607e-02 9.27745044e-01 -5.63145161e-01
-7.04591393e-01 -4.18815911e-02 1.43577652e-02 -7.84933448e-01
-1.99721143e-01 -7.98991382e-01 -1.47418189e+00 -4.83529001e-01
-7.38065019e-02 4.22626674e-01 2.00886980e-01 9.58480954e-01
4.40525487e-02 6.03091300e-01 7.57037282e-01 -5.98478377e-01
-9.90201175e-01 -3.94022077e-01 -5.79400301e-01 2.29039919e-02
5.69385946e-01 -8.51645231e-01 -2.10178405e-01 -1.84555084e-01] | [4.012010097503662, 1.711187720298767] |
f6aff2d5-5952-4595-b9cc-fcbe48270aa0 | attention2angiogan-synthesizing-fluorescein | 2007.09191 | null | https://arxiv.org/abs/2007.09191v1 | https://arxiv.org/pdf/2007.09191v1.pdf | Attention2AngioGAN: Synthesizing Fluorescein Angiography from Retinal Fundus Images using Generative Adversarial Networks | Fluorescein Angiography (FA) is a technique that employs the designated camera for Fundus photography incorporating excitation and barrier filters. FA also requires fluorescein dye that is injected intravenously, which might cause adverse effects ranging from nausea, vomiting to even fatal anaphylaxis. Currently, no other fast and non-invasive technique exists that can generate FA without coupling with Fundus photography. To eradicate the need for an invasive FA extraction procedure, we introduce an Attention-based Generative network that can synthesize Fluorescein Angiography from Fundus images. The proposed gan incorporates multiple attention based skip connections in generators and comprises novel residual blocks for both generators and discriminators. It utilizes reconstruction, feature-matching, and perceptual loss along with adversarial training to produces realistic Angiograms that is hard for experts to distinguish from real ones. Our experiments confirm that the proposed architecture surpasses recent state-of-the-art generative networks for fundus-to-angio translation task. | ['Sharif Amit Kamran', 'Stewart Lee Zuckerbrod', 'Khondker Fariha Hossain', 'Alireza Tavakkoli'] | 2020-07-17 | null | null | null | null | ['fundus-to-angiography-generation'] | ['computer-vision'] | [ 1.63583577e-01 3.17068577e-01 2.75112420e-01 -1.40283167e-01
-7.57307947e-01 -8.56830537e-01 2.17460185e-01 -5.88521063e-01
-2.08122090e-01 1.07165813e+00 1.42268658e-01 -5.50825536e-01
4.32917207e-01 -7.10848927e-01 -6.66422009e-01 -7.08950222e-01
5.14730752e-01 -2.09011942e-01 5.39547242e-02 1.89703956e-01
-7.20415860e-02 7.33778894e-01 -1.21009243e+00 2.27873981e-01
1.19377327e+00 1.06391633e+00 -5.50273806e-02 8.79655957e-01
3.89878452e-02 1.12465441e+00 -7.89302349e-01 -7.43497550e-01
5.48989117e-01 -1.17650628e+00 -3.96572739e-01 6.04284694e-04
6.02802873e-01 -8.93878222e-01 -3.08201700e-01 9.65141535e-01
8.68597865e-01 -8.49193260e-02 6.52505696e-01 -1.02823663e+00
-1.29568529e+00 1.00597754e-01 -5.94359398e-01 3.87433827e-01
1.37434125e-01 7.97463179e-01 2.41193548e-01 -3.18969280e-01
4.63570982e-01 7.82472193e-01 4.90178347e-01 9.14779067e-01
-8.45836341e-01 -6.04804933e-01 -3.55736613e-01 -1.77335843e-01
-8.60494494e-01 -3.18487585e-01 6.96781695e-01 -5.77490211e-01
6.74787521e-01 1.88289031e-01 8.74977827e-01 1.22117996e+00
4.78611529e-01 4.28552032e-01 1.40772188e+00 -2.66016275e-01
-7.60614425e-02 3.13840300e-01 -6.59742773e-01 7.52685964e-01
2.62594938e-01 4.71275985e-01 2.30475143e-01 7.92241246e-02
1.41838193e+00 -5.47387712e-02 -5.56305349e-01 -4.89013419e-02
-7.66156435e-01 7.99932659e-01 6.04843795e-01 -7.72525519e-02
-5.98784268e-01 3.07955801e-01 1.28665581e-01 1.47829309e-01
-1.77094966e-01 7.59072721e-01 1.63258374e-01 -7.66268373e-02
-6.56669736e-01 -4.07090455e-01 3.11154604e-01 9.22225058e-01
2.34606132e-01 5.02848327e-01 -3.48301500e-01 3.81964952e-01
-8.68589431e-02 4.48338240e-01 7.98289955e-01 -1.01376283e+00
1.76530614e-01 5.08587241e-01 3.43695432e-01 -5.40820241e-01
-2.39294067e-01 -6.68332100e-01 -8.71805608e-01 5.64456463e-01
2.87182093e-01 -5.42831361e-01 -1.18777633e+00 1.59953845e+00
1.70947552e-01 4.55912977e-01 -3.84098254e-02 1.17024708e+00
1.21249485e+00 1.25796393e-01 3.23671579e-01 -7.15229511e-02
1.23503888e+00 -1.16841483e+00 -5.52215457e-01 9.22821537e-02
1.41266033e-01 -9.46029007e-01 1.31485260e+00 2.36582384e-01
-1.43584347e+00 -7.69390464e-01 -1.05039048e+00 -2.64519721e-01
-4.82317507e-02 4.51280326e-01 7.83884168e-01 1.06043279e+00
-1.06924212e+00 2.60798514e-01 -6.41000211e-01 -6.88971952e-02
6.42569482e-01 1.75564468e-01 -3.66141528e-01 2.69416183e-01
-1.01294994e+00 9.23594475e-01 -1.20733924e-01 1.85298309e-01
-8.85572195e-01 -8.75453174e-01 -7.40696251e-01 2.67932005e-02
-2.63362136e-02 -1.54363525e+00 1.12344325e+00 -1.21101105e+00
-2.21089935e+00 8.11333776e-01 1.03057452e-01 -5.18507957e-01
8.52999806e-01 -1.81303129e-01 -5.31930327e-01 5.47093272e-01
-2.20741659e-01 8.76266837e-01 1.11431789e+00 -6.65256560e-01
-3.50800246e-01 1.43406779e-01 3.42896163e-01 -3.46606001e-02
2.69271046e-01 -6.37449622e-02 1.81865796e-01 -6.70735836e-01
-5.67966580e-01 -8.12916517e-01 -1.73085526e-01 2.85573334e-01
-4.06630814e-01 2.95476913e-01 2.67509162e-01 -5.62689781e-01
8.49720478e-01 -1.94955838e+00 -4.73907590e-01 -1.17612630e-01
3.30399603e-01 7.20348418e-01 -3.61165941e-01 1.58301458e-01
-2.11318210e-01 2.83086360e-01 -3.01931780e-02 1.66322142e-01
-4.14701730e-01 -3.15477073e-01 -5.04151285e-01 5.16607165e-01
5.56241632e-01 1.19388807e+00 -9.57059562e-01 -2.84428537e-01
5.30140936e-01 7.07360685e-01 -7.31388927e-01 5.53822279e-01
1.40750393e-01 9.84015107e-01 -3.30637932e-01 6.79991663e-01
7.32082188e-01 -4.46652353e-01 -2.33790860e-01 -3.31197798e-01
5.41398972e-02 4.24892306e-02 -5.40440202e-01 1.63823080e+00
-6.64701462e-01 5.03562689e-01 -4.22815293e-01 -4.87844735e-01
7.08140969e-01 4.90051925e-01 -3.88278104e-02 -5.75125158e-01
6.36416674e-01 2.97423422e-01 2.28643209e-01 -7.47014344e-01
-6.22062162e-02 -3.42233092e-01 3.17047089e-01 2.25794345e-01
8.43144879e-02 -9.03148875e-02 -1.05993487e-02 -2.25459635e-01
1.00558579e+00 3.94469798e-01 3.32945049e-01 2.58751452e-01
7.23661959e-01 -2.31903851e-01 3.66417915e-01 6.76523328e-01
-1.90465048e-01 9.90203500e-01 3.78237456e-01 -4.19967473e-01
-9.30247784e-01 -1.20059013e+00 3.78750525e-02 1.59111857e-01
1.91243321e-01 2.79570311e-01 -7.37980008e-01 -8.72218966e-01
-1.60010636e-01 5.12709439e-01 -7.08593071e-01 -2.19782218e-01
-3.03437620e-01 -4.30199653e-01 7.62209356e-01 7.33437240e-01
7.51817942e-01 -9.84515429e-01 -8.55373681e-01 3.43027890e-01
-1.39986336e-01 -8.72425675e-01 -6.22639537e-01 -5.60985029e-01
-7.24480927e-01 -1.28468311e+00 -1.13444591e+00 -5.95223784e-01
8.52265716e-01 9.72640440e-02 9.54309762e-01 -1.97338149e-01
-7.36136019e-01 2.06836864e-01 -1.81549221e-01 -5.22692382e-01
-6.34873509e-01 -2.61063516e-01 -3.91437531e-01 1.99085400e-01
3.27741861e-01 -8.25583100e-01 -1.45139039e+00 1.20133810e-01
-7.59681463e-01 -3.76351736e-02 1.09262753e+00 8.76412988e-01
6.96317911e-01 -5.38925529e-01 8.46239150e-01 -1.05546665e+00
7.07046986e-01 -3.30975801e-01 -8.12309861e-01 1.90651059e-01
-2.20599189e-01 -1.94281116e-01 1.17154276e+00 -5.33106565e-01
-9.68250573e-01 2.60026474e-02 -3.14036049e-02 -6.87202275e-01
-3.65939140e-01 -7.74825066e-02 1.89772949e-01 -4.25039232e-01
9.30810392e-01 2.21727878e-01 -3.06927413e-02 -3.02288204e-01
6.20021641e-01 7.33107865e-01 8.99314404e-01 1.14410192e-01
5.89270294e-01 5.93761921e-01 1.75531000e-01 -3.83427024e-01
-6.12943113e-01 7.39982575e-02 -2.37720475e-01 -4.17077579e-02
9.17326212e-01 -1.03515470e+00 -9.55267072e-01 4.29159105e-01
-1.13499594e+00 5.94891943e-02 -4.75225836e-01 7.09105074e-01
-6.62077665e-01 2.68721670e-01 -7.04821110e-01 -5.29779017e-01
-5.42023838e-01 -1.23342288e+00 6.87547982e-01 6.78815961e-01
-9.76014510e-02 -7.71086335e-01 -6.51707053e-02 3.29417497e-01
8.05651784e-01 7.94936121e-01 7.95100152e-01 9.56543759e-02
-9.59802091e-01 -2.10529819e-01 -3.28955948e-01 7.47126400e-01
4.72730279e-01 3.19512114e-02 -1.13886690e+00 -2.35443816e-01
1.78810544e-02 -3.35075140e-01 6.67839825e-01 6.28408730e-01
1.14441049e+00 -3.56378198e-01 7.01613054e-02 1.19479775e+00
1.43469894e+00 6.51031554e-01 1.10332859e+00 -1.02320332e-02
5.29540002e-01 1.05057970e-01 2.79059215e-03 7.75833577e-02
1.34784225e-02 7.29215145e-02 3.12743157e-01 -7.32408762e-01
-7.73536444e-01 -4.41713065e-01 6.32839352e-02 2.82895148e-01
-3.43702495e-01 -4.15872514e-01 -1.74287081e-01 4.89233404e-01
-1.08329749e+00 -9.47911918e-01 1.09190047e-01 1.96432161e+00
7.87236869e-01 -2.67226011e-01 4.02979217e-02 -4.73640442e-01
6.92852795e-01 -2.72445321e-01 -8.36292982e-01 -6.09574437e-01
2.54968386e-02 8.02755952e-01 5.36536217e-01 2.25258812e-01
-1.07415783e+00 6.80169165e-01 6.61095476e+00 2.40874961e-01
-1.73621905e+00 1.79273874e-01 6.06549323e-01 -1.35887131e-01
-3.67421269e-01 -1.94409579e-01 -2.57393897e-01 7.37980425e-01
7.84148037e-01 -1.00036375e-01 1.62609622e-01 8.18674862e-01
2.07470030e-01 -4.12653275e-02 -9.69777405e-01 1.09747815e+00
-8.88252817e-03 -1.47993910e+00 3.27568024e-01 -2.89001882e-01
7.90427268e-01 -3.67518604e-01 4.51707542e-01 -1.25754699e-01
3.37737709e-01 -1.31973791e+00 2.54294664e-01 6.95623398e-01
1.44506800e+00 -5.92537344e-01 9.25978899e-01 -1.82622522e-01
-6.28519535e-01 3.70405763e-02 -3.74398440e-01 2.69996077e-01
3.23953599e-01 1.58180296e-01 -9.37139213e-01 4.21130568e-01
1.16998918e-01 4.40956622e-01 -5.08126915e-01 1.50367785e+00
-6.77913547e-01 4.49126363e-01 3.60404737e-02 2.49611020e-01
3.92425984e-01 -3.22659224e-01 5.30448854e-01 7.33203053e-01
7.75044441e-01 8.30382295e-03 -4.44880873e-01 1.28278470e+00
-1.73403174e-01 7.48183057e-02 -6.93462253e-01 6.64302856e-02
3.31731826e-01 1.13491654e+00 -3.62696409e-01 -2.06863448e-01
-5.23540497e-01 1.17511368e+00 -1.40513614e-01 5.21067560e-01
-1.09076405e+00 -7.13497519e-01 3.82212967e-01 3.69797975e-01
2.83650398e-01 4.11711752e-01 7.23234052e-03 -1.09438407e+00
6.54431572e-03 -6.09448850e-01 2.15612262e-01 -1.25324869e+00
-1.32224083e+00 1.15808868e+00 -5.13539732e-01 -1.78498435e+00
-4.02169675e-01 -3.08198035e-01 -8.40041459e-01 1.35287857e+00
-2.02107286e+00 -1.35369170e+00 -7.05407560e-01 6.88334584e-01
3.07792537e-02 -2.48886228e-01 8.65479350e-01 3.12824667e-01
-2.97495246e-01 9.55226421e-01 -3.24396163e-01 8.46769735e-02
9.27026212e-01 -1.07126808e+00 2.48047203e-01 1.06865358e+00
-3.74185711e-01 6.91412508e-01 2.48436555e-01 -3.77597839e-01
-8.01009119e-01 -1.28902912e+00 4.02301133e-01 -2.73429841e-01
2.91866601e-01 2.54645735e-01 -4.12947416e-01 6.46744072e-01
6.22516096e-01 3.24289769e-01 7.58175910e-01 -1.01969004e+00
-2.08367214e-01 -2.74958551e-01 -1.45906579e+00 5.95966220e-01
7.51049399e-01 -4.34660643e-01 -4.20639783e-01 2.91684091e-01
4.78552043e-01 -7.31525362e-01 -6.48921192e-01 1.40302613e-01
5.65394938e-01 -1.10360813e+00 7.27724969e-01 -8.63218248e-01
7.78911233e-01 -3.04962516e-01 5.66637814e-01 -1.34254634e+00
-3.88268501e-01 -1.38636506e+00 3.20919871e-01 9.00595307e-01
3.88629675e-01 -1.10252452e+00 5.91457009e-01 5.66704631e-01
-2.60010540e-01 -4.91270363e-01 -5.65644324e-01 -8.05007041e-01
6.89032376e-02 3.90954792e-01 7.57906616e-01 6.89097166e-01
-5.53803146e-01 4.00510281e-02 -6.06770575e-01 2.01913074e-01
3.90999258e-01 3.02033454e-01 6.82255924e-01 -7.82551110e-01
-4.27177936e-01 -3.06966305e-01 -6.18754804e-01 -9.08828497e-01
-1.73314124e-01 -8.53831649e-01 -5.02468467e-01 -1.37989187e+00
-2.42313609e-01 -2.30351582e-01 -2.84159422e-01 3.57800692e-01
-7.10234046e-02 6.14945531e-01 6.57344908e-02 -8.90761018e-02
1.18115395e-01 3.29217672e-01 1.91399133e+00 4.55636121e-02
-2.80592412e-01 -3.55923176e-02 -1.15252090e+00 6.03014112e-01
7.92279005e-01 -3.13682854e-01 -8.40541899e-01 -3.58012527e-01
-4.14829105e-02 3.11319858e-01 5.15453815e-01 -1.02437627e+00
5.66547364e-02 2.47053951e-02 5.16938150e-01 -5.44087924e-02
5.72927110e-02 -6.84310853e-01 5.88902354e-01 3.63347858e-01
-1.71019435e-01 9.19233412e-02 1.42411143e-01 4.63799864e-01
-4.33190703e-01 -5.02233915e-02 1.21365607e+00 -4.63222861e-01
-3.19455177e-01 3.26714516e-01 -4.01629537e-01 2.26812959e-01
1.18036950e+00 -3.29289168e-01 -7.00215399e-01 -5.83992243e-01
-7.52457082e-01 -2.07709938e-01 4.91873384e-01 1.54939324e-01
7.62072682e-01 -1.15944695e+00 -7.99686372e-01 5.81325948e-01
-1.23468861e-01 -2.06919327e-01 7.91256726e-01 1.03088617e+00
-1.09717572e+00 3.92786831e-01 -6.85069442e-01 -1.49854288e-01
-8.13218474e-01 6.19560599e-01 7.38134563e-01 9.51739252e-02
-7.74539232e-01 9.77667391e-01 4.18626368e-01 2.60865986e-01
-1.01494402e-01 -4.47665066e-01 -3.48150730e-02 -4.55490470e-01
6.35849893e-01 1.46058947e-01 2.26862356e-02 -1.20672315e-01
-5.54939173e-02 6.97517335e-01 -1.13032327e-03 3.39456201e-01
9.55197334e-01 -1.10888839e-01 2.20905513e-01 -2.31549352e-01
9.49607909e-01 9.43724290e-02 -1.33818734e+00 2.25862935e-01
-1.07900906e+00 -8.37176561e-01 -1.46827137e-03 -1.33044744e+00
-1.45586061e+00 1.11223531e+00 8.22117805e-01 -6.81086555e-02
1.43558335e+00 -4.91531879e-01 1.15842259e+00 -2.92245388e-01
2.27015093e-01 -4.63143140e-01 -1.10556103e-01 -3.46323639e-01
1.09244788e+00 -1.03737938e+00 -3.46096814e-01 -5.10878146e-01
-7.70172834e-01 1.04106569e+00 6.65842950e-01 -4.51471806e-01
4.24164236e-01 1.90798193e-01 5.50950110e-01 1.82404220e-01
-4.05087084e-01 -8.02932307e-02 3.48273903e-01 8.68926644e-01
2.58597136e-01 -3.65238152e-02 -3.92963290e-01 5.80621719e-01
-2.73266584e-01 4.95242059e-01 9.30696368e-01 4.88633931e-01
2.49152541e-01 -1.03399181e+00 1.92611381e-01 4.37540680e-01
-8.44916046e-01 -2.85350502e-01 -1.46476820e-01 9.43768084e-01
2.72434920e-01 8.21368814e-01 1.71071221e-03 7.29228184e-02
2.40453869e-01 -1.67281762e-01 7.12702096e-01 -3.12203020e-01
-6.93040907e-01 1.53314248e-01 -9.77893099e-02 -7.06018090e-01
-5.43522596e-01 -8.26874375e-02 -6.91302776e-01 -9.69254598e-02
-1.67219996e-01 -2.50145495e-01 1.67514056e-01 2.77769476e-01
7.07437158e-01 6.95768476e-01 6.63239539e-01 -1.89682975e-01
-4.68629092e-01 -8.70705903e-01 -6.81541502e-01 3.85986358e-01
6.63557231e-01 -5.39309204e-01 -5.10851264e-01 3.28798443e-01] | [15.595406532287598, -3.772972583770752] |
5aa6675c-892c-4437-8071-70d6c69d193d | machine-translation-with-cross-lingual-word | 1912.10167 | null | https://arxiv.org/abs/1912.10167v2 | https://arxiv.org/pdf/1912.10167v2.pdf | Machine Translation with Cross-lingual Word Embeddings | Learning word embeddings using distributional information is a task that has been studied by many researchers, and a lot of studies are reported in the literature. On the contrary, less studies were done for the case of multiple languages. The idea is to focus on a single representation for a pair of languages such that semantically similar words are closer to one another in the induced representation irrespective of the language. In this way, when data are missing for a particular language, classifiers from another language can be used. | ['Evan Kaplan', 'Marco Berlot'] | 2019-12-10 | null | null | null | null | ['learning-word-embeddings'] | ['methodology'] | [-1.53401494e-01 -1.87703878e-01 -3.92003059e-01 -4.29580539e-01
-3.76094244e-02 -3.15111339e-01 7.95447171e-01 8.56054008e-01
-7.82038510e-01 5.16533077e-01 3.56966704e-01 -2.62364950e-02
1.47612885e-01 -8.93606961e-01 -2.01780066e-01 -7.05230713e-01
2.55048513e-01 2.27743387e-01 3.62392157e-01 -3.86041731e-01
1.96239665e-01 2.62817621e-01 -1.27621186e+00 -8.97337683e-04
4.56886590e-01 2.47465417e-01 2.43486121e-01 1.83145446e-03
-6.32068753e-01 2.14587286e-01 -5.74853003e-01 -4.37709779e-01
1.72714159e-01 -5.27315617e-01 -3.21811974e-01 -8.76464471e-02
3.76052931e-02 3.24613154e-01 -1.64678514e-01 1.23312128e+00
4.16934490e-01 1.26907572e-01 7.29486704e-01 -8.95568490e-01
-9.58548486e-01 5.37973285e-01 -5.35200059e-01 6.03082366e-02
2.72674382e-01 -3.99551481e-01 1.10614836e+00 -7.48071253e-01
3.44717503e-01 1.41968334e+00 2.99822867e-01 3.76626104e-01
-1.20454192e+00 -5.76105356e-01 3.56759340e-01 1.00320622e-01
-1.48982227e+00 1.49831429e-01 8.55976939e-01 -4.85473245e-01
4.20887560e-01 -2.67156959e-01 5.87372839e-01 1.00776184e+00
2.14022547e-01 4.14914906e-01 1.12772441e+00 -6.70252144e-01
1.54311642e-01 7.12349653e-01 2.61121839e-01 2.01569453e-01
6.69339240e-01 -2.28818491e-01 -2.68133134e-01 7.19583705e-02
1.51063710e-01 1.33576438e-01 -2.43975893e-01 -7.85415053e-01
-9.19527352e-01 1.12520313e+00 3.70068640e-01 1.17215025e+00
-1.87895074e-01 -2.90439039e-01 6.58324242e-01 2.45761827e-01
6.39845967e-01 5.48316956e-01 -1.71555206e-01 9.24497191e-03
-7.09476292e-01 2.81940728e-01 7.58732140e-01 4.03785616e-01
7.93944299e-01 -1.12889782e-01 2.49045298e-01 6.85779631e-01
4.88829404e-01 2.58456677e-01 8.36984038e-01 9.92357656e-02
2.79044420e-01 7.33567238e-01 -2.23975435e-01 -1.36033463e+00
-2.30921164e-01 -2.78196156e-01 -6.30785942e-01 -5.10251056e-03
1.70793325e-01 -8.90429597e-04 -4.42831516e-01 1.86077821e+00
2.02951893e-01 2.54374057e-01 4.24048305e-01 8.59441102e-01
6.82580411e-01 6.05115950e-01 6.82664663e-02 -4.69673127e-02
1.25393665e+00 -7.36729801e-01 -7.79045939e-01 -4.57039028e-01
8.84942949e-01 -8.13511491e-01 9.79530513e-01 2.55768239e-01
-2.86001295e-01 -6.00960493e-01 -1.10747242e+00 -1.53805450e-01
-9.73765075e-01 -3.77126664e-01 3.94974798e-01 7.70118475e-01
-7.67503381e-01 2.04514191e-01 -3.49983186e-01 -8.58218968e-01
-3.55779007e-02 -9.62068588e-02 -4.72563475e-01 -3.52408975e-01
-1.41022933e+00 1.32448113e+00 6.49318993e-01 -2.58600444e-01
-3.42218071e-01 -3.87338817e-01 -9.11003888e-01 -2.65947163e-01
2.64680795e-02 -3.56859803e-01 6.04559898e-01 -1.14575124e+00
-9.61060166e-01 9.36895967e-01 -1.28872186e-01 -4.61578846e-01
2.01912388e-01 -2.85348594e-01 -7.65776038e-01 -3.64062309e-01
1.61370113e-01 1.48561880e-01 7.79475987e-01 -1.21878493e+00
-3.60434383e-01 -7.16602266e-01 2.34519035e-01 2.03144595e-01
-8.35735142e-01 1.89451694e-01 -2.34064177e-01 -7.86296487e-01
-1.15704007e-01 -6.86218262e-01 -1.37826011e-01 -1.81414522e-02
-5.88272251e-02 -5.93379676e-01 8.15825462e-01 -5.40118515e-01
1.34072173e+00 -2.45884418e+00 3.63395602e-01 -2.01978764e-04
-1.41665682e-01 5.03780067e-01 -2.59767503e-01 5.85353255e-01
-2.42044538e-01 2.84831911e-01 -3.27101052e-01 -2.48072281e-01
-5.58785200e-02 3.54061484e-01 -3.69367748e-01 6.01660430e-01
1.18207470e-01 2.67102987e-01 -1.01537466e+00 -3.16912979e-01
2.34865785e-01 5.26214242e-01 -4.27865028e-01 2.57908612e-01
1.41562715e-01 1.20930776e-01 -3.79557401e-01 8.97062495e-02
6.12474203e-01 4.20826852e-01 4.48327512e-01 -1.62460461e-01
-2.96906620e-01 2.69221574e-01 -1.29795051e+00 1.49532306e+00
-8.14740658e-01 4.95192379e-01 -1.55662149e-01 -1.29439580e+00
1.16430938e+00 4.14865583e-01 1.18134402e-01 -5.05387723e-01
2.45756894e-01 3.50727260e-01 3.45853001e-01 -5.63417852e-01
3.35958451e-01 -6.00868702e-01 -2.59068698e-01 4.72367913e-01
-7.15715736e-02 -1.99792683e-01 3.47498089e-01 -2.14726657e-01
6.53370619e-01 -1.02263637e-01 6.63978934e-01 -1.17153801e-01
5.52984476e-01 -3.13144416e-01 4.00281847e-01 3.18001598e-01
-1.46954700e-01 3.37474972e-01 6.06606528e-02 -4.06805277e-01
-9.95812774e-01 -7.59354413e-01 -4.38788682e-01 9.78117824e-01
2.28239566e-01 -6.07133508e-01 -3.59425277e-01 -5.11364043e-01
1.80527404e-01 7.37514615e-01 -7.39432216e-01 -3.34468096e-01
-4.16479081e-01 -3.98342878e-01 4.41379935e-01 3.58023137e-01
2.64210135e-01 -8.63731027e-01 -4.36554521e-01 1.00825012e-01
2.41688550e-01 -8.68569791e-01 -3.95601124e-01 9.91793275e-02
-8.58519435e-01 -9.12609458e-01 -9.28057969e-01 -9.71896648e-01
5.59369504e-01 3.58979970e-01 1.01107991e+00 -5.77166351e-03
6.12929948e-02 6.13134354e-02 -5.70154130e-01 -5.96666873e-01
-4.19681311e-01 2.02548131e-01 4.11934733e-01 2.58631140e-01
8.76190901e-01 -3.90310436e-01 4.09506785e-04 1.40609015e-02
-1.10919249e+00 -3.57964456e-01 5.73958993e-01 5.38604259e-01
2.73150384e-01 -7.60094523e-02 6.32241249e-01 -8.23269725e-01
8.72837007e-01 -7.84773946e-01 -2.47541949e-01 2.91949570e-01
-4.46551472e-01 3.67451668e-01 7.43604004e-01 -4.09825534e-01
-7.04417527e-01 -2.04979420e-01 3.56672518e-02 -1.63962871e-01
-3.96350175e-01 7.27420747e-01 -2.87087142e-01 2.26843864e-01
3.22547495e-01 1.20368376e-01 -1.13090873e-01 -5.67319155e-01
4.12473232e-01 9.41718519e-01 -1.67330235e-01 -3.66835326e-01
6.33378744e-01 1.69856295e-01 -3.03551286e-01 -8.84196877e-01
-7.16527164e-01 -5.35917521e-01 -7.19239533e-01 1.70114309e-01
9.93110955e-01 -6.60805702e-01 1.40149638e-01 7.45721236e-02
-1.22677290e+00 4.91749316e-01 -1.27841830e-01 9.00948346e-01
8.46166909e-02 4.27998424e-01 3.07205886e-01 -7.62390912e-01
5.82959317e-03 -1.01906049e+00 5.92521608e-01 3.16013962e-01
-4.80420440e-01 -1.54886949e+00 4.65623945e-01 -2.27254033e-01
3.61296833e-01 -2.25825146e-01 1.18584359e+00 -1.23519337e+00
1.96365207e-01 -4.16392863e-01 6.45781984e-04 6.38615906e-01
5.58285475e-01 -2.03678071e-01 -7.28684783e-01 -3.88049573e-01
2.97582239e-01 -7.80337676e-02 8.18742931e-01 -3.65380645e-02
7.12302804e-01 -7.29575530e-02 -4.36089039e-01 1.04280550e-03
1.46285689e+00 2.00977653e-01 2.97508866e-01 3.36364657e-01
4.65249956e-01 6.85837388e-01 3.98472190e-01 1.05280548e-01
2.28033081e-01 6.48231566e-01 2.51997501e-01 6.48100004e-02
9.26551074e-02 -2.68748105e-01 4.15653497e-01 1.02321184e+00
4.12009954e-01 -3.43187034e-01 -8.95431161e-01 1.08050418e+00
-1.46480894e+00 -5.89185536e-01 1.21433124e-01 2.35575700e+00
6.75337911e-01 7.12772980e-02 1.66428443e-02 7.35358745e-02
9.49133456e-01 6.43122792e-01 -1.17511712e-01 -7.31325686e-01
-1.90617353e-01 2.50538886e-01 1.52058780e-01 2.78829426e-01
-7.87945390e-01 8.58790934e-01 5.87162256e+00 5.89877188e-01
-1.30316842e+00 1.10419663e-02 2.11273313e-01 3.31819765e-02
-5.95767021e-01 2.33322740e-01 -5.71699262e-01 4.67524588e-01
6.38729036e-01 -4.69403416e-01 8.28113686e-03 6.56309962e-01
1.84395894e-01 -2.01164231e-01 -9.81125593e-01 8.91865611e-01
4.59084481e-01 -5.85040867e-01 3.97878677e-01 -1.05884925e-01
4.57483858e-01 -2.07698628e-01 -8.48774984e-02 3.85296881e-01
-1.71722863e-02 -1.12184393e+00 2.44847476e-01 4.06759560e-01
1.84810594e-01 -9.13732231e-01 1.11406946e+00 5.13753235e-01
-1.15623724e+00 1.91666305e-01 -5.99323094e-01 -8.68667439e-02
2.95571119e-01 7.90083766e-01 -5.24683952e-01 8.08758199e-01
4.78406042e-01 9.07609284e-01 -6.27159238e-01 9.72119510e-01
-4.99804437e-01 5.64314663e-01 -1.17354728e-01 -3.46241951e-01
2.42443040e-01 -4.89572465e-01 4.60446715e-01 1.11870086e+00
4.16047543e-01 -3.78087908e-01 4.37590778e-01 6.50406837e-01
-1.20318010e-01 7.75269866e-01 -1.22530210e+00 -2.95473099e-01
2.91167945e-01 1.07222450e+00 -3.05057764e-01 -2.34774575e-01
-7.33111680e-01 7.97773898e-01 4.09758508e-01 -9.99521986e-02
-5.22641361e-01 -4.79930758e-01 8.64851654e-01 1.95115656e-01
1.50422454e-01 -4.02669609e-01 -1.57489285e-01 -8.78561616e-01
-2.95826867e-02 -4.70355332e-01 3.51067662e-01 -5.32887518e-01
-1.58634317e+00 6.31364942e-01 3.77262533e-02 -1.08327413e+00
6.10163249e-02 -6.59382224e-01 -5.63072920e-01 1.06554592e+00
-1.51501369e+00 -6.00539505e-01 1.28720701e-01 3.45763862e-01
5.19536018e-01 -3.42512429e-01 1.12649035e+00 2.95862556e-01
-4.46970582e-01 3.03261489e-01 2.39702433e-01 2.59856015e-01
9.53479946e-01 -9.23725963e-01 -7.64795765e-03 6.96084559e-01
6.04361713e-01 9.46642160e-01 6.89956427e-01 -6.31341517e-01
-9.90211487e-01 -8.97257090e-01 1.51586354e+00 -3.48219275e-02
8.35468113e-01 -1.24900989e-01 -1.27002394e+00 3.25166643e-01
8.31582189e-01 1.25376642e-01 9.56464887e-01 4.22276288e-01
-6.26879275e-01 -1.41623244e-01 -9.84675705e-01 7.40600049e-01
4.80364561e-01 -7.20864892e-01 -9.45493281e-01 1.06097080e-01
4.02929753e-01 1.80190161e-01 -6.23434246e-01 3.31390277e-03
2.84755439e-01 -7.11040616e-01 4.74729031e-01 -6.30701303e-01
6.70721903e-02 -4.61022913e-01 -3.18791598e-01 -1.80442476e+00
2.01265961e-02 1.87461153e-01 3.59062016e-01 1.44125986e+00
1.41722053e-01 -7.34911144e-01 -9.64707043e-03 2.10575983e-01
1.68407366e-01 -6.06475055e-01 -9.83385980e-01 -9.37411010e-01
4.54586148e-01 -3.85680228e-01 4.95673448e-01 1.03198147e+00
6.31277859e-02 5.95581412e-01 -1.51930034e-01 8.64954200e-03
9.59226862e-03 1.91829294e-01 4.16909575e-01 -1.40059435e+00
3.11038494e-01 -3.62252593e-01 -8.01563144e-01 -6.53874278e-01
4.64901537e-01 -1.18785655e+00 2.11012477e-04 -1.53836119e+00
1.92292303e-01 -2.76341259e-01 -4.68793958e-01 3.23386222e-01
-1.87639371e-01 1.51300743e-01 2.81126142e-01 -9.95691493e-02
-2.62032539e-01 6.76801205e-01 7.55006671e-01 -8.66152644e-02
1.16507955e-01 -3.24006289e-01 -7.79300213e-01 8.68854105e-01
9.74807143e-01 -7.25575447e-01 -3.41818303e-01 -5.22706926e-01
1.46661818e-01 -6.85027361e-01 2.31535330e-01 -9.83406067e-01
1.77314132e-02 -1.08448356e-01 -1.08866863e-01 -1.67595416e-01
1.36409417e-01 -1.02422392e+00 -1.73469812e-01 1.67386681e-01
-4.30536747e-01 4.80237752e-01 1.89630955e-01 7.11343944e-01
-6.02651477e-01 -6.51954114e-01 8.25019479e-01 -1.23603597e-01
-8.23916674e-01 6.13221489e-02 -4.42415595e-01 3.67456943e-01
1.23242831e+00 -1.80927143e-01 2.34991208e-01 -1.43995553e-01
-2.08162218e-01 -1.13021150e-01 5.27426839e-01 1.01591933e+00
4.87582088e-01 -1.46292329e+00 -8.13356400e-01 1.04878418e-01
6.51533186e-01 -3.41764897e-01 -2.29450375e-01 4.36078936e-01
-1.75893843e-01 3.91415536e-01 -8.84522200e-02 -3.03129762e-01
-1.15141869e+00 6.76197231e-01 1.94813102e-01 -1.11176431e-01
-4.38133776e-01 3.80177319e-01 2.46497989e-01 -5.56220293e-01
-1.00325033e-01 -1.89352021e-01 -5.92662871e-01 5.68492055e-01
5.69140851e-01 -1.67085007e-01 5.87656125e-02 -1.01487207e+00
-5.84209323e-01 7.54659951e-01 -2.00640887e-01 -1.76134065e-01
1.34698021e+00 5.05023338e-02 -2.70635486e-01 1.14481127e+00
1.41167808e+00 2.54084378e-01 -3.22636306e-01 -4.53380078e-01
3.26308966e-01 -5.87223351e-01 4.66019195e-03 -2.28466272e-01
-8.61312866e-01 1.14843094e+00 7.74512529e-01 2.09774420e-01
6.94020271e-01 1.15997434e-01 3.47120732e-01 3.98350768e-02
3.11807305e-01 -1.11497200e+00 -1.34914264e-01 8.18455577e-01
8.34067285e-01 -1.06954241e+00 -4.26370166e-02 -1.15600415e-02
-6.33713722e-01 1.20338416e+00 4.56400424e-01 -3.51560950e-01
8.69673908e-01 -3.42740566e-02 5.68645932e-02 8.91417712e-02
-2.06799552e-01 -2.87150323e-01 1.68675274e-01 7.07964599e-01
1.01306725e+00 -3.00561916e-02 -7.27073967e-01 6.47945225e-01
5.37483655e-02 -9.94438529e-02 3.32033485e-01 8.41414332e-01
-4.34686422e-01 -1.53390563e+00 -1.10368602e-01 1.62247077e-01
-1.41382620e-01 -2.20338121e-01 -4.89836812e-01 8.44415903e-01
3.99941564e-01 8.06476116e-01 1.35078177e-01 -3.50561514e-02
3.97493392e-01 4.20425951e-01 2.39408568e-01 -1.14870250e+00
-6.06559813e-01 -4.95171815e-01 -2.71484315e-01 5.36240563e-02
-7.38590419e-01 -5.52832484e-01 -1.15316474e+00 -1.24913491e-02
-2.86082596e-01 4.08220708e-01 6.49554908e-01 1.14569116e+00
1.72038376e-01 4.61556762e-01 4.71210122e-01 -2.87262917e-01
-2.10244924e-01 -9.64596152e-01 -8.69371951e-01 6.72028363e-01
1.36788294e-01 -8.24781775e-01 -3.68873596e-01 -4.34286416e-01] | [10.485784530639648, 8.7276029586792] |
6d78c389-ce4b-44fb-a9f7-9de3345fbe17 | the-university-of-arizona-at-semeval-2021 | null | null | https://aclanthology.org/2021.semeval-1.56 | https://aclanthology.org/2021.semeval-1.56.pdf | The University of Arizona at SemEval-2021 Task 10: Applying Self-training, Active Learning and Data Augmentation to Source-free Domain Adaptation | This paper describes our systems for negation detection and time expression recognition in SemEval 2021 Task 10, Source-Free Domain Adaptation for Semantic Processing. We show that self-training, active learning and data augmentation techniques can improve the generalization ability of the model on the unlabeled target domain data without accessing source domain data. We also perform detailed ablation studies and error analyses for our time expression recognition systems to identify the source of the performance improvement and give constructive feedback on the temporal normalization annotation guidelines. | ['Steven Bethard', 'Yiyun Zhao', 'Xin Su'] | 2021-08-01 | null | null | null | semeval-2021 | ['source-free-domain-adaptation', 'negation-detection'] | ['computer-vision', 'natural-language-processing'] | [ 2.67260939e-01 4.15640354e-01 -6.78180575e-01 -1.20627391e+00
-7.64830410e-01 -7.40499735e-01 4.14915830e-01 2.79486209e-01
-9.87656474e-01 9.26230669e-01 -5.66557609e-03 -2.42926225e-01
8.48831385e-02 -3.54598641e-01 -7.03260228e-02 -2.63461292e-01
-3.75912994e-01 5.80777824e-01 1.61750555e-01 -5.17573416e-01
-9.82037783e-02 4.71599162e-01 -1.02407634e+00 4.22535211e-01
8.61101866e-01 1.05033720e+00 -6.16512418e-01 5.31750679e-01
-4.19893205e-01 1.12185192e+00 -1.01929021e+00 -3.55270743e-01
-1.30589828e-01 -1.02827594e-01 -1.02958143e+00 -4.21289206e-01
2.11693734e-01 -2.33184025e-01 -7.47762695e-02 7.62945294e-01
3.73374045e-01 5.21800399e-01 4.62390870e-01 -1.54857671e+00
-4.69769031e-01 5.76714575e-01 1.13394205e-02 7.41315305e-01
1.98408455e-01 5.77887520e-04 4.93370712e-01 -8.28292847e-01
8.62925351e-01 8.85734200e-01 6.84080660e-01 1.11894953e+00
-1.20185018e+00 -8.77827823e-01 1.93802118e-01 3.87008160e-01
-1.25796461e+00 -1.16084731e+00 5.05631447e-01 -4.43939865e-02
2.04889822e+00 9.79291555e-03 4.21468318e-01 1.61889124e+00
-2.33240843e-01 1.11188209e+00 6.61601126e-01 -6.98458493e-01
5.74790239e-01 1.35321870e-01 6.95736825e-01 6.47151291e-01
-2.82783359e-01 2.81739414e-01 -9.96339381e-01 -4.23851937e-01
3.31440747e-01 -5.73775768e-01 1.58874422e-01 -1.84023887e-01
-6.26786351e-01 5.38220465e-01 9.73304436e-02 5.27412891e-01
-1.16665371e-01 -3.30775380e-02 1.18390286e+00 7.00133681e-01
8.67223263e-01 5.27651429e-01 -1.21877360e+00 -5.82153976e-01
-7.43824959e-01 -1.09718688e-01 6.86982989e-01 1.07718778e+00
3.18835974e-01 4.44477320e-01 -2.57345110e-01 9.92786884e-01
1.47351832e-03 1.75096095e-01 7.03930497e-01 -8.70059371e-01
4.81204629e-01 8.61123085e-01 1.39134333e-01 2.23615672e-03
-5.28804719e-01 6.80881664e-02 -3.25685926e-02 1.88760281e-01
3.91645074e-01 -3.54424059e-01 -1.13253629e+00 2.01767659e+00
-3.83714169e-01 -1.72347158e-01 4.06914830e-01 3.02008659e-01
5.52325785e-01 2.20946461e-01 1.00920010e+00 -6.74440801e-01
6.81679785e-01 -7.43575096e-01 -1.35374749e+00 -5.67796409e-01
1.58982623e+00 -2.38609716e-01 9.30662513e-01 1.83768421e-01
-9.79113936e-01 -2.54083425e-01 -1.28480148e+00 -1.13315642e-01
-7.91009843e-01 -3.37603390e-02 1.28800952e+00 6.47210300e-01
-6.83904171e-01 5.55624962e-01 -1.34295011e+00 -6.31791234e-01
6.61151469e-01 7.03946948e-01 -5.06720304e-01 3.61762255e-01
-1.75646043e+00 1.29830468e+00 7.67244101e-01 -1.67556614e-01
-4.50159967e-01 -5.72646141e-01 -1.06850147e+00 -1.50495470e-01
1.69827998e-01 -9.94801372e-02 1.84012771e+00 -1.57795036e+00
-1.43255150e+00 1.35114753e+00 -2.69768178e-01 -8.85311842e-01
1.03383556e-01 -2.69242674e-01 -1.28112447e+00 4.13608253e-02
1.03681823e-02 7.02868044e-01 4.74486351e-01 5.73359989e-02
-2.61792660e-01 -4.38701242e-01 -2.57898271e-01 3.36651206e-01
-5.84938645e-01 4.73821223e-01 -1.80375442e-01 -4.81108397e-01
-1.64652586e-01 -5.97760618e-01 6.24372019e-03 -1.68031510e-02
3.03887397e-01 -5.51326811e-01 1.03905892e+00 -5.03129184e-01
1.04279041e+00 -2.71035671e+00 -3.45069796e-01 1.21230721e-01
-2.13018328e-01 3.08562309e-01 -1.04672082e-01 -1.70176789e-01
-8.35052490e-01 4.23323587e-02 -1.77904904e-01 -3.95980299e-01
-6.88356236e-02 4.45457846e-01 -5.18371046e-01 3.76937360e-01
6.24740660e-01 1.13586402e+00 -8.69610846e-01 -4.40045178e-01
-5.92430085e-02 -4.43665460e-02 -2.54654616e-01 1.29600018e-01
-3.65301639e-01 1.57963514e-01 -1.11444838e-01 8.30317259e-01
2.18985379e-01 2.81982958e-01 1.62641838e-01 -7.98312575e-02
1.38187781e-01 4.73572850e-01 -3.87986153e-01 2.21060514e+00
-2.56526560e-01 6.86192751e-01 -6.21874556e-02 -1.17061317e+00
8.40295672e-01 7.71667004e-01 4.89167362e-01 -1.41001320e+00
2.19426736e-01 4.20827627e-01 -9.99242067e-02 -3.63936692e-01
2.89577782e-01 -1.70678765e-01 -4.53938425e-01 3.12953331e-02
4.89474863e-01 2.77836114e-01 3.15426826e-01 1.73236921e-01
1.36696506e+00 3.76316547e-01 4.43841249e-01 -2.32583448e-01
2.21161529e-01 2.90255487e-01 8.20098937e-01 5.25230527e-01
-9.39599454e-01 -1.14728101e-01 4.07306373e-01 -1.63773239e-01
-8.06369066e-01 -1.11220145e+00 2.58404985e-02 1.56781685e+00
-3.90997916e-01 -6.42165899e-01 -3.37558001e-01 -1.22791910e+00
-3.25948387e-01 1.36225533e+00 -7.21561551e-01 -7.52525926e-01
-5.31334460e-01 -7.72228539e-01 1.14770520e+00 1.11987054e+00
4.09621030e-01 -1.07491589e+00 -4.14741129e-01 1.99881926e-01
-7.88977444e-02 -1.11011600e+00 -3.09060086e-02 1.17139566e+00
-1.21027207e+00 -8.93837035e-01 -1.05418071e-01 -7.87112772e-01
5.64485848e-01 -9.37976778e-01 1.00056279e+00 -3.19415122e-01
-3.08437973e-01 5.89988887e-01 -3.21811110e-01 -7.59036601e-01
-2.65492737e-01 5.09709641e-02 2.38172472e-01 -7.56438613e-01
1.12216961e+00 -3.54202539e-01 8.67686346e-02 1.86563253e-01
-5.35710096e-01 -4.41610485e-01 5.49654057e-03 7.92192638e-01
3.37093621e-01 -1.47660732e-01 6.29855931e-01 -1.04947591e+00
7.27579594e-01 -2.17443943e-01 -4.63614106e-01 3.04836303e-01
-7.41129398e-01 1.23991892e-01 1.74992278e-01 -6.11498177e-01
-1.41468859e+00 4.41633731e-01 -9.36332121e-02 -2.37936720e-01
-1.71424314e-01 3.02232593e-01 -2.30389848e-01 2.20433548e-01
1.41887999e+00 -1.65211529e-01 -1.12472117e-01 -1.28280893e-01
2.87196815e-01 4.87370819e-01 7.43349791e-01 -6.40271366e-01
2.82502323e-01 2.00567111e-01 -4.91492838e-01 -4.67118084e-01
-1.00609207e+00 -5.20553112e-01 -8.97212386e-01 2.17728078e-01
6.47091806e-01 -1.01337659e+00 -2.67850727e-01 4.48601872e-01
-1.01561832e+00 -7.22222090e-01 -8.31815302e-01 3.89930815e-01
-9.40737724e-01 -5.40180355e-02 -4.79632139e-01 -8.44604671e-01
-4.52817887e-01 -3.48008513e-01 8.32337022e-01 9.01916176e-02
-1.22735131e+00 -1.25009429e+00 1.43573061e-01 -2.08497606e-02
5.22144020e-01 -8.06980282e-02 1.04852486e+00 -1.40832138e+00
2.35965967e-01 -4.21769023e-01 3.23692083e-01 5.21582067e-01
-2.78973505e-02 -1.10193640e-01 -1.20552838e+00 -5.14838882e-02
1.82371408e-01 -7.24018276e-01 5.14560044e-01 -6.89553320e-02
1.01287091e+00 -9.51981638e-03 -3.63291800e-01 4.31877047e-01
1.01391363e+00 6.69832170e-01 6.52861476e-01 3.33675891e-01
-1.98137462e-02 3.58700842e-01 9.70931768e-01 1.62491314e-02
-4.02442366e-01 4.99047339e-01 -2.67026484e-01 -2.93542385e-01
1.26028508e-01 2.39042655e-01 5.95828891e-01 2.06935152e-01
2.98199087e-01 -2.26831913e-01 -1.37335241e+00 6.29937768e-01
-1.83161485e+00 -8.90973508e-01 2.79397935e-01 1.95085573e+00
1.05017722e+00 6.71979785e-01 -1.12543128e-01 3.78228724e-01
3.70116413e-01 -2.62427658e-01 -7.76238739e-01 -8.51653576e-01
-5.54080784e-01 7.07930267e-01 5.93524218e-01 2.79242307e-01
-1.16794944e+00 1.28765810e+00 7.92902422e+00 5.75464964e-01
-1.11247468e+00 3.92920822e-01 3.79247963e-01 -3.50492448e-01
4.38015938e-01 1.92287546e-02 -5.49036741e-01 -7.85409659e-02
1.58271587e+00 -5.34307718e-01 6.98850676e-02 9.92664397e-01
5.38995781e-04 -1.13659568e-01 -1.73191357e+00 1.05407882e+00
-6.94769844e-02 -9.36199009e-01 -4.92476404e-01 -5.27093410e-01
2.27726221e-01 2.45217547e-01 -2.48910859e-01 9.38377321e-01
1.43978447e-01 -9.41103756e-01 6.06313467e-01 4.37611669e-01
7.79002309e-01 -6.22719646e-01 5.74699640e-01 2.64785260e-01
-5.75157821e-01 -2.04246208e-01 2.48301879e-01 -2.34773844e-01
9.92775559e-02 -7.82068595e-02 -1.00109017e+00 1.54250056e-01
4.81101990e-01 7.40617096e-01 -8.08446229e-01 6.58362627e-01
-3.51845264e-01 9.05843258e-01 -5.06000102e-01 1.12304360e-01
-2.00463608e-01 5.12099743e-01 4.43002790e-01 1.25399125e+00
-2.86989301e-01 2.15709373e-01 -2.80924708e-01 5.83731055e-01
9.84557196e-02 3.19454610e-01 -5.25763750e-01 -8.84778738e-01
3.86227846e-01 7.20430911e-01 -7.10105836e-01 -6.19472861e-01
-3.20192516e-01 1.31518030e+00 4.91092503e-01 6.17983043e-01
-8.70001137e-01 -6.12727404e-01 5.26930809e-01 -1.51788160e-01
-6.77167550e-02 -2.97536820e-01 -4.35760796e-01 -1.24111736e+00
-1.37070537e-01 -5.24585664e-01 9.67591286e-01 -1.10592949e+00
-1.27850914e+00 3.39023769e-01 1.71664685e-01 -9.32224751e-01
-7.28889465e-01 -7.34720230e-01 -5.09360313e-01 9.38953698e-01
-6.99911535e-01 -4.87720907e-01 9.23004374e-02 7.05485106e-01
3.02568018e-01 -7.19203115e-01 1.64182818e+00 4.88310575e-01
-5.48401356e-01 1.07952368e+00 -5.02837971e-02 6.59649491e-01
1.03425372e+00 -1.02072406e+00 5.26360393e-01 8.54963422e-01
-1.41033128e-01 9.55908418e-01 5.69575548e-01 -6.68370426e-01
-8.04890811e-01 -8.46161187e-01 1.04329348e+00 -6.74465477e-01
8.40754628e-01 -5.82196236e-01 -7.91344404e-01 1.16921258e+00
6.64543509e-02 -1.25532225e-01 9.31469977e-01 5.88944733e-01
-6.18449688e-01 -7.38821551e-02 -1.51659620e+00 3.46276820e-01
1.18497241e+00 -9.30425525e-01 -1.04741502e+00 2.59180695e-01
6.41511381e-01 -3.25917631e-01 -8.37834418e-01 8.12904716e-01
3.14453281e-02 -3.33095938e-01 5.48743129e-01 -1.17438269e+00
-1.48279279e-01 2.38577634e-01 -6.83089197e-02 -9.19607162e-01
-1.42660588e-01 -3.97066414e-01 -3.79432559e-01 1.25136614e+00
7.66742408e-01 -4.71619457e-01 1.01692688e+00 1.05361521e+00
-2.21567288e-01 -4.29322980e-02 -9.90604579e-01 -8.61073256e-01
4.26161662e-02 -8.81497324e-01 -3.55542712e-02 1.35670424e+00
6.17789924e-01 7.97590613e-01 3.11008990e-01 -4.07939762e-01
2.51016259e-01 -5.00072598e-01 -5.50071411e-02 -9.14601207e-01
1.54010385e-01 -6.17008246e-02 -6.36702001e-01 -4.43570852e-01
6.71041310e-01 -9.09182310e-01 -9.14482996e-02 -1.15394843e+00
-3.57928038e-01 -1.67191654e-01 -4.70734268e-01 1.43443561e+00
3.87077838e-01 8.00927803e-02 -4.91851956e-01 2.66793855e-02
-8.91186357e-01 5.54715753e-01 2.75353312e-01 -3.30262661e-01
-2.86508590e-01 -3.76056641e-01 -5.41497409e-01 6.67650163e-01
9.44637895e-01 -5.51616967e-01 -9.38411534e-01 -3.85924071e-01
3.16145480e-01 -1.86352938e-01 -1.47677157e-02 -8.47933292e-01
1.58019766e-01 -2.20069457e-02 6.36699259e-01 -4.37287867e-01
4.56002384e-01 -7.64648139e-01 -3.26165527e-01 2.27334648e-01
-7.23234057e-01 -9.40515026e-02 1.09421170e+00 9.31758136e-02
-3.40875715e-01 -1.13342032e-01 9.67476845e-01 -1.51379220e-02
-1.54061985e+00 -2.40200688e-03 -7.90139735e-01 3.34364921e-01
9.05328751e-01 -2.93991625e-01 -4.05196100e-01 6.49469718e-02
-1.64747977e+00 2.16464162e-01 2.83055961e-01 7.44273305e-01
2.92866796e-01 -1.38187385e+00 -2.03445628e-01 2.38817587e-01
7.28141546e-01 -4.57361370e-01 -9.11556184e-02 8.06754708e-01
-1.56307593e-01 4.62415665e-01 -4.85815555e-01 -4.13480669e-01
-1.43322539e+00 4.54065263e-01 7.63125896e-01 7.89575949e-02
-2.51951426e-01 9.95878875e-01 -3.18948179e-01 -4.30860043e-01
7.01412678e-01 -2.06505850e-01 -2.08941802e-01 1.93819880e-01
3.78677666e-01 4.00953814e-02 6.63242698e-01 -2.27328643e-01
-8.89376104e-01 -3.59469175e-01 -3.45120281e-01 -3.84390295e-01
1.10902917e+00 3.00836992e-02 -5.03820553e-02 8.64749491e-01
1.40065753e+00 -3.97992134e-01 -3.92818213e-01 -3.07419568e-01
7.31134415e-01 2.87857741e-01 1.43584162e-01 -1.45194161e+00
-5.75679719e-01 4.23733562e-01 1.20748782e+00 -2.65262753e-01
1.34532130e+00 -6.68456405e-02 1.92184851e-01 8.66641164e-01
1.46760717e-01 -1.74446857e+00 -1.59505963e-01 9.80522513e-01
4.57778782e-01 -1.07598543e+00 -2.06950590e-01 -4.12408978e-01
-5.19999266e-01 9.87381935e-01 1.01678586e+00 1.50329202e-01
4.56111163e-01 7.64334619e-01 4.33173507e-01 -2.61970758e-01
-1.28460896e+00 6.48552999e-02 -1.30449250e-01 6.64007068e-01
8.47166896e-01 -2.49215364e-01 -2.76146829e-01 9.95278120e-01
2.47814767e-02 4.62917507e-01 2.91125290e-02 1.34454465e+00
1.24438755e-01 -1.09919298e+00 1.54369906e-01 2.33146742e-01
-4.41223651e-01 -2.31062725e-01 -8.56519938e-01 7.90611029e-01
1.26136586e-01 5.82958817e-01 4.96965438e-01 -1.28283218e-01
7.57984996e-01 1.09176648e+00 8.31851363e-01 -8.89902353e-01
-5.61436594e-01 -2.36250669e-01 9.15534377e-01 -8.04897130e-01
-4.10732090e-01 -5.57338357e-01 -2.02347422e+00 2.75234908e-01
-8.39913264e-02 2.90378481e-01 5.61951160e-01 1.03307974e+00
2.85034955e-01 3.47043365e-01 1.62305776e-02 2.27112740e-01
-4.36878085e-01 -9.47811186e-01 -4.00615573e-01 6.24184668e-01
4.29987758e-02 -5.36982358e-01 -5.07065833e-01 2.08188131e-01] | [10.530580520629883, 8.029118537902832] |
102b902b-4630-4821-9ff5-a662c6f46873 | hotnas-hierarchical-optimal-transport-for | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yang_HOTNAS_Hierarchical_Optimal_Transport_for_Neural_Architecture_Search_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yang_HOTNAS_Hierarchical_Optimal_Transport_for_Neural_Architecture_Search_CVPR_2023_paper.pdf | HOTNAS: Hierarchical Optimal Transport for Neural Architecture Search | Instead of searching the entire network directly, current NAS approaches increasingly search for multiple relatively small cells to reduce search costs. A major challenge is to jointly measure the similarity of cell micro-architectures and the difference in macro-architectures between different cell-based networks. Recently, optimal transport (OT) has been successfully applied to NAS as it can capture the operational and structural similarity across various networks. However, existing OT-based NAS methods either ignore the cell similarity or focus solely on searching for a single cell architecture. To address these issues, we propose a hierarchical optimal transport metric called HOTNN for measuring the similarity of different networks. In HOTNN, the cell-level similarity computes the OT distance between cells in various networks by considering the similarity of each node and the differences in the information flow costs between node pairs within each cell in terms of operational and structural information. The network-level similarity calculates OT distance between networks by considering both the cell-level similarity and the variation in the global position of each cell within their respective networks. We then explore HOTNN in a Bayesian optimization framework called HOTNAS, and demonstrate its efficacy in diverse tasks. Extensive experiments demonstrate that HOTNAS can discover network architectures with better performance in multiple modular cell-based search spaces. | ['Hongteng Xu', 'Yong liu', 'Jiechao Yang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['architecture-search', 'bayesian-optimization'] | ['methodology', 'methodology'] | [-4.21992421e-01 -4.96645331e-01 -3.89167368e-01 1.52017117e-01
-4.26115304e-01 -6.12124860e-01 2.62702763e-01 2.63420343e-01
-1.09447822e-01 7.14009404e-01 -2.14077383e-02 -6.34253547e-02
-8.22728157e-01 -8.70221674e-01 -2.05034181e-01 -9.69617665e-01
-2.62149155e-01 7.15636909e-01 6.12507820e-01 -4.37578969e-02
5.57984114e-01 7.14870930e-01 -7.33383894e-01 -2.09139735e-01
6.88324809e-01 1.26762724e+00 2.82801449e-01 5.62760115e-01
-2.94318676e-01 2.96046555e-01 -7.20045865e-01 6.72838837e-02
3.11253130e-01 -2.09001869e-01 -6.38149142e-01 -1.88973784e-01
9.03410390e-02 2.49456540e-01 -7.03741670e-01 9.13111806e-01
5.93595624e-01 1.41970515e-01 6.24368548e-01 -1.35401881e+00
-1.82837158e-01 7.14260340e-01 -3.69349778e-01 7.37923563e-01
-4.10551056e-02 1.98118433e-01 1.34746075e+00 -5.43215156e-01
6.24894917e-01 1.25304079e+00 4.28808957e-01 6.44131843e-03
-1.30354393e+00 -5.02073467e-01 3.26011360e-01 2.86152959e-01
-1.83411992e+00 -4.37118709e-01 5.41837871e-01 -1.88052893e-01
8.98656964e-01 3.50027412e-01 6.74425006e-01 3.71331722e-01
3.63702923e-01 3.47851008e-01 7.20147014e-01 5.41813225e-02
4.01656181e-01 -3.64013731e-01 -1.13002747e-01 5.59635997e-01
3.04670781e-01 -2.95139909e-01 -7.66642094e-01 -8.29885006e-02
7.61808217e-01 1.11469449e-02 -4.85434324e-01 -2.87894934e-01
-1.54424679e+00 3.75050962e-01 7.95273900e-01 5.57497442e-01
-1.48439825e-01 4.22690541e-01 3.10663104e-01 2.26865336e-01
1.25773828e-02 6.96588039e-01 -5.18665433e-01 -6.34926781e-02
-8.73039484e-01 -2.20240746e-02 1.07608151e+00 1.03400147e+00
8.59367073e-01 -1.66937441e-01 -4.61552560e-01 6.37398064e-01
3.68034363e-01 3.75765301e-02 2.55869716e-01 -1.17155671e+00
3.56669426e-01 8.30634952e-01 -3.30600381e-01 -1.27802980e+00
-4.47663814e-01 -9.05363262e-01 -1.00649309e+00 -5.52280307e-01
4.33636248e-01 -1.20351747e-01 -3.18061084e-01 1.62818241e+00
4.00706172e-01 6.45114601e-01 -1.66382760e-01 6.01080179e-01
4.76245642e-01 5.99484444e-01 -5.34773409e-01 -5.09974182e-01
9.51161742e-01 -1.22958982e+00 -2.45049104e-01 9.82196406e-02
6.63190663e-01 -6.67977035e-01 5.66447139e-01 -5.49933091e-02
-1.10577726e+00 -4.94889617e-02 -9.92919564e-01 5.39796650e-01
-4.12212640e-01 -3.80089223e-01 -2.53153294e-02 5.93388617e-01
-1.23808420e+00 5.50272465e-01 -5.85855007e-01 -4.77728635e-01
3.12390387e-01 4.70683485e-01 2.46023029e-01 -1.56249285e-01
-6.70036614e-01 3.55349123e-01 8.23500156e-02 1.24342591e-01
-9.23639953e-01 -8.10731769e-01 -5.24461627e-01 4.47383076e-01
9.58357930e-01 -8.07392240e-01 6.24290705e-01 -9.69046205e-02
-1.29724550e+00 3.97305097e-03 -4.77716893e-01 -2.36846551e-01
5.85368797e-02 6.88786745e-01 -3.38678449e-01 -8.11120272e-02
1.42342985e-01 6.76562369e-01 1.40182331e-01 -9.96736467e-01
-6.82665527e-01 -2.09570006e-01 1.02957904e-01 2.18584642e-01
-3.72765630e-01 -1.31943569e-01 -9.21992183e-01 -6.11364484e-01
4.50285137e-01 -9.14773464e-01 -3.95039797e-01 -2.70614717e-02
-8.91711891e-01 -2.86656201e-01 7.67667472e-01 1.45079836e-01
1.43922973e+00 -1.73384702e+00 4.26714361e-01 8.62920403e-01
5.05980253e-01 -1.86043933e-01 -2.25155815e-01 2.17417330e-01
6.47063792e-01 4.32965904e-01 1.09094635e-01 -1.02053180e-01
-1.18028611e-01 2.61282802e-01 2.52817512e-01 2.74159491e-01
-2.69148886e-01 6.23225927e-01 -9.95476961e-01 -4.07058477e-01
-2.56637782e-01 2.72286534e-01 -5.80628395e-01 -6.03252649e-02
-6.80930763e-02 3.27145487e-01 -5.02495587e-01 1.10207009e+00
3.46809924e-01 -5.04303396e-01 5.02379000e-01 -4.46291983e-01
3.56787778e-02 2.31335178e-01 -1.17763126e+00 1.12426949e+00
-1.76718652e-01 8.00730228e-01 2.63229668e-01 -1.01533413e+00
8.09013426e-01 6.24093153e-02 7.41829157e-01 -5.32449961e-01
2.01009028e-02 5.07454991e-01 3.38248938e-01 1.09498061e-01
1.82114646e-01 3.00522029e-01 -5.50150648e-02 4.46995795e-01
-3.85268092e-01 1.82632320e-02 4.98451054e-01 1.17164105e-01
1.59384966e+00 -9.18025792e-01 -2.89480258e-02 -5.91396153e-01
7.38788724e-01 -4.50441211e-01 9.70160306e-01 7.68961430e-01
-4.66089129e-01 3.28998148e-01 6.49178743e-01 -1.77099764e-01
-6.05511606e-01 -1.12200737e+00 3.16976123e-02 7.78165281e-01
8.49351287e-01 -5.93704820e-01 -5.23219526e-01 -6.53791130e-01
2.11181417e-01 -2.71479897e-02 -2.01997802e-01 8.39152858e-02
-6.16870999e-01 -6.86389804e-01 5.73376000e-01 3.35632443e-01
4.32527959e-01 -3.57547671e-01 -1.20393693e-01 1.57880172e-01
-1.84627131e-01 -1.21283746e+00 -9.95841503e-01 -1.13149090e-02
-9.15753424e-01 -9.81930912e-01 -3.44002694e-01 -6.45931363e-01
5.85821450e-01 4.78050768e-01 9.53040481e-01 4.68185544e-01
-2.53583163e-01 9.76613015e-02 3.96517813e-02 1.92446038e-01
2.84143928e-02 6.75784051e-01 2.34299809e-01 7.45919999e-03
-2.65978277e-01 -5.42106330e-01 -7.12389290e-01 1.08466232e+00
-3.25408250e-01 -4.82426196e-01 3.93244803e-01 5.41867077e-01
9.49158788e-01 5.79304457e-01 4.92814124e-01 -3.67360622e-01
5.33018410e-01 -8.27828586e-01 -5.84228516e-01 4.39676613e-01
-8.84961426e-01 1.77946046e-01 3.49393785e-01 -2.05167577e-01
-5.05466402e-01 -5.87930918e-01 3.42900991e-01 -3.16325009e-01
3.42575580e-01 7.54986107e-01 -2.56283253e-01 -3.20204109e-01
5.95970191e-02 -7.01010693e-04 -1.66350335e-01 -2.80174553e-01
-3.11105043e-01 2.54458517e-01 7.58659989e-02 -6.84234321e-01
6.57740235e-01 6.59017444e-01 5.61501086e-01 -5.50165057e-01
-5.14857173e-01 -5.58010757e-01 -3.64188671e-01 -2.98567086e-01
3.47718149e-01 -6.43750131e-01 -9.48982120e-01 2.50616729e-01
-8.26997221e-01 -4.07754242e-01 5.61696589e-02 3.10060084e-01
-8.12401697e-02 3.57761472e-01 -6.81870639e-01 -4.86665338e-01
-5.95236756e-02 -1.50521576e+00 6.93855107e-01 1.83306634e-01
-1.99606821e-01 -1.28219867e+00 -8.09224695e-02 2.09819794e-01
5.80403686e-01 -1.67490810e-01 1.35257566e+00 -5.82301855e-01
-1.41603768e+00 7.79432878e-02 -6.74634039e-01 -3.26909751e-01
3.97875905e-01 2.86527783e-01 -1.49587825e-01 -5.43744683e-01
-4.80027229e-01 3.26461405e-01 6.59556687e-01 7.03230381e-01
1.14852118e+00 -1.98106319e-01 -8.88817966e-01 7.07535028e-01
1.49568188e+00 2.03029558e-01 3.57104480e-01 3.01917225e-01
6.85719967e-01 5.32150865e-01 3.04109991e-01 2.90393233e-01
5.60573220e-01 9.74008322e-01 6.11966431e-01 4.41196114e-01
-6.08540885e-02 2.36185566e-01 3.73874813e-01 1.21457052e+00
2.18592510e-01 -7.78921902e-01 -9.43880022e-01 5.81680834e-01
-1.86043382e+00 -5.51437736e-01 3.40642184e-02 1.92417371e+00
4.23031569e-01 3.88814896e-01 5.56760915e-02 -2.15031162e-01
9.47367430e-01 1.95141181e-01 -8.99140477e-01 -8.19851086e-02
-3.66815865e-01 -4.58065391e-01 6.52383864e-01 4.47560668e-01
-4.17169064e-01 4.97397453e-01 6.72117758e+00 1.42443740e+00
-8.88307631e-01 -1.27550378e-01 8.38452399e-01 -4.51243162e-01
-2.88716376e-01 1.80876121e-01 -1.11155057e+00 7.92038381e-01
7.63243794e-01 -2.67524481e-01 7.51238227e-01 4.49564248e-01
2.46934474e-01 -9.70991999e-02 -1.16711771e+00 9.14390385e-01
-2.38508880e-01 -1.89968431e+00 1.34163141e-01 5.66002607e-01
9.45816875e-01 3.40754092e-01 -1.43936902e-01 -1.53752603e-02
1.31220132e-01 -9.43083882e-01 5.78180194e-01 5.58074951e-01
4.11337465e-01 -6.07817113e-01 8.65666509e-01 1.60288244e-01
-1.80398631e+00 -3.55284452e-01 -1.17262572e-01 1.11132093e-01
1.85685605e-01 7.33019292e-01 -7.43800342e-01 3.90883178e-01
7.66379952e-01 7.25179911e-01 -2.57840693e-01 1.40802503e+00
5.99498868e-01 1.85658991e-01 -5.97360492e-01 -1.66754797e-01
1.83759466e-01 -3.41106266e-01 1.03510821e+00 8.70851874e-01
8.05363953e-01 -4.33193654e-01 4.23655689e-01 8.86326909e-01
-4.80319828e-01 -1.94457933e-01 -1.60720557e-01 7.04273731e-02
1.35303462e+00 1.15172267e+00 -1.01651466e+00 1.01398773e-01
-1.65402964e-01 2.21286923e-01 1.90527707e-01 4.12757605e-01
-6.02340341e-01 -4.13758546e-01 1.12522006e+00 1.66012391e-01
2.92532772e-01 -3.06351691e-01 -4.91993368e-01 -5.59546530e-01
-1.59756303e-01 -5.38631737e-01 4.78322387e-01 -4.26890194e-01
-1.26802361e+00 4.95212585e-01 -1.82968244e-01 -9.48372900e-01
3.42037588e-01 -3.77387017e-01 -7.98801601e-01 5.10150075e-01
-1.37780082e+00 -4.54726189e-01 -4.65092152e-01 3.29006165e-01
7.31888890e-01 -4.58181471e-01 1.90625548e-01 2.23184749e-01
-1.20921433e+00 6.72215760e-01 4.14891958e-01 -4.12360840e-02
4.15234715e-01 -9.80909288e-01 1.15952149e-01 5.58733046e-01
-5.57716712e-02 5.44546247e-01 2.39077210e-01 -5.89751244e-01
-1.34269941e+00 -9.54731345e-01 7.31250703e-01 -1.66466027e-01
5.92553556e-01 -5.70491292e-02 -5.82204700e-01 1.07847713e-01
3.23069245e-02 5.25338575e-02 5.91565251e-01 1.03515431e-01
-8.62511918e-02 -6.41505480e-01 -8.83671224e-01 8.98993552e-01
1.27368724e+00 -5.70805430e-01 5.02296925e-01 1.66872442e-01
8.20299447e-01 -1.10996678e-01 -1.09712863e+00 5.78725636e-01
4.35215563e-01 -9.29235041e-01 1.00186038e+00 -3.77488956e-02
-3.60565424e-01 -7.16273308e-01 -4.89904523e-01 -1.19257951e+00
-4.01768029e-01 -7.39050806e-01 -2.93456376e-01 1.04912186e+00
5.09203136e-01 -7.01541364e-01 1.11704087e+00 2.32491866e-01
-1.12784274e-01 -1.44129002e+00 -1.20075071e+00 -9.83304262e-01
-3.27166736e-01 -1.12817265e-01 1.10229218e+00 7.42715001e-01
-8.08484182e-02 4.55005318e-01 2.07637414e-01 3.28679323e-01
4.90984380e-01 5.48097529e-02 5.10181725e-01 -1.33365738e+00
-2.51068622e-01 -1.11399770e+00 -3.44584018e-01 -1.33203113e+00
1.31389946e-01 -8.55057061e-01 -6.64256513e-02 -1.22716272e+00
2.66993225e-01 -9.45755005e-01 -5.49687862e-01 -1.24958798e-01
2.05128431e-01 7.54123330e-02 1.39689043e-01 5.80683112e-01
-1.03055215e+00 3.99111271e-01 1.17534721e+00 -2.85928279e-01
-1.82704180e-01 -1.09463446e-01 -6.21453822e-01 6.05140448e-01
7.47246683e-01 -4.22813594e-01 -4.09522951e-01 -5.12245476e-01
4.20971990e-01 -6.61128685e-02 -3.69910747e-02 -1.25563383e+00
8.04073632e-01 -3.46540540e-01 -2.75015235e-02 -6.56723619e-01
4.27824736e-01 -7.63198853e-01 2.39319503e-01 4.56344843e-01
-8.06532577e-02 2.93780386e-01 -2.67728627e-01 1.12616587e+00
-1.53670743e-01 1.35103017e-01 5.20030618e-01 1.24938905e-01
-2.93448806e-01 6.09345913e-01 -5.01715660e-01 8.37965980e-02
1.03099740e+00 -7.79803455e-01 -5.95113516e-01 -2.86894679e-01
-5.10248840e-01 8.91246438e-01 5.80004513e-01 -9.16212127e-02
4.50048000e-01 -1.18213153e+00 -3.12520444e-01 -5.71097881e-02
-9.79463160e-02 7.96169192e-02 1.16966628e-01 1.38890028e+00
-5.36097586e-01 6.00164235e-01 3.78703326e-02 -8.79401505e-01
-1.25367951e+00 -9.29645915e-03 6.84071302e-01 -6.24791622e-01
3.96194935e-01 1.06785870e+00 2.08383232e-01 -3.55610967e-01
3.68631721e-01 -3.71495903e-01 3.52936387e-02 -3.72843109e-02
-1.20907061e-01 1.05069435e+00 3.93869020e-02 -5.80025434e-01
-4.36738133e-01 7.84662127e-01 2.95227896e-02 1.85023621e-01
9.19357896e-01 -5.83794415e-01 -6.18779063e-01 2.70563245e-01
1.13345361e+00 -2.60875169e-02 -1.06202149e+00 -4.92930472e-01
2.42808953e-01 -6.01204455e-01 1.98505372e-01 -3.10758978e-01
-1.47764814e+00 3.04712266e-01 1.56823620e-01 2.75733829e-01
1.02147448e+00 9.27681997e-02 1.02916110e+00 7.36497700e-01
5.42217433e-01 -1.19572353e+00 5.50261378e-01 7.70126224e-01
2.96936512e-01 -6.92187130e-01 5.58437780e-02 -7.21873164e-01
-3.18723992e-02 1.02164984e+00 8.38592589e-01 2.90657640e-01
1.03459132e+00 1.13413103e-01 -5.42350233e-01 -3.31066161e-01
-9.44767058e-01 -9.47260931e-02 5.74810579e-02 3.42043251e-01
4.83541936e-02 -2.31115576e-02 1.96166262e-01 2.42056161e-01
1.27927050e-01 -6.21113479e-01 3.95947427e-01 6.75219119e-01
-5.91772199e-01 -1.28485775e+00 -7.99469724e-02 7.43840575e-01
-1.03723079e-01 -8.82451087e-02 -3.95563960e-01 3.17046076e-01
6.16478883e-02 9.59635615e-01 4.42996085e-01 -5.70189059e-01
4.78537381e-02 -4.58824813e-01 2.26698205e-01 -5.99865973e-01
-3.00990760e-01 2.12118298e-01 1.47713870e-01 -5.65178275e-01
-6.96539730e-02 -5.82562745e-01 -1.22471201e+00 -7.89399207e-01
-5.20478427e-01 2.60688841e-01 3.59013617e-01 1.00362313e+00
7.57872581e-01 5.91932118e-01 7.43837774e-01 -6.63584054e-01
-2.72120714e-01 -3.44862759e-01 -7.30192363e-01 -3.46681088e-01
2.46201351e-01 -7.01211989e-01 -2.56109834e-01 -6.30387306e-01] | [7.194499969482422, 5.992859363555908] |
b8a6116a-f8da-4831-b16a-5e51cd4bf576 | ontology-based-information-integration-a | 1909.13762 | null | https://arxiv.org/abs/1909.13762v1 | https://arxiv.org/pdf/1909.13762v1.pdf | Ontology Based Information Integration: A Survey | An ontology makes a special vocabulary which describes the domain of interest and the meaning of the term on that vocabulary. Based on the precision of the specification, the concept of the ontology contains several data and conceptual models. The notion of ontology has emerged into wide ranges of applications including database integration, peer-to-peer systems, e-commerce, semantic web, etc. It can be considered as a practical tool for conceptualizing things which are expressed in computer format. This paper is devoted to ontology matching as a mean or information integration. Several matching solutions have been presented from various areas such as databases, information systems and artificial intelligence. All of them take advantages of different attributes of ontology like, structures, data instances, semantics and labels and its other valuable properties. The solutions have some common techniques and cope with similar problems, but use different methods for combining and exploiting their results. Information integration is among the first classes of applications at which matching was considered as a probable solution. Information integration contains many fields including, data integration, schema integration, catalogue integration and semantic integration. We cover these notions in term of ontology in our proposed paper. | ['Farajollah Tahernezhad-Javazm', 'Maliheh Heydarpour Shahrezaei', 'Maryam Alizadeh'] | 2019-09-26 | null | null | null | null | ['ontology-matching'] | ['knowledge-base'] | [-3.58701013e-02 6.29108548e-02 -3.81070614e-01 -4.14500177e-01
8.46507400e-02 -6.29747808e-01 9.56018031e-01 6.95338488e-01
-3.18630189e-01 5.03061891e-01 1.65773138e-01 -2.24833302e-02
-9.72661912e-01 -1.28758717e+00 9.39437151e-02 -2.52742380e-01
1.12058528e-01 7.35079169e-01 6.52877867e-01 -7.58186460e-01
4.08868104e-01 4.85736668e-01 -2.35737300e+00 3.09974372e-01
7.06236064e-01 1.18883467e+00 2.80643225e-01 -1.38430148e-01
-1.27979612e+00 4.79855269e-01 -5.26198924e-01 -4.86400664e-01
-8.53170920e-03 -7.48798028e-02 -1.06932056e+00 -1.07656255e-01
-3.18984121e-01 5.67944884e-01 3.59599560e-01 1.34397554e+00
1.01752713e-01 1.91719756e-02 4.58496571e-01 -1.61961901e+00
-4.95505780e-01 4.76803154e-01 2.96448953e-02 -8.46309438e-02
8.22135329e-01 -8.51612568e-01 6.38957262e-01 -4.85582530e-01
8.87621701e-01 1.47040761e+00 4.01271731e-01 2.83465475e-01
-6.62962258e-01 -3.55061382e-01 -2.20287010e-01 4.59655911e-01
-1.45336485e+00 -1.17675997e-01 4.41336423e-01 -4.84074324e-01
8.73133600e-01 4.93697405e-01 6.16470277e-01 6.54998124e-02
1.88526586e-01 1.10330828e-01 9.34391081e-01 -9.18221354e-01
3.57561022e-01 7.88560927e-01 6.43227458e-01 1.35596737e-01
6.91168725e-01 -4.50083882e-01 -9.74993259e-02 -4.15778130e-01
2.43411571e-01 2.95839697e-01 -4.96351868e-02 -4.89425391e-01
-8.87825668e-01 5.38609624e-01 4.24337909e-02 1.09420168e+00
-2.59458393e-01 -3.70519787e-01 5.03717542e-01 3.81582737e-01
-1.85332045e-01 2.33466700e-01 -4.27808583e-01 1.83548480e-01
-9.89419520e-02 4.11336690e-01 1.15404093e+00 1.39253426e+00
9.34023261e-01 -3.58104944e-01 6.61248565e-01 8.68565738e-01
5.69373727e-01 2.69393951e-01 9.29258525e-01 -9.33385491e-01
1.00697689e-01 1.27881098e+00 1.87381446e-01 -8.79862368e-01
-3.82591814e-01 9.08737406e-02 -3.33107293e-01 2.24873662e-01
-1.21794909e-01 4.73233074e-01 -4.53454584e-01 1.43510306e+00
6.28573596e-01 -2.78780252e-01 6.33435428e-01 3.72225523e-01
1.24977064e+00 4.56719697e-01 4.75744218e-01 -2.48626500e-01
2.00624728e+00 -3.06982309e-01 -1.23823297e+00 3.10218513e-01
5.85071445e-01 -1.10447264e+00 3.47717583e-01 2.73646623e-01
-8.17694426e-01 -4.35795218e-01 -9.48761940e-01 -1.44995570e-01
-1.42143869e+00 -6.67902231e-01 5.43985903e-01 6.57023311e-01
-8.98639560e-01 4.72918868e-01 -8.36918876e-02 -1.19727695e+00
-2.40365684e-01 3.77683342e-01 -5.08227527e-01 1.20011605e-01
-1.43608117e+00 1.13498831e+00 1.18581736e+00 -4.80591506e-01
1.28701329e-01 -1.49155140e-01 -7.53819644e-01 1.27534688e-01
5.56033134e-01 -7.63475478e-01 8.41722310e-01 -8.27422678e-01
-8.71412039e-01 1.14861012e+00 -1.44016547e-02 -4.58335578e-01
-5.23465015e-02 2.97147006e-01 -1.42840517e+00 -2.38322407e-01
4.40771520e-01 8.70426558e-03 -3.58704943e-03 -1.03728306e+00
-9.92100120e-01 -6.48840010e-01 3.08594167e-01 6.74842298e-02
-1.70045137e-01 5.22403896e-01 -1.85865045e-01 -2.49015629e-01
4.02265579e-01 -2.65390813e-01 -1.70944969e-03 -2.57365704e-01
2.71054715e-01 -5.76248646e-01 7.64821231e-01 -1.44955397e-01
1.28542602e+00 -1.95785344e+00 1.04413433e-02 5.06263793e-01
-1.97514314e-02 2.44951099e-01 4.41245258e-01 1.14764726e+00
2.31698528e-02 3.98291469e-01 -1.42961547e-01 4.97671664e-01
3.69286448e-01 5.22903979e-01 -1.56917170e-01 -1.87092006e-01
-5.29827237e-01 4.00621980e-01 -5.73419452e-01 -6.60980582e-01
4.13406909e-01 4.44647312e-01 -1.01202093e-01 -7.89706260e-02
-4.44236845e-01 -8.50679725e-02 -8.44181061e-01 6.80182636e-01
3.95701885e-01 1.56914815e-03 7.44940698e-01 -1.99018836e-01
-3.48764002e-01 9.38024297e-02 -1.69221318e+00 1.69633424e+00
-4.99828130e-01 -8.06342065e-02 -1.70973465e-02 -1.18528342e+00
1.21872234e+00 1.01463175e+00 7.05069363e-01 -7.71100819e-01
3.16948742e-01 8.29094648e-01 -1.58735678e-01 -8.88648868e-01
4.88471150e-01 -2.40626723e-01 1.65235121e-02 2.93868721e-01
1.38192311e-01 -6.80913180e-02 6.15391731e-01 -2.07704663e-01
5.85624933e-01 2.87895471e-01 1.24695337e+00 -5.78977406e-01
1.08071101e+00 1.12672880e-01 4.73812431e-01 1.62656114e-01
2.38160491e-01 -1.58986434e-01 3.74097377e-02 -7.33421147e-01
-1.04801607e+00 -8.13923061e-01 -5.27005017e-01 7.35402048e-01
4.61341619e-01 -6.24253452e-01 -5.96133173e-01 -4.32012370e-03
1.68748887e-03 5.59423745e-01 -1.50559708e-01 1.95620090e-01
-2.19300196e-01 -3.97411764e-01 4.18341979e-02 -2.05172598e-02
5.73147953e-01 -1.20051098e+00 -6.43190086e-01 5.41700482e-01
5.17740287e-03 -1.13330293e+00 6.33322239e-01 -2.48787299e-01
-1.04228139e+00 -1.22250795e+00 -5.32473251e-02 -6.75612986e-01
1.59345120e-01 3.76046628e-01 1.29382002e+00 6.02870762e-01
-1.37045383e-01 4.36077923e-01 -6.53073907e-01 -9.97136831e-01
-4.91031319e-01 -3.36188264e-02 3.16430219e-02 -1.43518686e-01
9.83240187e-01 -6.92306697e-01 -9.60703492e-02 3.71031582e-01
-1.46264040e+00 -3.68427098e-01 9.95338038e-02 2.70227134e-01
4.15209204e-01 5.74970543e-01 7.23331392e-01 -8.44629407e-01
6.65878117e-01 -7.92179167e-01 -7.26311684e-01 7.27799058e-01
-6.57304287e-01 2.24431351e-01 2.21422479e-01 2.20318601e-01
-1.10731316e+00 -4.44446594e-01 -1.15147516e-01 3.55010659e-01
-5.27557254e-01 6.23038113e-01 -8.87185752e-01 7.39830136e-02
3.16905290e-01 5.66103496e-02 -8.34793821e-02 -9.68846917e-01
2.15222716e-01 9.98710215e-01 2.81109601e-01 -8.22790086e-01
3.56758118e-01 5.38252413e-01 1.37538984e-01 -8.00597727e-01
-4.36562359e-01 -8.70401084e-01 -6.28418505e-01 -1.88257713e-02
1.01847804e+00 -4.81087834e-01 -7.55249977e-01 7.48680392e-03
-1.20343685e+00 7.53002167e-01 -4.54575598e-01 4.92440075e-01
-5.11241853e-01 2.96241164e-01 2.06283674e-01 -9.45831478e-01
-2.93760896e-01 -8.42998862e-01 4.31504309e-01 3.87225449e-01
-2.22226351e-01 -1.23999453e+00 -3.76331136e-02 2.37580225e-01
3.91675800e-01 2.12813884e-01 1.23047340e+00 -1.08891618e+00
-3.90245348e-01 -3.85189503e-01 1.24129625e-02 6.31664023e-02
4.67188299e-01 -2.34981984e-01 -5.99585772e-01 2.27496937e-01
6.01927610e-03 3.19588304e-01 5.94491921e-02 -3.08470458e-01
6.06241524e-01 -1.49108917e-01 -6.49351895e-01 1.47811398e-01
2.05635595e+00 9.85534608e-01 8.88877511e-01 9.27510917e-01
7.66726434e-02 1.07361519e+00 8.10518503e-01 3.20810139e-01
3.18677634e-01 1.16351807e+00 2.13453889e-01 3.50685954e-01
1.23120464e-01 1.56976625e-01 -3.48382801e-01 7.55032599e-01
-3.66038054e-01 -1.70817435e-01 -1.14265072e+00 4.52379346e-01
-2.08352995e+00 -1.27791941e+00 -2.14057922e-01 2.60583162e+00
4.92250174e-01 -2.55539715e-02 1.46054596e-01 3.24799061e-01
8.79853547e-01 -2.77503490e-01 5.02451770e-02 -6.66331828e-01
-1.20133653e-01 1.58951744e-01 9.72749218e-02 4.67189848e-01
-6.59246624e-01 6.51324272e-01 5.79780149e+00 6.83177531e-01
-6.79634392e-01 5.48435152e-01 -4.38497543e-01 6.58080101e-01
-5.31739235e-01 3.12096983e-01 -8.92724931e-01 4.82073814e-01
8.61781478e-01 -7.90325403e-01 3.88023823e-01 6.95681810e-01
-1.52527750e-01 -1.08215332e-01 -8.19450557e-01 7.76654243e-01
-3.62198949e-02 -1.41333568e+00 3.57825339e-01 1.78566724e-01
4.38447773e-01 -3.14656466e-01 -7.53367126e-01 -1.37858108e-01
1.47386730e-01 -7.37371325e-01 6.39260769e-01 8.45616460e-01
3.62610549e-01 -6.86791658e-01 1.14121759e+00 1.72983244e-01
-1.51109350e+00 -1.13459095e-01 -4.07392800e-01 -1.57051757e-01
4.03345734e-01 3.46918315e-01 -2.19535932e-01 1.44679272e+00
8.15765083e-01 3.61894757e-01 -4.33490574e-02 1.29518425e+00
1.99650466e-01 -4.37203288e-01 -3.53051871e-01 -1.70582324e-01
-1.80041775e-01 -6.29571199e-01 2.77184486e-01 7.97326565e-01
6.25016749e-01 3.34225357e-01 1.03964396e-01 6.37745380e-01
2.49063939e-01 7.55921841e-01 -7.99647808e-01 1.07365055e-02
7.42519438e-01 8.52986097e-01 -6.34501457e-01 -7.79022098e-01
-7.34496951e-01 1.62709847e-01 -3.18622619e-01 5.46442270e-02
-3.48214686e-01 -7.40307212e-01 8.36259902e-01 3.50953311e-01
-2.19840854e-01 -3.88396671e-03 2.61827130e-02 -1.00895953e+00
5.14393412e-02 -6.27406776e-01 6.42401516e-01 -7.34773695e-01
-1.03638399e+00 6.38812721e-01 5.59309006e-01 -1.48051798e+00
-3.30089748e-01 -7.26385891e-01 -1.57700852e-01 9.01719213e-01
-1.22425485e+00 -9.52231646e-01 -4.27769542e-01 5.80666184e-01
5.95230423e-02 -4.77193952e-01 1.33115089e+00 5.80970109e-01
6.95764720e-02 -4.69331264e-01 1.02944098e-01 -1.75131038e-01
4.48711902e-01 -9.87168789e-01 -1.14938416e-01 4.23488289e-01
1.53130675e-02 8.12878013e-01 7.17710972e-01 -4.93408650e-01
-8.92583191e-01 -2.79610634e-01 1.48165607e+00 -1.31512403e-01
7.21481919e-01 9.78491083e-02 -9.38635647e-01 7.03069329e-01
4.43471909e-01 -2.86425084e-01 9.45772529e-01 8.42038449e-03
-3.38217795e-01 -5.28721988e-01 -1.43335736e+00 2.86668152e-01
1.02447712e+00 -3.26227188e-01 -1.17801559e+00 3.87455136e-01
7.49281585e-01 1.15039274e-01 -1.45409214e+00 3.12939525e-01
7.32100904e-01 -1.18804574e+00 1.06230009e+00 -6.94199741e-01
-3.47542584e-01 -6.05910778e-01 -5.16417265e-01 -6.74550951e-01
-1.12617746e-01 -2.46559411e-01 4.18891072e-01 1.53759682e+00
1.17414832e-01 -1.17048764e+00 4.77673113e-02 6.79832697e-01
3.51970717e-02 -1.44765405e-02 -9.91909087e-01 -8.76698554e-01
-3.28854799e-01 -3.55272532e-01 1.59368646e+00 1.10872900e+00
3.77635241e-01 1.58192471e-01 1.48236468e-01 -7.01744631e-02
5.99119484e-01 2.18382716e-01 6.13687992e-01 -2.27171946e+00
1.77802622e-01 -5.60306191e-01 -1.17945611e+00 -1.78936377e-01
-1.42101929e-01 -8.35965753e-01 -7.73579180e-01 -2.04909587e+00
-1.89642176e-01 -7.93425500e-01 -3.57637584e-01 2.88269877e-01
6.30712092e-01 -1.23286076e-01 2.94620216e-01 5.06117105e-01
-4.25858319e-01 -8.64506289e-02 8.69872153e-01 3.95476967e-02
6.34766221e-02 -5.74795157e-02 -5.23174822e-01 8.42110276e-01
8.65097940e-01 -5.76531112e-01 -5.72232842e-01 -2.63385803e-01
5.53294122e-01 -3.85069549e-02 -4.31956574e-02 -1.17386472e+00
3.76903415e-01 -5.00717223e-01 -3.63860905e-01 -1.12867363e-01
3.84490490e-01 -1.71808708e+00 1.07793844e+00 3.92291993e-01
1.18687063e-01 1.04772657e-01 -1.40019879e-01 1.20664336e-01
-6.56200588e-01 -1.03712916e+00 5.65265894e-01 -6.32993519e-01
-1.36425722e+00 -8.14803541e-02 -1.12176634e-01 -5.61432205e-02
1.20571125e+00 -5.26261508e-01 -3.09327602e-01 9.40143317e-02
-9.04704332e-01 1.65065471e-02 7.54760206e-01 6.89602077e-01
-1.30141620e-04 -1.46339250e+00 -8.18409696e-02 -4.08792235e-02
5.38402081e-01 -3.64468127e-01 1.40803568e-02 3.55886251e-01
-5.40390372e-01 7.88240194e-01 -6.93045557e-01 -2.50742882e-01
-1.24658656e+00 9.48928773e-01 4.56319094e-01 -9.50213671e-02
-3.77789617e-01 -1.03558101e-01 -8.60852450e-02 -5.22615373e-01
1.28069326e-01 -1.22809634e-01 -8.83126795e-01 1.98560640e-01
7.34892845e-01 4.04411554e-01 1.68081954e-01 -1.00246453e+00
-5.93098402e-01 1.20999682e+00 5.01670718e-01 -1.99614599e-01
1.03214800e+00 -3.34781349e-01 -8.73121202e-01 6.65597796e-01
8.32190871e-01 2.07979809e-02 2.22111627e-01 -5.10566831e-01
8.25725615e-01 -3.31025004e-01 -3.92887950e-01 -7.21368551e-01
-5.17882168e-01 4.81180638e-01 6.69903874e-01 1.06896949e+00
1.08312464e+00 1.35281682e-01 3.38937610e-01 4.06430453e-01
1.06971002e+00 -1.27853811e+00 -8.27751517e-01 2.67449707e-01
8.18912268e-01 -9.13099408e-01 7.88149834e-02 -8.34555328e-01
-1.43896669e-01 1.40615559e+00 1.28577620e-01 2.83283710e-01
1.06612778e+00 2.67088026e-01 2.18678445e-01 -6.22050405e-01
-4.98759657e-01 -6.89530611e-01 2.07233399e-01 9.03504908e-01
5.52351832e-01 1.27244247e-02 -1.35671580e+00 5.19534588e-01
-1.21409178e-01 4.53208089e-01 2.84678698e-01 1.21926641e+00
-1.00013781e+00 -1.89178336e+00 -6.44941449e-01 -2.49065049e-02
-4.64779884e-01 2.30382815e-01 -9.83404070e-02 9.82165158e-01
7.88347542e-01 1.00685525e+00 2.06442073e-01 -4.51849625e-02
7.10789740e-01 3.90133470e-01 2.94455647e-01 -6.07521892e-01
-4.96663809e-01 -3.98883104e-01 2.33526692e-01 -3.19467217e-01
-1.16794515e+00 -3.05791378e-01 -1.62741399e+00 -2.36657873e-01
-1.78452209e-01 8.37203503e-01 1.05813527e+00 1.21699035e+00
2.90545493e-01 3.21356088e-01 4.50729486e-03 -1.92840219e-01
-1.38285868e-02 -7.04145610e-01 -9.17103708e-01 7.07913280e-01
-2.98182696e-01 -9.57125962e-01 7.12654740e-02 1.87166985e-02] | [9.177143096923828, 7.958730697631836] |
24a13d82-1a5a-44f5-a91d-60d5eab8290d | self-supervised-learning-of-face | 1903.01000 | null | http://arxiv.org/abs/1903.01000v1 | http://arxiv.org/pdf/1903.01000v1.pdf | Self-Supervised Learning of Face Representations for Video Face Clustering | Analyzing the story behind TV series and movies often requires understanding
who the characters are and what they are doing. With improving deep face
models, this may seem like a solved problem. However, as face detectors get
better, clustering/identification needs to be revisited to address increasing
diversity in facial appearance. In this paper, we address video face clustering
using unsupervised methods. Our emphasis is on distilling the essential
information, identity, from the representations obtained using deep pre-trained
face networks. We propose a self-supervised Siamese network that can be trained
without the need for video/track based supervision, and thus can also be
applied to image collections. We evaluate our proposed method on three video
face clustering datasets. The experiments show that our methods outperform
current state-of-the-art methods on all datasets. Video face clustering is
lacking a common benchmark as current works are often evaluated with different
metrics and/or different sets of face tracks. | ['Rainer Stiefelhagen', 'Vivek Sharma', 'M. Saquib Sarfraz', 'Makarand Tapaswi'] | 2019-03-03 | null | null | null | null | ['face-clustering'] | ['computer-vision'] | [-1.61508337e-01 -4.32782769e-01 -8.48085657e-02 -6.00736916e-01
-4.52886701e-01 -6.17659032e-01 6.70451760e-01 -3.73080641e-01
-1.67812213e-01 2.10304424e-01 2.13311985e-02 4.32277352e-01
-1.06394731e-01 -2.78966993e-01 -6.60289347e-01 -8.73746932e-01
-4.24839258e-02 7.69753158e-01 2.60929409e-02 -1.10951355e-02
9.64586288e-02 4.06562507e-01 -2.03321004e+00 4.67757851e-01
3.17279845e-01 1.12431324e+00 -3.88356186e-02 4.20196980e-01
-2.33621880e-01 1.07118189e+00 -4.76186454e-01 -6.33386314e-01
3.24033886e-01 -5.39504528e-01 -7.82046139e-01 5.45891047e-01
1.05927110e+00 -2.05808192e-01 -3.54229957e-01 1.20474374e+00
3.97965074e-01 1.30919650e-01 8.51272583e-01 -1.51965976e+00
-5.68663061e-01 5.40238559e-01 -8.80505204e-01 1.61087602e-01
3.72799724e-01 -2.65608191e-01 6.28852427e-01 -8.87036979e-01
8.08349550e-01 1.47020125e+00 7.87867785e-01 8.39421690e-01
-1.04924917e+00 -9.41885293e-01 1.01497456e-01 5.37190437e-01
-1.66855741e+00 -9.98722970e-01 8.98867965e-01 -6.24084771e-01
3.09302479e-01 -1.49568329e-02 2.88562804e-01 1.24862134e+00
-5.49594939e-01 8.11878443e-01 7.81728566e-01 -3.36044520e-01
1.18696108e-01 2.74962842e-01 8.49278197e-02 8.11467409e-01
-1.18744504e-02 -3.81249577e-01 -5.54915905e-01 1.23720415e-01
4.50346649e-01 1.97746828e-01 -1.57171026e-01 -6.22454226e-01
-6.43608987e-01 7.68141091e-01 5.69447726e-02 4.67166215e-01
-2.05718949e-01 1.04046568e-01 4.39275384e-01 1.91909343e-01
4.56995934e-01 -6.13005422e-02 -2.27481723e-01 1.32552043e-01
-1.53307438e+00 3.22086774e-02 8.25060010e-01 8.38877141e-01
7.79180169e-01 -2.26683822e-02 8.03308040e-02 9.68613803e-01
3.84895474e-01 2.78828472e-01 3.44381452e-01 -1.27003837e+00
5.31811304e-02 6.19020998e-01 -1.74501047e-01 -1.15700793e+00
-1.93592727e-01 -7.58235008e-02 -8.33291650e-01 1.97674245e-01
6.08292103e-01 2.44582025e-03 -9.42272842e-01 1.77764082e+00
3.26000035e-01 5.52176297e-01 -1.78649008e-01 7.58466780e-01
9.82645214e-01 3.89664829e-01 -1.33781835e-01 -2.90494621e-01
1.27698970e+00 -9.65184450e-01 -8.46063495e-01 1.19771473e-01
2.42920324e-01 -8.01225722e-01 6.72107935e-01 4.62473810e-01
-8.85487795e-01 -6.95649087e-01 -7.18538821e-01 1.56666160e-01
-3.50326389e-01 5.78188717e-01 4.79029298e-01 1.14482629e+00
-1.38595986e+00 5.44937551e-01 -7.64225841e-01 -8.07111800e-01
8.53694856e-01 7.31088519e-01 -7.21613765e-01 -2.05847114e-01
-5.61142683e-01 3.86210084e-01 9.74847749e-02 9.68794525e-02
-1.13177121e+00 -5.47051132e-01 -6.25537574e-01 1.50772575e-02
3.69681567e-01 -2.95188606e-01 9.59416926e-01 -1.65962577e+00
-1.49939132e+00 1.19150960e+00 -2.87491977e-01 -2.44830966e-01
3.28861088e-01 -1.04227349e-01 -5.65806389e-01 4.54915971e-01
5.39147854e-02 8.65400136e-01 1.23683512e+00 -1.59442270e+00
-5.29816747e-01 -6.00717843e-01 -1.62496209e-01 -1.34138376e-01
-7.80790627e-01 4.23729509e-01 -9.22168791e-01 -4.64371711e-01
-7.54815191e-02 -1.11929166e+00 2.89716125e-01 7.85815641e-02
-1.93330318e-01 -3.77562672e-01 1.24536633e+00 -5.82864821e-01
1.17816162e+00 -2.31184292e+00 1.76632553e-01 2.25414604e-01
2.25524127e-01 2.83965766e-01 1.58533771e-02 2.79675871e-01
-2.61058718e-01 -4.96870931e-03 -2.84208536e-01 -8.21808338e-01
-9.77472886e-02 1.19649440e-01 1.18783057e-01 7.16333747e-01
1.29371524e-01 3.05448502e-01 -5.46402872e-01 -8.20011199e-01
4.24389318e-02 6.23703837e-01 -6.87036455e-01 8.52821022e-02
-6.57159016e-02 4.54773277e-01 -5.35538197e-02 8.47736418e-01
9.01719451e-01 -3.20162475e-01 4.56738293e-01 -2.36622140e-01
2.27892980e-01 -4.63976115e-01 -1.45900333e+00 1.80889571e+00
1.12250829e-02 8.35080504e-01 4.76631403e-01 -1.20920694e+00
5.54957390e-01 3.43166441e-01 7.20708013e-01 -2.89783031e-01
3.98077011e-01 -1.56556010e-01 3.41239423e-02 -6.52768493e-01
1.19465463e-01 1.62647411e-01 3.88438284e-01 5.87865293e-01
5.30780613e-01 5.58035374e-01 5.37603676e-01 3.10033679e-01
8.21015537e-01 1.60087034e-01 -2.96457201e-01 -4.88808006e-01
8.11775148e-01 -1.83027893e-01 5.77804923e-01 3.02155584e-01
-3.36311877e-01 7.30789483e-01 4.27472979e-01 -4.57016975e-01
-7.24271834e-01 -8.69851112e-01 -4.83975224e-02 1.33340919e+00
-1.40737012e-01 -5.55250466e-01 -1.29580426e+00 -7.70601988e-01
-1.99785680e-01 7.87284598e-02 -9.22012508e-01 1.25038445e-01
-3.91430378e-01 -6.52039170e-01 5.74626625e-01 4.07098323e-01
4.02231753e-01 -9.58085418e-01 -4.66550104e-02 -3.12933326e-01
-8.95407274e-02 -1.20374453e+00 -5.28498769e-01 -2.82609522e-01
-5.19181788e-01 -1.38133299e+00 -6.58869326e-01 -1.09227145e+00
8.69416833e-01 4.19272810e-01 1.14588857e+00 3.09034109e-01
-3.15370679e-01 7.75739789e-01 -3.50040406e-01 -2.52791673e-01
-2.00870901e-01 5.33731701e-03 3.98231506e-01 7.84044802e-01
6.31357074e-01 -5.39205432e-01 -5.87227523e-01 3.85554403e-01
-9.06740308e-01 -5.86436749e-01 2.34058946e-01 5.20233154e-01
2.19838187e-01 3.35878670e-01 3.01201969e-01 -1.02881718e+00
1.97030440e-01 -4.83834863e-01 -3.48212302e-01 5.03125906e-01
-1.70002818e-01 -6.11779168e-02 5.65294147e-01 -4.01085198e-01
-1.17452610e+00 5.77136695e-01 2.08666295e-01 -1.10095477e+00
-3.93879473e-01 -1.06461644e-01 -3.60518038e-01 -2.25940585e-01
4.75481123e-01 6.43993914e-02 3.08588147e-01 -5.05163014e-01
2.79957235e-01 6.52917862e-01 5.54333925e-01 -5.71498156e-01
8.05695772e-01 9.46591139e-01 -2.38193065e-01 -1.02999258e+00
-6.20944560e-01 -7.15029836e-01 -9.59275186e-01 -5.58322191e-01
1.03312719e+00 -9.42724824e-01 -1.02526796e+00 5.40362000e-01
-9.49991822e-01 8.30193534e-02 2.30409548e-01 2.72775531e-01
-3.68857235e-01 5.96932173e-01 -6.23374641e-01 -8.07410359e-01
1.67177208e-02 -1.24127138e+00 9.71644402e-01 2.35761434e-01
-1.81864843e-01 -9.41386938e-01 -3.39413714e-03 5.96417546e-01
1.40237421e-01 1.17333390e-01 4.11564559e-01 -5.36888063e-01
-3.54851454e-01 -4.65379581e-02 -8.84099603e-02 3.78113121e-01
3.15765202e-01 5.75215042e-01 -1.32734168e+00 -3.35242122e-01
6.04652837e-02 -4.27562803e-01 9.59835112e-01 2.98129678e-01
1.40123475e+00 -2.28256792e-01 -3.91270667e-01 6.97493851e-01
1.22436345e+00 -9.19998661e-02 5.96352458e-01 -5.13769835e-02
9.10094559e-01 1.08157897e+00 7.59418681e-02 3.38875920e-01
4.63433236e-01 6.03655636e-01 2.85076231e-01 5.08951768e-02
-1.92656636e-01 -5.70581332e-02 6.04582787e-01 6.27418756e-01
-2.14502782e-01 -8.46611410e-02 -8.79964292e-01 4.25670862e-01
-1.82840860e+00 -1.41477334e+00 -1.27887607e-01 1.98236358e+00
5.66417813e-01 -3.14520031e-01 5.37139773e-01 1.63570777e-01
9.72178459e-01 -1.70984432e-01 -4.43459094e-01 1.67400733e-01
-1.24888837e-01 1.91483393e-01 1.00759566e-01 8.73075947e-02
-1.46147680e+00 1.02593195e+00 6.18089533e+00 9.54146922e-01
-1.11949611e+00 3.25434953e-01 8.65618169e-01 -2.57959634e-01
3.29914004e-01 -2.21412003e-01 -7.67221153e-01 4.42127019e-01
8.17270339e-01 1.93170443e-01 4.53001887e-01 8.89761090e-01
-1.23709701e-01 1.72720496e-02 -1.37285841e+00 1.48605108e+00
7.15530992e-01 -1.18554103e+00 2.71584429e-02 4.73752320e-02
9.24502909e-01 -2.81278253e-01 2.97375828e-01 1.33442834e-01
2.22035259e-01 -1.16244912e+00 6.97436631e-01 5.42670488e-01
5.43449700e-01 -7.66452789e-01 5.60072243e-01 -2.86099352e-02
-1.21625590e+00 -2.39640981e-01 -2.48322383e-01 3.10344428e-01
-2.55135268e-01 8.85248408e-02 -3.52021486e-01 3.65124315e-01
1.21046841e+00 9.86472607e-01 -8.23050141e-01 7.00038910e-01
2.27267936e-01 5.04868150e-01 -2.82802224e-01 2.68199772e-01
1.42692953e-01 -3.49549353e-01 -5.38014807e-02 1.02249384e+00
2.30482534e-01 5.24107255e-02 2.07946897e-01 5.45020461e-01
-3.32703173e-01 2.10820720e-01 -5.46501637e-01 -1.36350900e-01
2.86231250e-01 1.62655735e+00 -1.10586810e+00 -3.02428961e-01
-6.75033271e-01 9.76536930e-01 3.18635374e-01 2.04622716e-01
-8.03005934e-01 1.58986747e-01 8.36355507e-01 3.82150352e-01
5.06850779e-01 -4.59478907e-02 3.65287751e-01 -1.05093420e+00
-1.19834512e-01 -1.04703653e+00 5.20173430e-01 -5.36880851e-01
-1.48903656e+00 7.28408635e-01 -1.32157788e-01 -1.14306104e+00
-1.54694542e-01 -6.83449268e-01 -5.74705839e-01 7.68681318e-02
-1.11397314e+00 -1.15354753e+00 -3.69590670e-01 1.02114570e+00
5.12994170e-01 -6.30240858e-01 6.45380616e-01 7.60933578e-01
-9.49124873e-01 8.71666610e-01 2.48868138e-01 6.14622295e-01
9.57173347e-01 -8.24718714e-01 -2.09369421e-01 6.20814860e-01
5.80506384e-01 6.56002998e-01 4.55731869e-01 -3.02720070e-01
-1.55531299e+00 -1.03718615e+00 2.94087708e-01 -7.68331349e-01
4.92969751e-01 -6.16323054e-01 -7.32270360e-01 5.45091152e-01
5.57286680e-01 1.11565948e-01 9.76038516e-01 2.06063673e-01
-5.25038838e-01 -3.95550787e-01 -1.28001845e+00 2.88861752e-01
1.23572028e+00 -6.08865678e-01 -1.34318367e-01 4.83309805e-01
6.70895651e-02 1.54204130e-01 -6.11163020e-01 1.16105802e-01
5.37097991e-01 -1.16651475e+00 8.41214299e-01 -6.32772326e-01
3.58193815e-01 -4.50496644e-01 -2.56637424e-01 -8.24437559e-01
-2.75050730e-01 -5.51819623e-01 -3.67235616e-02 1.65830982e+00
1.12508640e-01 4.92853522e-02 1.22700989e+00 4.07651365e-01
3.38131875e-01 -1.90808982e-01 -7.49638200e-01 -6.70626283e-01
-1.96515188e-01 -1.50494784e-01 4.51692879e-01 1.33487689e+00
-3.01827490e-01 2.94383287e-01 -3.82371694e-01 1.82198182e-01
9.04237449e-01 -6.53512701e-02 8.94028664e-01 -1.48893559e+00
4.31925245e-02 -6.03426814e-01 -6.59331441e-01 -3.52267176e-01
6.23968422e-01 -9.61800039e-01 -2.16037855e-01 -1.01418269e+00
4.40979689e-01 -6.48289770e-02 -1.29367337e-01 3.51954550e-01
1.06855206e-01 6.04754746e-01 1.26742259e-01 2.86393315e-01
-9.99817729e-01 5.48623323e-01 5.52241862e-01 -2.80299991e-01
1.75348118e-01 -2.13630959e-01 -5.70663929e-01 9.80180025e-01
7.01768577e-01 -4.82115120e-01 -4.69182163e-01 -4.22005624e-01
-1.53342903e-01 -3.60668600e-01 2.50033915e-01 -1.39306557e+00
6.81560576e-01 2.89048672e-01 6.30725443e-01 -4.39287841e-01
4.10676032e-01 -9.37338173e-01 2.94844836e-01 1.28577814e-01
-4.95245196e-02 2.06914887e-01 4.47728373e-02 5.75808764e-01
-3.52363914e-01 -1.58846170e-01 9.26783383e-01 -2.54749864e-01
-7.11375892e-01 5.69461644e-01 -3.03259999e-01 1.31428111e-02
1.23410201e+00 -2.72943288e-01 1.25635356e-01 -4.68690664e-01
-8.97720397e-01 2.11542115e-01 6.68565750e-01 6.11661375e-01
3.66226196e-01 -1.37673020e+00 -6.08253241e-01 1.21924475e-01
4.90154400e-02 -5.24065793e-01 4.38670248e-01 6.15656495e-01
-5.23272276e-01 9.15116444e-02 -2.67287046e-01 -8.91722143e-01
-1.64459193e+00 7.80859590e-01 3.50224346e-01 2.96177506e-01
-8.85579810e-02 9.94076192e-01 2.76193738e-01 -4.64141399e-01
6.10527456e-01 2.55923748e-01 -5.92246771e-01 5.66852272e-01
7.13257134e-01 2.61564672e-01 -9.49315205e-02 -1.18382168e+00
-4.97502923e-01 9.42792833e-01 -2.41267219e-01 1.40193284e-01
1.53086734e+00 -4.52308059e-02 -3.94620359e-01 2.43490264e-01
1.36690581e+00 -1.67710453e-01 -1.22479510e+00 -9.68081802e-02
1.79859549e-01 -4.55790579e-01 -8.47264081e-02 -1.71593457e-01
-1.80557895e+00 8.97104442e-01 9.39223409e-01 1.11932337e-01
1.24365699e+00 7.30743334e-02 3.50364000e-01 3.92970890e-01
1.73194602e-01 -1.36827302e+00 3.49286675e-01 2.50647277e-01
5.18688500e-01 -1.48828161e+00 -1.04957782e-01 -4.38270330e-01
-7.08066761e-01 1.12763274e+00 7.22999156e-01 -1.55320331e-01
1.05476975e+00 2.41122007e-01 2.15521067e-01 -4.54328984e-01
-5.18406808e-01 -3.86840373e-01 2.03649804e-01 5.00380397e-01
5.77544451e-01 -3.11732113e-01 2.30666488e-01 4.33253795e-01
-1.22631319e-01 -1.16677478e-01 2.62212545e-01 6.51937604e-01
-1.69311523e-01 -1.32391155e+00 -5.39333463e-01 3.81038874e-01
-6.57087743e-01 1.49592325e-01 -7.76564181e-01 6.90531492e-01
3.46980274e-01 1.17148316e+00 2.22376481e-01 -3.49729687e-01
-1.03964180e-01 2.81635612e-01 7.15435088e-01 -5.66073954e-01
-4.56525922e-01 1.29984962e-02 -2.88566947e-01 -5.14292955e-01
-1.12557232e+00 -1.01647854e+00 -1.02200592e+00 -5.46199620e-01
-1.38943598e-01 1.30753741e-01 5.28971255e-01 7.98046589e-01
3.35909247e-01 1.69641897e-01 8.01391542e-01 -9.27331090e-01
-1.67206265e-02 -8.01038861e-01 -6.39414847e-01 8.17669332e-01
6.84886873e-02 -8.47509205e-01 -3.35405320e-01 5.18640935e-01] | [13.495491981506348, 1.0624017715454102] |
a5e134e3-4884-486a-9b28-4b5d00e7ecff | learning-the-unlearnable-adversarial | 2303.15127 | null | https://arxiv.org/abs/2303.15127v1 | https://arxiv.org/pdf/2303.15127v1.pdf | Learning the Unlearnable: Adversarial Augmentations Suppress Unlearnable Example Attacks | Unlearnable example attacks are data poisoning techniques that can be used to safeguard public data against unauthorized use for training deep learning models. These methods add stealthy perturbations to the original image, thereby making it difficult for deep learning models to learn from these training data effectively. Current research suggests that adversarial training can, to a certain degree, mitigate the impact of unlearnable example attacks, while common data augmentation methods are not effective against such poisons. Adversarial training, however, demands considerable computational resources and can result in non-trivial accuracy loss. In this paper, we introduce the UEraser method, which outperforms current defenses against different types of state-of-the-art unlearnable example attacks through a combination of effective data augmentation policies and loss-maximizing adversarial augmentations. In stark contrast to the current SOTA adversarial training methods, UEraser uses adversarial augmentations, which extends beyond the confines of $ \ell_p $ perturbation budget assumed by current unlearning attacks and defenses. It also helps to improve the model's generalization ability, thus protecting against accuracy loss. UEraser wipes out the unlearning effect with error-maximizing data augmentations, thus restoring trained model accuracies. Interestingly, UEraser-Lite, a fast variant without adversarial augmentations, is also highly effective in preserving clean accuracies. On challenging unlearnable CIFAR-10, CIFAR-100, SVHN, and ImageNet-subset datasets produced with various attacks, it achieves results that are comparable to those obtained during clean training. We also demonstrate its efficacy against possible adaptive attacks. Our code is open source and available to the deep learning community: https://github.com/lafeat/ueraser. | ['Cheng-Zhong Xu', 'Kejiang Ye', 'Juanjuan Zhao', 'Xitong Gao', 'Tianrui Qin'] | 2023-03-27 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 2.89814863e-02 1.01116158e-01 -1.33351803e-01 3.11684906e-02
-8.55691254e-01 -1.29691756e+00 5.87720513e-01 1.13810316e-01
-7.76252747e-01 8.31868112e-01 -2.04433240e-02 -7.22857058e-01
2.16541678e-01 -9.75789249e-01 -1.16050565e+00 -8.28709722e-01
-9.90243256e-02 5.58794737e-02 -4.88774665e-02 -3.71755421e-01
-6.84166253e-02 7.46624410e-01 -1.04269552e+00 1.30002901e-01
6.12517416e-01 5.28496504e-01 -6.62820697e-01 7.08640933e-01
3.05092722e-01 9.74277496e-01 -1.07044101e+00 -8.01315010e-01
7.04635561e-01 -5.40833212e-02 -8.04888546e-01 -5.35358787e-01
8.81829202e-01 -8.47570837e-01 -8.61014962e-01 1.41288865e+00
5.36179960e-01 -8.94109234e-02 2.38091826e-01 -1.62959957e+00
-9.21784282e-01 8.01310956e-01 -4.21740204e-01 3.50144088e-01
-1.00779399e-01 6.81072354e-01 6.36259258e-01 -3.53753775e-01
3.16534013e-01 1.27096355e+00 6.26297891e-01 1.23795378e+00
-1.20790851e+00 -1.52545762e+00 -2.44455077e-02 -1.06430076e-01
-1.07625449e+00 -5.01083434e-01 5.74229658e-01 -1.33018076e-01
6.75309181e-01 6.06751263e-01 2.10771322e-01 1.76890755e+00
9.41180512e-02 6.99689090e-01 1.21151412e+00 -1.87774956e-01
2.31532291e-01 1.38815701e-01 2.09704280e-01 6.32802665e-01
6.65209115e-01 5.34184575e-01 -2.06539139e-01 -7.93032944e-01
5.31269014e-01 1.12815574e-01 -5.78048587e-01 -1.55262694e-01
-5.38117468e-01 9.81909633e-01 7.32788861e-01 8.62155482e-02
-9.52086523e-02 5.79841077e-01 5.75690091e-01 4.94287401e-01
3.63545686e-01 8.40750217e-01 -5.82399309e-01 2.49504134e-01
-3.65937203e-01 4.53542471e-01 7.48730540e-01 5.79702139e-01
4.87906903e-01 5.14246762e-01 1.36068240e-02 3.26729953e-01
-1.06004514e-02 7.00553358e-01 3.89228165e-01 -8.82713556e-01
3.49808693e-01 5.81281222e-02 -4.62788194e-02 -7.95860589e-01
-1.98739041e-02 -5.64628422e-01 -7.38152683e-01 8.09752584e-01
5.80749929e-01 -5.41477799e-01 -1.02963066e+00 2.19098926e+00
3.01727295e-01 4.14499223e-01 3.68564874e-01 4.97493774e-01
4.60241675e-01 4.22838002e-01 4.78382647e-01 1.79964676e-01
1.29146898e+00 -6.15489841e-01 -4.73347217e-01 -1.30733609e-01
6.07152522e-01 -3.24618191e-01 1.18558133e+00 3.70528191e-01
-8.43809545e-01 7.87817966e-03 -1.14518976e+00 1.14725754e-01
-7.32228816e-01 -8.92144799e-01 6.44011617e-01 1.35629511e+00
-7.73137808e-01 5.17255783e-01 -9.27421033e-01 1.16311513e-01
1.15716875e+00 4.45607334e-01 -4.93105263e-01 -3.99876200e-02
-1.42761433e+00 9.09732997e-01 2.79814392e-01 -4.71988976e-01
-1.60647762e+00 -1.10561991e+00 -8.47053409e-01 8.23345780e-02
3.15554887e-01 -4.15678114e-01 1.14278555e+00 -9.90081966e-01
-7.46381819e-01 6.76294506e-01 5.47195554e-01 -1.16703522e+00
6.83052540e-01 -4.59966958e-01 -4.83227640e-01 1.71335906e-01
-2.57269114e-01 6.44754469e-01 1.04876637e+00 -1.55036628e+00
-1.02353379e-01 -2.56081760e-01 3.87755185e-01 -7.49365985e-02
-9.88336027e-01 7.03313127e-02 2.20962316e-01 -9.36193168e-01
-8.34824979e-01 -8.96309972e-01 -4.02172744e-01 -7.07821846e-02
-6.31394804e-01 2.51913905e-01 1.36479461e+00 -5.47225595e-01
9.01275873e-01 -2.14057493e+00 -4.62649733e-01 1.02430694e-01
4.63592023e-01 1.03619921e+00 -3.96326125e-01 2.05071583e-01
-2.01878235e-01 7.47027218e-01 -4.06405360e-01 -1.64000183e-01
-1.35519980e-02 2.97888219e-01 -8.66859734e-01 6.98379219e-01
-1.23827107e-01 9.02528584e-01 -8.29080164e-01 1.11694045e-01
4.48915809e-02 6.19555414e-01 -7.64762163e-01 5.63623197e-02
-2.90217698e-01 3.40171695e-01 -3.62537682e-01 5.50229788e-01
7.64396966e-01 1.78682998e-01 -1.79700673e-01 1.15606770e-01
4.97081608e-01 1.40385583e-01 -5.96763492e-01 1.01410556e+00
-2.22405866e-01 5.67036211e-01 1.23221248e-01 -6.21594667e-01
5.24162531e-01 3.40947539e-01 1.06139287e-01 -4.23826456e-01
3.57718080e-01 1.05811171e-01 1.66004840e-02 -2.14111105e-01
2.86342561e-01 -5.19060753e-02 -1.52777985e-01 5.17508030e-01
7.94970989e-02 1.48415998e-01 -3.64134490e-01 4.64716852e-01
1.56753528e+00 -3.44928741e-01 8.91434699e-02 -1.03675611e-01
2.05535099e-01 -3.67525257e-02 5.02720714e-01 1.09378588e+00
-4.51352984e-01 1.54085726e-01 2.74551004e-01 -4.73294288e-01
-1.22302890e+00 -1.11105025e+00 -1.39969200e-01 1.04212475e+00
4.81156483e-02 -2.48511046e-01 -1.17631292e+00 -1.40182459e+00
2.65209377e-01 1.08143973e+00 -7.42007911e-01 -6.65734589e-01
-5.08048475e-01 -8.26237261e-01 1.58196902e+00 3.66265327e-01
7.82152057e-01 -1.07040000e+00 -2.70557612e-01 -9.69977304e-02
8.61983970e-02 -9.81304049e-01 -4.81242836e-01 2.02818289e-01
-6.66573763e-01 -1.30915380e+00 -3.23444963e-01 -3.13034505e-01
6.94847822e-01 1.27139270e-01 8.89687598e-01 5.42097569e-01
-3.11672270e-01 5.94835818e-01 -2.67374039e-01 -9.25554514e-01
-9.01446164e-01 -2.68896502e-02 3.34450752e-01 -2.77174026e-01
4.49912310e-01 -6.80447936e-01 -3.96283299e-01 2.75376886e-02
-1.40504324e+00 -6.08781457e-01 2.29271263e-01 7.51524150e-01
2.42828742e-01 2.15231165e-01 7.94010997e-01 -1.25042653e+00
6.63310528e-01 -6.57151043e-01 -5.28053820e-01 5.65283969e-02
-4.78421837e-01 1.48797194e-02 9.79319572e-01 -9.80204463e-01
-6.28781259e-01 -3.67274970e-01 -5.19739091e-01 -9.09964681e-01
-3.45672429e-01 -1.92289770e-01 -3.61057222e-01 -6.15806222e-01
1.26088107e+00 1.58048183e-01 -1.26150399e-02 -4.95626420e-01
3.96834791e-01 4.63936836e-01 7.06790745e-01 -6.68196440e-01
1.45794964e+00 5.13352454e-01 -1.60723567e-01 -6.06619298e-01
-7.82865524e-01 1.56986833e-01 6.98814839e-02 1.93023935e-01
6.94690466e-01 -7.30067015e-01 -7.11214423e-01 8.72515202e-01
-8.25248837e-01 -5.46657205e-01 -4.46482688e-01 9.03744996e-02
-1.67367801e-01 3.23108941e-01 -8.41128469e-01 -6.12241268e-01
-6.71789944e-01 -9.82428193e-01 2.94545650e-01 1.09674945e-01
8.41101557e-02 -9.53090549e-01 -9.98447239e-02 4.78548467e-01
5.52106500e-01 7.63307989e-01 8.49672735e-01 -1.44287312e+00
-4.98961806e-01 -5.10688663e-01 2.21521243e-01 7.97583044e-01
-6.73912317e-02 -3.05613846e-01 -1.27455473e+00 -7.45546043e-01
2.59551406e-01 -7.61571705e-01 8.24590623e-01 -6.68016151e-02
1.54004705e+00 -1.20364225e+00 -4.35170531e-02 8.42793167e-01
1.41939986e+00 2.54127860e-01 8.85127485e-01 6.33520663e-01
8.27925622e-01 5.66238239e-02 1.41574621e-01 2.06893623e-01
-8.88318643e-02 1.16514951e-01 1.10732436e+00 -2.05553606e-01
5.64114675e-02 -3.57546031e-01 5.05497754e-01 -1.74621746e-01
2.63909101e-01 -3.54734004e-01 -8.25228333e-01 3.87795061e-01
-1.25837409e+00 -1.29880178e+00 1.08329527e-01 2.11918855e+00
1.09740841e+00 2.64164597e-01 8.18756297e-02 1.35288760e-01
4.81198341e-01 2.82556802e-01 -9.18516636e-01 -5.96313000e-01
-1.80814773e-01 5.29208302e-01 1.12703848e+00 5.15700877e-01
-1.29943776e+00 1.13275611e+00 5.85072041e+00 9.20166850e-01
-9.83038187e-01 4.77843404e-01 8.85018766e-01 -4.64574665e-01
-4.25342441e-01 -2.01935068e-01 -7.30601788e-01 3.71429771e-01
1.09589803e+00 -1.65122688e-01 7.64629126e-01 1.14557254e+00
-3.46214235e-01 5.27809978e-01 -8.54896724e-01 5.53465366e-01
-2.52005998e-02 -1.53595769e+00 1.98763981e-01 2.62974888e-01
5.65861523e-01 3.31847757e-01 4.44440186e-01 4.80530322e-01
1.15658772e+00 -1.32724035e+00 5.43237925e-01 -1.15108274e-01
8.18992198e-01 -1.33812487e+00 4.65850711e-01 4.91937369e-01
-4.49582726e-01 -3.03248197e-01 -3.52065295e-01 2.93384314e-01
-2.81265080e-01 1.52471572e-01 -6.31327868e-01 4.56605181e-02
7.62633204e-01 -4.98144180e-02 -6.86232388e-01 6.23133123e-01
-4.41769123e-01 1.10224283e+00 -2.09155798e-01 2.39644110e-01
4.93795216e-01 4.69641745e-01 8.06279421e-01 1.06963634e+00
-2.03603968e-01 2.28465483e-01 7.61143044e-02 7.61139929e-01
-8.23840976e-01 -3.47546756e-01 -9.67782438e-01 -9.49949622e-02
8.84182215e-01 9.52914953e-01 -1.19022004e-01 -2.61640966e-01
-1.77323003e-03 8.18414748e-01 3.92591834e-01 4.16157901e-01
-9.99291718e-01 -3.69653314e-01 1.24591684e+00 4.61040363e-02
-1.24750525e-01 -6.39616624e-02 -2.04823583e-01 -1.12886572e+00
-4.09017533e-01 -1.67167461e+00 5.11356950e-01 -3.30748439e-01
-1.39274609e+00 5.99006236e-01 -6.83883429e-02 -7.88741469e-01
1.22838959e-01 -4.88119513e-01 -6.43740058e-01 8.28858197e-01
-1.38832617e+00 -1.12300253e+00 1.58633530e-01 1.01929605e+00
2.65181243e-01 -3.81280959e-01 1.01660073e+00 9.09873694e-02
-5.57961643e-01 1.20745647e+00 7.80601948e-02 7.56155431e-01
5.75655997e-01 -1.09260285e+00 7.25146711e-01 1.27206731e+00
2.26481095e-01 8.93090546e-01 8.98426950e-01 -6.53265953e-01
-1.25883746e+00 -1.42493010e+00 1.27936378e-01 -6.64648652e-01
7.80497372e-01 -4.56996918e-01 -1.26716638e+00 1.05117071e+00
3.01638454e-01 2.65325487e-01 7.96608508e-01 -3.07104290e-01
-1.18371463e+00 1.19065642e-01 -1.87481606e+00 9.32685614e-01
7.93823004e-01 -6.39495850e-01 -4.04762954e-01 4.73403275e-01
1.15611815e+00 -3.82652342e-01 -7.17772961e-01 3.44662577e-01
2.37757280e-01 -6.75722301e-01 1.27615881e+00 -1.06919801e+00
1.51869178e-01 7.87016749e-02 -1.85376301e-01 -1.20942795e+00
-7.61277527e-02 -8.27886522e-01 -3.75215292e-01 1.29303491e+00
2.20661804e-01 -1.05584371e+00 7.73107052e-01 7.40851402e-01
2.13343024e-01 -4.55448985e-01 -7.48677611e-01 -9.23713207e-01
7.27191210e-01 -4.19111609e-01 8.94158721e-01 1.46958685e+00
-5.15957117e-01 -3.25193703e-01 -5.05178511e-01 6.03745818e-01
1.18309987e+00 -8.10366929e-01 9.77465034e-01 -6.63644195e-01
-3.79473388e-01 -2.69795716e-01 -2.89487958e-01 -2.84373313e-01
3.59137148e-01 -8.00062954e-01 -3.98885578e-01 -7.32605338e-01
-9.10953060e-02 -6.28367484e-01 -4.10487294e-01 1.21561086e+00
-2.60674775e-01 5.87945104e-01 4.34607178e-01 2.06138626e-01
-1.46930078e-02 2.50407070e-01 9.09548461e-01 -2.63458967e-01
6.82153851e-02 -5.26711568e-02 -1.20053470e+00 7.92344093e-01
1.24916863e+00 -9.65343237e-01 -5.21828651e-01 -3.98866326e-01
-1.37104809e-01 -5.23903847e-01 7.87610769e-01 -1.04277396e+00
-7.80813769e-02 -1.83124825e-01 3.07950616e-01 5.29168323e-02
1.14443608e-01 -8.60361516e-01 4.20276932e-02 8.90830576e-01
-5.52568138e-01 -4.17553559e-02 6.21909976e-01 5.58404863e-01
3.22860271e-01 -2.32652962e-01 1.29237688e+00 -2.68974066e-01
-3.14687431e-01 6.75271094e-01 -2.13427946e-01 3.54107976e-01
1.18103814e+00 2.64358163e-01 -1.03913522e+00 -2.66493291e-01
-5.13268173e-01 9.32950824e-02 7.69676626e-01 3.63251567e-01
5.48440516e-01 -1.11204445e+00 -6.96990490e-01 2.08048761e-01
-2.02633783e-01 -2.17763454e-01 1.57810509e-01 9.13334563e-02
-4.65801656e-01 -2.53080875e-01 -3.25356841e-01 6.15351088e-02
-1.56141436e+00 1.09506702e+00 6.63973331e-01 -2.85122901e-01
-6.44067764e-01 9.58847344e-01 2.59994298e-01 -4.49447125e-01
4.75574791e-01 2.98709214e-01 1.31025106e-01 -4.78067130e-01
9.96273875e-01 2.62853920e-01 5.76365879e-03 -4.40504640e-01
-2.40439102e-01 -3.12005699e-01 -4.73934621e-01 1.21907704e-02
1.21223342e+00 3.25191289e-01 1.13779798e-01 -3.56741726e-01
1.27312970e+00 3.99987817e-01 -1.29436231e+00 -1.91712916e-01
-4.06158835e-01 -4.75376725e-01 4.72286791e-02 -1.02834117e+00
-1.19867325e+00 7.24970996e-01 6.08871520e-01 3.17931682e-01
1.14024985e+00 -3.23039204e-01 1.05285597e+00 4.83796209e-01
3.59966636e-01 -4.18396205e-01 1.92491427e-01 2.79540062e-01
7.17572391e-01 -1.07828736e+00 -1.64005369e-01 2.78344639e-02
-5.87554216e-01 6.10552490e-01 8.35727215e-01 -4.18425173e-01
5.29896736e-01 7.26811230e-01 1.98032230e-01 1.75901279e-02
-6.19792044e-01 3.19725364e-01 -1.50480404e-01 8.58035862e-01
-2.32159227e-01 4.24813293e-03 1.68766767e-01 3.74791205e-01
-3.34471971e-01 -5.01521647e-01 6.70464516e-01 1.01848209e+00
-3.93275857e-01 -1.21674252e+00 -8.16823542e-01 3.02780479e-01
-1.26580155e+00 -2.54071206e-01 -5.18888950e-01 9.76774037e-01
8.86304975e-02 9.43102598e-01 -2.88770676e-01 -5.84967911e-01
6.92320392e-02 -2.86436137e-02 1.98498309e-01 -4.37848538e-01
-1.14376378e+00 -5.10184288e-01 2.28758305e-02 -5.54850161e-01
6.76692501e-02 -3.27644318e-01 -1.26631713e+00 -9.65875089e-01
-2.95249760e-01 7.39029497e-02 5.70213556e-01 5.98295212e-01
2.50659883e-01 3.47694427e-01 7.70681381e-01 -5.70822179e-01
-1.08622777e+00 -5.56339800e-01 -3.58031899e-01 6.81570828e-01
6.54156268e-01 -2.52596229e-01 -6.31119370e-01 -1.60334945e-01] | [5.716926097869873, 7.777214050292969] |
f6611886-13a6-471e-a81f-285d977d29a5 | croano-a-crowd-annotation-platform-for | null | null | https://aclanthology.org/2021.emnlp-demo.32 | https://aclanthology.org/2021.emnlp-demo.32.pdf | CroAno : A Crowd Annotation Platform for Improving Label Consistency of Chinese NER Dataset | In this paper, we introduce CroAno, a web-based crowd annotation platform for the Chinese named entity recognition (NER). Besides some basic features for crowd annotation like fast tagging and data management, CroAno provides a systematic solution for improving label consistency of Chinese NER dataset. 1) Disagreement Adjudicator: CroAno uses a multi-dimensional highlight mode to visualize instance-level inconsistent entities and makes the revision process user-friendly. 2) Inconsistency Detector: CroAno employs a detector to locate corpus-level label inconsistency and provides users an interface to correct inconsistent entities in batches. 3) Prediction Error Analyzer: We deconstruct the entity prediction error of the model to six fine-grained entity error types. Users can employ this error system to detect corpus-level inconsistency from a model perspective. To validate the effectiveness of our platform, we use CroAno to revise two public datasets. In the two revised datasets, we get an improvement of +1.96% and +2.57% F1 respectively in model performance. | ['Yafei Shi', 'Shengping Liu', 'Jun Zhao', 'Kang Liu', 'Jing Wan', 'Yubo Chen', 'Zhen Gan', 'Zhucong Li', 'Baoli Zhang'] | null | null | null | null | emnlp-acl-2021-11 | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-5.74996352e-01 3.60492207e-02 1.69251904e-01 -4.23474848e-01
-8.86950850e-01 -8.19634497e-01 1.42207086e-01 4.55343395e-01
-7.45447040e-01 8.18873227e-01 1.57317594e-01 -1.19045310e-01
3.70466173e-01 -6.75652623e-01 -2.72856712e-01 -3.51452351e-01
4.12159950e-01 4.49368179e-01 5.37247062e-01 -7.15225637e-02
4.37994450e-01 6.40108660e-02 -1.26047838e+00 5.46965480e-01
1.32948554e+00 7.01415718e-01 9.49217752e-02 5.43677986e-01
-6.02711737e-01 8.74336302e-01 -1.09650970e+00 -6.41640246e-01
-1.23916283e-01 -1.52285919e-01 -1.12213039e+00 -3.41021270e-01
1.20417237e-01 3.58856976e-01 4.42037344e-01 1.24697411e+00
6.59016371e-01 -1.29596487e-01 3.63139778e-01 -1.14634061e+00
-9.04364586e-01 7.80472517e-01 -2.73906499e-01 1.51851729e-01
4.42313254e-01 -1.10206552e-01 8.04821789e-01 -1.19517994e+00
1.01310360e+00 7.73129702e-01 1.06647444e+00 5.99901855e-01
-6.34035587e-01 -8.07401896e-01 -4.25788574e-02 1.65151611e-01
-1.58063281e+00 -4.04803693e-01 1.97189018e-01 -6.13302767e-01
9.85061944e-01 4.42966938e-01 6.81784302e-02 5.79838693e-01
5.07524125e-02 4.01380569e-01 1.05118179e+00 -5.43271363e-01
4.30627704e-01 3.14580202e-01 4.99870151e-01 5.90679228e-01
5.93977332e-01 -8.48375738e-01 -4.51434702e-01 -2.91581094e-01
-3.86632904e-02 -1.16758272e-01 1.18022365e-02 3.53131622e-01
-9.26131487e-01 4.25303966e-01 1.77219942e-01 6.15813911e-01
-3.48205566e-01 -3.05867940e-01 5.52808344e-01 6.59169853e-02
5.44882178e-01 6.63568974e-01 -7.81590223e-01 -2.28010073e-01
-5.29797256e-01 6.08463548e-02 8.46593440e-01 1.25961721e+00
6.11893058e-01 -2.46228859e-01 -4.34623778e-01 9.56159592e-01
4.09732789e-01 4.57830817e-01 7.15616703e-01 -7.60496855e-01
6.65293515e-01 1.03882182e+00 5.84255755e-01 -1.04005039e+00
-7.02511549e-01 -1.46879703e-01 -6.96238637e-01 -2.30606928e-01
4.24406528e-01 -5.41362524e-01 -4.89542276e-01 1.53206158e+00
4.67361271e-01 -1.39870688e-01 1.76670372e-01 7.76620865e-01
1.23498011e+00 3.94124627e-01 6.41047180e-01 -2.00042292e-01
1.75178540e+00 -9.49224591e-01 -1.07300258e+00 2.45600849e-01
1.50028169e+00 -8.89423609e-01 1.06603861e+00 8.81105810e-02
-7.86255002e-01 -3.27529758e-01 -6.62268341e-01 -1.68127790e-01
-7.71572948e-01 5.66017926e-01 3.12783748e-01 7.42865562e-01
-7.08072364e-01 3.10905665e-01 -6.01041496e-01 -5.33002377e-01
7.56395087e-02 1.52783245e-01 -5.84734797e-01 2.75947422e-01
-1.40155768e+00 8.54360700e-01 4.60415989e-01 1.36868462e-01
2.09203079e-01 -6.62614107e-01 -8.11283052e-01 1.51269704e-01
3.35057497e-01 -2.12115183e-01 1.38473451e+00 -6.81002676e-01
-1.11134136e+00 9.34609830e-01 -4.64952230e-01 2.07256209e-02
3.44943196e-01 -2.63474375e-01 -8.98527443e-01 -2.73211062e-01
6.55372977e-01 2.30674595e-01 -1.37888297e-01 -9.88526106e-01
-1.02393723e+00 -3.04425985e-01 -3.69647145e-01 6.38662353e-02
-2.10242867e-01 5.53299487e-01 -5.94414175e-01 -5.99959373e-01
-1.06777161e-01 -8.39468360e-01 -1.61384895e-01 -5.48618853e-01
-5.07097602e-01 -6.16674781e-01 5.39495587e-01 -9.75910962e-01
1.89946330e+00 -1.90201652e+00 -6.14243269e-01 1.38588116e-01
3.67142439e-01 5.56340635e-01 1.54298944e-02 4.51481909e-01
1.11078605e-01 7.68742561e-01 -1.99356437e-01 -4.39519167e-01
-5.03153838e-02 -2.91671902e-01 2.88919240e-01 9.29621533e-02
2.19288856e-01 7.19354928e-01 -9.95077372e-01 -6.90657496e-01
-3.74023557e-01 9.13444906e-02 -3.56139332e-01 5.78817241e-02
2.56975517e-02 3.20718735e-01 -4.16839570e-01 7.19651580e-01
1.01756382e+00 -3.05796951e-01 5.20427883e-01 -1.04609042e-01
-5.67344844e-01 2.19874009e-01 -1.48525822e+00 1.42521453e+00
-2.79693484e-01 5.71962178e-01 -1.37733787e-01 -1.44285530e-01
1.16177452e+00 4.54076797e-01 1.12642758e-01 -8.56196105e-01
-5.69337495e-02 3.51127774e-01 -4.49252188e-01 -9.27333891e-01
1.12410545e+00 3.17337036e-01 -4.29760993e-01 2.88213253e-01
-1.40488431e-01 5.82990646e-01 4.28161681e-01 3.34999949e-01
1.26580083e+00 -9.09016952e-02 4.60778415e-01 -4.16375965e-01
6.09390795e-01 4.34947699e-01 1.12483788e+00 6.91535294e-01
-3.85688901e-01 4.67535049e-01 4.26165491e-01 -5.11354983e-01
-8.16342354e-01 -2.65712947e-01 -3.39980610e-02 1.12853479e+00
1.40311301e-01 -8.06481481e-01 -9.07639980e-01 -1.09737444e+00
-2.48708829e-01 6.77744210e-01 -4.62993354e-01 5.05676031e-01
-4.68211591e-01 -7.21710265e-01 9.98593748e-01 5.09414911e-01
7.40873933e-01 -1.00129545e+00 -3.00215304e-01 2.86098644e-02
-5.08841693e-01 -1.05234289e+00 -5.57218432e-01 -2.35417280e-02
-3.92698586e-01 -1.26333594e+00 -3.23506594e-01 -1.02821374e+00
7.71248817e-01 -8.42200499e-03 1.20722806e+00 4.09092277e-01
-3.78666408e-02 4.29149112e-03 -5.58874607e-01 -4.34949011e-01
-3.76957536e-01 2.93302923e-01 1.82134584e-02 -3.92863184e-01
7.32076764e-01 1.26216590e-01 -3.26246738e-01 6.61295116e-01
-6.84114218e-01 -2.57217064e-02 -3.24377827e-02 6.08356714e-01
4.11951482e-01 -5.91770299e-02 8.08752894e-01 -1.52197945e+00
9.00776684e-01 -6.48003161e-01 -7.65331328e-01 6.56454027e-01
-8.49195063e-01 7.17970133e-02 7.02856302e-01 -6.86712265e-02
-1.35766494e+00 1.62762076e-01 -1.88303292e-01 3.28856707e-01
-4.44287509e-01 7.05002725e-01 -2.63128817e-01 4.24543291e-01
7.94959664e-01 -2.56037056e-01 -7.55058825e-01 -8.89710546e-01
1.02096520e-01 1.22846973e+00 5.25532186e-01 -3.38163972e-01
1.98096395e-01 -2.33771279e-02 -5.80717325e-01 -3.57985139e-01
-7.01892257e-01 -7.05496967e-01 -7.38904417e-01 -9.36686844e-02
1.08427978e+00 -1.12435269e+00 -1.00389338e+00 5.02382994e-01
-1.43939447e+00 -1.35673001e-01 7.25296140e-02 2.23582029e-01
3.58244181e-01 3.27765256e-01 -6.95475399e-01 -7.95502365e-01
-6.84315562e-01 -7.47482300e-01 8.31592977e-01 7.44259953e-01
-2.16027573e-01 -8.61137092e-01 3.44379395e-01 3.66663396e-01
3.14936936e-01 6.64671883e-02 4.40140367e-01 -1.18810987e+00
8.29917714e-02 -1.92303509e-01 -1.80606976e-01 -1.64545357e-01
-1.18937790e-01 2.81217963e-01 -9.84371245e-01 2.17970297e-01
-4.86563176e-01 3.94847542e-02 1.94224477e-01 -2.42665857e-01
9.01069641e-01 -2.38900408e-01 -3.75785500e-01 7.07033873e-02
1.30062187e+00 4.20258164e-01 6.45898163e-01 6.49515331e-01
5.94777763e-01 4.40500796e-01 7.49903500e-01 4.01329994e-01
9.48962569e-01 7.07383931e-01 -1.02194756e-01 4.75417972e-02
1.98562533e-01 -2.78329492e-01 -2.97819600e-02 1.15411711e+00
-6.07919656e-02 -4.49246436e-01 -1.32424748e+00 3.94613504e-01
-2.11788821e+00 -9.11185861e-01 -9.57731247e-01 1.84842908e+00
8.75521541e-01 -3.45528454e-01 -7.15314597e-02 -2.23289162e-01
1.27368152e+00 -3.44412744e-01 5.43519519e-02 -3.34115773e-01
-1.86989993e-01 -1.93840474e-01 5.67785501e-01 4.03178900e-01
-1.07274079e+00 1.10717666e+00 5.72482967e+00 7.11822510e-01
-1.03511870e+00 4.43665653e-01 5.60771525e-01 4.97644693e-01
-1.68290257e-01 5.76610379e-02 -1.27212059e+00 9.37044859e-01
8.15080523e-01 1.69540234e-02 -2.45476037e-01 9.43318188e-01
2.94212848e-01 -3.31542343e-01 -6.73715770e-01 9.71116483e-01
-1.52287111e-01 -1.66242909e+00 -4.21501845e-01 -2.17951402e-01
8.38061333e-01 2.87113916e-02 -7.09276557e-01 6.21629536e-01
4.91348386e-01 -7.48106301e-01 8.17303896e-01 8.43287170e-01
6.21745646e-01 -5.65207899e-01 1.32364023e+00 4.18394715e-01
-1.20929968e+00 1.84400395e-01 -4.58011180e-01 -1.61051787e-02
1.42538816e-01 5.92587471e-01 -9.35655475e-01 7.20776975e-01
9.13332283e-01 2.73285359e-01 -9.44087565e-01 1.22197676e+00
-1.28623977e-01 6.74149156e-01 -1.95220754e-01 -2.61745751e-01
-1.99960306e-01 1.91682559e-02 1.68547750e-01 1.70565772e+00
2.93762147e-01 2.87477225e-01 -1.32340282e-01 5.14997423e-01
-5.44704378e-01 5.34795761e-01 -1.65191323e-01 2.79335856e-01
1.14175129e+00 1.51340389e+00 -6.55846596e-01 -5.07494926e-01
-2.75502682e-01 9.52162504e-01 6.40111268e-01 6.80285245e-02
-7.37989724e-01 -9.08036888e-01 1.58750474e-01 -1.77063897e-01
2.24893153e-01 1.28019997e-03 -4.55283850e-01 -1.34362376e+00
2.33911484e-01 -8.28724146e-01 6.01172209e-01 -9.19177115e-01
-1.10938120e+00 9.41875875e-01 -5.31166852e-01 -1.02823782e+00
4.39624228e-02 -2.70647138e-01 -8.46987247e-01 8.53572130e-01
-1.09002376e+00 -9.32246208e-01 -5.96365631e-01 2.36836866e-01
4.65648383e-01 3.81484488e-03 1.02423203e+00 7.16495097e-01
-9.97367859e-01 8.87417376e-01 -1.58950165e-02 5.64151704e-01
1.06934416e+00 -1.35039747e+00 4.73456532e-01 8.96049559e-01
-1.33494616e-01 7.73051143e-01 3.61435562e-01 -1.05039763e+00
-8.00950289e-01 -1.14945614e+00 1.92303526e+00 -8.23899031e-01
3.66155773e-01 -3.74491513e-01 -9.65118408e-01 4.65030938e-01
8.23849067e-02 2.07912356e-01 9.55929697e-01 2.86193013e-01
-2.83839792e-01 1.14101440e-01 -1.36713493e+00 3.51892203e-01
7.39734352e-01 -3.47060800e-01 -4.25156772e-01 1.76185176e-01
6.30348265e-01 -5.02723753e-01 -1.14266515e+00 1.40278757e-01
3.60545129e-01 -6.62402868e-01 -3.47727141e-03 -7.80605793e-01
6.27370775e-02 -8.96829247e-01 4.39754911e-02 -9.95699942e-01
-3.96724015e-01 -5.26632249e-01 3.22154403e-01 1.96197855e+00
9.47065055e-01 -4.35043812e-01 4.51083452e-01 1.20078754e+00
-1.69456393e-01 -3.12047094e-01 -8.06210339e-01 -4.88904566e-01
-1.85479477e-01 -4.03914303e-01 9.32995737e-01 1.40620422e+00
4.57025617e-01 3.59879702e-01 -3.52646887e-01 4.48068559e-01
-8.38012025e-02 -4.14717734e-01 7.00035810e-01 -1.27499127e+00
2.20302358e-01 -1.32340025e-02 2.11922638e-02 -4.65702027e-01
-1.49715737e-01 -7.48715043e-01 1.64630994e-01 -1.46145606e+00
2.66791105e-01 -9.32115436e-01 -5.28375357e-02 6.51707470e-01
-5.06464839e-01 2.18381852e-01 2.90926516e-01 6.37576938e-01
-1.41823006e+00 3.04562016e-03 4.55086917e-01 3.55446875e-01
-3.51162612e-01 -2.47889385e-01 -6.54035807e-01 6.51811123e-01
8.82661104e-01 -9.58278954e-01 4.49258566e-01 -6.96089387e-01
8.11663091e-01 -3.33026260e-01 -2.20149264e-01 -8.34443927e-01
6.19783819e-01 -6.93615973e-02 3.09643686e-01 -4.33512002e-01
-6.40045047e-01 -6.75441027e-01 3.92450809e-01 2.24695936e-01
-3.19306254e-01 5.61660826e-01 1.89082902e-02 3.22877288e-01
-1.01432927e-01 -6.80918038e-01 5.09918094e-01 1.65088419e-02
-8.45401883e-01 -1.72972471e-01 -6.12911284e-01 2.81609863e-01
8.11238825e-01 1.25466734e-01 -8.28979254e-01 -2.70899311e-02
-6.12799585e-01 4.29652572e-01 5.87133884e-01 3.26539397e-01
-1.08430915e-01 -1.27673781e+00 -4.70748842e-01 -1.32321775e-01
4.75902736e-01 -9.03016403e-02 2.87353307e-01 7.78164983e-01
-8.12454462e-01 3.54590148e-01 1.71940085e-02 -3.41973543e-01
-1.40289068e+00 2.22668171e-01 2.40194410e-01 -5.43250442e-01
-1.66440472e-01 5.78116834e-01 -4.85487521e-01 -9.30111468e-01
5.71635105e-02 -2.39236921e-01 -5.59668005e-01 1.75549872e-02
7.80439258e-01 7.17300117e-01 3.80870134e-01 -6.61427796e-01
-7.58643746e-01 1.75614223e-01 1.49826288e-01 7.24624330e-03
1.11089563e+00 -3.66135538e-01 -2.01596141e-01 2.79348433e-01
7.34815121e-01 6.86030626e-01 -6.49083912e-01 -4.25761938e-02
6.94952726e-01 -1.19662642e-01 -4.24803585e-01 -1.19510424e+00
-7.29309082e-01 2.67574817e-01 6.62641406e-01 4.52885777e-01
6.94871426e-01 -1.43436879e-01 4.54906791e-01 5.73722720e-01
2.12618515e-01 -1.54191041e+00 -6.13036454e-01 9.48331177e-01
5.27148247e-01 -1.49177241e+00 -2.58560836e-01 -4.50295955e-01
-1.01505554e+00 1.01394951e+00 8.50851715e-01 4.01321352e-01
5.82151294e-01 3.10959518e-01 4.14044231e-01 -3.13879550e-01
-7.28846371e-01 -6.92690313e-02 2.58221328e-01 5.50880849e-01
8.18314910e-01 1.92304224e-01 -9.30280268e-01 1.42609894e+00
4.09977064e-02 1.78873986e-01 4.14981008e-01 8.66940141e-01
-4.26705867e-01 -1.03048491e+00 -4.52942550e-01 1.36697009e-01
-7.29590118e-01 -1.02014579e-01 -3.90110046e-01 6.47868574e-01
4.54425484e-01 1.40671504e+00 1.50453746e-01 -3.91593337e-01
5.62285483e-01 2.56415784e-01 -5.54498792e-01 -6.39995515e-01
-1.22801805e+00 -2.77905017e-01 5.94603837e-01 -4.61261779e-01
-3.50524604e-01 -3.66950274e-01 -1.61791575e+00 -2.90227473e-01
-7.08757102e-01 7.06576884e-01 9.16100323e-01 9.52256858e-01
1.03358924e+00 1.45387962e-01 4.45574731e-01 -1.28996506e-01
3.83150019e-02 -1.15627110e+00 -3.17095160e-01 5.63783526e-01
-3.52899164e-01 -2.96750993e-01 -4.82809357e-02 1.56012729e-01] | [9.604256629943848, 9.4447021484375] |
334e610c-c888-40fa-b7f4-4cabbb122f08 | combining-label-propagation-and-simple-models-1 | 2010.13993 | null | https://arxiv.org/abs/2010.13993v2 | https://arxiv.org/pdf/2010.13993v2.pdf | Combining Label Propagation and Simple Models Out-performs Graph Neural Networks | Graph Neural Networks (GNNs) are the predominant technique for learning over graphs. However, there is relatively little understanding of why GNNs are successful in practice and whether they are necessary for good performance. Here, we show that for many standard transductive node classification benchmarks, we can exceed or match the performance of state-of-the-art GNNs by combining shallow models that ignore the graph structure with two simple post-processing steps that exploit correlation in the label structure: (i) an "error correlation" that spreads residual errors in training data to correct errors in test data and (ii) a "prediction correlation" that smooths the predictions on the test data. We call this overall procedure Correct and Smooth (C&S), and the post-processing steps are implemented via simple modifications to standard label propagation techniques from early graph-based semi-supervised learning methods. Our approach exceeds or nearly matches the performance of state-of-the-art GNNs on a wide variety of benchmarks, with just a small fraction of the parameters and orders of magnitude faster runtime. For instance, we exceed the best known GNN performance on the OGB-Products dataset with 137 times fewer parameters and greater than 100 times less training time. The performance of our methods highlights how directly incorporating label information into the learning algorithm (as was done in traditional techniques) yields easy and substantial performance gains. We can also incorporate our techniques into big GNN models, providing modest gains. Our code for the OGB results is at https://github.com/Chillee/CorrectAndSmooth. | ['Austin R. Benson', 'Ser-Nam Lim', 'Abhay Singh', 'Horace He', 'Qian Huang'] | 2020-10-27 | combining-label-propagation-and-simple-models | https://openreview.net/forum?id=8E1-f3VhX1o | https://openreview.net/pdf?id=8E1-f3VhX1o | iclr-2021-1 | ['node-classification-on-non-homophilic'] | ['graphs'] | [ 2.46255621e-01 4.13322598e-01 -4.73936856e-01 -5.68233609e-01
-5.94294429e-01 -6.03669345e-01 5.65164566e-01 5.48698962e-01
-2.97167003e-01 7.71672308e-01 -9.33016539e-02 -8.32957864e-01
-1.94459215e-01 -1.00043213e+00 -9.41422641e-01 -5.56364298e-01
-3.51225525e-01 6.46935165e-01 5.37836611e-01 -2.87365377e-01
-8.23912099e-02 3.45091164e-01 -1.04072988e+00 8.45679827e-03
6.32474184e-01 7.77725279e-01 -4.71476465e-01 9.71911311e-01
-2.13994935e-01 1.14895213e+00 -1.18631765e-01 -5.09339035e-01
2.81190962e-01 -3.41630995e-01 -1.10505879e+00 -1.59882918e-01
5.47593594e-01 -6.00748062e-02 -3.63254815e-01 9.96029615e-01
2.43539497e-01 1.12541720e-01 5.97245395e-01 -1.33192420e+00
-5.79282165e-01 1.11615336e+00 -6.36215389e-01 4.98901904e-02
-4.57614586e-02 4.33668382e-02 1.14845920e+00 -4.84237194e-01
5.88326275e-01 9.88069534e-01 1.21244824e+00 4.60032433e-01
-1.62354982e+00 -5.57782352e-01 1.86515838e-01 -2.27798209e-01
-1.20295846e+00 -3.89893383e-01 5.55558145e-01 -3.63167316e-01
1.14598918e+00 1.28522202e-01 6.51094794e-01 7.60206640e-01
6.35149330e-02 6.97437406e-01 9.45783138e-01 -5.08578718e-01
1.38025805e-01 2.39133388e-01 5.87538362e-01 1.12364721e+00
3.73725653e-01 1.75413579e-01 -2.41884872e-01 -1.53233707e-01
4.75027472e-01 5.90907447e-02 -1.25251397e-01 -4.82254267e-01
-8.12328696e-01 1.05900395e+00 9.18898404e-01 1.73188061e-01
-2.74221282e-02 6.24620140e-01 4.88817573e-01 4.48646635e-01
7.99594998e-01 3.31894636e-01 -4.80982929e-01 8.69621709e-02
-1.05058920e+00 -9.36203003e-02 1.25752223e+00 8.78723502e-01
1.10016227e+00 1.47125140e-01 1.67817652e-01 7.92370558e-01
3.89159083e-01 4.83335629e-02 -7.88840652e-03 -8.42215061e-01
1.90415606e-01 6.49245918e-01 -4.82390165e-01 -8.37094367e-01
-6.84915364e-01 -7.76298523e-01 -7.14223027e-01 2.47683540e-01
7.72198796e-01 -3.77792835e-01 -1.24793291e+00 1.76769567e+00
3.04505289e-01 8.07355121e-02 -2.81092614e-01 4.44159746e-01
7.93975830e-01 5.89064002e-01 1.47166505e-01 6.66627213e-02
9.19380307e-01 -1.18507993e+00 -8.93269032e-02 -5.31473398e-01
1.34969771e+00 -4.34085369e-01 7.79854596e-01 2.94104785e-01
-9.27004576e-01 -1.45369291e-01 -1.15157163e+00 -1.29729196e-01
-6.05285883e-01 -3.44864368e-01 1.20683765e+00 7.79065549e-01
-1.56537962e+00 1.15563214e+00 -7.33729959e-01 -6.52009010e-01
4.93160248e-01 5.97444057e-01 -3.47523779e-01 -2.69070297e-01
-9.84663963e-01 7.66572475e-01 4.16520447e-01 -9.88668650e-02
-8.57870042e-01 -7.71641552e-01 -8.82764637e-01 -7.19672516e-02
5.31402051e-01 -5.48377335e-01 1.21072209e+00 -1.06397319e+00
-1.18013191e+00 8.15313995e-01 1.38825968e-01 -5.00667155e-01
3.44030023e-01 1.91315621e-01 -1.36342809e-01 -2.62660474e-01
-2.56304741e-01 7.61206806e-01 3.82445395e-01 -1.24489522e+00
-3.86910677e-01 -2.08993316e-01 1.16179958e-01 7.58968592e-02
-3.36649954e-01 -2.29138076e-01 -4.78606701e-01 -3.25888097e-01
1.11487754e-01 -9.76111710e-01 -2.96483457e-01 -8.71775374e-02
-5.21961570e-01 -3.58162433e-01 4.29715514e-01 -5.13489783e-01
1.14893770e+00 -1.80173337e+00 -1.39265269e-01 6.07676268e-01
6.56385064e-01 2.79263258e-01 -2.75723130e-01 5.35760820e-01
-1.59625515e-01 3.41703236e-01 -2.93575794e-01 -3.50982159e-01
4.08873260e-02 2.47405395e-01 1.05815232e-01 5.22450805e-01
2.33780127e-02 1.12379885e+00 -1.09236753e+00 -3.16550553e-01
-1.05019405e-01 3.05747658e-01 -4.58909810e-01 -1.90390423e-01
-3.19729090e-01 -8.00736770e-02 -4.71041650e-02 7.10536659e-01
3.55567455e-01 -8.01825047e-01 4.63544011e-01 7.65675306e-02
3.41961980e-01 7.05640912e-01 -9.48531151e-01 1.31649184e+00
-1.67254791e-01 6.15386128e-01 1.18742272e-01 -1.22339487e+00
8.28038335e-01 -9.29097906e-02 2.50242770e-01 -3.88167858e-01
1.34552121e-01 2.92780876e-01 2.04176798e-01 1.30793706e-01
2.54081607e-01 -8.79653320e-02 2.58097351e-01 7.22789645e-01
2.46669292e-01 -2.50663459e-02 4.07169491e-01 5.73306918e-01
1.59556448e+00 8.32260400e-02 2.08599254e-01 -3.66273731e-01
2.31032018e-02 1.89126655e-01 3.78261238e-01 9.72277403e-01
-1.27701774e-01 2.17337921e-01 8.07883918e-01 -4.34428513e-01
-9.81590450e-01 -7.43792474e-01 1.73751518e-01 1.44735444e+00
-2.14545071e-01 -6.10589087e-01 -7.04127908e-01 -1.00728250e+00
1.72402889e-01 5.74462414e-01 -7.50338495e-01 -1.84034050e-01
-3.81945282e-01 -1.02433908e+00 7.70661831e-01 6.84176862e-01
1.73084944e-01 -9.24987435e-01 5.48416495e-01 1.07036330e-01
4.82269853e-01 -7.61046469e-01 -1.88259259e-01 7.54054666e-01
-1.15100706e+00 -9.24610078e-01 -3.12476635e-01 -9.02022183e-01
8.40583026e-01 9.89232138e-02 1.39744103e+00 6.39554322e-01
-5.79394698e-02 8.12409818e-02 -1.53529167e-01 -2.59507120e-01
-6.42651737e-01 5.40808380e-01 -2.70623773e-01 -4.89269257e-01
3.72835159e-01 -7.58659363e-01 -2.81638384e-01 8.96706879e-02
-5.50738037e-01 1.18744135e-01 5.30624986e-01 8.18328500e-01
3.03973526e-01 3.41285802e-02 5.30153275e-01 -1.63966703e+00
5.27158201e-01 -6.81835115e-01 -5.36932647e-01 1.82349473e-01
-1.35448551e+00 2.03343660e-01 5.50562918e-01 -2.74097949e-01
-3.84274304e-01 -1.09117664e-01 -2.49239504e-01 -4.47014421e-02
7.40739238e-03 7.39774406e-01 4.33744699e-01 -4.93513316e-01
9.76041436e-01 -2.04428777e-01 1.26750216e-01 -2.42770210e-01
3.13717127e-01 4.06757295e-01 1.38335258e-01 -5.84928930e-01
8.37700367e-01 2.77427852e-01 2.55512565e-01 -3.64172429e-01
-9.28598583e-01 -3.89120370e-01 -5.02074420e-01 -2.07517788e-01
4.50042963e-01 -6.87507749e-01 -6.34923577e-01 5.27607739e-01
-7.27739811e-01 -1.04333627e+00 -2.70076245e-01 2.61460692e-01
-2.18542919e-01 2.09292442e-01 -1.22439969e+00 -5.27065992e-01
-4.77653235e-01 -8.35722864e-01 5.56060255e-01 1.99517682e-01
-1.03257913e-02 -1.64600253e+00 1.52479768e-01 2.34655127e-01
5.22576392e-01 2.34486647e-02 1.06060243e+00 -1.01542914e+00
-4.99318361e-01 -2.99718320e-01 -5.03736615e-01 5.26069999e-01
1.21753654e-02 1.22518539e-01 -9.10438657e-01 -4.06889737e-01
-6.44152701e-01 -5.11634767e-01 1.17932320e+00 2.95878738e-01
8.39822114e-01 -9.17661339e-02 -5.87247312e-01 6.60753965e-01
1.48115695e+00 -2.22141027e-01 4.91739064e-01 1.67401388e-01
1.16840088e+00 4.29165393e-01 1.66492209e-01 -2.41651282e-01
6.07115030e-01 2.20157668e-01 3.10330659e-01 -4.59989518e-01
-2.57354915e-01 -4.32884037e-01 3.61802071e-01 9.82422233e-01
2.73959581e-02 -3.95880908e-01 -1.22430110e+00 5.87109864e-01
-2.00892735e+00 -5.70771754e-01 -4.44189936e-01 2.09750628e+00
9.57819641e-01 5.74541450e-01 1.67695254e-01 1.58983797e-01
6.23682737e-01 2.39020035e-01 -5.68907022e-01 -6.17026806e-01
3.04741889e-01 3.16056281e-01 1.06386030e+00 8.19466114e-01
-1.04104745e+00 1.11685967e+00 7.03804731e+00 9.74487424e-01
-1.00848389e+00 1.15592018e-01 7.47895956e-01 1.35038331e-01
-4.11849231e-01 4.74562287e-01 -8.76457512e-01 4.32851762e-02
1.29496431e+00 7.30317831e-02 8.12664628e-01 9.29782987e-01
-2.94875383e-01 -1.13938116e-02 -1.27313721e+00 4.25450027e-01
2.36113928e-03 -1.43231416e+00 -2.91240185e-01 2.01717630e-01
7.83612013e-01 6.82069361e-01 -4.15088713e-01 6.29206538e-01
1.14345932e+00 -1.11592591e+00 3.93082529e-01 2.36833125e-01
6.74163461e-01 -5.09949267e-01 6.90854967e-01 3.94780636e-01
-1.02848530e+00 1.33109801e-02 -4.21290159e-01 -1.89986661e-01
-1.20682023e-01 9.20066357e-01 -1.24958551e+00 3.56871039e-01
4.69490796e-01 8.72510672e-01 -8.71313393e-01 8.99701357e-01
-3.85378957e-01 1.20625567e+00 -5.39267004e-01 -3.49606305e-01
3.93616676e-01 -1.13806821e-01 2.75722325e-01 1.41281152e+00
8.54038899e-06 -2.59846956e-01 2.37755299e-01 6.66527629e-01
-4.49635595e-01 7.91055560e-02 -6.36925280e-01 -2.61592269e-01
3.62235427e-01 1.52503133e+00 -1.00493717e+00 -4.42458630e-01
-4.11127031e-01 4.73718286e-01 8.66039634e-01 3.92434716e-01
-6.65166020e-01 -4.92690206e-01 1.15086570e-01 3.15860659e-01
2.87651896e-01 -6.75180703e-02 -4.00248349e-01 -8.27146053e-01
-3.07784677e-01 -8.42604458e-01 6.03165686e-01 -6.45447612e-01
-1.41430199e+00 4.09680367e-01 -3.04216862e-01 -5.35867691e-01
-1.78565279e-01 -7.80770779e-01 -6.04488075e-01 6.19438708e-01
-1.46422720e+00 -1.21645069e+00 -3.65322471e-01 3.85217279e-01
-1.20571656e-02 1.44964248e-01 6.57881439e-01 3.86945009e-01
-5.36705673e-01 6.37781739e-01 7.74964839e-02 4.12682325e-01
7.23179042e-01 -1.55858994e+00 7.38746941e-01 8.11366498e-01
2.93739051e-01 5.49048543e-01 5.30454636e-01 -8.34412396e-01
-1.15350890e+00 -1.11530936e+00 8.70807409e-01 -3.87496740e-01
9.27857399e-01 -6.25031948e-01 -1.01628304e+00 1.19845581e+00
1.24282239e-03 2.25870863e-01 5.48378587e-01 7.48225808e-01
-6.16120160e-01 -1.90776959e-01 -8.60821843e-01 5.08747697e-01
1.24365258e+00 -4.44818556e-01 -5.47328107e-02 5.96715212e-01
7.38633990e-01 -2.75770307e-01 -8.49677503e-01 4.63632554e-01
3.99724245e-01 -8.00466657e-01 6.54621363e-01 -7.10643828e-01
2.55692929e-01 -1.46089524e-01 -1.68129876e-02 -1.39036012e+00
-5.07944107e-01 -5.92002392e-01 -1.24101289e-01 1.23974180e+00
9.95532751e-01 -9.47644055e-01 1.02935803e+00 6.03327513e-01
-8.37853923e-02 -9.67909276e-01 -3.40448022e-01 -6.05527580e-01
2.11121455e-01 -5.18604219e-01 3.24952543e-01 1.14238429e+00
3.37745287e-02 4.46908444e-01 -1.75129041e-01 -1.10359207e-01
8.02192569e-01 -1.48900434e-01 7.31445074e-01 -1.45413363e+00
-5.04618168e-01 -4.21252072e-01 -4.01777446e-01 -9.68667209e-01
2.33712256e-01 -1.58843434e+00 5.26525825e-02 -1.77467096e+00
3.43740731e-01 -9.35418010e-01 -4.00780350e-01 1.08903790e+00
-6.61227629e-02 4.71597403e-01 -3.54226977e-02 7.69507587e-02
-6.73339307e-01 -2.98576299e-02 8.71866405e-01 -5.51657155e-02
-1.68210760e-01 -9.18256640e-02 -1.04297853e+00 7.75396585e-01
8.33643794e-01 -7.95820057e-01 -4.52119321e-01 -3.96785825e-01
6.35566354e-01 -2.30949000e-01 3.13193500e-01 -9.05585587e-01
4.35799450e-01 1.30796611e-01 3.06482285e-01 -1.38741970e-01
-1.30893858e-02 -3.97024542e-01 2.98705809e-02 4.68796849e-01
-4.66612071e-01 -1.38196975e-01 1.68593556e-01 5.22103190e-01
2.84027159e-01 -3.86079460e-01 7.14398623e-01 -5.99758960e-02
-4.31987137e-01 3.56277406e-01 -1.92609295e-01 1.48912013e-01
6.44263208e-01 1.26050459e-02 -8.20999503e-01 -5.40321112e-01
-5.98938286e-01 3.23478222e-01 6.38641000e-01 6.78822473e-02
1.23420171e-01 -1.03810990e+00 -5.88468492e-01 8.11628103e-02
-1.32111326e-01 3.63573735e-03 -2.43959874e-01 1.06340897e+00
-6.86778903e-01 1.31163627e-01 1.93769380e-01 -5.20232975e-01
-1.10200727e+00 3.99147987e-01 5.85141242e-01 -5.98496616e-01
-4.81875092e-01 1.09063148e+00 -1.51144385e-01 -9.44205403e-01
2.32245222e-01 -2.49797449e-01 1.63178667e-01 -1.20221160e-01
5.94222210e-02 3.77594322e-01 1.71649247e-01 -9.63871181e-02
-2.83290535e-01 2.70111799e-01 -3.33819270e-01 1.27648056e-01
1.57438958e+00 1.32005215e-01 -3.27836365e-01 4.41590339e-01
1.16492486e+00 4.85181101e-02 -1.11649406e+00 -3.44987333e-01
1.33516833e-01 1.29685536e-01 3.22442353e-01 -9.93372142e-01
-1.12895298e+00 7.90187478e-01 1.77498594e-01 5.48936844e-01
8.17772925e-01 2.34905735e-01 6.01058483e-01 4.98255879e-01
3.59379768e-01 -1.01779997e+00 -2.09226280e-01 4.81580466e-01
3.34542841e-01 -1.29107106e+00 3.52396518e-01 -6.30715609e-01
-3.04061115e-01 9.38126564e-01 4.13310021e-01 -3.11660618e-01
8.02388310e-01 4.76959229e-01 1.80653129e-02 -4.03586954e-01
-1.01806808e+00 -8.70531723e-02 1.32692546e-01 4.58431393e-01
5.40001452e-01 1.29466861e-01 -1.11488894e-01 2.69134164e-01
-1.34200007e-01 -2.92301029e-02 4.66398746e-01 8.24808002e-01
-5.32894194e-01 -1.09602463e+00 8.49766061e-02 8.37564349e-01
-3.12407315e-01 -4.34791625e-01 -5.24148643e-01 9.15313363e-01
-2.49147207e-01 9.70494092e-01 -1.66372672e-01 -6.92606986e-01
1.09871430e-02 2.34943911e-01 4.04034734e-01 -6.78911686e-01
-7.22210407e-01 -1.56636059e-01 5.27069032e-01 -6.13241851e-01
-2.82108933e-01 -4.20160353e-01 -1.44214892e+00 -8.51958454e-01
-6.10423028e-01 5.78616448e-02 7.51233160e-01 8.22198212e-01
2.60015547e-01 5.48002183e-01 2.97709525e-01 -7.29043305e-01
-5.73032439e-01 -9.41888690e-01 -6.23430490e-01 1.12290084e-01
1.24824539e-01 -2.83241093e-01 -7.86509871e-01 -3.24204564e-01] | [6.999476432800293, 6.1305975914001465] |
18941652-a192-446b-b919-2edb1ccd7cb4 | spectnet-end-to-end-audio-signal | 2211.09352 | null | https://arxiv.org/abs/2211.09352v1 | https://arxiv.org/pdf/2211.09352v1.pdf | SpectNet : End-to-End Audio Signal Classification Using Learnable Spectrograms | Pattern recognition from audio signals is an active research topic encompassing audio tagging, acoustic scene classification, music classification, and other areas. Spectrogram and mel-frequency cepstral coefficients (MFCC) are among the most commonly used features for audio signal analysis and classification. Recently, deep convolutional neural networks (CNN) have been successfully used for audio classification problems using spectrogram-based 2D features. In this paper, we present SpectNet, an integrated front-end layer that extracts spectrogram features within a CNN architecture that can be used for audio pattern recognition tasks. The front-end layer utilizes learnable gammatone filters that are initialized using mel-scale filters. The proposed layer outputs a 2D spectrogram image which can be fed into a 2D CNN for classification. The parameters of the entire network, including the front-end filterbank, can be updated via back-propagation. This training scheme allows for fine-tuning the spectrogram-image features according to the target audio dataset. The proposed method is evaluated in two different audio signal classification tasks: heart sound anomaly detection and acoustic scene classification. The proposed method shows a significant 1.02\% improvement in MACC for the heart sound classification task and 2.11\% improvement in accuracy for the acoustic scene classification task compared to the classical spectrogram image features. The source code of our experiments can be found at \url{https://github.com/mHealthBuet/SpectNet} | ['Taufiq Hasan', 'Md. Istiaq Ansari'] | 2022-11-17 | null | null | null | null | ['audio-tagging', 'sound-classification', 'scene-classification', 'music-classification'] | ['audio', 'audio', 'computer-vision', 'music'] | [ 2.09597543e-01 -4.83242005e-01 3.11695367e-01 -2.92196900e-01
-6.90700114e-01 -3.44209373e-01 -5.36271222e-02 3.36026460e-01
-5.41863382e-01 2.74928361e-01 6.69107400e-03 -6.23955913e-02
4.26875129e-02 -4.35467690e-01 -4.00551319e-01 -6.14356756e-01
-3.21090251e-01 -3.36422354e-01 2.45921850e-01 1.11916460e-01
2.52445042e-01 5.20114303e-01 -1.85650349e+00 6.25060499e-01
1.39336705e-01 1.72413290e+00 5.76141700e-02 1.19928503e+00
2.35902831e-01 5.87475419e-01 -7.53614545e-01 1.24431677e-01
-6.72940165e-03 -4.13755625e-01 -5.68140626e-01 -2.07449138e-01
3.57743233e-01 -3.96785466e-03 -2.85740346e-01 9.35540915e-01
9.90127802e-01 3.06866884e-01 3.82933497e-01 -1.11318791e+00
5.57862082e-03 5.66546142e-01 -6.25589564e-02 7.62138784e-01
2.90913373e-01 -7.78564194e-04 1.00704885e+00 -9.03930843e-01
-1.09994680e-01 1.02562094e+00 7.96412945e-01 2.85625160e-01
-8.40473533e-01 -9.97104406e-01 -3.67323101e-01 6.28731370e-01
-1.51218355e+00 -5.31493902e-01 1.11797178e+00 -5.66154122e-01
9.69838381e-01 3.59463215e-01 7.99196184e-01 5.82050085e-01
3.35633188e-01 5.00203848e-01 7.41949618e-01 -6.03561580e-01
1.96795017e-01 -4.54358011e-02 1.61347806e-01 5.88985980e-01
-3.72134894e-01 1.14360124e-01 -7.72490323e-01 -1.66755527e-01
7.39691556e-01 -2.54604548e-01 -2.84284204e-01 3.46428305e-01
-9.35194254e-01 7.12484717e-01 3.70152473e-01 5.49423754e-01
-3.60765010e-01 5.54688454e-01 8.09529126e-01 4.94179875e-01
4.62872684e-01 4.08739120e-01 -4.28128362e-01 -4.11767721e-01
-8.36248279e-01 2.30048239e-01 4.47339445e-01 9.83918905e-02
2.91793346e-01 5.98260403e-01 -1.84336510e-02 1.30797553e+00
2.68905014e-01 3.16591322e-01 8.17164183e-01 -8.33087623e-01
1.57561719e-01 2.65154898e-01 -2.59948254e-01 -1.02635944e+00
-7.51475930e-01 -6.17321551e-01 -8.18035543e-01 1.27272576e-01
2.90475011e-01 -2.98018783e-01 -5.54130197e-01 1.41933453e+00
3.15225303e-01 6.75731361e-01 -2.03617975e-01 1.00799072e+00
9.50084805e-01 7.19847381e-01 -1.17199719e-01 8.45982656e-02
1.48058593e+00 -5.82748175e-01 -6.64932191e-01 2.21087635e-01
3.52922708e-01 -1.10359812e+00 1.06144965e+00 6.26157820e-01
-8.59938383e-01 -1.03201938e+00 -1.05495477e+00 1.50623128e-01
-3.52704853e-01 5.38322985e-01 1.50883868e-01 7.49351501e-01
-8.15665603e-01 6.80174768e-01 -8.31567049e-01 -8.01403001e-02
4.29163009e-01 6.24935091e-01 -1.28283530e-01 5.19277811e-01
-1.26193500e+00 2.39645943e-01 6.05206847e-01 1.10944144e-01
-9.20460045e-01 -7.18249023e-01 -8.58859360e-01 2.59014875e-01
-1.07197516e-01 -3.14191133e-01 1.44347823e+00 -8.61578524e-01
-1.71151221e+00 6.13145113e-01 8.99442732e-02 -7.08809376e-01
-5.26719466e-02 -2.22121283e-01 -7.92853892e-01 2.32828990e-01
-2.76073933e-01 4.85822916e-01 1.25598323e+00 -3.72802973e-01
-8.99729252e-01 -1.24141656e-01 -2.66265780e-01 1.34477958e-01
-4.61689383e-01 2.07558215e-01 -1.88818239e-02 -1.04100370e+00
1.21403569e-02 -7.87416220e-01 2.60313511e-01 -3.27053279e-01
-3.86309743e-01 -2.72942722e-01 8.28106046e-01 -6.74501061e-01
1.35895038e+00 -2.65521789e+00 -2.34471872e-01 8.35052282e-02
-1.14812627e-01 4.29265022e-01 -1.34889856e-01 1.14644997e-01
-3.35489124e-01 -1.48228496e-01 -9.38606411e-02 -1.76917478e-01
3.46113602e-03 -4.42008704e-01 -2.59305567e-01 4.03501660e-01
2.62781441e-01 4.34188485e-01 -5.51799238e-01 -3.17121632e-02
4.47541893e-01 6.45073593e-01 -5.71993113e-01 1.54662684e-01
1.12298198e-01 5.40540338e-01 -2.63781305e-02 3.02635223e-01
4.57730055e-01 2.76813239e-01 -1.35599092e-01 -3.30808610e-01
-1.33627892e-01 4.94452000e-01 -1.45035851e+00 1.67427015e+00
-6.71125650e-01 8.67121398e-01 -6.47124415e-03 -9.78670359e-01
1.07146370e+00 7.38069177e-01 5.02021194e-01 -4.17629719e-01
3.83948684e-01 2.99464643e-01 3.49829912e-01 -5.40736735e-01
2.72971272e-01 -1.61943138e-01 -1.47478655e-01 2.91698128e-01
3.47989231e-01 -2.08344907e-01 -9.58634838e-02 -3.89183819e-01
9.30304527e-01 -4.70215440e-01 1.85953140e-01 6.27626991e-03
9.39076066e-01 -4.79187101e-01 4.43358928e-01 4.98964846e-01
-7.50619769e-02 5.27124882e-01 1.59447059e-01 -6.50142014e-01
-8.54814589e-01 -6.99557543e-01 -2.78005511e-01 1.28144431e+00
-4.39951301e-01 -4.89514530e-01 -6.88156188e-01 -2.95741737e-01
-7.39021748e-02 3.07358325e-01 -3.81757051e-01 -2.43323594e-01
-5.38698852e-01 -4.00268435e-01 1.14858770e+00 5.24169087e-01
4.93955553e-01 -1.43645322e+00 -7.24735200e-01 4.09338295e-01
-6.57120794e-02 -9.56022859e-01 -3.87770176e-01 3.86856854e-01
-7.33414769e-01 -8.77315402e-01 -4.57944304e-01 -7.70397604e-01
1.09407138e-02 -3.20771635e-01 5.34138739e-01 -4.32283580e-02
-6.16023898e-01 3.78317505e-01 -3.27062070e-01 -7.76769161e-01
-3.94852191e-01 2.40007773e-01 3.90123904e-01 4.37999219e-01
2.45180786e-01 -8.84325802e-01 -7.29672372e-01 1.57748666e-02
-7.84921288e-01 -3.56654286e-01 6.93935901e-02 7.71965086e-01
6.14251018e-01 1.84993252e-01 9.91351247e-01 -3.92701715e-01
7.84469187e-01 -7.46472925e-02 -6.25163317e-01 -4.06289399e-01
3.40266190e-02 -4.50667977e-01 7.69703269e-01 -4.49614555e-01
-5.11351824e-01 2.80174077e-01 -7.75239229e-01 -6.37070358e-01
-6.93712592e-01 4.06929106e-01 9.42183957e-02 6.03636354e-02
5.51946700e-01 1.80056170e-01 -2.40473703e-01 -8.51003230e-01
-1.08703420e-01 1.01683009e+00 5.59900820e-01 -6.68590367e-02
3.60988319e-01 2.14459971e-01 -1.25413135e-01 -1.09752893e+00
-7.71275759e-01 -6.56991065e-01 -6.00763440e-01 -4.37810302e-01
9.90318537e-01 -8.73219550e-01 -8.31907809e-01 8.00764441e-01
-9.76233482e-01 -1.94231376e-01 -3.49816799e-01 8.55928242e-01
-5.01857936e-01 9.13283154e-02 -6.45951986e-01 -9.48662043e-01
-6.27239704e-01 -9.99491453e-01 9.17677641e-01 2.01462328e-01
-4.58890229e-01 -8.57533574e-01 1.39775276e-01 1.03978299e-01
3.94703925e-01 4.15084437e-02 1.00255215e+00 -8.54203045e-01
3.72396559e-02 -2.77204961e-01 2.08667204e-01 7.34641016e-01
1.60397127e-01 -2.40755767e-01 -1.58962381e+00 -2.05313548e-01
2.30834913e-02 -2.89365292e-01 8.51570010e-01 6.90200090e-01
1.63851750e+00 -1.47309467e-01 2.60654449e-01 6.18624866e-01
1.04131389e+00 5.97744226e-01 3.61503184e-01 5.85659333e-02
4.81418937e-01 9.95071605e-02 4.29717064e-01 6.89206660e-01
-1.37255728e-01 8.18649769e-01 3.48212421e-01 -3.34022641e-02
-3.50532442e-01 -5.14182150e-02 2.72632420e-01 1.09839630e+00
1.05529711e-01 8.50964561e-02 -8.56275618e-01 3.83757442e-01
-1.48276412e+00 -9.33089674e-01 -5.08977007e-03 2.11134982e+00
6.95031583e-01 1.51515901e-01 3.40641320e-01 1.13944197e+00
7.52742350e-01 9.38589498e-02 -3.53960007e-01 -6.56224430e-01
3.13665301e-01 8.50837767e-01 -1.27072692e-01 3.43713552e-01
-1.45773339e+00 6.43920839e-01 5.04547644e+00 9.52872574e-01
-1.72832215e+00 2.60977149e-01 3.64573956e-01 -2.56258845e-01
4.77453500e-01 -5.21200776e-01 -4.90719020e-01 3.71935576e-01
1.20011294e+00 2.11426243e-01 3.51031572e-01 6.49912238e-01
4.58851665e-01 1.11203445e-02 -9.22859550e-01 1.45594871e+00
-1.73155442e-01 -1.23837209e+00 -1.52254149e-01 -3.42756003e-01
1.88965499e-01 4.04309370e-02 2.76633292e-01 1.31403044e-01
-5.10920703e-01 -8.00400615e-01 7.70074964e-01 3.16468000e-01
9.21616852e-01 -1.14944804e+00 7.62681961e-01 -7.49526778e-03
-1.42717087e+00 -4.38560218e-01 -1.71655163e-01 -2.64327854e-01
6.40601991e-03 5.99489808e-01 -1.18523538e+00 2.48072937e-01
1.01227438e+00 6.15129232e-01 -3.61025363e-01 1.49629796e+00
6.92565292e-02 1.25117791e+00 -3.28283101e-01 6.52088299e-02
3.33627798e-02 4.46605891e-01 7.09107637e-01 1.46758270e+00
5.22298217e-01 -1.24830648e-01 1.59294959e-02 3.96724403e-01
-1.28403172e-01 3.41764599e-01 -2.98064619e-01 -1.10780001e-01
3.74622434e-01 1.26467276e+00 -6.94619477e-01 -1.84137315e-01
-1.69603117e-02 6.13919556e-01 -2.20920652e-01 5.08228578e-02
-8.18014443e-01 -1.08504546e+00 7.62277722e-01 4.37307656e-02
4.80522126e-01 -7.75816143e-02 -5.64471148e-02 -7.93457448e-01
-1.71334431e-01 -8.32925737e-01 4.73480493e-01 -5.86529374e-01
-9.67497587e-01 9.00299072e-01 -3.37131858e-01 -1.32354879e+00
-2.99351782e-01 -6.52884781e-01 -7.17537224e-01 6.99657083e-01
-1.28624976e+00 -6.19458020e-01 -1.65417030e-01 8.74141037e-01
7.32305169e-01 -4.68867302e-01 1.07550776e+00 7.06337333e-01
-5.52019238e-01 6.25783265e-01 -1.07150510e-01 4.26168293e-01
6.23710692e-01 -1.10907876e+00 2.49749020e-01 5.54786682e-01
5.83748341e-01 1.55847877e-01 3.23688716e-01 -1.56353191e-01
-8.59883428e-01 -1.30294633e+00 7.97885537e-01 1.75892204e-01
5.77891290e-01 -4.52768683e-01 -7.52495646e-01 1.66912913e-01
4.73780781e-02 1.87357023e-01 1.06381047e+00 -1.49377227e-01
-1.65958747e-01 -5.19933581e-01 -7.83638179e-01 -3.47772278e-02
3.71583194e-01 -8.23622346e-01 -3.21502775e-01 7.13655502e-02
4.48797852e-01 -4.52377021e-01 -9.67149496e-01 3.17205191e-01
6.41406596e-01 -8.30388308e-01 7.63327181e-01 -3.56706589e-01
9.01060924e-02 -4.54637706e-01 -2.72453725e-01 -1.26734686e+00
-2.09695324e-01 -5.99826574e-01 -2.13617552e-02 9.96657133e-01
3.04792196e-01 -4.37474400e-01 3.66473228e-01 -2.80798435e-01
-3.75966489e-01 -5.56821704e-01 -1.23318088e+00 -5.27850747e-01
-3.75909686e-01 -1.13068295e+00 4.77733523e-01 7.04708278e-01
-1.32414419e-02 5.06943107e-01 -3.42164874e-01 2.96488166e-01
2.29435027e-01 -8.62932578e-02 4.74342436e-01 -1.31181943e+00
-3.91482979e-01 -4.06865537e-01 -9.15813446e-01 -5.64248741e-01
-1.65461265e-02 -1.01404321e+00 -5.07892892e-02 -9.63734686e-01
-4.17969555e-01 -2.25089267e-01 -7.82193780e-01 5.50603628e-01
1.98151648e-01 8.00596952e-01 3.86586368e-01 -1.71868175e-01
-3.10297370e-01 3.39696586e-01 8.20711672e-01 -1.11122131e-01
-4.65522945e-01 6.11865461e-01 -2.63695978e-02 8.26648176e-01
1.16302967e+00 -5.88049650e-01 -2.60966837e-01 -1.60709605e-01
-1.53318718e-01 1.06855527e-01 6.13326728e-01 -1.59779131e+00
1.02278255e-01 4.59393024e-01 4.93556023e-01 -4.80374962e-01
7.33958900e-01 -6.83921933e-01 1.40466899e-01 5.97942352e-01
-6.06495380e-01 -3.00181378e-02 6.35316730e-01 3.53395373e-01
-5.95368147e-01 -3.41434747e-01 8.28158081e-01 1.09650493e-01
-4.77994800e-01 1.23313777e-01 -9.02958691e-01 -3.11271995e-01
5.04119217e-01 -1.97923034e-02 2.26725474e-01 -4.26211745e-01
-1.23735964e+00 -5.83993614e-01 -4.65199828e-01 3.83143425e-01
7.77880132e-01 -1.52085805e+00 -6.33045912e-01 4.42713290e-01
-2.35623121e-01 -3.21758598e-01 5.49533904e-01 8.99528205e-01
-4.32660520e-01 4.78258193e-01 -3.81106824e-01 -7.79422998e-01
-1.56820095e+00 1.33521846e-02 6.42085969e-01 9.95164514e-02
-4.12247986e-01 1.03569436e+00 -5.08075617e-02 -1.38809875e-01
6.86401725e-01 -7.46733606e-01 -4.70345974e-01 1.17942557e-01
6.65439665e-01 4.34934407e-01 4.18927163e-01 -5.28367281e-01
-4.32248861e-01 5.14073730e-01 1.76622719e-01 -1.19221963e-01
1.48634672e+00 8.66862088e-02 1.19378632e-02 7.72430003e-01
1.33016562e+00 -1.12046540e-01 -7.12094843e-01 -9.27892774e-02
-1.48022413e-01 -2.09836408e-01 3.58585924e-01 -7.30476975e-01
-1.34691572e+00 1.49252474e+00 1.22950315e+00 5.12391865e-01
1.54910827e+00 -2.83656329e-01 7.67714798e-01 2.70439744e-01
-1.04147658e-01 -9.35131609e-01 2.21799657e-01 6.23311162e-01
9.43365455e-01 -6.44049764e-01 -3.98618490e-01 1.91957131e-02
-4.38999116e-01 1.46549177e+00 3.63437831e-01 -3.09719205e-01
1.10947800e+00 2.95140326e-01 3.36132467e-01 -1.16728902e-01
-5.91781199e-01 -1.64165050e-01 5.85564613e-01 4.11828518e-01
6.30383253e-01 1.62908086e-03 1.63627803e-01 8.78596842e-01
-5.79298437e-01 -9.70169902e-02 1.87411204e-01 6.45343900e-01
-4.53190088e-01 -8.54920328e-01 -5.46817601e-01 3.99293274e-01
-9.37124968e-01 -9.21220109e-02 -2.74148822e-01 2.17853740e-01
4.37090039e-01 9.81170774e-01 3.52986783e-01 -7.45856166e-01
3.20531398e-01 5.51737249e-01 2.87167102e-01 -6.23338223e-01
-1.07000995e+00 4.83853459e-01 -3.85437049e-02 -2.92456388e-01
-3.28756601e-01 -4.67583627e-01 -1.25865018e+00 2.36138985e-01
-4.09068882e-01 2.93925732e-01 7.39398420e-01 6.88998044e-01
3.08633298e-01 9.07863975e-01 7.62294292e-01 -8.84033084e-01
-2.67736048e-01 -1.16354609e+00 -6.39590383e-01 1.06700629e-01
5.74176610e-01 -4.76086289e-01 -1.33205622e-01 4.03767556e-01] | [15.222073554992676, 5.2752604484558105] |
42d44a6f-021d-4a96-88e4-526f2cd8e787 | sequential-estimation-of-nonparametric | 2012.06287 | null | https://arxiv.org/abs/2012.06287v2 | https://arxiv.org/pdf/2012.06287v2.pdf | Sequential estimation of Spearman rank correlation using Hermite series estimators | In this article we describe a new Hermite series based sequential estimator for the Spearman rank correlation coefficient and provide algorithms applicable in both the stationary and non-stationary settings. To treat the non-stationary setting, we introduce a novel, exponentially weighted estimator for the Spearman rank correlation, which allows the local nonparametric correlation of a bivariate data stream to be tracked. To the best of our knowledge this is the first algorithm to be proposed for estimating a time varying Spearman rank correlation that does not rely on a moving window approach. We explore the practical effectiveness of the Hermite series based estimators through real data and simulation studies demonstrating good practical performance. The simulation studies in particular reveal competitive performance compared to an existing algorithm. The potential applications of this work are manifold. The Hermite series based Spearman rank correlation estimator can be applied to fast and robust online calculation of correlation which may vary over time. Possible machine learning applications include, amongst others, fast feature selection and hierarchical clustering on massive data sets. | ['Melvin Varughese', 'Michael Stephanou'] | 2020-12-11 | null | null | null | null | ['sequential-correlation-estimation', 'data-summarization'] | ['miscellaneous', 'miscellaneous'] | [-1.95844650e-01 -7.22153962e-01 1.74782231e-01 -2.03774273e-01
-9.42789972e-01 -5.69930553e-01 3.80656958e-01 1.80954993e-01
-3.32771689e-01 8.91512632e-01 -3.71290118e-01 -1.88697666e-01
-7.13741899e-01 -5.59319496e-01 -9.50795114e-02 -9.60591853e-01
-1.07629585e+00 4.45455760e-01 5.76111555e-01 -6.79484606e-02
2.95635372e-01 6.15649760e-01 -1.52802324e+00 -2.72482663e-01
5.74688554e-01 1.38694608e+00 -1.33860976e-01 1.11734760e+00
2.62521625e-01 3.77236724e-01 -5.77035308e-01 2.24306975e-02
4.57739979e-01 -5.38762450e-01 -1.66222617e-01 -2.11882949e-01
-2.16716468e-01 -5.19902864e-03 1.02134332e-01 5.61260760e-01
5.73205113e-01 5.07514775e-01 6.59468055e-01 -1.46079767e+00
-2.17963066e-02 3.41654867e-01 -9.77099299e-01 6.88937128e-01
5.88098884e-01 -5.22149920e-01 9.38684523e-01 -9.79991376e-01
9.47301015e-02 9.10260499e-01 1.11368322e+00 -1.41218185e-01
-1.01907492e+00 -1.08908308e+00 -4.29251105e-01 2.17532858e-01
-1.52162874e+00 -1.24547780e-01 4.26608682e-01 -3.22879761e-01
4.76948947e-01 3.39928210e-01 8.31629097e-01 6.13504469e-01
2.50011355e-01 4.49322671e-01 1.36253572e+00 -4.50322837e-01
3.18267077e-01 -1.59120098e-01 1.44471228e-01 6.04360521e-01
4.59738046e-01 2.90530890e-01 -8.18344414e-01 -7.22443879e-01
6.94438100e-01 -7.43755773e-02 6.49266541e-02 -5.91483653e-01
-1.03538942e+00 7.81074524e-01 -1.69638067e-01 3.19300234e-01
-4.03316379e-01 2.24877998e-01 6.14920855e-01 5.45475006e-01
7.91000307e-01 1.36502147e-01 -6.66836798e-01 -6.13281488e-01
-1.12491930e+00 1.48356080e-01 1.25763166e+00 8.91756713e-01
2.26648688e-01 -3.50876361e-01 -4.89423517e-03 5.46180010e-01
1.50980204e-01 4.84038383e-01 3.33276004e-01 -7.28789270e-01
1.49249300e-01 -1.14008978e-01 2.08207816e-01 -7.31023908e-01
-7.10445464e-01 -4.65061456e-01 -7.80417562e-01 -3.19044024e-01
6.58080518e-01 -3.51403981e-01 1.28014967e-01 1.26826239e+00
4.77439404e-01 8.51337135e-01 -2.63153892e-02 4.47764844e-01
3.47425818e-01 4.31351632e-01 -3.58847976e-01 -6.79727077e-01
1.26522863e+00 -4.16510940e-01 -7.00572014e-01 6.27599120e-01
5.11859238e-01 -1.02813637e+00 3.22485149e-01 5.10072410e-01
-9.70772862e-01 5.89231178e-02 -1.02703702e+00 6.44962013e-01
-2.40612790e-01 -1.19395152e-01 1.04902196e+00 8.51771593e-01
-8.73368382e-01 6.97088420e-01 -1.12455869e+00 -6.25222087e-01
7.11624399e-02 3.89292806e-01 -7.80680701e-02 3.09825003e-01
-1.19974017e+00 5.48032045e-01 8.52949545e-02 1.27708420e-01
-7.81232938e-02 -7.48692036e-01 -5.51220953e-01 8.09649453e-02
5.76820411e-02 -7.30893970e-01 1.10545111e+00 -6.59496725e-01
-1.51985836e+00 3.75532150e-01 -3.43635708e-01 -7.92468488e-01
6.25780940e-01 -5.42408004e-02 -4.58222181e-01 9.83067080e-02
2.23196343e-01 -4.60661292e-01 7.51937389e-01 -6.94620848e-01
-8.42295647e-01 -2.06768081e-01 -7.26436317e-01 -1.35023564e-01
-2.13997334e-01 1.61398232e-01 -9.39337835e-02 -6.06737614e-01
2.26375654e-01 -9.72155929e-01 -2.60559887e-01 -2.47086778e-01
1.86904445e-01 -2.76877940e-01 5.91768920e-01 -2.17094183e-01
1.49934447e+00 -2.05577731e+00 -6.28856182e-01 7.61485100e-01
-2.98682060e-02 -2.26530924e-01 6.18905239e-02 9.73421574e-01
-3.88093404e-02 -2.43705243e-01 -2.53604531e-01 5.38878143e-02
-2.22776264e-01 1.84484955e-03 -1.75166145e-01 1.14825952e+00
1.11122675e-01 3.54237497e-01 -1.23207796e+00 -4.85784739e-01
-4.98124547e-02 4.69551593e-01 -3.36913019e-01 2.25237776e-02
8.67256641e-01 2.17116535e-01 -3.73769343e-01 3.20551544e-01
8.75656307e-01 -1.56033009e-01 1.93003729e-01 4.47415635e-02
-2.42502093e-01 -2.46480495e-01 -1.49358821e+00 1.07443142e+00
-6.66606367e-01 9.81539488e-01 -4.43159789e-03 -1.16814852e+00
1.16191900e+00 2.83514678e-01 1.16560054e+00 -3.89641970e-01
8.56130719e-02 5.05072057e-01 -6.45134225e-03 -2.67110676e-01
4.71721143e-01 -5.43481350e-01 -2.18960702e-01 5.28228998e-01
-1.90088004e-01 1.20682111e-02 5.43410718e-01 2.08680138e-01
1.54508305e+00 -1.15355678e-01 7.83055186e-01 -6.39653981e-01
7.71027684e-01 -2.44911253e-01 3.00011843e-01 8.18959534e-01
-4.31515217e-01 3.27094138e-01 4.71190393e-01 -2.89179564e-01
-7.41600215e-01 -1.41733670e+00 -5.56885064e-01 1.02920055e+00
1.11338161e-01 -3.30330282e-01 -7.96372220e-02 -9.06976759e-02
2.87403852e-01 9.02709290e-02 -5.57368159e-01 1.24986082e-01
-3.26504678e-01 -1.03808308e+00 3.67550433e-01 5.06378651e-01
3.15286100e-01 -3.29033673e-01 -6.28662527e-01 4.14337069e-01
2.82938659e-01 -1.08261168e+00 -4.32107955e-01 5.49002111e-01
-1.13507974e+00 -1.39441788e+00 -6.12531006e-01 -6.45770729e-01
4.98140812e-01 3.32557708e-01 9.81053710e-01 -1.02094024e-01
-3.03926617e-01 8.34464431e-01 -5.11912584e-01 -2.84861803e-01
9.89488885e-02 -1.51607305e-01 2.09825382e-01 7.93789923e-02
6.08449280e-01 -5.31226039e-01 -7.34152973e-01 6.94268823e-01
-7.47755408e-01 -8.09709430e-01 3.15815538e-01 1.18990088e+00
4.84425515e-01 4.05458868e-01 8.05498183e-01 -8.73756111e-01
1.06206572e+00 -7.94895232e-01 -8.24797571e-01 -1.78772599e-01
-6.55686617e-01 2.03361392e-01 6.36343062e-01 -4.52596456e-01
-7.98480749e-01 2.37826496e-01 4.89521831e-01 1.32129818e-01
4.86217976e-01 4.04886127e-01 6.19899750e-01 -5.97206540e-02
4.34338361e-01 7.27325827e-02 3.27255167e-02 -8.18307847e-02
1.09155849e-01 3.72431308e-01 5.56570411e-01 -4.71618742e-01
1.04591918e+00 9.02253926e-01 8.55056167e-01 -1.01027250e+00
-6.28219783e-01 -1.48666120e+00 -7.29328573e-01 2.22627409e-02
2.91891903e-01 -7.90859044e-01 -1.14554989e+00 3.93882617e-02
-9.80369151e-01 -1.12411723e-01 -1.51903376e-01 8.45731616e-01
-8.49896669e-01 4.97647315e-01 -5.48355162e-01 -1.39318705e+00
-1.74611658e-01 -2.50033677e-01 8.74957561e-01 2.35858257e-03
-4.07055527e-01 -1.45635951e+00 4.08317596e-01 -2.67515033e-01
2.93286383e-01 4.06329364e-01 1.31043315e-01 -8.01136732e-01
-1.83112279e-01 -6.72105968e-01 -3.75235021e-01 5.69485053e-02
5.35202771e-02 3.31903845e-01 -4.64053452e-01 -2.91215271e-01
-3.03800199e-02 3.41696829e-01 5.90513885e-01 6.59228861e-01
7.10850477e-01 -8.47091505e-05 -3.17536503e-01 4.14105982e-01
1.20491898e+00 1.40739188e-01 2.77255535e-01 5.43811202e-01
-2.91062854e-02 5.23103833e-01 1.18137503e+00 1.34141850e+00
1.42779648e-01 4.27046865e-01 -7.41782859e-02 1.59460649e-01
5.28136015e-01 1.70361117e-01 4.09309536e-01 1.20108223e+00
-1.19576380e-01 1.36642724e-01 -6.51617646e-01 4.83440757e-01
-2.02042246e+00 -1.37249041e+00 -7.67286181e-01 2.71717191e+00
4.54540014e-01 -1.62145004e-01 3.94519210e-01 2.57397950e-01
8.66000116e-01 -1.89622760e-01 -1.97315648e-01 -5.60885966e-01
-3.66808102e-02 5.42510033e-01 1.06546569e+00 2.13487387e-01
-1.20043087e+00 3.10623139e-01 6.89967060e+00 7.11464942e-01
-5.34695923e-01 1.43825084e-01 1.63018480e-02 -2.54458725e-01
2.51677930e-01 1.98518306e-01 -5.16303420e-01 6.74521923e-01
1.28908634e+00 -6.16359770e-01 2.79980898e-01 7.69001067e-01
5.36754906e-01 -5.66377759e-01 -9.12283003e-01 1.41945136e+00
-2.34680444e-01 -6.80187941e-01 -9.83299732e-01 1.37508318e-01
9.26527679e-01 -1.68568999e-01 5.97504899e-02 7.76716322e-02
3.15984458e-01 -7.10044384e-01 4.52291816e-02 6.33967280e-01
2.88592756e-01 -1.03801870e+00 1.05056727e+00 -2.65804268e-02
-1.81997108e+00 -4.02691886e-02 -2.38666534e-01 -2.48806462e-01
1.41285270e-01 1.11041045e+00 -8.66899133e-01 4.32970315e-01
7.48045921e-01 7.82276094e-01 -3.35520059e-01 1.55100453e+00
2.98678815e-01 5.17938912e-01 -7.92021632e-01 -5.49878061e-01
4.03583720e-02 -3.14504921e-01 3.73401254e-01 1.32023942e+00
1.01778090e+00 4.02417704e-02 -1.04223778e-02 1.04239397e-02
4.03277487e-01 4.90555674e-01 -4.87926394e-01 1.96823627e-01
7.17827678e-01 1.13458347e+00 -1.06978190e+00 -2.31045023e-01
-5.77738166e-01 7.64611840e-01 -2.42698431e-01 2.51035273e-01
-7.38689184e-01 -6.70674562e-01 4.48368877e-01 2.87874848e-01
4.56436455e-01 -7.21652448e-01 -4.58141029e-01 -9.20518398e-01
4.39137816e-02 -2.65021622e-01 8.92698348e-01 -1.10497467e-01
-1.48741591e+00 1.87307119e-01 3.42497677e-01 -1.56436396e+00
-5.17704606e-01 -4.22159225e-01 -5.61060846e-01 6.11918747e-01
-1.19071841e+00 -2.49183774e-01 -1.66733339e-01 5.12451828e-01
1.75648034e-01 -2.32633471e-01 5.14541864e-01 1.75242290e-01
-3.30141872e-01 3.58143598e-01 8.47512841e-01 -3.28136355e-01
8.32957804e-01 -1.38549197e+00 6.83799833e-02 6.59151971e-01
-1.70910448e-01 6.48632228e-01 1.20110023e+00 -5.56073487e-01
-1.39489019e+00 -5.97804189e-01 8.32468510e-01 -4.95030582e-01
1.52905369e+00 -1.81231126e-01 -6.06228828e-01 9.86272395e-02
-2.81543314e-01 3.97558957e-01 1.12302458e+00 3.30525070e-01
-5.03490232e-02 -4.02681708e-01 -1.09743309e+00 2.41039827e-01
7.76362836e-01 -4.22910571e-01 -1.34677738e-01 2.80763447e-01
-2.38290235e-01 4.43021245e-02 -1.23753476e+00 2.86075354e-01
1.13329577e+00 -1.07282639e+00 8.21565211e-01 -9.71143171e-02
-2.31615975e-01 -3.13811392e-01 -6.00577295e-02 -1.09027410e+00
-1.32278457e-01 -1.18169928e+00 -2.05602467e-01 1.04859149e+00
2.08739534e-01 -1.05592287e+00 8.34728837e-01 3.01509202e-01
4.83109504e-01 -6.41866148e-01 -1.31289411e+00 -1.44895625e+00
-3.00001949e-01 -6.29578352e-01 3.84655654e-01 6.91890180e-01
2.76846468e-01 -5.65070510e-02 -3.91051233e-01 1.86181009e-01
8.19346547e-01 1.28713712e-01 7.48651147e-01 -1.54278159e+00
-4.56594735e-01 -3.03655386e-01 -7.84635484e-01 -8.86183918e-01
-8.18990916e-03 -5.52585006e-01 1.98025152e-01 -9.62905049e-01
1.31402716e-01 -5.20370543e-01 -5.72434545e-01 -4.02969241e-01
1.14312414e-02 2.65548170e-01 -2.55276769e-01 6.96193799e-02
-8.05783153e-01 5.15062690e-01 7.24330962e-01 5.22806883e-01
-3.69641870e-01 4.43182975e-01 -1.92716077e-01 4.57629800e-01
8.10729206e-01 -7.51195490e-01 -3.14541668e-01 5.93962848e-01
3.34775269e-01 2.01518714e-01 -2.02224590e-02 -9.04218733e-01
4.48191941e-01 1.33337170e-01 3.40218514e-01 -7.65996337e-01
3.14857066e-02 -9.73896861e-01 9.20011550e-02 5.26470482e-01
-3.99086587e-02 4.60065424e-01 4.90675941e-02 8.99668455e-01
-3.09419334e-01 -1.43453926e-01 7.51431704e-01 2.47165024e-01
-4.77810621e-01 2.04218730e-01 -5.60237408e-01 -9.31959227e-02
1.23752308e+00 -2.77165860e-01 1.41436346e-02 -6.81723356e-01
-5.48435271e-01 3.14253271e-01 5.72603345e-02 2.37220265e-02
4.57367688e-01 -1.41267335e+00 -5.13302922e-01 5.85873090e-02
2.88254190e-02 -6.27019584e-01 1.05694637e-01 1.33289075e+00
-4.98589277e-01 5.11136055e-01 6.81145713e-02 -5.20007014e-01
-1.43756545e+00 4.37889576e-01 -1.24065280e-01 -5.52305520e-01
-2.13509321e-01 5.62786639e-01 -6.34656698e-02 7.73598999e-02
-1.26112550e-01 -2.34726459e-01 1.53423741e-01 4.27597135e-01
5.72431505e-01 1.02625942e+00 5.16231209e-02 -3.50479543e-01
-6.94213867e-01 6.53030574e-01 3.14418435e-01 -3.63229543e-01
1.41926670e+00 -4.16360646e-01 -8.86873454e-02 8.74356508e-01
1.46634758e+00 1.75721824e-01 -1.14925826e+00 -5.76575249e-02
5.80814779e-01 -5.54790676e-01 -3.03558141e-01 1.98277682e-02
-6.24493062e-01 4.25883770e-01 7.87979662e-01 6.21590436e-01
1.30460370e+00 -2.04758234e-02 6.50916278e-01 3.19752097e-01
4.36906576e-01 -1.32363844e+00 1.21347405e-01 4.74020988e-01
5.23954570e-01 -1.16531992e+00 4.28919464e-01 -4.40500915e-01
-3.71720850e-01 1.55616796e+00 -3.18114571e-02 -6.33336067e-01
1.04353678e+00 4.28767860e-01 -2.24932030e-01 4.63451967e-02
-9.97535169e-01 -5.37101150e-01 2.02760845e-01 7.90101349e-01
8.69922161e-01 2.24153280e-01 -9.01044905e-01 3.37110937e-01
-2.80958474e-01 -1.94033518e-01 6.86688185e-01 8.81225407e-01
-2.16184467e-01 -1.20765281e+00 -5.43800712e-01 7.76860118e-01
-5.49104691e-01 -7.18682408e-02 6.52214661e-02 9.75822628e-01
-5.78987062e-01 1.15158129e+00 2.94435978e-01 -3.11635081e-02
2.21500635e-01 -8.53893161e-02 3.88423562e-01 -2.41249710e-01
-4.64375973e-01 2.04700083e-01 9.33408514e-02 -4.87696946e-01
-5.09449422e-01 -1.25568306e+00 -1.08300698e+00 -4.43862438e-01
-4.50779885e-01 5.59399903e-01 8.00793409e-01 6.79210722e-01
-8.24038759e-02 2.39260346e-01 1.07640696e+00 -5.68803608e-01
-5.90271711e-01 -6.44234717e-01 -1.23008108e+00 1.96161419e-01
3.19896162e-01 -8.03286314e-01 -6.60611272e-01 -3.28922063e-01] | [7.1565985679626465, 3.9749016761779785] |
7f5c91e2-571a-48f8-a6f2-9127980a27dd | weakly-supervised-segmentation-with-multi | 2007.01152 | null | https://arxiv.org/abs/2007.01152v3 | https://arxiv.org/pdf/2007.01152v3.pdf | Learning to Segment from Scribbles using Multi-scale Adversarial Attention Gates | Large, fine-grained image segmentation datasets, annotated at pixel-level, are difficult to obtain, particularly in medical imaging, where annotations also require expert knowledge. Weakly-supervised learning can train models by relying on weaker forms of annotation, such as scribbles. Here, we learn to segment using scribble annotations in an adversarial game. With unpaired segmentation masks, we train a multi-scale GAN to generate realistic segmentation masks at multiple resolutions, while we use scribbles to learn their correct position in the image. Central to the model's success is a novel attention gating mechanism, which we condition with adversarial signals to act as a shape prior, resulting in better object localization at multiple scales. Subject to adversarial conditioning, the segmentor learns attention maps that are semantic, suppress the noisy activations outside the objects, and reduce the vanishing gradient problem in the deeper layers of the segmentor. We evaluated our model on several medical (ACDC, LVSC, CHAOS) and non-medical (PPSS) datasets, and we report performance levels matching those achieved by models trained with fully annotated segmentation masks. We also demonstrate extensions in a variety of settings: semi-supervised learning; combining multiple scribble sources (a crowdsourcing scenario) and multi-task learning (combining scribble and mask supervision). We release expert-made scribble annotations for the ACDC dataset, and the code used for the experiments, at https://vios-s.github.io/multiscale-adversarial-attention-gates | ['Gabriele Valvano', 'Sotirios A. Tsaftaris', 'Andrea Leo'] | 2020-07-02 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [ 5.03706157e-01 5.81321597e-01 1.29777208e-01 -3.58507901e-01
-1.32028282e+00 -8.88662398e-01 1.78270683e-01 4.88127805e-02
-5.67164719e-01 6.90406978e-01 -5.27909398e-03 -2.52846152e-01
5.26607990e-01 -5.10713935e-01 -1.13283515e+00 -8.23312044e-01
3.10591072e-01 6.10962808e-01 5.32046258e-01 -1.59744933e-01
-2.38433242e-01 4.52581674e-01 -8.22847486e-01 5.15674233e-01
8.33644629e-01 9.22773063e-01 2.80234426e-01 8.40437591e-01
3.13604593e-01 7.34784484e-01 -5.66632807e-01 -4.85997379e-01
4.93032694e-01 -3.97234797e-01 -8.26531947e-01 2.10956171e-01
4.25125301e-01 -3.47112685e-01 -1.39702260e-01 9.86832023e-01
5.27851641e-01 -1.66298196e-01 4.85716045e-01 -8.68686795e-01
-6.88369155e-01 2.98220932e-01 -6.58642530e-01 9.91196930e-02
9.32340473e-02 5.78875065e-01 7.84794331e-01 -5.67182243e-01
5.38316250e-01 9.11186755e-01 7.86625326e-01 8.32838893e-01
-1.47641277e+00 -4.56299126e-01 -8.83167610e-03 -3.93759876e-01
-1.19545448e+00 -2.97767699e-01 6.61585391e-01 -5.62428415e-01
3.83576453e-01 3.12072009e-01 4.13386166e-01 1.42015517e+00
2.31338710e-01 8.89721274e-01 1.45492375e+00 -3.04187804e-01
3.31182927e-01 5.91211319e-02 -1.94115326e-01 8.90563786e-01
-6.71431869e-02 1.14541359e-01 -3.29421610e-02 -1.77022293e-01
1.27240360e+00 9.33256298e-02 -4.53794569e-01 -2.04512730e-01
-1.12410843e+00 8.30343366e-01 7.52445459e-01 2.02251121e-01
-3.48642796e-01 2.64710575e-01 1.96828857e-01 -2.57522985e-02
4.41751450e-01 5.79454839e-01 -5.31807005e-01 3.36228907e-01
-9.58538294e-01 2.35641235e-03 5.27387440e-01 8.56022656e-01
6.55553877e-01 -2.58349106e-02 -5.16186416e-01 6.72660470e-01
6.81945756e-02 4.20904785e-01 5.04488468e-01 -1.07901442e+00
1.90287307e-01 2.90717274e-01 2.18761951e-01 -4.54773098e-01
-3.99181783e-01 -5.67508280e-01 -7.56782651e-01 6.02679253e-01
7.99375176e-01 -3.51941943e-01 -1.56602311e+00 1.63531804e+00
3.06984901e-01 1.73213825e-01 -1.74767375e-01 1.25712955e+00
7.88331270e-01 2.67056108e-01 1.50627971e-01 1.75191820e-01
1.35071015e+00 -1.11981595e+00 -4.80587751e-01 -5.03531039e-01
5.48082411e-01 -5.80932260e-01 1.41622841e+00 2.86559999e-01
-1.15212762e+00 -3.95559460e-01 -8.02793980e-01 -1.85081318e-01
-1.65556043e-01 1.29402712e-01 4.15587753e-01 4.87237841e-01
-1.17320931e+00 5.84470987e-01 -1.25726092e+00 2.12626327e-02
1.07538188e+00 3.68605554e-01 -2.56255776e-01 3.76006514e-02
-9.78034079e-01 7.54084826e-01 2.57348679e-02 4.75913547e-02
-1.36767697e+00 -6.41292572e-01 -8.90547931e-01 -5.50677106e-02
3.14527780e-01 -6.93223417e-01 1.00544000e+00 -1.60558271e+00
-1.22238040e+00 1.28535247e+00 2.99065728e-02 -5.88632405e-01
8.44542682e-01 2.66711395e-02 1.60448208e-01 4.04822111e-01
2.95620650e-01 1.12225878e+00 9.45164323e-01 -1.46408308e+00
-5.77686653e-02 -3.59939754e-01 1.26395315e-01 2.83223372e-02
2.81016678e-01 -5.77775985e-02 -4.59666133e-01 -9.40062702e-01
-2.83095151e-01 -1.08178091e+00 -7.70603478e-01 1.79718480e-01
-6.84121370e-01 2.12264925e-01 5.32933772e-01 -8.77756953e-01
5.08951843e-01 -2.26598811e+00 1.01948768e-01 3.82906944e-02
2.26119608e-01 2.31479093e-01 -2.14902624e-01 -1.84541196e-01
-8.97310092e-04 2.07587332e-01 -8.21294129e-01 -6.52551353e-01
-3.63743573e-01 3.88350666e-01 -1.88501671e-01 5.20894766e-01
3.78021479e-01 1.43171835e+00 -8.90946090e-01 -5.52431285e-01
1.95166215e-01 3.80899251e-01 -6.21067107e-01 1.14642426e-01
-4.22821432e-01 1.13684106e+00 -5.25851905e-01 6.92261934e-01
3.32135201e-01 -4.75543618e-01 -1.86153173e-01 -9.70424935e-02
3.54012311e-01 -3.01200807e-01 -7.14481115e-01 1.95987451e+00
-4.54884559e-01 4.27235872e-01 4.72449064e-01 -9.12352145e-01
4.57942158e-01 3.82747591e-01 4.54305768e-01 -3.25761676e-01
2.55970269e-01 2.64703959e-01 -3.99176590e-03 -3.97341788e-01
-1.49505973e-01 -2.65262306e-01 -6.68166205e-02 1.34634659e-01
1.64508477e-01 -4.72885430e-01 -9.86429006e-02 2.07859397e-01
1.19409943e+00 2.65367597e-01 -1.16500169e-01 -2.50397831e-01
1.66587442e-01 4.48971167e-02 5.55761874e-01 9.30118680e-01
-2.99289525e-01 1.27677453e+00 5.93648732e-01 -3.99956763e-01
-1.02648556e+00 -1.14725137e+00 -2.16264024e-01 9.48521256e-01
4.36491743e-02 1.52515724e-01 -1.12413788e+00 -1.07002878e+00
-2.22074136e-01 4.04560924e-01 -9.29185092e-01 6.58704862e-02
-5.44042826e-01 -6.97510540e-01 7.87123322e-01 6.18786931e-01
2.62659580e-01 -1.36230302e+00 -7.04888999e-01 2.34092787e-01
-1.33562773e-01 -1.30791867e+00 -7.18101621e-01 4.59814191e-01
-6.31618738e-01 -1.11175787e+00 -1.12263811e+00 -8.34681749e-01
1.01044953e+00 -4.92435396e-01 1.19400024e+00 2.39291877e-01
-5.92234850e-01 2.22512200e-01 -3.06874096e-01 -4.91720706e-01
-4.19081450e-01 -4.73837219e-02 -5.22364378e-01 -3.29336263e-02
-2.42104158e-01 -5.96709609e-01 -7.87903309e-01 3.26939821e-01
-1.07567894e+00 1.68130100e-01 7.23939836e-01 1.15369892e+00
9.90775049e-01 -4.91812110e-01 5.71890295e-01 -1.35217702e+00
1.68039843e-01 -2.40844488e-01 -5.94517708e-01 5.68672381e-02
-1.36611983e-02 -1.35203168e-01 6.61500871e-01 -4.92911041e-01
-7.27326512e-01 3.94621760e-01 -5.73652327e-01 -6.85550809e-01
-4.64047134e-01 1.95755120e-02 -1.29051447e-01 -3.13710064e-01
9.26262617e-01 3.17152627e-02 1.11432029e-02 -2.23098844e-01
3.79167378e-01 2.57965267e-01 6.80095434e-01 -6.41324103e-01
6.08217835e-01 8.55694771e-01 -2.69476414e-01 -3.51355970e-01
-1.16078627e+00 -1.31046727e-01 -5.78620195e-01 3.68687958e-02
1.44201195e+00 -9.37841475e-01 -3.68197113e-01 3.85672241e-01
-1.15154552e+00 -1.10191393e+00 -6.60946369e-01 2.14965343e-01
-6.80265665e-01 -4.51050922e-02 -9.59894896e-01 -2.63182640e-01
-4.42058772e-01 -1.55438578e+00 1.51773691e+00 2.40078002e-01
-1.28225416e-01 -1.05909514e+00 -2.80526012e-01 6.53838217e-01
2.27730900e-01 8.44102204e-01 6.03524864e-01 -7.61814594e-01
-6.46644890e-01 -4.46105562e-02 -1.57597482e-01 5.28838575e-01
-2.45176945e-02 -4.59153175e-01 -1.17221618e+00 -2.57333249e-01
1.09884690e-03 -7.11833358e-01 9.03340459e-01 7.41793752e-01
1.63824427e+00 -1.84900299e-01 -2.97096193e-01 9.03670669e-01
1.20185149e+00 -6.58343658e-02 5.81098437e-01 -6.88862503e-02
9.77833509e-01 4.59645391e-01 2.20947146e-01 4.04558703e-02
8.08494445e-03 4.32904899e-01 4.88753408e-01 -8.83342266e-01
-3.40503216e-01 -1.57343313e-01 -1.44073945e-02 2.23085180e-01
4.94945794e-02 -2.15566844e-01 -8.09618592e-01 6.63820922e-01
-1.63449669e+00 -4.68186855e-01 -7.15025440e-02 1.80541587e+00
1.12214708e+00 3.12413305e-01 -1.26413582e-02 -3.20566893e-01
5.54373324e-01 1.03361728e-02 -7.93427229e-01 -1.71864122e-01
-1.60695896e-01 5.41447282e-01 7.89475083e-01 7.79921770e-01
-1.21247220e+00 1.17503810e+00 5.49079323e+00 9.00913239e-01
-1.22768784e+00 7.00649440e-01 1.32447004e+00 -1.06223710e-01
-2.74366796e-01 -2.42909178e-01 -3.37998897e-01 5.37245572e-01
4.83002394e-01 6.66877151e-01 4.53830153e-01 6.21419966e-01
2.18423203e-01 -4.78871316e-02 -1.12613225e+00 6.67058110e-01
-9.39531848e-02 -1.39712477e+00 -2.98596174e-01 -6.41297316e-03
1.06595218e+00 4.53861654e-01 -2.22027991e-02 7.33005479e-02
6.80703819e-01 -1.25360751e+00 6.95961833e-01 2.76963592e-01
1.09232223e+00 -3.42039526e-01 7.28683293e-01 2.94578999e-01
-5.65394461e-01 3.03978324e-01 -1.30984467e-02 3.12679589e-01
4.15662885e-01 6.02419496e-01 -6.27501369e-01 2.47463867e-01
5.54642260e-01 3.68009925e-01 -6.48708880e-01 6.88727736e-01
-6.13811493e-01 7.26666391e-01 -3.08498204e-01 4.13310319e-01
3.50542694e-01 -3.66605781e-02 4.43618715e-01 1.12748241e+00
-2.12781683e-01 1.61655143e-01 3.50104541e-01 1.12518537e+00
-3.57251048e-01 -2.12459877e-01 -3.14070165e-01 3.31696779e-01
6.23870753e-02 1.41357243e+00 -1.03431296e+00 -2.46742830e-01
-2.76227385e-01 1.37970591e+00 3.31316680e-01 6.96413457e-01
-1.02467489e+00 -8.41361582e-02 2.48080149e-01 3.36452097e-01
2.83893108e-01 1.15313932e-01 -7.82907128e-01 -1.02749538e+00
-3.99088934e-02 -8.81015420e-01 2.47345448e-01 -9.46716964e-01
-1.27416325e+00 7.57971048e-01 -2.56210536e-01 -7.23977506e-01
-3.98444161e-02 -5.72407007e-01 -7.28895783e-01 9.70701396e-01
-1.33798850e+00 -1.43987346e+00 -2.34786794e-01 6.22959912e-01
4.93216425e-01 2.48026043e-01 7.10214794e-01 3.46346855e-01
-2.32121333e-01 5.53199351e-01 -3.17220718e-01 5.55663764e-01
6.65593684e-01 -1.44314325e+00 3.45428884e-01 7.15215921e-01
2.66191274e-01 3.68008792e-01 5.74038446e-01 -5.72876155e-01
-7.70774066e-01 -1.34215355e+00 1.14993356e-01 -7.73613334e-01
5.26133418e-01 -6.42758250e-01 -8.84024203e-01 9.84770954e-01
1.25522405e-01 7.80863822e-01 4.35068190e-01 -3.91656637e-01
-3.03486399e-02 2.19213352e-01 -1.46698940e+00 4.72634315e-01
7.96477020e-01 -4.71492976e-01 -3.31649691e-01 7.13856459e-01
9.41637456e-01 -8.80883455e-01 -7.49886751e-01 3.57330471e-01
2.72292286e-01 -7.89763451e-01 9.45092559e-01 -4.55448806e-01
5.24228692e-01 -2.90196836e-01 1.95126459e-01 -1.28811336e+00
4.99037467e-02 -5.59408605e-01 3.34147990e-01 9.25816000e-01
6.45992398e-01 -6.54127300e-01 9.31265414e-01 6.14795983e-01
-4.62682039e-01 -1.11004782e+00 -9.33075726e-01 -3.87805253e-01
3.97362024e-01 -1.77914590e-01 2.06986457e-01 8.57684135e-01
-6.06842041e-01 7.94630721e-02 -6.78390935e-02 2.72122979e-01
4.85744894e-01 -1.22173980e-01 4.46813941e-01 -6.26289129e-01
-7.29895175e-01 -3.14734340e-01 -9.85331237e-02 -8.98446262e-01
3.16832572e-01 -9.30046797e-01 2.26551518e-01 -1.45994937e+00
-1.18191317e-01 -6.20663226e-01 -1.00183003e-01 8.97475362e-01
-3.30587476e-01 8.01015139e-01 1.94016516e-01 1.94962561e-01
-4.98020589e-01 2.11997643e-01 1.65568995e+00 -1.81809425e-01
-4.64110896e-02 6.79589063e-02 -6.55704975e-01 9.66834962e-01
6.63830042e-01 -5.24940073e-01 -1.93827510e-01 -6.37914956e-01
-2.80810356e-01 1.84844613e-01 8.55410874e-01 -7.58688986e-01
-3.91884744e-02 1.19446859e-01 5.91027856e-01 5.42718917e-02
2.92624027e-01 -7.50352323e-01 -1.52351633e-02 4.22611058e-01
-4.35927182e-01 -3.93771976e-01 2.78277427e-01 3.87767166e-01
4.20161150e-02 -2.17295095e-01 1.09826970e+00 -4.41981107e-01
-1.25778288e-01 4.41808552e-01 -2.01069206e-01 4.24476564e-01
1.01944113e+00 4.30663303e-02 -1.60016984e-01 -2.43399560e-01
-1.25120306e+00 3.18692416e-01 7.13940978e-01 5.85889518e-02
4.24372107e-01 -8.89220834e-01 -6.66749120e-01 1.93145394e-01
-1.38422519e-01 6.25774562e-01 1.34704053e-01 9.65817332e-01
-6.78083360e-01 -5.45319617e-02 -1.78278256e-02 -6.54442191e-01
-7.80138016e-01 4.83109087e-01 7.21181989e-01 -4.71148640e-01
-6.01130426e-01 9.42388594e-01 7.19216108e-01 -5.17307043e-01
1.51469141e-01 -4.85668480e-01 3.59670460e-01 -4.50171471e-01
2.41036758e-01 -2.25485772e-01 9.55317393e-02 -3.44883889e-01
-2.98163623e-01 3.61096591e-01 7.90786892e-02 -2.22685322e-01
1.10486317e+00 3.29145454e-02 2.05232635e-01 2.06814989e-01
1.04463840e+00 1.32571921e-01 -1.77042389e+00 2.02837307e-02
-2.80574352e-01 -1.11569420e-01 3.09471823e-02 -1.20322454e+00
-1.25940382e+00 8.03771615e-01 6.95044160e-01 8.16641096e-03
1.02435815e+00 3.24821800e-01 8.24257612e-01 -2.03317866e-01
3.06393147e-01 -6.68680191e-01 1.28426522e-01 1.33705020e-01
9.43350554e-01 -1.59714949e+00 -3.70882332e-01 -4.42750365e-01
-8.42226624e-01 5.70237637e-01 6.69846416e-01 -3.59144717e-01
3.88802618e-01 6.88473761e-01 3.51888090e-01 -1.36072725e-01
-2.45570749e-01 -2.70159692e-01 2.60844499e-01 6.61622703e-01
3.47686321e-01 1.50531605e-01 5.32512851e-02 7.35007644e-01
4.87414747e-02 -3.75325754e-02 4.13935453e-01 7.81558752e-01
-4.46476452e-02 -9.34816360e-01 -3.57069194e-01 5.17230928e-01
-9.60252106e-01 -3.36724877e-01 -3.45017582e-01 5.25125265e-01
3.16360533e-01 6.92204416e-01 -7.28524383e-03 2.17371359e-01
1.74484089e-01 -2.67431349e-01 4.69478488e-01 -8.82924855e-01
-8.65662277e-01 3.24818432e-01 -1.50041834e-01 -6.64157629e-01
-1.77457288e-01 -5.27365208e-01 -1.52166677e+00 3.07242393e-01
-7.79654831e-02 4.40342426e-02 4.31847900e-01 1.03571594e+00
1.87665790e-01 7.93774366e-01 1.89381123e-01 -1.02473354e+00
-1.92783296e-01 -9.61069942e-01 -4.38186020e-01 7.87889421e-01
4.34355974e-01 -2.41753846e-01 -2.45915785e-01 4.19275284e-01] | [14.491504669189453, -2.0460047721862793] |
874d78fa-3c2c-4f8f-a49d-2808e65188d9 | clustering-aware-negative-sampling-for | 2305.09892 | null | https://arxiv.org/abs/2305.09892v1 | https://arxiv.org/pdf/2305.09892v1.pdf | Clustering-Aware Negative Sampling for Unsupervised Sentence Representation | Contrastive learning has been widely studied in sentence representation learning. However, earlier works mainly focus on the construction of positive examples, while in-batch samples are often simply treated as negative examples. This approach overlooks the importance of selecting appropriate negative examples, potentially leading to a scarcity of hard negatives and the inclusion of false negatives. To address these issues, we propose ClusterNS (Clustering-aware Negative Sampling), a novel method that incorporates cluster information into contrastive learning for unsupervised sentence representation learning. We apply a modified K-means clustering algorithm to supply hard negatives and recognize in-batch false negatives during training, aiming to solve the two issues in one unified framework. Experiments on semantic textual similarity (STS) tasks demonstrate that our proposed ClusterNS compares favorably with baselines in unsupervised sentence representation learning. Our code has been made publicly available. | ['Rui Wang', 'Xiaojun Quan', 'Tao Yang', 'Fanqi Wan', 'Jinghao Deng'] | 2023-05-17 | null | null | null | null | ['semantic-textual-similarity'] | ['natural-language-processing'] | [ 6.35225534e-01 6.69540092e-02 -3.83970916e-01 -6.75231040e-01
-1.16663814e+00 -5.69463849e-01 7.63852119e-01 4.29436356e-01
-6.65387154e-01 6.47253931e-01 3.29053372e-01 -3.13900083e-01
1.03425831e-01 -7.24263906e-01 -4.39664781e-01 -6.10229731e-01
2.51119494e-01 4.52909589e-01 3.91848497e-02 -2.83899635e-01
5.41829646e-01 -9.86378267e-03 -1.60600173e+00 8.37252438e-01
1.03387511e+00 6.78027213e-01 3.29176784e-01 2.83932894e-01
-4.05914396e-01 9.84197736e-01 -8.98275137e-01 -5.18096864e-01
2.52315961e-02 -7.08126128e-01 -9.24351573e-01 1.47107020e-01
4.30320650e-01 6.61372989e-02 3.16075683e-01 1.18722260e+00
4.92513239e-01 3.43199641e-01 8.13844681e-01 -1.13831270e+00
-6.47190690e-01 9.11368668e-01 -7.00013280e-01 3.04934621e-01
2.25045815e-01 -4.83823903e-02 1.35765815e+00 -1.25286222e+00
4.09743816e-01 1.34590602e+00 6.78389132e-01 7.13649631e-01
-1.28509247e+00 -3.98049384e-01 1.59107015e-01 1.54799432e-01
-1.18073034e+00 -6.46940410e-01 8.97939682e-01 -2.62959927e-01
1.10741115e+00 5.18665791e-01 4.61094856e-01 1.13348258e+00
-2.85108358e-01 1.13169050e+00 9.59239483e-01 -8.00404072e-01
5.70162177e-01 3.36969733e-01 2.63867944e-01 5.31369150e-02
2.58973747e-01 -2.16905758e-01 -3.17239612e-01 -3.84331912e-01
3.46604168e-01 2.00892150e-01 -1.57164752e-01 -4.79160994e-01
-8.95926356e-01 1.11587870e+00 3.78717929e-01 4.87402648e-01
-2.70049542e-01 -1.45339027e-01 8.39994371e-01 4.85663295e-01
8.18277895e-01 5.77028096e-01 -4.74157482e-01 -9.10147950e-02
-8.07824194e-01 1.93532463e-02 5.81110358e-01 7.61506021e-01
7.33086109e-01 3.98312099e-02 -4.65403348e-01 1.35867739e+00
7.23201334e-02 2.59018987e-01 9.14621890e-01 -5.94505668e-01
7.29544044e-01 7.19412386e-01 -8.82215351e-02 -1.02783775e+00
-1.08918920e-01 -4.70220804e-01 -8.76701653e-01 -2.50237375e-01
-5.07937884e-03 -4.86665173e-03 -5.75313330e-01 1.57010710e+00
4.24704887e-02 1.17279656e-01 3.41046900e-01 8.48828495e-01
7.62748599e-01 3.80704492e-01 3.22995812e-01 -5.17280579e-01
8.68972778e-01 -1.07253861e+00 -6.19693875e-01 -3.59440565e-01
1.19491220e+00 -7.29953527e-01 1.43836236e+00 3.39403033e-01
-8.97563338e-01 -4.51188982e-01 -7.35201120e-01 1.90785661e-01
-2.07793355e-01 2.88478643e-01 7.03604579e-01 7.64047086e-01
-1.00936759e+00 6.89473748e-01 -3.95762533e-01 -3.72543007e-01
5.43211699e-01 3.81147057e-01 -8.24129060e-02 -1.32797971e-01
-9.48576272e-01 6.02068186e-01 5.33221543e-01 7.86230154e-03
-2.96002030e-01 -5.23629248e-01 -8.96022499e-01 2.43250951e-02
4.70095634e-01 -4.64239061e-01 1.38107693e+00 -1.64871967e+00
-1.19261611e+00 7.22138643e-01 -3.73883009e-01 -4.75722224e-01
3.52660656e-01 -3.14414859e-01 -3.00398737e-01 2.07242578e-01
2.71888584e-01 5.18612921e-01 8.71311665e-01 -1.36060166e+00
-2.09377289e-01 -2.49940395e-01 -2.00243652e-01 3.69544089e-01
-7.79725671e-01 2.65699942e-02 -1.07709199e-01 -8.44023407e-01
1.46478608e-01 -5.77078879e-01 -5.98669767e-01 -3.55599433e-01
-4.69653457e-01 -6.62842095e-01 5.83399415e-01 7.64283002e-04
1.21195984e+00 -2.16542387e+00 -3.43770206e-01 3.18005770e-01
3.56500186e-02 4.24335152e-01 -3.46862525e-01 3.87486368e-01
-4.33646530e-01 2.44087785e-01 -3.79090250e-01 -5.51683247e-01
-1.68767780e-01 9.14847776e-02 -4.74624306e-01 1.95445180e-01
6.23554468e-01 6.90391064e-01 -1.42302513e+00 -6.57903850e-01
2.84629345e-01 1.08274519e-01 -7.44454265e-01 1.53912872e-01
-1.71590209e-01 7.44097233e-02 -2.97084928e-01 3.51245075e-01
6.58662796e-01 -4.37084436e-01 5.74798763e-01 6.42491225e-03
2.77638197e-01 6.82071805e-01 -1.02792537e+00 1.34069657e+00
-3.33017021e-01 2.06784949e-01 -4.53492820e-01 -1.47293508e+00
9.48984921e-01 2.62866050e-01 1.82900816e-01 -7.58168399e-01
1.02857791e-01 1.19169772e-01 -8.09239522e-02 -4.51290935e-01
8.36741388e-01 -4.98547494e-01 2.67186705e-02 6.55421615e-01
-5.50940335e-02 -1.83784842e-01 3.19222480e-01 5.42095125e-01
8.62290144e-01 -8.83658156e-02 4.35438037e-01 -1.58365831e-01
6.43561602e-01 7.80046806e-02 4.89058554e-01 9.81401443e-01
-1.61839724e-01 9.37958300e-01 5.26537299e-01 -9.69111845e-02
-8.34972978e-01 -1.06120360e+00 1.14700466e-01 1.23200846e+00
-2.07293838e-01 -7.58989751e-01 -5.14653623e-01 -1.12108302e+00
-2.19569102e-01 8.07788968e-01 -5.71760535e-01 -3.68928432e-01
-4.00802612e-01 -8.38264763e-01 2.53005356e-01 6.81826949e-01
-9.04136002e-02 -1.21798098e+00 -9.96221453e-02 7.05959722e-02
-2.86241025e-01 -5.85405588e-01 -2.62539446e-01 4.07796472e-01
-9.71260726e-01 -1.08946693e+00 -4.10844237e-01 -1.02522695e+00
1.12485683e+00 7.81002462e-01 1.42633545e+00 3.12543452e-01
-1.26956508e-01 4.43774909e-01 -7.88567066e-01 -2.60697216e-01
-5.73539257e-01 2.73876134e-02 1.37107119e-01 -5.67752942e-02
7.97029734e-01 -3.11064571e-01 -4.31631595e-01 -1.42843634e-01
-9.42747235e-01 -3.28958213e-01 5.26923120e-01 1.08555508e+00
4.92274255e-01 -2.67020226e-01 1.22385967e+00 -1.43249023e+00
7.92162180e-01 -7.56215811e-01 -1.01684459e-01 3.49067539e-01
-5.94209671e-01 -9.17621031e-02 1.04378307e+00 -5.78571081e-01
-9.88081455e-01 -3.84090915e-02 -1.02467142e-01 -3.24313253e-01
-1.35436282e-01 4.87129658e-01 9.03757438e-02 2.96877623e-01
8.87955785e-01 3.13471168e-01 8.44078287e-02 -2.80532837e-01
2.92550355e-01 8.40285718e-01 2.33608738e-01 -5.96483231e-01
6.09522402e-01 2.93607175e-01 -5.67544103e-01 -8.88792515e-01
-1.10000563e+00 -7.74037540e-01 -6.21364772e-01 -6.14791773e-02
1.78997993e-01 -1.01062250e+00 -1.85292795e-01 -2.31717855e-01
-8.03811550e-01 -1.36028633e-01 -5.61251402e-01 4.32689250e-01
-3.88015926e-01 9.04574215e-01 -4.09217209e-01 -1.02842426e+00
-5.14978886e-01 -6.91840112e-01 9.81685698e-01 -2.12574396e-02
-7.06706703e-01 -9.54922140e-01 3.44696920e-03 2.91756630e-01
2.76094764e-01 -3.90250802e-01 8.48437309e-01 -1.22725415e+00
1.79058760e-02 -3.12955052e-01 -2.26305760e-02 6.19327962e-01
1.91773623e-01 -1.93009943e-01 -1.15581095e+00 -3.76601905e-01
1.00087896e-01 -8.22107136e-01 1.15909231e+00 2.10227177e-01
1.47230291e+00 -3.41288656e-01 -1.93729058e-01 -5.56230880e-02
1.23985744e+00 -2.10067436e-01 5.74694812e-01 1.72490776e-01
5.68981409e-01 8.20613921e-01 8.74926567e-01 7.62187183e-01
1.67926863e-01 2.48424977e-01 9.78794545e-02 -1.46020502e-01
2.68836111e-01 -3.99527065e-02 2.95311034e-01 9.68565285e-01
3.26767683e-01 -1.36723310e-01 -6.56060815e-01 7.36960411e-01
-1.82349026e+00 -1.14567256e+00 -3.42304677e-01 2.26785755e+00
1.10982597e+00 1.68762475e-01 1.81794092e-01 4.37134683e-01
8.34142506e-01 1.43959120e-01 -1.18659392e-01 -3.06247264e-01
-2.39188209e-01 3.37263316e-01 3.63066618e-04 1.80848509e-01
-1.21449459e+00 1.00966120e+00 6.12843370e+00 9.65110779e-01
-1.12058640e+00 -3.79124209e-02 8.21825087e-01 -2.26199403e-01
-6.80381477e-01 4.22407277e-02 -4.96088803e-01 3.84267211e-01
7.54008830e-01 8.42051767e-03 -9.01528895e-02 1.05673146e+00
2.65668482e-01 -3.67203471e-03 -1.04292583e+00 9.53150928e-01
3.57007295e-01 -1.22261155e+00 2.99091429e-01 -5.32063127e-01
8.20504844e-01 -2.06524320e-02 8.76469985e-02 6.50748909e-01
2.68401504e-01 -7.40907013e-01 5.26317060e-01 1.58673137e-01
3.72651666e-01 -9.03285146e-01 8.02942276e-01 3.08434337e-01
-8.11013520e-01 -1.08521588e-01 -6.76761448e-01 -2.62345314e-01
-1.82028830e-01 8.17110479e-01 -1.12349355e+00 3.74214083e-01
6.70017004e-01 9.67237115e-01 -8.96713078e-01 9.85381007e-01
-9.89131704e-02 9.76986349e-01 -8.91475603e-02 -3.20923209e-01
1.08334228e-01 -1.67136967e-01 1.45751521e-01 1.43854809e+00
-9.97647643e-02 -1.88670486e-01 3.07404757e-01 8.58940840e-01
-1.02985315e-01 5.94007552e-01 -6.83935523e-01 -4.50299643e-02
5.71288168e-01 1.17702246e+00 -8.95265639e-01 -7.35041261e-01
-4.86607462e-01 8.76797855e-01 6.89499319e-01 1.78273723e-01
-2.63313711e-01 -3.52167368e-01 2.04770774e-01 -1.45962194e-01
4.67505515e-01 2.56858319e-01 -5.94505966e-01 -1.22984350e+00
2.41190791e-01 -8.41840565e-01 5.68339229e-01 -3.25910717e-01
-1.87867522e+00 3.06064159e-01 -2.61993945e-01 -1.50987411e+00
-4.03514177e-01 -2.51865298e-01 -9.04368341e-01 4.87683773e-01
-1.40066791e+00 -6.29677653e-01 7.59794563e-02 3.78462315e-01
7.07327068e-01 -5.39353304e-02 8.39882374e-01 8.48591626e-02
-5.32167077e-01 8.94204736e-01 3.07371378e-01 8.46902281e-02
9.09586489e-01 -1.34431863e+00 2.68186808e-01 7.36675322e-01
6.39532432e-02 9.89687204e-01 4.31575924e-01 -6.42661273e-01
-9.51919734e-01 -1.24831522e+00 1.21900117e+00 -4.06515688e-01
4.67101157e-01 -4.35691774e-01 -1.20014989e+00 4.18826818e-01
8.98387134e-02 -4.81433794e-02 1.29233193e+00 2.42969289e-01
-4.55282420e-01 -3.07998713e-02 -1.14885187e+00 6.08722568e-01
8.04690719e-01 -5.24283528e-01 -7.42972672e-01 5.76647699e-01
6.46375656e-01 2.79943347e-01 -4.39694583e-01 3.57437670e-01
1.37250442e-02 -1.00669169e+00 9.44760501e-01 -6.16735101e-01
5.41624427e-01 -2.35065818e-01 -2.89871842e-02 -1.18468535e+00
-1.67309582e-01 -2.28231430e-01 1.49824530e-01 1.49849927e+00
4.88902807e-01 -5.61200559e-01 8.29634368e-01 3.07737619e-01
-1.91149995e-01 -6.35538161e-01 -6.24224961e-01 -7.80606985e-01
2.57058591e-01 -5.16363502e-01 2.48044312e-01 1.45395100e+00
4.85647321e-01 6.73004448e-01 -1.87792256e-01 -2.55628794e-01
4.66944158e-01 -1.72575004e-02 6.10219181e-01 -1.24723315e+00
-3.64627242e-01 -4.04294014e-01 -1.30497813e-01 -9.89920139e-01
3.53149235e-01 -1.15792179e+00 3.59227210e-01 -1.37116945e+00
5.91591418e-01 -5.23635149e-01 -3.09711307e-01 3.85416955e-01
-5.99481821e-01 4.40618277e-01 2.27049403e-02 3.04368585e-01
-1.10077906e+00 6.74853265e-01 7.80710220e-01 -1.08811982e-01
-1.42201811e-01 -1.12624578e-02 -9.78080094e-01 6.07253313e-01
9.32282388e-01 -4.94440645e-01 -6.08069122e-01 -2.01407135e-01
2.53190309e-01 -4.28180099e-01 1.46072432e-01 -7.54713953e-01
2.10567135e-02 -1.51728719e-01 4.60100323e-01 -7.60875940e-01
6.66870028e-02 -6.47512972e-01 -6.99000239e-01 3.86711836e-01
-7.79342830e-01 1.99956223e-02 -1.34643555e-01 7.09340513e-01
-3.22572142e-01 -6.47068322e-01 5.33391595e-01 -3.52762699e-01
-5.66844106e-01 -2.05121428e-01 -5.70402801e-01 2.98654020e-01
7.20644951e-01 -2.58893490e-01 -2.10451826e-01 -2.63385296e-01
-3.63867283e-01 2.38579199e-01 4.43049699e-01 5.62867105e-01
7.76274920e-01 -1.14810479e+00 -6.01070166e-01 2.48712450e-01
4.31973368e-01 1.06283903e-01 1.80693135e-01 8.76628578e-01
-2.20995732e-02 1.90729588e-01 4.76828456e-01 -6.94894075e-01
-1.17942667e+00 6.08688831e-01 -1.16329147e-02 -1.72902927e-01
-4.84819025e-01 9.00755942e-01 1.25383541e-01 -7.83908784e-01
3.16474199e-01 6.00322150e-03 -5.58356702e-01 1.48692116e-01
6.50359750e-01 5.28979162e-03 2.00479239e-01 -4.23392624e-01
-2.51075208e-01 7.03666806e-02 -4.70721841e-01 2.13294238e-01
1.39030278e+00 1.65178217e-02 -3.00686240e-01 6.94591463e-01
1.31226981e+00 -1.37880057e-01 -6.48073912e-01 -3.35314095e-01
5.06993651e-01 -4.57176805e-01 -4.01098549e-01 -5.42627096e-01
-7.29560971e-01 7.99540102e-01 3.16877276e-01 1.63560703e-01
8.53558540e-01 2.99092680e-02 5.43859184e-01 7.34888792e-01
1.59448572e-02 -1.37307787e+00 5.44715345e-01 6.63846016e-01
4.43650097e-01 -1.51438701e+00 2.03514919e-02 -6.30087733e-01
-8.81002069e-01 1.01226604e+00 8.36269975e-01 -3.00660551e-01
3.07877690e-01 -3.72965448e-02 1.25701874e-01 -6.22380264e-02
-8.51031661e-01 -1.22750998e-01 7.76316002e-02 6.32448018e-01
1.01600540e+00 9.78232548e-03 -5.81106961e-01 6.98939323e-01
-2.05824286e-01 -3.61482382e-01 5.47518492e-01 1.31734097e+00
-4.11976308e-01 -1.28458524e+00 -2.08589435e-01 8.80585194e-01
-4.81787413e-01 -4.68978703e-01 -7.87706077e-01 4.11014706e-01
-3.07072192e-01 1.19137323e+00 3.12307626e-01 -3.12820792e-01
9.94083658e-02 2.08526582e-01 3.57022554e-01 -1.21593571e+00
-8.79767656e-01 1.69027168e-02 2.31334016e-01 -1.17271915e-01
-5.62647521e-01 -6.76340520e-01 -1.22585106e+00 2.44384274e-01
-7.03989685e-01 5.35161912e-01 1.60344824e-01 1.05178189e+00
3.49944085e-01 2.72117108e-01 9.06880200e-01 -7.38368392e-01
-9.75552201e-01 -1.13940430e+00 -4.34721559e-01 8.76372576e-01
2.44514987e-01 -3.31484616e-01 -6.01252019e-01 -1.26879945e-01] | [10.939155578613281, 8.613167762756348] |
a347ad3c-cc5c-4baa-9d97-a6120131fea6 | 3d-to-2d-distillation-for-indoor-scene | 2104.02243 | null | https://arxiv.org/abs/2104.02243v2 | https://arxiv.org/pdf/2104.02243v2.pdf | 3D-to-2D Distillation for Indoor Scene Parsing | Indoor scene semantic parsing from RGB images is very challenging due to occlusions, object distortion, and viewpoint variations. Going beyond prior works that leverage geometry information, typically paired depth maps, we present a new approach, a 3D-to-2D distillation framework, that enables us to leverage 3D features extracted from large-scale 3D data repository (e.g., ScanNet-v2) to enhance 2D features extracted from RGB images. Our work has three novel contributions. First, we distill 3D knowledge from a pretrained 3D network to supervise a 2D network to learn simulated 3D features from 2D features during the training, so the 2D network can infer without requiring 3D data. Second, we design a two-stage dimension normalization scheme to calibrate the 2D and 3D features for better integration. Third, we design a semantic-aware adversarial training model to extend our framework for training with unpaired 3D data. Extensive experiments on various datasets, ScanNet-V2, S3DIS, and NYU-v2, demonstrate the superiority of our approach. Also, experimental results show that our 3D-to-2D distillation improves the model generalization. | ['Chi-Wing Fu', 'Xiaojuan Qi', 'Zhengzhe Liu'] | 2021-04-06 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liu_3D-to-2D_Distillation_for_Indoor_Scene_Parsing_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liu_3D-to-2D_Distillation_for_Indoor_Scene_Parsing_CVPR_2021_paper.pdf | cvpr-2021-1 | ['scene-parsing'] | ['computer-vision'] | [ 1.21987581e-01 3.73716623e-01 1.38851598e-01 -8.44170272e-01
-5.22087991e-01 -6.69221461e-01 3.66596401e-01 -4.99899209e-01
-4.45308715e-01 4.19211447e-01 1.59249723e-01 -4.20149744e-01
2.58118331e-01 -1.14559007e+00 -1.14107704e+00 -5.41089833e-01
1.83489412e-01 6.98793307e-02 3.62156898e-01 -3.11809659e-01
8.89791846e-02 7.37452686e-01 -1.29479265e+00 -8.58113319e-02
8.22937489e-01 1.41471386e+00 1.14630304e-01 5.45212090e-01
-3.38630468e-01 5.79621196e-01 -4.46889967e-01 -1.11063309e-01
8.11450720e-01 -2.89807349e-01 -5.02101421e-01 -5.51900864e-02
4.31660950e-01 -7.14525580e-01 -7.38790512e-01 9.62359667e-01
6.11866713e-01 9.33535770e-02 4.37875241e-01 -1.30210960e+00
-6.93464875e-01 2.75164898e-02 -4.21620131e-01 -3.89942348e-01
4.10916120e-01 1.91980898e-01 4.18474138e-01 -6.73740029e-01
7.05409467e-01 1.53567469e+00 7.66558170e-01 8.26061130e-01
-8.75392616e-01 -8.47248435e-01 3.67804855e-01 -6.69708177e-02
-1.28560162e+00 9.46571119e-03 1.43324900e+00 -1.87989026e-01
9.11992610e-01 1.00705884e-01 7.05988109e-01 1.39596236e+00
-2.34688982e-01 8.50259602e-01 1.34251988e+00 -2.02803850e-01
4.87300217e-01 -9.17074904e-02 -1.95063919e-01 7.49622107e-01
6.96400851e-02 3.22874397e-01 -2.78863549e-01 2.27308065e-01
1.12326968e+00 1.72661647e-01 -1.08672351e-01 -8.61151695e-01
-7.33504236e-01 8.13146651e-01 1.07144630e+00 -1.34956539e-01
-1.26024425e-01 2.83175230e-01 1.94616467e-01 2.55906999e-01
4.82273281e-01 2.00728998e-01 -7.55436361e-01 -6.57941326e-02
-3.09055805e-01 2.44475991e-01 5.92382967e-01 1.33989131e+00
1.12055647e+00 -2.22898707e-01 3.02428544e-01 5.10524213e-01
4.42795157e-01 8.80400658e-01 4.13149297e-01 -1.06114769e+00
8.13688755e-01 6.60872400e-01 -1.73326939e-01 -8.52776825e-01
-6.05603874e-01 3.03549729e-02 -7.77819872e-01 4.35074806e-01
1.20516412e-01 -5.20484522e-03 -1.21819222e+00 1.78637266e+00
4.79144245e-01 1.07020594e-01 3.41297001e-01 1.06122446e+00
9.98884857e-01 1.24349967e-01 -2.21226871e-01 5.81474483e-01
7.87610829e-01 -8.94054353e-01 -2.76278973e-01 -3.51706237e-01
5.85409880e-01 -2.64272213e-01 1.16140783e+00 5.00423871e-02
-6.03224814e-01 -6.20304763e-01 -1.23001587e+00 -4.68315035e-01
-7.06630528e-01 -3.88944536e-01 8.19382250e-01 7.26249874e-01
-7.86662519e-01 4.96201962e-01 -1.05118823e+00 -3.47527444e-01
8.37402046e-01 3.57374340e-01 -6.18673325e-01 -4.18286771e-01
-1.14660358e+00 6.37201250e-01 5.85646570e-01 3.48494090e-02
-7.49768496e-01 -5.74570715e-01 -1.38504755e+00 -5.06545722e-01
2.56763816e-01 -7.51624882e-01 9.44846392e-01 -3.93620312e-01
-1.64845705e+00 9.29568708e-01 4.85468172e-02 -2.05151498e-01
7.74288952e-01 -3.32309186e-01 1.07258320e-01 1.46663234e-01
1.44707458e-02 8.72804761e-01 4.79199439e-01 -1.50323915e+00
-2.42126763e-01 -8.31983924e-01 4.48801219e-01 3.65752637e-01
-5.12158014e-02 -7.55801976e-01 -6.33087993e-01 -5.32417774e-01
8.38198483e-01 -7.81939149e-01 -4.83499706e-01 3.19037229e-01
-7.88817286e-01 2.06432506e-01 9.08288062e-01 -5.23193121e-01
2.51800120e-01 -2.22510982e+00 1.06031001e-01 2.27073133e-01
1.42691359e-01 8.13925639e-02 -1.50707290e-01 -9.91142839e-02
1.10639356e-01 1.74073622e-01 -3.50532889e-01 -7.07487226e-01
2.20491439e-01 6.86754942e-01 -3.32895160e-01 4.96828169e-01
5.15763402e-01 1.20495844e+00 -9.10301685e-01 -2.78661430e-01
7.79001415e-01 6.80506885e-01 -7.80489206e-01 4.25524741e-01
-1.21753916e-01 8.76691401e-01 -8.59764397e-01 8.34950268e-01
1.15357697e+00 9.98922810e-02 -4.24034595e-01 -2.23589301e-01
1.74837872e-01 4.25987780e-01 -9.92669106e-01 2.58330750e+00
-6.37119949e-01 1.54958695e-01 -2.94070572e-01 -9.86221373e-01
1.12254024e+00 -2.31765598e-01 3.74225646e-01 -1.07242870e+00
3.51778895e-01 1.41755551e-01 -6.57753289e-01 -5.85987628e-01
3.02961200e-01 -1.64212547e-02 -4.58081603e-01 8.13748538e-02
1.30100533e-01 -8.93229008e-01 -5.83971739e-01 8.64404999e-03
1.17526424e+00 6.94134176e-01 -6.23604767e-02 3.10382158e-01
3.80738080e-01 -1.64743021e-01 6.53414905e-01 7.26496816e-01
-2.82787710e-01 9.89978909e-01 4.22651023e-01 -4.88685846e-01
-1.06464005e+00 -1.32981277e+00 -6.98049217e-02 3.37102026e-01
6.41136050e-01 -5.20242751e-02 -6.97550058e-01 -1.13424742e+00
2.04016343e-01 5.86872339e-01 -8.44978273e-01 -3.31220537e-01
-6.62571907e-01 -4.59327102e-01 6.29336894e-01 9.17714953e-01
1.19404352e+00 -4.99381185e-01 -7.64349699e-01 -8.24555680e-02
8.40674862e-02 -1.47837937e+00 -1.38499802e-02 4.58088756e-01
-9.87483203e-01 -1.07710743e+00 -4.77901906e-01 -5.22343814e-01
7.75261462e-01 2.64752418e-01 8.37466061e-01 -2.62092084e-01
-2.06916243e-01 5.30939996e-01 -6.38946056e-01 -5.24286926e-01
-8.40400755e-02 3.85518223e-02 3.29367397e-03 -5.30315638e-01
4.75575894e-01 -9.14424121e-01 -4.88130361e-01 3.00374418e-01
-7.63189495e-01 2.29554519e-01 5.42159438e-01 5.47277868e-01
9.32292342e-01 -1.81768447e-01 6.95502236e-02 -7.14564979e-01
-1.21010445e-01 -2.54663110e-01 -5.45935988e-01 -1.81854546e-01
-3.01119894e-01 1.54069856e-01 6.12655938e-01 -3.06852013e-01
-9.00547922e-01 4.28318083e-01 -4.84655261e-01 -7.61236906e-01
-5.53672612e-01 -9.36529785e-02 -8.90883505e-01 -3.57868552e-01
5.23923755e-01 2.11816967e-01 -1.03330858e-01 -5.96589327e-01
7.45386064e-01 5.38201213e-01 7.30800033e-01 -6.36721671e-01
1.38138306e+00 7.78397501e-01 1.79754987e-01 -4.23142582e-01
-8.74528170e-01 -2.21526712e-01 -1.00726128e+00 1.06851928e-01
1.10305572e+00 -1.18699217e+00 -4.86721456e-01 7.51553595e-01
-1.21276689e+00 -4.97772068e-01 -4.98886168e-01 5.05578876e-01
-7.29535282e-01 1.24817289e-01 -3.34825128e-01 -5.59678495e-01
-2.49413736e-02 -1.21778810e+00 1.37268770e+00 3.52556318e-01
2.73953021e-01 -9.96733487e-01 -1.42829493e-01 3.22393656e-01
3.10990781e-01 9.35548186e-01 6.54251397e-01 -5.44403613e-01
-7.65430331e-01 -1.82019800e-01 -3.88907343e-01 6.11861944e-01
3.91723633e-01 -6.03507519e-01 -1.33768106e+00 8.05486888e-02
2.39774168e-01 -6.41626418e-01 6.97827220e-01 -2.65784096e-02
1.42474806e+00 6.70877621e-02 -1.77605197e-01 1.52248490e+00
1.32631314e+00 -1.17470734e-02 5.60258269e-01 6.02873921e-01
1.23198259e+00 2.88479656e-01 6.02500141e-01 2.13962689e-01
8.05025339e-01 4.06311452e-01 1.03411067e+00 -2.90598065e-01
-3.28932077e-01 -7.34932005e-01 -4.62805852e-02 7.41559863e-01
7.75524229e-02 9.56565961e-02 -6.47064507e-01 2.03976393e-01
-1.54145980e+00 -2.45113015e-01 1.99967965e-01 1.84185481e+00
6.03453159e-01 2.40802839e-01 -3.48333299e-01 1.55988306e-01
3.13125014e-01 2.73012727e-01 -1.10182130e+00 -1.84503570e-02
-2.38538757e-01 4.33173627e-01 8.11820567e-01 4.68365014e-01
-1.16282880e+00 1.14414573e+00 5.05882502e+00 3.27810973e-01
-1.17527914e+00 5.54431370e-03 4.31369513e-01 -1.44434363e-01
-4.97731715e-01 -1.42797202e-01 -5.64720333e-01 3.41833293e-01
3.41356725e-01 3.12759340e-01 4.11574632e-01 1.11677539e+00
-7.57673457e-02 1.66607335e-01 -1.11873186e+00 1.15779388e+00
1.17810294e-01 -1.01432180e+00 -1.56335011e-02 2.14076281e-01
6.56891406e-01 2.39767984e-01 -1.65047254e-02 4.69821662e-01
4.93587136e-01 -8.47801983e-01 8.78362000e-01 2.66034603e-01
8.84310544e-01 -7.41819620e-01 5.94236910e-01 2.99631357e-01
-1.09096670e+00 7.97740147e-02 -5.27424216e-01 -6.84766248e-02
1.26091823e-01 5.28763950e-01 -5.71762145e-01 1.00270510e+00
9.04675186e-01 9.20086980e-01 -5.64885795e-01 5.77589154e-01
-7.71079063e-01 9.05833468e-02 -5.83913803e-01 2.59105742e-01
2.14270443e-01 -5.75102083e-02 2.20755056e-01 6.53985977e-01
3.24224979e-01 2.41060555e-01 -4.18729223e-02 1.06200790e+00
-2.21824735e-01 -3.45047593e-01 -9.41360533e-01 4.50342506e-01
6.12149954e-01 9.34603214e-01 -6.77076638e-01 -1.60359945e-02
-4.05797243e-01 1.43779266e+00 3.42393935e-01 3.37877184e-01
-9.53997552e-01 -7.48838305e-01 8.93054962e-01 -2.93300152e-01
5.43921411e-01 -4.90045577e-01 -4.54816639e-01 -1.24846613e+00
3.57655525e-01 -3.11534643e-01 -6.96800575e-02 -9.02554274e-01
-1.17912006e+00 5.91265142e-01 -1.74533743e-02 -1.15138960e+00
-2.08768368e-01 -8.87134194e-01 -4.31333363e-01 9.00292099e-01
-1.94483066e+00 -1.32618237e+00 -8.84122312e-01 8.12745452e-01
1.86005175e-01 1.69815749e-01 7.47621298e-01 2.39290714e-01
-3.31391275e-01 6.37047350e-01 -2.60352552e-01 5.52386165e-01
4.11050856e-01 -1.30343127e+00 9.30367529e-01 6.85713649e-01
-1.69552684e-01 2.46547997e-01 2.07586512e-01 -3.59721363e-01
-1.63513064e+00 -1.37731183e+00 1.90898240e-01 -7.65645623e-01
2.34863400e-01 -7.86804378e-01 -6.14352524e-01 6.90624356e-01
-5.70763886e-01 4.75417703e-01 5.67149162e-01 -3.47263455e-01
-6.57086492e-01 -1.10259168e-01 -1.66783202e+00 3.24124247e-01
1.77432561e+00 -7.26593554e-01 -5.98969042e-01 9.93519425e-02
1.57170081e+00 -9.95685697e-01 -9.14791107e-01 7.10524738e-01
3.92469496e-01 -1.09927964e+00 1.30826426e+00 -4.00692195e-01
3.68586421e-01 -1.80095926e-01 -8.57758522e-01 -1.26914716e+00
2.66524523e-01 -2.63164461e-01 -7.79550821e-02 1.19296539e+00
-5.75711690e-02 -8.23357880e-01 1.09268308e+00 7.28271008e-01
-3.52727175e-01 -6.33068621e-01 -1.20569086e+00 -9.48638320e-01
2.99426883e-01 -8.47815931e-01 1.20537281e+00 8.57439876e-01
-8.64471614e-01 -6.70560747e-02 -1.22930827e-02 5.25116444e-01
7.31961071e-01 1.55878812e-01 1.22936833e+00 -1.05399990e+00
-1.21449210e-01 -1.34030670e-01 -9.07328963e-01 -1.58898163e+00
2.75483817e-01 -8.64348531e-01 5.40718026e-02 -1.50646186e+00
-3.92964274e-01 -6.90894663e-01 -2.60902792e-01 5.47835886e-01
-2.56191492e-02 3.19250613e-01 1.81680292e-01 -1.44946799e-01
-3.47303063e-01 1.02339697e+00 1.57770455e+00 -1.54078290e-01
-3.13913286e-01 -1.64104715e-01 -7.06707001e-01 7.34838545e-01
6.15220726e-01 -3.84749711e-01 -6.01551771e-01 -8.69775295e-01
-1.74113020e-01 -3.71951342e-01 6.51781261e-01 -1.16719651e+00
-1.12476662e-01 -1.29130468e-01 7.44330168e-01 -7.47653067e-01
5.98771870e-01 -1.04334462e+00 -4.33681428e-01 1.30248562e-01
8.07220936e-02 -3.04100662e-01 2.47475550e-01 4.67666328e-01
-2.33859406e-04 1.56707406e-01 5.14300942e-01 -2.64328420e-01
-9.30579484e-01 5.82828224e-01 4.78871167e-01 1.35385290e-01
9.32192743e-01 -3.01054358e-01 -1.61185294e-01 -1.54810816e-01
-4.39438999e-01 2.23066017e-01 7.66361892e-01 6.28421009e-01
7.75209904e-01 -1.61138606e+00 -1.86090440e-01 6.70061529e-01
2.22114667e-01 1.19106829e+00 1.63717628e-01 2.71725178e-01
-7.20891654e-01 1.48815840e-01 -2.42480859e-01 -8.76879454e-01
-6.31776273e-01 5.64584553e-01 3.35371405e-01 1.61843851e-01
-8.40742528e-01 1.20872021e+00 3.19743931e-01 -1.25951862e+00
3.38918746e-01 -7.61011541e-01 4.02964026e-01 -4.94217068e-01
2.58760959e-01 -2.30353442e-03 -2.27070134e-03 -5.24431765e-01
-5.84656358e-01 9.76789892e-01 2.41840526e-01 -1.01370692e-01
1.57257056e+00 -2.77088463e-01 1.82115421e-01 3.13144296e-01
1.61775923e+00 -2.94260144e-01 -1.76055813e+00 -2.57407904e-01
-5.65013170e-01 -7.26872861e-01 2.81509757e-02 -5.50999522e-01
-1.31740117e+00 1.04729640e+00 7.18728304e-01 -2.19062507e-01
1.28858829e+00 1.08223170e-01 9.32991087e-01 6.18632793e-01
7.60909259e-01 -7.67244220e-01 1.44210979e-01 5.94610035e-01
6.32437825e-01 -1.43436646e+00 -1.70404494e-01 -4.48076099e-01
-2.74111927e-01 9.48215187e-01 8.47188950e-01 -1.94842428e-01
6.87723041e-01 1.61800742e-01 3.40458184e-01 -4.30612490e-02
3.89847234e-02 -1.70059085e-01 -1.12802729e-01 1.06193888e+00
-3.56626391e-01 -1.00338466e-01 5.22574544e-01 6.83252037e-01
-6.06603563e-01 -8.93073231e-02 2.14343846e-01 1.25878191e+00
-9.88539010e-02 -1.13267469e+00 -2.45549604e-01 -1.90813139e-01
1.91229269e-01 7.78642818e-02 -3.37067485e-01 9.80826974e-01
4.51908350e-01 6.34680569e-01 1.34704292e-01 -9.37396467e-01
6.80551410e-01 -1.09280467e-01 7.94885755e-01 -4.46486622e-01
2.57118102e-02 -3.61096472e-01 -4.63116854e-01 -1.13967097e+00
-3.15044522e-01 -2.81170398e-01 -1.50194204e+00 -8.71084929e-02
1.62494723e-02 -3.32284451e-01 1.30590522e+00 9.14573014e-01
3.05722266e-01 6.93647683e-01 9.28461552e-01 -1.22053051e+00
-3.87137443e-01 -8.50477278e-01 -4.11405921e-01 5.52446127e-01
5.36673963e-01 -8.02642524e-01 -4.34186816e-01 -2.14224890e-01] | [8.379837036132812, -2.9104576110839844] |
3c13b429-fe0d-4648-a582-4ce2f20b3f8b | evaluation-and-comparison-of-deep-learning | 2112.10390 | null | https://arxiv.org/abs/2112.10390v1 | https://arxiv.org/pdf/2112.10390v1.pdf | Evaluation and Comparison of Deep Learning Methods for Pavement Crack Identification with Visual Images | Compared with contact detection techniques, pavement crack identification with visual images via deep learning algorithms has the advantages of not being limited by the material of object to be detected, fast speed and low cost. The fundamental frameworks and typical model architectures of transfer learning (TL), encoder-decoder (ED), generative adversarial networks (GAN), and their common modules were first reviewed, and then the evolution of convolutional neural network (CNN) backbone models and GAN models were summarized. The crack classification, segmentation performance, and effect were tested on the SDNET2018 and CFD public data sets. In the aspect of patch sample classification, the fine-tuned TL models can be equivalent to or even slightly better than the ED models in accuracy, and the predicting time is faster; In the aspect of accurate crack location, both ED and GAN algorithms can achieve pixel-level segmentation and is expected to be detected in real time on low computing power platform. Furthermore, a weakly supervised learning framework of combined TL-SSGAN and its performance enhancement measures are proposed, which can maintain comparable crack identification performance with that of the supervised learning, while greatly reducing the number of labeled samples required. | ['Kai-Liang Lu'] | 2021-12-20 | null | null | null | null | ['contact-detection'] | ['robots'] | [ 1.75335124e-01 9.36294049e-02 5.15865069e-03 1.42587572e-02
-9.13905799e-01 -1.19532801e-01 5.06222481e-04 -3.10116798e-01
-1.73423752e-01 6.00069344e-01 -4.25566018e-01 -2.27513120e-01
4.27862972e-01 -1.38543522e+00 -6.67388797e-01 -9.93753970e-01
2.73726702e-01 1.93542868e-01 6.28988802e-01 -1.19278960e-01
3.50749940e-01 4.37064320e-01 -1.42732835e+00 2.53629476e-01
1.04149616e+00 1.38190663e+00 9.32282954e-02 5.90416253e-01
1.96948245e-01 1.07446706e+00 -5.25049269e-01 -3.19977969e-01
-5.92299774e-02 -3.13946009e-01 -6.68914258e-01 -9.65694189e-02
3.41906011e-01 -5.37715971e-01 -6.29548132e-01 9.19200897e-01
8.58862579e-01 -3.41714829e-01 8.75851393e-01 -1.05962276e+00
-8.28992844e-01 4.89607215e-01 -7.10092962e-01 5.58125926e-03
2.87916325e-02 4.35928941e-01 6.17549837e-01 -9.07244742e-01
2.51897365e-01 1.04275942e+00 1.21688128e+00 5.73527217e-01
-8.65428805e-01 -7.17312336e-01 -2.68485069e-01 4.23420221e-01
-1.23972058e+00 7.21892565e-02 1.16371489e+00 -5.45150578e-01
5.32581031e-01 1.05245598e-01 6.28334284e-01 1.03747523e+00
2.76638478e-01 8.07036936e-01 9.47177052e-01 -4.09216404e-01
3.38543087e-01 -2.99775153e-01 9.89528000e-03 1.06087112e+00
3.22145164e-01 2.75823385e-01 -1.37407184e-02 1.86435327e-01
1.08390558e+00 -4.01542895e-02 -1.90668970e-01 5.04085980e-02
-8.24311852e-01 9.06576157e-01 7.73719370e-01 1.66669354e-01
-5.25669679e-02 4.98453856e-01 5.03932536e-01 2.39955142e-01
3.86020809e-01 6.50275350e-02 -5.33723757e-02 2.22263798e-01
-7.82048643e-01 -2.95244791e-02 3.79700929e-01 8.50395560e-01
9.16412771e-01 7.29953051e-01 -2.12026592e-02 8.91104758e-01
3.72616887e-01 1.01652658e+00 3.35742563e-01 -8.46792281e-01
3.07507843e-01 6.49500132e-01 -3.17511886e-01 -1.11408174e+00
-3.07609469e-01 -2.58549035e-01 -1.09993815e+00 6.43871963e-01
1.28958702e-01 -3.32634777e-01 -1.11231017e+00 1.03564298e+00
6.76095346e-03 2.70121306e-01 -7.52900615e-02 6.38186276e-01
1.06033087e+00 6.19707525e-01 -6.12471290e-02 3.64846051e-01
1.25805700e+00 -1.04050338e+00 -6.47225738e-01 -5.66388786e-01
5.18125176e-01 -5.63892126e-01 1.01497662e+00 1.94068685e-01
-9.26574826e-01 -8.13875616e-01 -1.51698875e+00 7.79182557e-03
-2.75810868e-01 4.65703487e-01 3.31834942e-01 8.69723558e-01
-6.96918368e-01 5.24879038e-01 -7.56661117e-01 -2.26132479e-02
9.75027263e-01 -3.44797671e-02 -7.19676539e-03 -3.13841164e-01
-1.20115554e+00 8.28610301e-01 1.99480042e-01 4.57823873e-01
-1.20801401e+00 -5.46903849e-01 -9.12207305e-01 -3.61378789e-02
-8.53406787e-02 -3.62356573e-01 8.71168315e-01 -6.72898352e-01
-1.40635514e+00 7.91982532e-01 3.81764472e-01 -4.00626332e-01
5.35579562e-01 -1.43645778e-01 -4.66126531e-01 4.95323598e-01
6.62974417e-02 3.98541182e-01 1.10283887e+00 -1.55741680e+00
-5.61120749e-01 9.09886230e-03 -1.56642213e-01 -3.80644202e-01
-1.84635650e-02 -4.16114450e-01 1.43723860e-01 -7.23167360e-01
8.49241242e-02 -6.16911829e-01 -1.12644911e-01 3.60737681e-01
-3.51993263e-01 -3.78849097e-02 1.51006627e+00 -9.61152375e-01
9.58085477e-01 -2.30054379e+00 -2.50978947e-01 1.81955546e-01
9.13937837e-02 4.25189793e-01 -2.14269222e-03 4.16299075e-01
-2.49355379e-02 3.13054100e-02 -1.03283191e+00 -1.04211882e-01
-3.86411190e-01 3.72370780e-01 1.01778299e-01 7.09310949e-01
4.16518539e-01 1.19236624e+00 -6.55825078e-01 -7.31646121e-01
4.06088710e-01 1.07784159e-01 -3.01212400e-01 2.54104763e-01
1.09300077e-01 3.53496581e-01 -2.87796706e-01 1.06154358e+00
9.70562339e-01 -2.75583677e-02 -4.50316727e-01 -5.57267189e-01
9.95121598e-02 -4.80981886e-01 -1.14341402e+00 1.31193697e+00
-6.11152709e-01 7.59825230e-01 1.07346602e-01 -1.12843955e+00
1.24505985e+00 3.77248108e-01 2.59141535e-01 -7.94663608e-01
3.10364008e-01 3.77982676e-01 -3.28613728e-01 -8.80247712e-01
2.55203210e-02 8.69102180e-02 -4.22603600e-02 4.13958222e-01
-2.33474731e-01 -2.18460575e-01 -2.08731145e-01 -3.33596319e-01
8.31479311e-01 -5.87354228e-03 -5.38377345e-01 -1.88014805e-01
3.97704422e-01 5.84192909e-02 2.08963647e-01 4.00637835e-01
7.23524392e-02 9.64229763e-01 2.38189027e-01 -3.72742444e-01
-1.15031660e+00 -1.05724072e+00 -2.27152169e-01 3.12557131e-01
5.94875336e-01 4.32931542e-01 -1.01207888e+00 -5.74051559e-01
-4.48433720e-02 3.05143833e-01 -8.40417504e-01 -5.11900425e-01
-8.69551361e-01 -7.48042047e-01 1.14512813e+00 1.06744933e+00
1.18003345e+00 -1.27521062e+00 -7.72625327e-01 8.87963176e-02
-1.18047453e-01 -1.07462764e+00 -2.79287457e-01 -3.92221361e-02
-7.54437923e-01 -1.49967730e+00 -7.84775496e-01 -1.39777732e+00
7.55192399e-01 -1.29256293e-01 8.00814867e-01 5.77825904e-01
-5.39374292e-01 5.59775457e-02 -6.06098413e-01 -3.25879037e-01
-5.21752417e-01 -1.86101459e-02 -6.93926215e-01 2.30677024e-01
-1.79715514e-01 -3.10980320e-01 -9.34743524e-01 3.02334160e-01
-9.08143282e-01 -7.66182244e-02 7.50421286e-01 1.23280859e+00
1.98600218e-01 1.35365337e-01 7.34255254e-01 -8.76441360e-01
2.83971429e-01 -2.18340680e-01 -1.35122314e-01 1.11872196e-01
-9.47833061e-01 -3.99936289e-01 3.28339845e-01 -1.95273682e-01
-1.25101626e+00 -2.09757816e-02 -6.76292062e-01 -3.14881027e-01
-8.94673765e-02 5.16324222e-01 -5.48527166e-02 -3.28886390e-01
6.94990933e-01 1.74480200e-01 2.81373203e-01 -4.29355830e-01
1.65140390e-01 8.12353253e-01 6.39287055e-01 -6.34705365e-01
9.75241423e-01 6.35053158e-01 -2.34591886e-02 -8.37214410e-01
-5.67846835e-01 1.38237691e-02 -7.58751810e-01 -6.39128149e-01
1.12550318e+00 -9.63488221e-01 -3.95691276e-01 1.50075293e+00
-1.07072628e+00 -5.12191653e-01 -5.12302279e-01 1.52093545e-01
-4.64103997e-01 7.12203801e-01 -1.08294773e+00 -7.13349223e-01
-7.25243390e-01 -1.14294410e+00 9.85521913e-01 3.23846340e-01
4.57192600e-01 -9.62921798e-01 -3.20353568e-01 4.35146898e-01
5.57802200e-01 8.21958899e-01 1.30836499e+00 9.35637131e-02
-3.50466132e-01 -6.36014521e-01 -4.80499506e-01 9.83412445e-01
9.38503817e-02 1.95938244e-01 -1.17485690e+00 -2.03678966e-01
-8.51451978e-02 -1.98633522e-01 9.13657069e-01 5.38530231e-01
1.19137335e+00 1.50021594e-02 -4.66159165e-01 2.68657178e-01
1.64820230e+00 3.74476969e-01 1.16331470e+00 1.57843739e-01
9.94364977e-01 2.82986730e-01 4.93771225e-01 -2.51465775e-02
1.28188953e-01 2.74055362e-01 1.02193332e+00 -8.71978939e-01
-5.56840241e-01 -2.24137381e-01 1.38001725e-01 7.81781852e-01
-2.66281366e-01 -1.12505585e-01 -9.45705712e-01 6.88307345e-01
-1.57674539e+00 -1.12916541e+00 -6.51721895e-01 1.73653591e+00
5.58976591e-01 8.00606608e-02 4.70646620e-02 6.66364193e-01
1.03911865e+00 1.00319095e-01 -5.69060445e-01 -2.34779716e-01
-3.23094487e-01 5.30182958e-01 7.41224527e-01 1.53371543e-01
-1.08517230e+00 6.50384247e-01 6.39556503e+00 1.14048696e+00
-1.12772202e+00 4.00146157e-01 8.92461658e-01 7.51392007e-01
-1.16644889e-01 -1.50513157e-01 -1.44174501e-01 6.10818982e-01
5.17933667e-01 6.77495241e-01 1.45869285e-01 9.35295224e-01
-2.87709534e-01 7.90079311e-02 -9.35969651e-01 8.85386407e-01
1.17955077e-02 -1.37234354e+00 -1.86203733e-01 -6.54618740e-02
7.77172208e-01 -4.29666787e-02 -1.19143510e-02 1.31875366e-01
-5.41302636e-02 -1.11887336e+00 9.19035256e-01 4.25999105e-01
1.23914671e+00 -9.19689894e-01 1.20575118e+00 1.43656790e-01
-1.24529243e+00 -4.40756738e-01 -3.74851733e-01 1.35834977e-01
1.62689090e-01 4.87325907e-01 -2.75787830e-01 7.69410312e-01
1.03272223e+00 9.09952104e-01 -6.56383395e-01 8.93799305e-01
-3.00167263e-01 1.03266656e+00 -1.12023912e-01 1.86838388e-01
3.14070672e-01 -8.32026675e-02 -5.02933934e-03 1.25200999e+00
3.38893235e-01 -2.27520525e-01 1.19777154e-02 9.13739085e-01
7.11053237e-02 -2.43561074e-01 -3.42546523e-01 5.34010530e-01
5.48842669e-01 1.07398558e+00 -7.54408479e-01 -1.83886319e-01
-4.21476215e-01 6.94745898e-01 -1.59320161e-01 7.60928122e-03
-1.20207465e+00 -5.02635181e-01 2.45953232e-01 3.70120764e-01
3.13677192e-01 1.31564498e-01 -7.54385412e-01 -6.10341847e-01
-1.88357472e-01 -6.14907682e-01 1.53638899e-01 -7.29962051e-01
-1.48625875e+00 4.45376158e-01 -3.60766053e-01 -1.53762043e+00
5.08483112e-01 -7.36658335e-01 -1.13854289e+00 4.52292651e-01
-1.52135110e+00 -1.63309097e+00 -7.18543530e-01 4.45358872e-01
6.95267916e-01 3.77962776e-02 4.80012059e-01 7.40020752e-01
-6.44803941e-01 7.74104714e-01 7.30595589e-02 8.84329259e-01
3.60677302e-01 -1.08751166e+00 4.10222620e-01 9.78156686e-01
-3.95710647e-01 -2.63454467e-01 1.60797074e-01 -5.40346682e-01
-9.60585833e-01 -1.48125911e+00 2.88105905e-02 -8.66372287e-02
2.05407158e-01 -9.00889784e-02 -1.11126769e+00 3.39222103e-01
3.18556756e-01 -9.85403778e-04 1.23521551e-01 -6.01954222e-01
-1.00210607e-01 -1.67411342e-01 -1.41505229e+00 2.04261169e-01
6.78114474e-01 -4.85667199e-01 -2.65443891e-01 1.81072965e-01
4.18721795e-01 -6.16368830e-01 -1.13296068e+00 5.56403100e-01
5.60101032e-01 -8.98990333e-01 1.11693943e+00 1.77105278e-01
8.99840653e-01 -2.67030716e-01 3.01513877e-02 -9.96385336e-01
-3.88220608e-01 1.27313301e-01 7.24701909e-03 1.38523972e+00
2.78010190e-01 -8.14854503e-01 9.37831521e-01 1.22588150e-01
-7.72526145e-01 -9.34784830e-01 -1.22757745e+00 -5.07263243e-01
3.88015985e-01 -2.19532549e-01 8.36715758e-01 8.74020338e-01
-4.76168424e-01 1.68317109e-01 -4.78931308e-01 4.12265271e-01
6.98605001e-01 1.05293721e-01 3.85885030e-01 -1.39005518e+00
6.19029254e-02 -4.19556856e-01 -4.83566076e-01 -7.52322018e-01
-1.97446972e-01 -6.46633863e-01 9.26373079e-02 -1.60764539e+00
-1.64417803e-01 -6.16446555e-01 -1.55311033e-01 6.15534663e-01
-5.61821721e-02 5.89700758e-01 -2.37113163e-01 1.75636709e-01
1.74143985e-01 5.22611141e-01 1.65995443e+00 -7.62178063e-01
3.96776110e-01 -1.56071084e-02 -4.78627086e-02 5.77134550e-01
7.33309805e-01 -5.11032999e-01 -4.84775454e-02 -4.87429261e-01
2.36191899e-02 6.31408095e-02 8.62547576e-01 -1.42871296e+00
2.53930748e-01 3.90235931e-01 3.71328413e-01 -7.02398419e-01
1.29084185e-01 -9.51823771e-01 1.80288345e-01 1.11995018e+00
3.46361697e-02 -2.32126713e-01 1.16657697e-01 6.37619615e-01
-3.64726484e-01 -6.02047324e-01 1.37406993e+00 -1.41557127e-01
-8.15270364e-01 2.54108995e-01 -5.98852813e-01 -4.62704115e-02
1.26196206e+00 -7.78924942e-01 -5.81358016e-01 -1.21943630e-01
-3.64637136e-01 -1.01226479e-01 4.39718753e-01 2.92048156e-01
9.19461668e-01 -1.30351567e+00 -8.19443583e-01 4.08186495e-01
4.83162664e-02 5.02504468e-01 6.85907006e-01 7.40010619e-01
-1.32086468e+00 -2.14526832e-01 -3.90852451e-01 -7.47582614e-01
-7.36894846e-01 3.76541823e-01 5.90183496e-01 2.03756206e-02
-7.69853294e-01 7.20745087e-01 -7.04580396e-02 -2.74573028e-01
-1.66137487e-01 -3.69127452e-01 -1.58355892e-01 -1.36722267e-01
4.74028327e-02 8.75763416e-01 2.39080042e-01 -4.64802206e-01
-9.90639850e-02 9.96665120e-01 1.86826870e-01 8.29768121e-01
1.42544961e+00 1.60015672e-01 -1.47684485e-01 2.24328251e-03
1.14164984e+00 -3.56767535e-01 -1.34216595e+00 -4.12985012e-02
-4.20309037e-01 -1.24682933e-01 1.57605141e-01 -6.28863811e-01
-1.97627807e+00 1.07231617e+00 1.35909319e+00 2.38264009e-01
1.12944150e+00 9.12745222e-02 1.37794113e+00 -2.05480993e-01
3.48883629e-01 -1.15496802e+00 3.71287495e-01 -3.03606614e-02
6.44998372e-01 -1.21075737e+00 -2.22477153e-01 -5.84940016e-01
-5.65729916e-01 1.22457850e+00 7.97549903e-01 -5.17292678e-01
7.59980559e-01 6.02982879e-01 1.80034488e-01 -7.12306201e-01
2.60536149e-02 1.65624946e-01 -1.22935936e-01 8.34080458e-01
-6.77240565e-02 -2.01154426e-01 1.15322046e-01 2.48552561e-01
3.74341905e-01 -4.05971974e-01 4.53563601e-01 8.94600749e-01
-5.09281397e-01 -7.30187535e-01 -3.52922946e-01 6.99167430e-01
-4.22670305e-01 1.01223722e-01 9.05819759e-02 9.13077414e-01
4.26506519e-01 9.82879043e-01 1.79577600e-02 -5.77286899e-01
4.57374424e-01 -3.84322017e-01 1.69472754e-01 -4.31183577e-01
-5.77538013e-01 -3.42459053e-01 -1.78055242e-01 -1.82326138e-01
-4.86355335e-01 -3.79891187e-01 -1.14930308e+00 -2.04126090e-01
-8.12539220e-01 -5.14185801e-02 3.10564339e-01 9.31238830e-01
3.65374759e-02 6.47968531e-01 9.35982823e-01 -8.66027832e-01
-3.08799416e-01 -1.13533878e+00 -7.79783189e-01 2.44564131e-01
2.01172441e-01 -6.53960407e-01 -3.85881543e-01 3.36668670e-01] | [7.506360054016113, 1.5549601316452026] |
60155ccb-71f4-458e-afef-48bf920d7a1c | derivation-of-information-theoretically | 2007.14042 | null | https://arxiv.org/abs/2007.14042v1 | https://arxiv.org/pdf/2007.14042v1.pdf | Derivation of Information-Theoretically Optimal Adversarial Attacks with Applications to Robust Machine Learning | We consider the theoretical problem of designing an optimal adversarial attack on a decision system that maximally degrades the achievable performance of the system as measured by the mutual information between the degraded signal and the label of interest. This problem is motivated by the existence of adversarial examples for machine learning classifiers. By adopting an information theoretic perspective, we seek to identify conditions under which adversarial vulnerability is unavoidable i.e. even optimally designed classifiers will be vulnerable to small adversarial perturbations. We present derivations of the optimal adversarial attacks for discrete and continuous signals of interest, i.e., finding the optimal perturbation distributions to minimize the mutual information between the degraded signal and a signal following a continuous or discrete distribution. In addition, we show that it is much harder to achieve adversarial attacks for minimizing mutual information when multiple redundant copies of the input signal are available. This provides additional support to the recently proposed ``feature compression" hypothesis as an explanation for the adversarial vulnerability of deep learning classifiers. We also report on results from computational experiments to illustrate our theoretical results. | ['Raghu Mudumbai', 'Jirong Yi', 'Weiyu Xu'] | 2020-07-28 | null | null | null | null | ['feature-compression'] | ['computer-vision'] | [ 7.21836984e-01 6.54397428e-01 3.18560869e-01 -2.37737447e-01
-1.03080285e+00 -1.03514695e+00 4.17446941e-01 -5.38985729e-02
-2.20664546e-01 5.81840634e-01 1.85399633e-02 -3.81559014e-01
-2.87850112e-01 -6.91518605e-01 -9.49576557e-01 -1.13331234e+00
-3.33914399e-01 -1.05802551e-01 -2.29691461e-01 -1.28681287e-01
-9.90214106e-03 7.76065648e-01 -1.35859120e+00 2.11051673e-01
5.22108793e-01 1.02086890e+00 -4.00765538e-01 9.62666214e-01
6.30322456e-01 6.61399782e-01 -1.08632910e+00 -4.67527002e-01
8.03062499e-01 -7.40476310e-01 -7.41594136e-01 6.42848834e-02
5.55489540e-01 -1.32298797e-01 -8.28660250e-01 1.69186890e+00
5.77501416e-01 6.45615952e-03 8.14196706e-01 -1.49846351e+00
-6.69905066e-01 9.08492386e-01 5.44109121e-02 3.46008003e-01
1.16428159e-01 1.28016293e-01 8.94778848e-01 -2.66350299e-01
3.20703208e-01 1.17786098e+00 4.70106632e-01 8.57989371e-01
-1.25785923e+00 -6.63094699e-01 -2.34478991e-02 9.85449255e-02
-1.27778733e+00 -6.72678471e-01 8.61325204e-01 -3.72750103e-01
4.11240309e-01 7.43791342e-01 -4.34027202e-02 1.25408769e+00
5.23425341e-01 5.93952239e-01 1.07007051e+00 -5.28650939e-01
3.42036605e-01 5.12920022e-01 -8.95995274e-02 5.51548719e-01
2.11112186e-01 5.53479314e-01 -1.29528463e-01 -3.74119282e-01
1.45049542e-01 -2.49889493e-01 -4.87792999e-01 -5.51660471e-02
-8.18167686e-01 8.61557305e-01 5.46958447e-01 4.67392534e-01
-5.58860116e-02 2.11187050e-01 3.10781568e-01 1.11995125e+00
2.12203234e-01 8.48585129e-01 -2.64936775e-01 4.17316884e-01
-1.79579645e-01 1.60036221e-01 1.05992305e+00 8.44178677e-01
2.31123909e-01 2.25303024e-01 5.81512004e-02 3.81165326e-01
1.50846004e-01 7.62796521e-01 1.64459839e-01 -9.63472366e-01
4.41048712e-01 -3.18491012e-02 -3.00120767e-02 -9.26122844e-01
-2.17708036e-01 -7.41706073e-01 -7.12653875e-01 3.92135203e-01
7.14062929e-01 -5.40243149e-01 -5.18617570e-01 2.16330338e+00
-7.26032555e-02 -4.69312631e-02 6.43377006e-01 6.70056701e-01
3.17492276e-01 3.61801654e-01 -3.37076008e-01 -5.08313060e-01
6.86564684e-01 -2.16440186e-01 -6.43214405e-01 -3.34794015e-01
4.96387184e-01 -7.43799567e-01 7.31947482e-01 2.12529257e-01
-8.72178316e-01 -1.85191765e-01 -1.51841772e+00 5.58227241e-01
-1.02324814e-01 -6.60830200e-01 2.54134625e-01 1.16451645e+00
-6.33134067e-01 8.51386726e-01 -5.70028305e-01 7.53353834e-02
3.21082532e-01 3.96982253e-01 -4.25471276e-01 -6.53647557e-02
-1.47547734e+00 8.98399293e-01 1.94792569e-01 1.34175252e-02
-1.15383554e+00 -5.58928370e-01 -6.50028050e-01 -3.14627960e-03
5.32567427e-02 -2.80866355e-01 1.24837387e+00 -1.51477313e+00
-1.07699001e+00 7.00175405e-01 3.05147052e-01 -7.10552812e-01
8.25819135e-01 9.16323662e-02 -5.05407631e-01 2.27505282e-01
-3.18579316e-01 -5.01865521e-03 1.12455225e+00 -1.33260143e+00
-2.99569070e-01 -4.20008570e-01 2.98592180e-01 1.97716370e-01
-5.82107663e-01 -3.00054438e-02 2.77056098e-01 -9.56219316e-01
4.39513214e-02 -1.14417779e+00 -3.29103053e-01 -7.22808018e-03
-6.89389348e-01 3.28328520e-01 8.51658702e-01 -3.72918963e-01
7.07746863e-01 -2.53792548e+00 1.94351941e-01 5.69204450e-01
2.17100441e-01 4.67350483e-02 -1.32382572e-01 2.90427595e-01
-5.64946830e-01 2.24557191e-01 -4.25079912e-01 2.83417180e-02
1.11073293e-01 2.14902624e-01 -6.60449266e-01 1.10946894e+00
6.55764192e-02 4.67248768e-01 -6.41385376e-01 -1.23290233e-02
-2.06427187e-01 2.51735002e-01 -4.99191403e-01 2.74677068e-01
1.05288871e-01 3.98431629e-01 -3.70727032e-01 3.84841919e-01
5.82988977e-01 2.95047998e-01 -8.31381243e-04 2.93765087e-02
5.86786449e-01 -2.14300409e-01 -1.04653454e+00 8.04485202e-01
-2.20481128e-01 9.93517458e-01 3.78824651e-01 -1.26963961e+00
6.91126764e-01 4.63162601e-01 3.40867281e-01 -2.14313030e-01
5.76041698e-01 4.86790650e-02 3.36964428e-01 -2.14094758e-01
-1.28841385e-01 -5.75240076e-01 -6.19084299e-01 5.25556564e-01
4.42057252e-02 -4.32772376e-02 -6.02391720e-01 1.50716469e-01
1.50379491e+00 -7.55782604e-01 8.80990401e-02 -3.50675553e-01
3.25018585e-01 -4.34084982e-01 5.63268483e-01 1.20317209e+00
-4.27325577e-01 3.70501727e-01 6.74637079e-01 8.66237581e-02
-1.01603496e+00 -1.27410257e+00 -3.83568048e-01 6.98599219e-01
1.44565433e-01 1.59180224e-01 -1.02918172e+00 -7.91126311e-01
2.08522171e-01 6.96744382e-01 -7.52028942e-01 -7.80131042e-01
-2.33353078e-01 -6.77961051e-01 9.94258463e-01 2.00293243e-01
4.26667839e-01 -6.72157705e-01 -1.63058519e-01 -1.13627367e-01
4.83238250e-02 -1.08838034e+00 -4.33140427e-01 5.72058976e-01
-6.32318914e-01 -1.11050761e+00 -2.83621043e-01 -6.04227543e-01
9.08333659e-01 1.18152117e-02 5.79923809e-01 -2.89236475e-02
-1.76621228e-01 4.55330044e-01 -2.93384880e-01 -6.01593137e-01
-1.18546498e+00 -2.98293948e-01 4.56075341e-01 4.71476689e-02
-8.81574377e-02 -5.79259872e-01 -1.99445441e-01 4.45471346e-01
-1.18626750e+00 -5.75518191e-01 3.75869542e-01 8.02607298e-01
2.39849731e-01 6.44376934e-01 8.22036684e-01 -8.95134032e-01
7.27306485e-01 -7.17949033e-01 -3.90684038e-01 1.34826943e-01
-4.13557917e-01 3.11023980e-01 1.17952979e+00 -5.93770325e-01
-6.48565352e-01 1.24309793e-01 -1.69532239e-01 -5.52551329e-01
-1.11897178e-01 1.20006546e-01 -7.36099720e-01 -5.75395346e-01
1.15405440e+00 1.63629025e-01 2.15204172e-02 -1.69690520e-01
4.46259201e-01 9.00812149e-01 7.14932382e-01 -4.48668838e-01
1.12561679e+00 2.35646963e-01 2.29429588e-01 -9.53471303e-01
-7.79206038e-01 3.81330475e-02 -4.45848048e-01 -2.90509313e-01
3.89638335e-01 -4.65602636e-01 -6.27673149e-01 6.05117142e-01
-9.05394077e-01 -4.54304218e-02 -4.13367927e-01 3.48287940e-01
-9.87404525e-01 4.09090608e-01 -4.12027895e-01 -8.18897307e-01
-4.32008840e-02 -1.08559513e+00 4.62005645e-01 -1.25909731e-01
5.59351481e-02 -9.93406415e-01 -2.89751381e-01 1.70693845e-01
1.63509607e-01 7.23891258e-01 8.62668633e-01 -1.16096449e+00
-2.76272386e-01 -8.47422540e-01 1.93387136e-01 8.62144351e-01
9.94029343e-02 -4.10084099e-01 -1.35238004e+00 -5.59437215e-01
5.33764899e-01 -2.01996416e-01 7.93374419e-01 2.23733895e-02
1.09847176e+00 -9.72434163e-01 -1.56253204e-01 7.02539563e-01
1.35423803e+00 3.55954885e-01 4.26578432e-01 4.99251820e-02
4.52518791e-01 6.46522701e-01 3.06113511e-01 2.52419233e-01
-4.96363491e-01 4.91751164e-01 6.81325793e-01 2.12922260e-01
3.29879194e-01 -1.71714146e-02 5.05745113e-01 4.30908412e-01
5.42866588e-01 -4.48199630e-01 -5.88039577e-01 4.07834679e-01
-1.47447658e+00 -1.20910501e+00 3.00060123e-01 2.48317885e+00
8.32294583e-01 4.10993397e-01 -8.76601636e-02 6.00269020e-01
8.14264596e-01 2.05356088e-02 -8.34227085e-01 -5.46763182e-01
-2.67796129e-01 4.23348360e-02 9.00100350e-01 7.54397690e-01
-1.38892019e+00 5.13394952e-01 7.00107765e+00 5.98991930e-01
-8.65468323e-01 5.74193597e-02 6.05057955e-01 -2.08944291e-01
-2.97181129e-01 -3.36618841e-01 -4.18332607e-01 4.52114552e-01
1.23162293e+00 -6.56757176e-01 4.49090183e-01 8.42051029e-01
-2.26029128e-01 4.48760837e-01 -1.36057675e+00 6.07485652e-01
4.79604229e-02 -1.01321781e+00 -2.69575536e-01 2.64840305e-01
6.93736136e-01 -1.74590036e-01 3.61295551e-01 -2.87012681e-02
6.04239881e-01 -1.03319097e+00 6.61993980e-01 4.12880063e-01
6.04444683e-01 -1.14792323e+00 7.35241592e-01 5.61782777e-01
-5.29251993e-01 -4.56167221e-01 -3.30923170e-01 1.29099667e-01
-3.02043408e-01 5.75650454e-01 -6.75569177e-01 2.69053906e-01
2.76367337e-01 2.43426308e-01 -2.91941553e-01 7.69869208e-01
-1.97236225e-01 6.19669139e-01 -3.67265016e-01 1.80710837e-01
1.22546144e-02 3.83435100e-01 1.08367860e+00 1.09066916e+00
1.11189447e-02 1.48622364e-01 -9.45387874e-04 5.39745688e-01
-2.86484331e-01 -2.17640966e-01 -1.18558264e+00 -5.52569255e-02
6.95022225e-01 6.80775225e-01 -5.19130707e-01 5.66304065e-02
1.77751705e-02 1.13977528e+00 -9.97009221e-03 2.62845308e-01
-6.70442045e-01 -7.68031001e-01 9.73168612e-01 -3.37760091e-01
1.49677679e-01 3.60433199e-02 -3.29577237e-01 -9.65713441e-01
1.40774727e-01 -1.17549574e+00 4.38217878e-01 -1.34030640e-01
-1.55261302e+00 6.37524545e-01 -2.21514687e-01 -1.37173307e+00
-3.92705679e-01 -3.59717786e-01 -5.41679323e-01 8.21422040e-01
-7.29982197e-01 -5.60706913e-01 2.04089120e-01 8.24428678e-01
1.44451246e-01 -4.13758606e-01 1.02862883e+00 9.02947155e-04
-4.96717185e-01 1.30928743e+00 5.06913185e-01 2.69571841e-01
4.04249310e-01 -1.32115865e+00 3.43197256e-01 1.22336411e+00
1.43634439e-01 3.05263907e-01 1.20948756e+00 -2.18637958e-01
-1.49834692e+00 -1.24052346e+00 3.81834447e-01 -6.15491688e-01
7.11910188e-01 -4.29884791e-01 -7.26721704e-01 7.31934905e-01
-1.06375955e-01 1.30579188e-01 8.83864284e-01 -2.44638950e-01
-5.36890566e-01 -3.20432968e-02 -1.75082695e+00 5.56433022e-01
1.01462126e+00 -8.34760368e-01 -5.91178298e-01 6.05164289e-01
8.96317840e-01 -2.60888964e-01 -9.62831855e-01 1.61594331e-01
4.07838136e-01 -6.99282944e-01 9.22101974e-01 -8.44572127e-01
2.35329628e-01 -9.61601827e-03 -7.14328587e-01 -1.65411091e+00
-1.09175093e-01 -9.26094711e-01 -6.10705912e-02 7.80285895e-01
4.62003767e-01 -7.95609891e-01 5.56066394e-01 5.49217463e-01
-6.59985701e-03 -4.85181063e-01 -1.30364537e+00 -1.08710206e+00
4.39256907e-01 -6.05775356e-01 2.68929213e-01 1.04106236e+00
2.46706501e-01 -1.05239384e-01 -2.71934211e-01 6.74652100e-01
8.76840234e-01 -2.89552927e-01 4.29677725e-01 -8.83129597e-01
-5.48900962e-01 -3.98013473e-01 -1.05020010e+00 -6.00873768e-01
4.51681554e-01 -9.87914681e-01 3.49711806e-01 -5.57477355e-01
-1.05387010e-01 -3.99397612e-01 -4.58506554e-01 2.25997746e-01
-5.92921115e-02 2.25006565e-01 4.10538077e-01 1.54698908e-01
-9.34412479e-02 1.70958385e-01 8.74983728e-01 -3.63331616e-01
4.90848944e-02 5.25885224e-01 -1.22749174e+00 6.99761569e-01
9.38610077e-01 -8.06917667e-01 -4.08963382e-01 -1.14187695e-01
3.15454379e-02 1.87107414e-01 2.54761428e-01 -1.02046001e+00
4.27791700e-02 -8.64395946e-02 2.41368517e-01 1.94566339e-01
1.76442504e-01 -1.06761825e+00 -2.31723823e-02 6.75763786e-01
-8.59448910e-01 -3.24926019e-01 9.80043188e-02 8.75294030e-01
6.97488524e-03 -4.38957185e-01 1.34624600e+00 9.21313167e-02
-1.68193430e-01 1.33993834e-01 -5.62655687e-01 5.83805405e-02
1.32604325e+00 9.08531323e-02 -3.26523066e-01 -6.69457495e-01
-1.09012496e+00 -8.16356242e-02 2.96923012e-01 4.25700814e-01
5.48115015e-01 -1.33461261e+00 -7.95682907e-01 4.41231877e-01
7.27984905e-02 -5.96706033e-01 6.45816233e-03 2.70749629e-01
-1.05570227e-01 2.47095808e-01 -1.13723792e-01 -1.55703977e-01
-1.57547796e+00 7.38407195e-01 8.56634319e-01 8.40958804e-02
-4.68247354e-01 9.82532203e-01 1.43739715e-01 1.51513051e-02
5.40643752e-01 6.80342764e-02 2.30681941e-01 -2.26210207e-01
6.01233482e-01 2.14424342e-01 2.26633489e-01 -6.37477398e-01
-2.70654440e-01 7.51094073e-02 -7.94751346e-02 -2.88968265e-01
9.14647639e-01 1.59927011e-02 1.72961041e-01 1.26430720e-01
1.72907186e+00 1.14083126e-01 -1.27043831e+00 -4.75345939e-01
-2.91878670e-01 -6.44600034e-01 1.11856185e-01 -5.32806993e-01
-1.13560641e+00 6.72481775e-01 7.45339096e-01 7.16498852e-01
1.28333640e+00 1.35183245e-01 3.67691547e-01 7.32672751e-01
2.81118691e-01 -7.47551739e-01 -5.36697134e-02 3.65516305e-01
1.11511207e+00 -8.61852646e-01 -2.61817157e-01 -2.24783897e-01
-4.84751642e-01 9.64913070e-01 2.21642256e-01 -3.64195049e-01
9.42707539e-01 4.31494564e-01 7.34193102e-02 8.72345418e-02
-7.26509333e-01 9.44107324e-02 2.01490805e-01 8.60782504e-01
-3.06262616e-02 3.13810080e-01 2.86458135e-02 4.92125094e-01
-6.59555495e-01 -7.32127845e-01 7.19336987e-01 9.27405119e-01
-5.68270266e-01 -8.39707851e-01 -6.61129951e-01 4.34529334e-01
-9.47970390e-01 9.16847587e-02 -8.61322343e-01 3.69865447e-01
-1.97171830e-02 1.29659021e+00 -9.54845175e-02 -7.81608760e-01
3.27192128e-01 -1.31130335e-03 5.34671783e-01 -4.70492363e-01
-3.90300184e-01 -4.08099353e-01 2.29047183e-02 -3.38239044e-01
9.23124049e-03 -7.55051792e-01 -7.97397494e-01 -4.92111683e-01
-3.74051690e-01 1.28757685e-01 5.15318811e-01 1.14436495e+00
1.06073767e-02 5.76060057e-01 1.41604638e+00 -4.30596143e-01
-1.26487076e+00 -7.77954698e-01 -7.20313549e-01 3.73052180e-01
8.89338434e-01 -3.11386168e-01 -1.06603658e+00 -4.91971243e-03] | [5.6996588706970215, 7.766674041748047] |
d9c270ea-e8c2-4cda-9cd2-9e41a9838031 | quantized-radio-map-estimation-using-tensor | 2303.01770 | null | https://arxiv.org/abs/2303.01770v1 | https://arxiv.org/pdf/2303.01770v1.pdf | Quantized Radio Map Estimation Using Tensor and Deep Generative Models | Spectrum cartography (SC), also known as radio map estimation (RME), aims at crafting multi-domain (e.g., frequency and space) radio power propagation maps from limited sensor measurements. While early methods often lacked theoretical support, recent works have demonstrated that radio maps can be provably recovered using low-dimensional models -- such as the block-term tensor decomposition (BTD) model and certain deep generative models (DGMs) -- of the high-dimensional multi-domain radio signals. However, these existing provable SC approaches assume that sensors send real-valued (full-resolution) measurements to the fusion center, which is unrealistic. This work puts forth a quantized SC framework that generalizes the BTD and DGM-based SC to scenarios where heavily quantized sensor measurements are used. A maximum likelihood estimation (MLE)-based SC framework under a Gaussian quantizer is proposed. Recoverability of the radio map using the MLE criterion are characterized under realistic conditions, e.g., imperfect radio map modeling and noisy measurements. Simulations and real-data experiments are used to showcase the effectiveness of the proposed approach. | ['Xiao Fu', 'Sagar Shrestha', 'Subash Timilsina'] | 2023-03-03 | null | null | null | null | ['spectrum-cartography'] | ['computer-vision'] | [ 3.99213821e-01 9.51306149e-02 -2.41617918e-01 -9.99422371e-02
-8.86064410e-01 -5.03183544e-01 5.68187475e-01 -3.11731458e-01
4.63152342e-02 9.19391811e-01 3.11247349e-01 -3.92701149e-01
-6.33068204e-01 -9.61461306e-01 -8.25734496e-01 -9.11774635e-01
-4.31087703e-01 1.53542027e-01 -1.59528121e-01 -2.11997241e-01
1.08234992e-03 2.64251590e-01 -9.60316062e-01 -4.99694109e-01
5.73511183e-01 1.03663111e+00 3.74930859e-01 9.31161344e-01
3.12162966e-01 8.31854284e-01 -7.52746582e-01 -1.13383755e-01
3.24412644e-01 -3.93679470e-01 -1.70899212e-01 -5.22520989e-02
-3.61847878e-01 -3.78000200e-01 -7.02170432e-01 1.31892693e+00
5.16907990e-01 -4.52095419e-01 4.88199919e-01 -1.57227337e+00
-6.56658828e-01 7.47115254e-01 -6.43582344e-01 4.13693767e-03
4.85586882e-01 -8.20906103e-01 6.61817253e-01 -5.58048964e-01
2.52833962e-01 1.10188746e+00 7.26783574e-01 9.74603593e-02
-1.19158626e+00 -9.68209028e-01 -4.35266286e-01 4.03669626e-02
-1.81795251e+00 -4.94930804e-01 8.93493414e-01 -5.20556904e-02
1.68418974e-01 3.89835149e-01 2.09116414e-01 8.77705097e-01
7.86478639e-01 5.94951928e-01 1.40962851e+00 -4.58402157e-01
3.92349571e-01 -3.74026992e-03 -4.28326190e-01 4.51593846e-01
5.54156661e-01 3.64222080e-01 -8.60601068e-01 -6.20320439e-01
7.62002945e-01 -1.30278900e-01 -4.23523098e-01 -3.25379848e-01
-1.45260251e+00 6.56899750e-01 2.59014100e-01 4.35044885e-01
-7.17237711e-01 3.68678689e-01 -2.34905496e-01 6.43637896e-01
4.29677784e-01 6.43383414e-02 1.09327525e-01 -1.01044038e-02
-1.26847637e+00 -5.68266958e-02 9.72926915e-01 1.51367998e+00
8.63183618e-01 2.82121807e-01 1.65077835e-01 3.35346907e-01
5.51605284e-01 1.31249833e+00 1.24593325e-01 -7.69157112e-01
5.38920164e-01 -3.80630523e-01 5.04759014e-01 -1.42020822e+00
-5.84957182e-01 -9.06644225e-01 -1.38028824e+00 -3.69290024e-01
-1.37614489e-01 -5.60941041e-01 -5.40681422e-01 1.96001840e+00
1.47188827e-01 6.25632226e-01 3.55423808e-01 9.35994983e-01
7.51226246e-02 7.46688068e-01 -6.72625303e-01 -4.98735189e-01
9.36426103e-01 2.97138114e-02 -1.16070044e+00 -2.06210822e-01
2.11693004e-01 -5.89085937e-01 1.96181372e-01 5.74755311e-01
-9.47572410e-01 -2.25411385e-01 -1.75942600e+00 6.38700426e-01
-9.78723988e-02 -1.83112055e-01 2.95280069e-01 1.15618026e+00
-1.35876393e+00 3.79128531e-02 -7.49297738e-01 2.47804419e-04
2.42182589e-03 2.33293518e-01 -9.82824937e-02 -4.96083856e-01
-1.49396932e+00 6.55854583e-01 2.01815605e-01 2.59463161e-01
-1.06723201e+00 -4.70526189e-01 -5.86064994e-01 -1.64231360e-01
3.72789979e-01 -5.78431070e-01 1.02245367e+00 -1.72779679e-01
-1.43410397e+00 -1.38203129e-01 -2.95425743e-01 -6.93512201e-01
1.63632557e-01 -4.83016148e-02 -7.98352420e-01 3.79338145e-01
2.97093004e-01 -1.14934504e-01 1.07909870e+00 -1.34423876e+00
-6.62915707e-01 -3.07254106e-01 1.69198122e-02 1.29725307e-01
-5.80526181e-02 -3.79812837e-01 1.87546611e-01 -6.91648722e-01
8.26996982e-01 -1.04472685e+00 -1.91809028e-01 -4.05851394e-01
-6.28529847e-01 5.50759256e-01 5.89374840e-01 -5.62890291e-01
1.07792258e+00 -2.10813260e+00 5.21680154e-02 6.58571243e-01
2.75422841e-01 -5.94178379e-01 2.17696786e-01 8.00573707e-01
2.84361243e-01 5.45541085e-02 4.53283079e-02 -3.35297912e-01
2.62208194e-01 2.23280773e-01 -5.95984578e-01 1.08757436e+00
-4.17329878e-01 4.78433311e-01 -8.93751681e-01 -2.72682637e-01
2.13956669e-01 4.68235254e-01 -2.54096538e-02 -2.27528185e-01
2.30698958e-01 4.73937809e-01 -6.61273301e-01 5.75814188e-01
1.07314956e+00 -2.75401920e-01 1.63383201e-01 -1.28820345e-01
-3.16732489e-02 -1.04199097e-01 -1.40279782e+00 1.82533371e+00
-7.30248153e-01 6.71698987e-01 6.42377257e-01 -8.24091554e-01
8.39889586e-01 5.80487132e-01 8.10967445e-01 -8.24135363e-01
9.52906907e-02 3.36673260e-01 -3.49419832e-01 1.28489256e-01
6.43460572e-01 -2.84909159e-01 -6.49803221e-01 5.89373469e-01
-1.06550073e-02 -9.79604870e-02 -3.49398464e-01 3.87872308e-01
1.36440432e+00 -3.50433856e-01 5.05046129e-01 -3.15099984e-01
2.99966991e-01 -3.06993008e-01 4.57356513e-01 1.05322444e+00
5.90566732e-02 2.35821456e-01 -7.04526082e-02 3.68140101e-01
-9.61019635e-01 -1.08891809e+00 -2.90805668e-01 7.91524172e-01
7.25413322e-01 -3.40405285e-01 -4.30531889e-01 2.56271482e-01
-5.28290831e-02 8.20211947e-01 -3.96638006e-01 -6.09026179e-02
-1.18700430e-01 -1.04408836e+00 9.43472147e-01 -5.85355908e-02
6.76109612e-01 1.89039662e-01 -4.98234808e-01 6.67945564e-01
-4.10878301e-01 -1.47347593e+00 7.42841363e-02 2.78797448e-01
-3.26830387e-01 -5.69806457e-01 -8.24774444e-01 -2.38961518e-01
3.46477002e-01 7.20660686e-01 5.75946808e-01 -4.53564167e-01
1.93279117e-01 8.34613264e-01 -6.86294258e-01 -4.96196091e-01
-3.73854756e-01 -3.75531554e-01 5.63873649e-01 3.48328978e-01
-1.06489891e-02 -9.55633104e-01 -3.76823187e-01 6.45428121e-01
-7.92905569e-01 7.43061155e-02 1.04253089e+00 6.19467199e-01
4.70646650e-01 6.67173982e-01 7.91249573e-01 -1.98526561e-01
7.72316337e-01 -8.42769206e-01 -6.84585392e-01 2.33715266e-01
-7.34562635e-01 2.48709574e-01 4.71810848e-01 -1.77568570e-01
-7.23841131e-01 -2.34914884e-01 2.97919750e-01 -2.12335676e-01
1.78993419e-01 5.99214315e-01 -2.99488544e-01 -5.30517459e-01
5.28933942e-01 6.76000595e-01 -4.07592386e-01 -1.05971649e-01
5.38034797e-01 9.76413071e-01 8.34352016e-01 -4.51804996e-01
1.32743645e+00 7.96092093e-01 4.46681291e-01 -9.14422274e-01
-5.49143672e-01 -3.60446900e-01 -3.07374090e-01 -2.97735810e-01
5.75353503e-01 -1.14014781e+00 -4.94767994e-01 1.50072992e-01
-9.77895856e-01 1.14411622e-01 3.36676270e-01 8.41894269e-01
-8.17217290e-01 3.55265915e-01 -1.71730727e-01 -1.27348518e+00
3.92866507e-02 -7.77633846e-01 1.19453263e+00 -5.92101105e-02
3.03627998e-01 -9.39920306e-01 1.64237022e-01 -4.33752313e-02
7.48907685e-01 5.55374622e-01 5.16569972e-01 -3.11276466e-01
-7.47565508e-01 -4.25245464e-01 -4.91751023e-02 -1.30577758e-01
7.55778654e-03 -9.78662312e-01 -8.92442286e-01 -5.37044883e-01
3.34250331e-01 -4.10800800e-02 1.94684014e-01 4.95806903e-01
5.47480047e-01 -4.16622400e-01 -5.97410202e-01 5.90849936e-01
1.58261657e+00 -3.02360076e-02 5.00155747e-01 1.27785757e-01
2.72778273e-01 -2.04181194e-01 7.86153674e-01 8.30680370e-01
5.09846091e-01 6.75417304e-01 6.59297764e-01 3.39907497e-01
1.17160104e-01 -2.88634419e-01 2.46942356e-01 8.98267031e-01
2.24608079e-01 -6.74663365e-01 -5.26676714e-01 3.20672274e-01
-1.61814988e+00 -8.29403222e-01 1.55383751e-01 2.13245654e+00
3.75670075e-01 1.55590057e-01 -2.40517944e-01 5.95498264e-01
9.10028219e-01 3.77513051e-01 -4.54805583e-01 2.20449507e-01
-2.87007898e-01 -1.87967420e-01 1.32093024e+00 5.16842127e-01
-8.13024580e-01 1.36466339e-01 6.00082493e+00 7.78846204e-01
-7.77094066e-01 3.76364678e-01 -3.04222137e-01 3.53494585e-01
-3.70190412e-01 -1.56613011e-02 -5.53271353e-01 4.85592544e-01
1.26189613e+00 -5.14911234e-01 7.10836649e-01 4.31559294e-01
1.18012562e-01 -2.01520741e-01 -7.96380877e-01 1.26742244e+00
1.06341355e-01 -1.22927701e+00 -5.93446314e-01 6.30851150e-01
5.35848558e-01 2.15726063e-01 8.56922343e-02 8.58359784e-02
5.52403986e-01 -6.62842333e-01 9.78969932e-01 6.86648965e-01
9.12837327e-01 -7.90452123e-01 8.65882456e-01 7.82554746e-01
-1.39614820e+00 -2.97036469e-01 -4.70209777e-01 2.36470208e-01
5.15892327e-01 1.02063191e+00 -1.08899069e+00 1.44486177e+00
3.60256910e-01 3.55214119e-01 -1.20436423e-01 9.39552248e-01
-4.94143330e-02 7.12758899e-01 -6.67483151e-01 -1.50990412e-01
1.86598405e-01 4.54922915e-02 9.58731651e-01 8.91210258e-01
9.27653909e-01 1.70569092e-01 1.71226978e-01 6.18301868e-01
8.58839825e-02 -5.01007199e-01 -5.89965820e-01 1.58380404e-01
1.15593171e+00 9.50038195e-01 -3.92861485e-01 1.15867577e-01
-2.63423413e-01 7.07100272e-01 -5.33100069e-01 5.12596428e-01
-8.48959684e-01 -5.32896101e-01 6.29599616e-02 3.08715948e-03
1.62192494e-01 -6.59352839e-01 -1.49981081e-01 -7.86274433e-01
-8.22563469e-02 -4.97712791e-01 -1.49008155e-01 -8.20888519e-01
-1.05535126e+00 1.58942193e-01 2.81241089e-01 -1.63904476e+00
-7.18411922e-01 2.04488579e-02 -1.52498782e-01 8.41858506e-01
-1.54724646e+00 -1.12445772e+00 -2.47200981e-01 8.16284359e-01
-6.38519004e-02 -2.47188628e-01 7.58084893e-01 3.27518761e-01
7.02149644e-02 3.50400895e-01 6.81389034e-01 -2.57938355e-02
1.55181214e-01 -1.17833710e+00 7.19079822e-02 9.65835273e-01
1.90974817e-01 3.84773731e-01 1.19644296e+00 -6.46408916e-01
-1.96011627e+00 -1.01409173e+00 3.38246167e-01 -2.48029884e-02
8.97510052e-01 -7.32883394e-01 -2.02109516e-01 5.03760219e-01
1.40165880e-01 -5.31098545e-02 7.44817793e-01 -2.82226562e-01
-9.07104183e-03 -1.53329685e-01 -1.17800796e+00 1.53077751e-01
8.12264681e-01 -7.17155874e-01 -3.61542195e-01 4.33477789e-01
5.07654488e-01 -2.75363564e-01 -1.16058946e+00 3.18900824e-01
5.51551282e-01 -6.68272018e-01 8.89841855e-01 3.80669981e-01
-5.23762703e-01 -6.22932792e-01 -1.00653708e+00 -1.38344181e+00
-2.40924791e-01 -1.07397068e+00 -3.11203420e-01 8.87083471e-01
3.20922464e-01 -6.21996760e-01 4.13448632e-01 -1.43942147e-01
-3.82447392e-02 -1.11943431e-01 -1.50216377e+00 -1.09957182e+00
-3.85168225e-01 -8.48574877e-01 9.38556850e-01 7.45274305e-01
2.18505897e-02 2.16141522e-01 -8.59489620e-01 1.10388076e+00
1.20979428e+00 -1.94906592e-01 8.60782146e-01 -1.17930162e+00
-4.01504517e-01 2.32531518e-01 -8.54718566e-01 -1.35944903e+00
8.01227987e-03 -6.39826119e-01 2.89437920e-01 -1.51011920e+00
-3.28625143e-01 -5.94307423e-01 -2.56582767e-01 -4.15358394e-02
5.96261144e-01 4.11143228e-02 1.13209911e-01 3.43087226e-01
-6.92497909e-01 6.21928811e-01 8.01448882e-01 2.47491691e-02
1.64776832e-01 4.82718438e-01 -6.99634910e-01 3.44431311e-01
6.31772041e-01 -6.72949433e-01 -8.09643388e-01 -3.32793802e-01
6.08803868e-01 9.22514498e-01 6.10503256e-01 -1.50557899e+00
7.14519799e-01 1.13797607e-02 2.15701103e-01 -8.40323806e-01
7.05326557e-01 -1.08354235e+00 4.60769981e-01 3.98872703e-01
-1.02316104e-02 -2.85232663e-01 -2.15593949e-01 1.40261900e+00
-5.66874221e-02 1.69387892e-01 3.96356523e-01 3.26867908e-01
-5.04852116e-01 6.55911639e-02 -2.83552915e-01 -3.16916883e-01
9.03625488e-01 -1.35597870e-01 -4.56329077e-01 -1.20537412e+00
-5.99143326e-01 1.51874617e-01 7.22143352e-02 1.55242592e-01
7.27132022e-01 -1.53726339e+00 -9.83313203e-01 4.03320193e-02
3.49875987e-02 -2.84764618e-01 8.22236165e-02 1.00167620e+00
-2.66744271e-02 7.56930053e-01 1.20316587e-01 -7.74716079e-01
-5.28053045e-01 3.73449713e-01 1.77203476e-01 -6.29488230e-02
-5.70353389e-01 4.30674046e-01 -1.22779369e-01 -2.10558578e-01
4.98564839e-02 -1.78633168e-01 2.86365241e-01 -4.71978873e-01
6.27723634e-01 4.09012675e-01 1.25432890e-02 -7.90226936e-01
-4.76359040e-01 5.17508626e-01 4.62571442e-01 -8.76776397e-01
1.07230461e+00 -8.86294723e-01 6.71709776e-02 4.28785503e-01
1.02286625e+00 3.14466536e-01 -9.08068538e-01 -6.49695039e-01
4.00771759e-02 -4.72297549e-01 4.82128918e-01 -5.71705520e-01
-5.54734349e-01 4.87104386e-01 6.13257825e-01 6.98030114e-01
1.21877694e+00 -2.91865389e-03 5.60649991e-01 5.54725349e-01
1.45890880e+00 -1.03080118e+00 -2.97963351e-01 7.13301301e-02
6.74472034e-01 -5.95206738e-01 2.20101625e-01 -2.16271043e-01
-1.35627136e-01 8.87152553e-01 -4.23453808e-01 4.53409702e-02
9.94800210e-01 5.19455969e-01 -3.50846678e-01 -3.80697735e-02
-4.99702185e-01 -1.28925396e-02 -3.71422549e-03 8.54667723e-01
-2.76223332e-01 3.81651402e-01 2.41859093e-01 6.89613879e-01
-6.42336845e-01 -5.93646318e-02 9.57679152e-01 1.17488790e+00
-7.78786123e-01 -5.62445879e-01 -9.58193541e-01 3.43304664e-01
-3.17765504e-01 1.06164493e-01 3.22621465e-02 6.75532043e-01
-8.05152208e-02 1.71363997e+00 -4.12107050e-01 -9.23607409e-01
2.42652949e-02 -4.49121982e-01 3.09223145e-01 -6.21673279e-02
6.42250359e-01 3.93739671e-01 4.18295786e-02 -4.60606128e-01
-7.08644152e-01 -4.98445481e-01 -9.73985493e-01 -5.09144127e-01
-5.57737052e-01 4.51456606e-01 1.15595591e+00 1.01352990e+00
1.45206288e-01 5.58417737e-01 1.23374021e+00 -6.30765140e-01
-8.19543660e-01 -8.78312111e-01 -1.29651678e+00 -5.95932543e-01
8.16699982e-01 -7.93035865e-01 -5.72916389e-01 -3.77431870e-01] | [6.39297342300415, 1.2513070106506348] |
ef5484bf-b2b9-4a16-a386-a07be08545ac | rainbow-keywords-efficient-incremental | 2203.16361 | null | https://arxiv.org/abs/2203.16361v2 | https://arxiv.org/pdf/2203.16361v2.pdf | Rainbow Keywords: Efficient Incremental Learning for Online Spoken Keyword Spotting | Catastrophic forgetting is a thorny challenge when updating keyword spotting (KWS) models after deployment. This problem will be more challenging if KWS models are further required for edge devices due to their limited memory. To alleviate such an issue, we propose a novel diversity-aware incremental learning method named Rainbow Keywords (RK). Specifically, the proposed RK approach introduces a diversity-aware sampler to select a diverse set from historical and incoming keywords by calculating classification uncertainty. As a result, the RK approach can incrementally learn new tasks without forgetting prior knowledge. Besides, the RK approach also proposes data augmentation and knowledge distillation loss function for efficient memory management on the edge device. Experimental results show that the proposed RK approach achieves 4.2% absolute improvement in terms of average accuracy over the best baseline on Google Speech Command dataset with less required memory. The scripts are available on GitHub. | ['Eng Siong Chng', 'Nana Hou', 'Yang Xiao'] | 2022-03-30 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 8.85887742e-02 -4.88839000e-02 -4.60125595e-01 -2.99158216e-01
-1.01916993e+00 -3.14635932e-01 3.85207027e-01 -7.19958395e-02
-4.90258425e-01 9.53026295e-01 2.19754055e-01 -4.55636412e-01
-1.59640610e-01 -3.42173427e-01 -8.45611513e-01 -5.75085938e-01
9.53053683e-02 2.15619266e-01 4.11944419e-01 6.61595836e-02
3.91230471e-02 2.08672792e-01 -1.38756597e+00 1.06252782e-01
1.14784956e+00 1.22161293e+00 6.69455111e-01 5.48902810e-01
-1.87215239e-01 5.83439946e-01 -5.77234447e-01 -2.90728629e-01
2.15730712e-01 2.62174755e-02 -5.45345664e-01 -2.02719674e-01
1.39868453e-01 -5.66919088e-01 -5.09984016e-01 7.38284945e-01
8.05086911e-01 2.33982340e-01 2.11427242e-01 -1.21736193e+00
-3.40523452e-01 1.16767287e+00 -5.20443678e-01 4.90913063e-01
-6.12948500e-02 1.16027035e-01 7.46282816e-01 -1.06554747e+00
1.80916458e-01 7.06425250e-01 8.18321288e-01 4.26784903e-01
-9.12293673e-01 -9.14435506e-01 7.32079804e-01 5.32953978e-01
-1.70107055e+00 -7.11837828e-01 7.23842502e-01 -2.28259377e-02
8.96379769e-01 3.29847813e-01 3.51379812e-01 1.11197150e+00
-1.72970787e-01 1.20422649e+00 7.88400650e-01 -2.51189977e-01
3.49219859e-01 4.89629626e-01 2.26272240e-01 3.34450096e-01
2.82323629e-01 -2.86605477e-01 -9.20388162e-01 -2.51633346e-01
4.39448990e-02 7.70503581e-02 -4.24144745e-01 -1.47634268e-01
-8.02861810e-01 2.64631450e-01 4.60963026e-02 1.12707689e-02
-5.49217880e-01 1.05749033e-01 4.02743667e-01 1.50118962e-01
5.57923973e-01 1.10229552e-01 -9.28139210e-01 -5.69174826e-01
-1.15274918e+00 8.12057331e-02 5.97363710e-01 1.06621945e+00
5.04711211e-01 2.30327651e-01 -4.15567577e-01 9.20448542e-01
4.33440059e-02 4.17407036e-01 7.16426909e-01 -4.48905706e-01
6.05160832e-01 3.16529006e-01 1.71109557e-01 -6.03462458e-01
-3.14273536e-01 -9.14368033e-01 -6.87597990e-01 -3.72923285e-01
-3.38232666e-01 -2.40663633e-01 -1.03070307e+00 1.71237385e+00
5.11888444e-01 7.39081562e-01 -7.08107054e-02 3.99015486e-01
7.08444357e-01 6.65013194e-01 7.46659115e-02 -6.45990610e-01
9.66200531e-01 -9.84448671e-01 -9.46768343e-01 -1.26126483e-01
4.39415157e-01 -5.42163074e-01 1.42432415e+00 5.05733967e-01
-8.86656225e-01 -3.07143390e-01 -1.20968139e+00 3.06230515e-01
-2.34546006e-01 4.15242165e-01 6.90291524e-01 8.04303288e-01
-9.18397486e-01 2.62030065e-01 -9.68564689e-01 -1.01492748e-01
5.89388430e-01 4.23813760e-01 2.62282789e-01 8.80238134e-03
-1.31416285e+00 4.70747560e-01 5.32367766e-01 -7.56078735e-02
-1.02665007e+00 -1.08289528e+00 -4.75912988e-01 1.04209252e-01
8.95063818e-01 -4.79742736e-01 1.60680652e+00 -3.99277151e-01
-1.75637376e+00 6.44111559e-02 -3.83640736e-01 -7.19544232e-01
4.51819688e-01 -5.44975221e-01 -7.23791182e-01 -2.42513642e-01
-2.08385050e-01 2.85708129e-01 1.24539030e+00 -1.11915517e+00
-9.25179899e-01 -2.48104215e-01 -5.91395833e-02 4.30443436e-01
-1.03806913e+00 -5.55417418e-01 -8.95818114e-01 -9.56245899e-01
-1.37397140e-01 -8.68504941e-01 -8.76559392e-02 -4.33061540e-01
-4.53442663e-01 -2.25763425e-01 1.11452639e+00 -9.72651124e-01
2.10789800e+00 -2.10935736e+00 -1.17678821e-01 -5.42498380e-02
4.24407795e-03 6.64022088e-01 1.98969290e-01 7.34782144e-02
3.53908211e-01 -4.66652811e-02 -2.84425348e-01 -5.73444128e-01
-1.47297919e-01 1.63601041e-02 -4.17000562e-01 4.42858227e-02
-2.39838690e-01 7.19531238e-01 -6.72110617e-01 -1.08118631e-01
-2.64466200e-02 4.27076489e-01 -4.14442539e-01 2.94510666e-02
-3.90246630e-01 2.68273234e-01 -2.64634281e-01 6.57958746e-01
9.25833046e-01 -1.40535489e-01 1.67392325e-02 -1.37423828e-01
4.43648621e-02 4.58347917e-01 -1.24860013e+00 1.89332986e+00
-9.79639053e-01 1.93250254e-01 -1.06525399e-01 -7.35119998e-01
5.81902444e-01 1.72119647e-01 1.13802649e-01 -5.69536209e-01
-2.11654883e-02 6.21269457e-02 -4.34302688e-01 -2.95835048e-01
8.56978714e-01 3.09614509e-01 -4.46048565e-02 3.93396109e-01
-5.60418442e-02 2.31131256e-01 -2.02195466e-01 4.54304129e-01
1.11329365e+00 -5.37920631e-02 1.18668400e-01 8.68187398e-02
4.63484168e-01 -2.56150365e-01 9.34456527e-01 8.69617820e-01
-2.64761597e-02 2.12052628e-01 2.11030841e-02 -2.09309205e-01
-5.14751673e-01 -9.07470345e-01 2.46163132e-03 1.04078603e+00
8.66158605e-02 -6.60438120e-01 -4.76366222e-01 -8.87394488e-01
1.12358384e-01 1.15465939e+00 -2.92301774e-01 -5.69861531e-01
-2.96526015e-01 -7.89873242e-01 4.56849158e-01 5.49552798e-01
7.04384327e-01 -5.34441888e-01 -3.69641721e-01 9.31966975e-02
-1.58906460e-01 -1.08592796e+00 -7.26347804e-01 1.91045865e-01
-8.52474689e-01 -5.17491996e-01 -7.81461596e-01 -3.91606838e-01
4.13693845e-01 4.84088987e-01 6.78082049e-01 -2.84664482e-01
-1.06754944e-01 5.71916282e-01 -5.01840889e-01 -4.77371305e-01
-1.31974444e-01 7.17711926e-01 3.74286205e-01 1.57614410e-01
2.06315845e-01 -6.97001696e-01 -7.51299024e-01 1.06258601e-01
-8.56190443e-01 -1.00052349e-01 7.28158951e-01 8.74323905e-01
6.89803600e-01 4.39351261e-01 9.60276484e-01 -8.74428034e-01
8.55798066e-01 -8.37100029e-01 -6.80846572e-01 5.73510706e-01
-1.21229410e+00 2.23645687e-01 6.15439594e-01 -8.11665773e-01
-1.40275669e+00 1.61416277e-01 -4.41220477e-02 -4.86272663e-01
3.88395637e-01 4.35320437e-01 -3.84584367e-01 3.52006592e-02
3.69107246e-01 5.75544119e-01 -5.09971380e-01 -8.02930057e-01
4.10381287e-01 1.13671327e+00 5.49254775e-01 -3.71281117e-01
6.36980474e-01 3.19322228e-01 -5.33270478e-01 -7.36927390e-01
-6.46107793e-01 -5.21104217e-01 -1.66938841e-01 5.41239008e-02
1.50063157e-01 -1.20377958e+00 -4.37474459e-01 6.27362370e-01
-7.93731749e-01 -2.01381311e-01 -3.23378295e-01 4.57447767e-01
-1.18739933e-01 3.58285755e-01 -1.66506052e-01 -1.11279750e+00
-7.79948294e-01 -8.77082586e-01 8.53625119e-01 3.49354565e-01
-4.69968766e-02 -6.48729146e-01 -2.69242346e-01 3.01204771e-01
6.64111733e-01 -3.67908448e-01 7.55590856e-01 -6.46000087e-01
-5.23444057e-01 -2.94544518e-01 5.05484976e-02 4.24054146e-01
3.30094844e-01 -5.05918145e-01 -1.01733315e+00 -4.03345644e-01
2.24141106e-02 -1.00008033e-01 1.02131832e+00 1.63722619e-01
1.52306962e+00 -3.93976390e-01 -5.34519553e-01 5.18844247e-01
1.07922935e+00 4.67671245e-01 2.72669405e-01 1.34320274e-01
6.67220175e-01 -4.25385125e-02 9.02924240e-01 9.65656817e-01
6.84687495e-01 8.19990575e-01 3.94270308e-02 4.46795106e-01
-3.87230039e-01 -5.28634787e-01 4.25655872e-01 9.67623889e-01
5.44697404e-01 -2.61475950e-01 -8.63871396e-01 6.97098851e-01
-1.82440817e+00 -5.06708145e-01 4.38627779e-01 2.64866614e+00
1.14698517e+00 3.75360608e-01 4.01113927e-02 2.78266311e-01
6.18974149e-01 9.17141661e-02 -1.24128091e+00 -2.78163645e-02
-4.17448804e-02 7.39222020e-02 7.45320261e-01 4.54434901e-01
-9.55358803e-01 1.08869326e+00 5.64459229e+00 1.30458808e+00
-1.21532667e+00 5.58103442e-01 8.43731165e-01 -3.41246605e-01
-4.07290608e-01 -4.94729877e-02 -1.30465484e+00 8.80498111e-01
1.09910941e+00 -4.10186946e-01 4.43670511e-01 1.01601899e+00
1.33668453e-01 -3.37987632e-01 -6.65089965e-01 1.35926127e+00
1.09375797e-01 -1.26942086e+00 -1.70560984e-03 -4.00097847e-01
6.97267056e-01 1.88985057e-02 3.58668178e-01 5.94665170e-01
2.33367205e-01 -4.79689419e-01 6.04272187e-01 5.94812751e-01
9.26001310e-01 -9.05600071e-01 4.13494080e-01 4.98413503e-01
-1.07211292e+00 -3.17520589e-01 -6.77756825e-03 1.81637689e-01
3.24520379e-01 9.57391918e-01 -1.19465804e+00 4.45401341e-01
9.39935029e-01 2.73964226e-01 -5.73791444e-01 1.16097283e+00
-1.58771902e-01 9.44607913e-01 -5.25741041e-01 3.05556264e-02
-1.96780711e-01 4.10957515e-01 6.98160529e-01 9.50794458e-01
4.23278123e-01 -8.96339118e-03 -8.80493876e-03 2.41465643e-01
-4.18013811e-01 -2.11411528e-02 -3.04697841e-01 -6.50125220e-02
1.08446968e+00 1.09843516e+00 -4.95057404e-01 -3.83048713e-01
-1.26242980e-01 1.39651203e+00 1.24871314e-01 2.98249960e-01
-9.52173948e-01 -4.54976678e-01 6.14192188e-01 6.96382448e-02
4.91587698e-01 -2.18002334e-01 -1.59155831e-01 -1.09446132e+00
4.35043544e-01 -6.22845650e-01 3.78256232e-01 -5.49421668e-01
-8.53580534e-01 5.94838977e-01 1.00895360e-01 -1.02269840e+00
-1.79043282e-02 1.21016532e-01 -3.37093890e-01 5.26358187e-01
-1.54598534e+00 -9.59508717e-01 -3.28339368e-01 3.71395320e-01
8.91548097e-01 -2.42375389e-01 5.16517401e-01 6.63605928e-01
-6.41450465e-01 1.22825575e+00 2.96702385e-01 -7.71858275e-01
6.65579617e-01 -8.96371365e-01 5.53833604e-01 8.89951408e-01
1.66362643e-01 5.32120824e-01 6.02580369e-01 -8.55454504e-01
-1.49252653e+00 -1.18778467e+00 5.90914965e-01 -2.46864319e-01
4.57813889e-01 -4.79549170e-01 -8.85323763e-01 4.79097754e-01
-5.58871478e-02 2.30964515e-02 5.46741486e-01 6.51959553e-02
-2.47006506e-01 -5.54516494e-01 -1.16812956e+00 5.94903946e-01
1.22104096e+00 -5.13819933e-01 -1.95207790e-01 4.32208061e-01
1.36114943e+00 -5.44202507e-01 -5.58511436e-01 6.38216496e-01
4.35217589e-01 -4.27136689e-01 7.44723141e-01 -2.92507827e-01
-5.88583410e-01 -3.88610154e-01 -1.89695552e-01 -1.29510093e+00
5.32192811e-02 -1.06723177e+00 -8.52763236e-01 1.58420134e+00
7.96828747e-01 -6.22960031e-01 8.16715956e-01 6.37880862e-01
-1.03229649e-01 -1.05576587e+00 -1.05111468e+00 -1.09945548e+00
-4.43858355e-01 -7.70751595e-01 8.06702018e-01 7.34805584e-01
-2.87822157e-01 3.29695523e-01 -8.16034496e-01 3.52241695e-01
3.98468435e-01 -2.59728462e-01 6.39588952e-01 -8.98799956e-01
-5.82973421e-01 5.46730161e-02 -2.96241380e-02 -1.33314645e+00
1.56801138e-02 -6.94793940e-01 -2.48417482e-02 -1.26200199e+00
3.96781527e-02 -8.38990629e-01 -6.31704450e-01 4.67568338e-01
-3.97147119e-01 -2.57520795e-01 -4.79493104e-02 1.52886361e-01
-9.89449441e-01 1.01273561e+00 5.06209731e-01 -1.96816921e-01
-6.08834982e-01 5.01861155e-01 -9.46134150e-01 4.90697682e-01
1.01001012e+00 -5.08496642e-01 -8.79933119e-01 -4.84671652e-01
2.01562822e-01 -2.58968472e-01 -1.48303688e-01 -1.10071743e+00
5.65119326e-01 1.43864557e-01 -2.27262941e-03 -8.67528737e-01
3.78267944e-01 -8.22301984e-01 2.47582346e-01 2.91655898e-01
-1.64791808e-01 5.12810387e-02 5.39306939e-01 1.05490780e+00
1.31516576e-01 -9.52465385e-02 4.15603042e-01 2.41183609e-01
-8.85461807e-01 4.80043530e-01 -1.16608910e-01 -1.29030451e-01
1.04787982e+00 6.92350641e-02 4.59660329e-02 -5.03695548e-01
-6.95984840e-01 4.49969411e-01 1.15098283e-01 7.49537051e-01
6.62553847e-01 -1.19387376e+00 -2.83623040e-01 -1.70516334e-02
1.87586918e-01 -6.87290402e-03 5.23635685e-01 8.70005488e-01
1.05834015e-01 4.14740771e-01 4.78912026e-01 -3.33194435e-01
-1.32426691e+00 3.63710046e-01 -7.37735853e-02 -2.98290908e-01
-4.46076542e-01 1.16769612e+00 -3.10625613e-01 -3.27274829e-01
7.89230168e-01 -2.86762923e-01 1.04276843e-01 -3.92772146e-02
7.72918344e-01 5.61324596e-01 5.64463675e-01 -9.79763716e-02
-3.82842630e-01 1.50976591e-02 -6.08703852e-01 -9.57278684e-02
1.12261987e+00 -5.30027509e-01 3.15609962e-01 7.05710769e-01
8.90728235e-01 9.62939411e-02 -1.40511549e+00 -7.77593076e-01
1.76008582e-01 -4.58330750e-01 4.36324477e-01 -1.06460953e+00
-9.69978750e-01 4.38432962e-01 1.05714202e+00 -3.30507398e-01
1.46003675e+00 -2.07345098e-01 1.25404680e+00 4.80924338e-01
5.14467537e-01 -1.34242761e+00 -1.20112859e-01 3.58691931e-01
7.17735291e-01 -1.24023652e+00 1.31172061e-01 -2.33895183e-01
-8.12360585e-01 5.10319948e-01 5.83660960e-01 6.13329291e-01
9.49595213e-01 3.22484314e-01 -3.15042466e-01 2.07309008e-01
-9.47375119e-01 -1.14025727e-01 1.58406749e-01 3.89636606e-01
-1.19207919e-01 1.06400728e-01 -4.51188087e-01 1.11239159e+00
5.43948598e-02 9.27104875e-02 1.75079033e-01 9.99614418e-01
-2.96516299e-01 -1.21051204e+00 -2.61090230e-03 7.73881197e-01
-4.48712707e-01 -4.97035861e-01 -6.73545301e-02 2.02311069e-01
1.71201564e-02 7.73534596e-01 -3.04693818e-01 -8.62348437e-01
2.68531471e-01 2.62306303e-01 1.77971497e-01 -5.19508839e-01
-3.11148435e-01 -1.70831457e-01 4.38918583e-02 -2.81581998e-01
1.95076510e-01 -6.06278718e-01 -1.12842250e+00 -3.94253731e-02
-6.52988493e-01 3.37671697e-01 9.39963043e-01 7.22305477e-01
1.07384026e+00 6.68840110e-01 8.16786826e-01 -4.38827008e-01
-6.67775989e-01 -9.83855724e-01 -3.51580352e-01 -2.73907393e-01
1.47877678e-01 -7.76723325e-01 -2.91508049e-01 -1.79578051e-01] | [9.941373825073242, 3.6127612590789795] |
09f2d25c-0db1-417d-89f7-cdb03e175a85 | noun-paraphrasing-based-on-a-variety-of | null | null | https://aclanthology.org/Y14-1073 | https://aclanthology.org/Y14-1073.pdf | Noun Paraphrasing Based on a Variety of Contexts | null | ['Tomoyuki Kajiwara', 'Kazuhide Yamamoto'] | 2014-12-01 | noun-paraphrasing-based-on-a-variety-of-1 | https://aclanthology.org/Y14-1073 | https://aclanthology.org/Y14-1073.pdf | paclic-2014-12 | ['lexical-analysis'] | ['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] | [-7.387618064880371, 3.639005184173584] |
6f51d1a6-8fd3-4cd4-9d3b-156dc7dc194c | compressing-neural-network-by-tensor-network | 2305.06058 | null | https://arxiv.org/abs/2305.06058v1 | https://arxiv.org/pdf/2305.06058v1.pdf | Compressing neural network by tensor network with exponentially fewer variational parameters | Neural network (NN) designed for challenging machine learning tasks is in general a highly nonlinear mapping that contains massive variational parameters. High complexity of NN, if unbounded or unconstrained, might unpredictably cause severe issues including over-fitting, loss of generalization power, and unbearable cost of hardware. In this work, we propose a general compression scheme that significantly reduces the variational parameters of NN by encoding them to multi-layer tensor networks (TN's) that contain exponentially-fewer free parameters. Superior compression performance of our scheme is demonstrated on several widely-recognized NN's (FC-2, LeNet-5, and VGG-16) and datasets (MNIST and CIFAR-10), surpassing the state-of-the-art method based on shallow tensor networks. For instance, about 10 million parameters in the three convolutional layers of VGG-16 are compressed in TN's with just $632$ parameters, while the testing accuracy on CIFAR-10 is surprisingly improved from $81.14\%$ by the original NN to $84.36\%$ after compression. Our work suggests TN as an exceptionally efficient mathematical structure for representing the variational parameters of NN's, which superiorly exploits the compressibility than the simple multi-way arrays. | ['Shi-Ju Ran', 'Ke Li', 'Peng-Fei Zhou', 'Yong Qing'] | 2023-05-10 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [-1.49753362e-01 4.43451367e-02 9.67874303e-02 -5.30525267e-01
-4.98618692e-01 -3.33844066e-01 -7.39778653e-02 -4.51294154e-01
-8.02080750e-01 7.02262044e-01 -1.82544142e-01 -4.23265368e-01
-2.95058370e-01 -6.28524005e-01 -9.94484603e-01 -8.92709494e-01
-5.19142151e-01 3.52684706e-01 -3.15534249e-02 -1.72880307e-01
1.41537204e-01 4.47038978e-01 -1.53033292e+00 1.25733852e-01
5.09948611e-01 1.64316010e+00 1.88959405e-01 6.07892513e-01
-1.74117967e-01 6.53566897e-01 -5.33123910e-01 -8.88466001e-01
4.92398024e-01 1.61741927e-01 -7.94429839e-01 -6.11166477e-01
9.44706023e-01 -4.42460239e-01 -8.01364183e-01 1.27237129e+00
2.16361046e-01 1.72214046e-01 5.18207371e-01 -9.60750520e-01
-7.85548270e-01 9.87722635e-01 -2.49765560e-01 2.91924328e-01
-7.18985438e-01 6.02996089e-02 1.05372977e+00 -7.03581631e-01
1.26527175e-01 1.19515002e+00 9.33046699e-01 7.97884285e-01
-9.25334394e-01 -8.86165738e-01 -9.39757153e-02 2.14158818e-01
-1.41598666e+00 -4.52799946e-01 6.24702990e-01 -2.67557472e-01
1.26232481e+00 3.87725770e-01 6.42356038e-01 9.00555491e-01
1.36801094e-01 5.29294908e-01 4.26412642e-01 8.22340623e-02
2.50635087e-01 -1.86018527e-01 2.95975119e-01 9.07497883e-01
5.17674565e-01 9.24341567e-03 -4.67661411e-01 -2.07280383e-01
1.00666058e+00 -1.24164008e-01 -1.18385807e-01 1.32622972e-01
-1.16624260e+00 7.55695760e-01 8.01819742e-01 1.02660358e-01
-2.20208049e-01 8.28706682e-01 6.54158294e-01 3.30358714e-01
1.65303722e-01 4.10804242e-01 -7.59287894e-01 -4.34076518e-01
-8.63902628e-01 4.62782443e-01 6.89557672e-01 9.91501629e-01
5.95883250e-01 6.81815445e-01 1.68516174e-01 1.03121984e+00
2.90919870e-01 4.97018576e-01 7.89255917e-01 -1.04688549e+00
6.84446633e-01 3.51027578e-01 -3.95918608e-01 -8.87935579e-01
-3.91239256e-01 -9.62589443e-01 -1.51146245e+00 -4.01674330e-01
2.56231874e-01 -1.46656260e-01 -9.59163666e-01 1.77965188e+00
-3.04640010e-02 3.46283391e-02 6.35708570e-02 8.51563752e-01
9.59764481e-01 7.59968102e-01 -1.22131236e-01 7.51010105e-02
1.30477977e+00 -9.24758792e-01 -6.06869459e-01 -2.47031704e-01
8.00935030e-01 -5.31194389e-01 9.79642928e-01 4.61834341e-01
-1.30466688e+00 -5.20771027e-01 -1.13611531e+00 -3.04358900e-01
-2.54208773e-01 1.83825299e-01 1.12225020e+00 6.28924251e-01
-1.04193211e+00 1.02789164e+00 -8.74968767e-01 4.46220368e-01
9.08506036e-01 7.43571162e-01 -2.44378909e-01 -8.55284482e-02
-1.16353035e+00 5.60681105e-01 5.56171596e-01 5.74640453e-01
-6.46360099e-01 -1.21892762e+00 -6.45830631e-01 2.38557026e-01
-1.12119921e-01 -4.90130246e-01 1.02714169e+00 -2.47090220e-01
-1.35927784e+00 3.25376362e-01 8.28354061e-02 -7.25059569e-01
2.03171968e-01 -2.62847953e-02 -4.32264149e-01 -1.64490640e-02
-5.24323642e-01 7.51018107e-01 7.69017637e-01 -5.71881890e-01
-2.35222355e-01 -6.19529545e-01 5.51601266e-03 -9.46910083e-02
-8.59559059e-01 -3.13785762e-01 -3.35954815e-01 -7.70707071e-01
5.76067507e-01 -1.03794646e+00 -3.37770313e-01 5.88811152e-02
-4.66957986e-01 -3.11012894e-01 8.06332469e-01 -5.30225277e-01
1.24035084e+00 -2.08496380e+00 1.08781591e-01 1.17442466e-01
6.29178286e-01 5.47245383e-01 -1.12963259e-01 -2.17451021e-01
5.80651313e-02 2.79455006e-01 -2.68317848e-01 -3.49945545e-01
1.44156694e-01 5.83772838e-01 -2.41468564e-01 4.55466002e-01
-8.20058584e-02 9.99347448e-01 -3.55235457e-01 -2.30021566e-01
-2.59739757e-01 6.95846915e-01 -8.84792328e-01 -5.25102466e-02
-1.29010558e-01 -3.26442897e-01 -2.78222561e-01 6.02002144e-01
8.55381131e-01 -3.84504169e-01 -1.39534384e-01 -7.47380614e-01
1.41658664e-01 3.36246878e-01 -8.34576964e-01 1.58903694e+00
-5.37623139e-03 7.66197860e-01 -5.64696826e-02 -1.26844871e+00
1.02325439e+00 1.18462630e-01 2.92564780e-01 -7.09885657e-01
2.22946629e-01 4.93539035e-01 2.66644865e-01 -4.69671965e-01
5.47997415e-01 9.50491801e-03 -5.11772744e-02 2.92510420e-01
3.71727616e-01 6.38359636e-02 1.98795974e-01 -9.78621282e-03
8.93262029e-01 -3.11880559e-01 -4.53742892e-01 -3.31176341e-01
1.88056052e-01 -3.33272994e-01 6.45116389e-01 5.82081735e-01
-2.13906810e-01 3.56882542e-01 6.06824160e-01 -7.89325833e-01
-1.47058547e+00 -7.36256301e-01 -4.62786525e-01 8.64790022e-01
-4.91214663e-01 -4.75915670e-01 -9.06631589e-01 -2.12256476e-01
1.04903556e-01 3.09345454e-01 -6.61565185e-01 -3.20183843e-01
-8.44220221e-01 -1.19098616e+00 1.24344277e+00 5.59752703e-01
9.39014614e-01 -5.83676577e-01 -4.09318507e-01 9.84660164e-02
-5.67274503e-02 -1.31783307e+00 -4.17887032e-01 3.27194661e-01
-1.47587276e+00 -6.32412672e-01 -6.24241590e-01 -7.06567705e-01
5.35672545e-01 -1.16002955e-01 8.85048866e-01 -3.66065738e-04
-1.29113495e-01 -5.79977989e-01 9.49692428e-02 -2.52169490e-01
-6.47798087e-03 2.20778212e-01 3.83035868e-01 -1.88702866e-01
3.47468644e-01 -8.02773297e-01 -6.05039477e-01 1.45257384e-01
-9.41228449e-01 -3.49518135e-02 8.13992620e-01 9.14204538e-01
7.35582113e-01 1.04896858e-01 2.20365167e-01 -9.01890516e-01
3.97112995e-01 -2.13215873e-01 -6.69615030e-01 7.81573504e-02
-7.61764526e-01 4.31883574e-01 9.31786358e-01 -7.36134946e-01
-4.76559460e-01 -1.83872476e-01 -1.83252558e-01 -9.36144829e-01
3.91092241e-01 6.52728140e-01 2.25693002e-01 -5.90460300e-01
7.17650652e-01 2.91625857e-01 -1.13272086e-01 -8.83086681e-01
2.07189750e-02 4.49728817e-01 8.07680845e-01 -6.45278215e-01
8.43615592e-01 3.49313207e-02 3.43931049e-01 -8.16230834e-01
-7.72614062e-01 1.22002937e-01 -3.32698822e-01 1.86457679e-01
2.92264074e-01 -7.63871014e-01 -1.10604811e+00 6.32027030e-01
-1.14473796e+00 -2.64766783e-01 -1.31443426e-01 6.77039862e-01
-1.55424446e-01 1.42506644e-01 -1.01688695e+00 -4.43894297e-01
-5.88282943e-01 -1.16843224e+00 5.04099131e-01 6.83329254e-02
3.58439982e-01 -8.27833235e-01 -3.46598506e-01 5.13878942e-01
8.99127483e-01 -1.51633620e-02 1.16914380e+00 -3.69405746e-01
-6.12061262e-01 -2.02684462e-01 -5.80373645e-01 8.57332826e-01
-6.03496253e-01 -1.42021611e-01 -1.01899636e+00 -3.52785259e-01
1.40207425e-01 -4.45387691e-01 9.79401767e-01 4.79754329e-01
1.74765623e+00 -7.63791859e-01 1.31219357e-01 1.45415246e+00
1.52753758e+00 1.03423156e-01 7.88627267e-01 -3.30951549e-02
1.06910717e+00 1.33729741e-01 -8.08729008e-02 3.03119481e-01
3.07903588e-01 3.92823189e-01 6.71691298e-01 1.77963674e-01
4.46348824e-02 -1.08414060e-02 1.29576653e-01 1.57931316e+00
-5.55168986e-01 1.70425866e-02 -7.53429651e-01 1.53938323e-01
-1.41313553e+00 -6.57825351e-01 -1.49013251e-01 1.93119574e+00
9.36972797e-01 2.70903170e-01 -3.23773712e-01 1.78811535e-01
2.48488516e-01 2.55911678e-01 -8.39378417e-01 -5.93470693e-01
-2.60024428e-01 2.03084931e-01 1.00535119e+00 1.47108659e-01
-9.51857388e-01 7.49757648e-01 6.46484566e+00 9.49767172e-01
-1.20358503e+00 1.29669964e-01 7.70358562e-01 -5.48746467e-01
-2.61850178e-01 -3.96647245e-01 -1.06643605e+00 4.67281193e-01
1.44454885e+00 1.14577331e-01 7.97722518e-01 8.76394689e-01
-3.04559857e-01 4.56587732e-01 -1.12312973e+00 1.54243743e+00
-1.00235835e-01 -1.70122075e+00 3.14128846e-01 3.22530538e-01
6.27904832e-01 5.22462070e-01 4.36193705e-01 4.98910189e-01
-5.72131462e-02 -1.35337055e+00 6.41367614e-01 4.29320663e-01
1.07644987e+00 -8.95403147e-01 7.87864506e-01 1.91987142e-01
-9.40041125e-01 -1.64386034e-01 -1.22770202e+00 9.61029064e-03
-2.61593103e-01 8.09518576e-01 -3.67859066e-01 6.69336841e-02
1.14435554e+00 4.70763475e-01 -5.62159181e-01 7.80374885e-01
4.85959858e-01 8.03062916e-01 -5.72562158e-01 -3.29213709e-01
5.82101583e-01 -2.07232684e-01 3.11568469e-01 1.06783128e+00
3.48854870e-01 3.16599488e-01 -5.40982723e-01 9.82501507e-01
-6.03697121e-01 -1.81451350e-01 -2.80776829e-01 -4.12730247e-01
4.26928103e-01 1.06277025e+00 -8.57662410e-02 -3.20778251e-01
-8.01607501e-03 5.98257124e-01 5.70969462e-01 4.21451271e-01
-9.23934758e-01 -5.93876958e-01 8.74735117e-01 -2.84434199e-01
4.71587688e-01 -4.07647103e-01 -4.30458754e-01 -1.19127619e+00
2.95504451e-01 -5.73352456e-01 6.30312413e-02 -4.63223517e-01
-1.01606464e+00 1.11801863e+00 -2.01150164e-01 -9.45993245e-01
-9.99500602e-02 -9.76275921e-01 -3.65285705e-05 8.79838526e-01
-1.43073320e+00 -7.47487903e-01 -1.75619200e-01 6.29084349e-01
1.90450579e-01 -5.11162102e-01 9.21588421e-01 6.91682696e-01
-8.74627471e-01 1.26628983e+00 2.56832093e-01 2.69882143e-01
6.00348376e-02 -8.92928302e-01 5.78475177e-01 5.38879693e-01
-9.20820162e-02 7.94973016e-01 4.89414603e-01 -1.51940182e-01
-1.84947586e+00 -1.21554613e+00 9.40097988e-01 7.76672736e-02
6.07189536e-01 -5.03811955e-01 -1.17401814e+00 6.72074080e-01
-4.20987397e-01 5.37777305e-01 5.98770440e-01 -4.54793051e-02
-6.96703970e-01 -4.57416147e-01 -1.18607366e+00 4.49197352e-01
1.17937207e+00 -5.17419577e-01 -1.31013155e-01 4.84187543e-01
9.99418080e-01 -7.23647594e-01 -1.33707726e+00 4.41266328e-01
7.20333517e-01 -7.54228771e-01 1.17514050e+00 -7.62520790e-01
5.41349471e-01 1.82186589e-01 -6.54913604e-01 -9.63275433e-01
-4.62262243e-01 -6.40265644e-01 -5.18395662e-01 6.97689116e-01
3.93887669e-01 -8.12495470e-01 1.10127628e+00 8.11523259e-01
-4.25748050e-01 -1.26168144e+00 -1.13634217e+00 -7.88008451e-01
2.26182207e-01 -6.69323325e-01 9.18558180e-01 8.54256749e-01
-3.97133321e-01 5.14226779e-02 -2.85521060e-01 1.22858770e-02
9.67436016e-01 -4.19083893e-01 2.57928848e-01 -1.21445417e+00
-3.79205167e-01 -5.96177220e-01 -7.02075958e-01 -1.26529086e+00
-3.56419757e-02 -9.33057487e-01 -3.74336123e-01 -9.04476643e-01
-5.31439495e-04 -9.93966460e-01 -3.72178316e-01 6.35624588e-01
4.29291070e-01 5.65525889e-01 -4.03226614e-02 5.06180823e-01
-1.21233389e-01 6.24845564e-01 1.51872277e+00 -1.78292006e-01
2.51317382e-01 -2.05906659e-01 -6.37175381e-01 6.04418159e-01
7.20213473e-01 -3.72331619e-01 -4.07911181e-01 -1.07275045e+00
4.76003587e-01 -2.79939044e-02 3.53944600e-01 -1.00344944e+00
4.97770727e-01 4.99849133e-02 3.67372215e-01 -6.52909040e-01
5.91983318e-01 -5.68654418e-01 1.97575048e-01 4.22594696e-01
-4.14605349e-01 2.18225092e-01 3.18416864e-01 3.01628232e-01
-2.40072325e-01 -3.06914091e-01 7.57148027e-01 -1.44453242e-01
-3.81245613e-01 8.35485816e-01 2.67427653e-01 1.33711427e-01
4.72975582e-01 -1.15965240e-01 -6.10017359e-01 1.00117780e-01
-3.80044341e-01 -1.54934824e-01 -1.57633170e-01 7.15419874e-02
9.90759730e-01 -1.44001555e+00 -6.44889653e-01 5.73649704e-01
-4.02042449e-01 4.52171892e-01 4.76188481e-01 7.07558095e-01
-8.82053912e-01 8.24359894e-01 -2.71215409e-01 -5.38727045e-01
-7.54803360e-01 1.11472853e-01 5.35249591e-01 -3.26449051e-02
-6.38356924e-01 1.51408470e+00 -1.08832464e-01 -4.10072714e-01
5.95444381e-01 -7.22917497e-01 5.71986660e-02 -3.53542715e-01
4.85434115e-01 4.31702226e-01 3.83277148e-01 -5.33341408e-01
-1.83690891e-01 5.51256001e-01 -2.31786817e-01 2.45944262e-01
1.68011498e+00 3.33677411e-01 -4.04066861e-01 2.73723215e-01
1.83981383e+00 -6.53546453e-01 -1.20708382e+00 -4.39975381e-01
-4.84989136e-01 -2.19558328e-01 5.53846896e-01 -4.39233005e-01
-1.77988684e+00 1.04083860e+00 5.83126485e-01 5.64725175e-02
9.61198270e-01 -1.18747137e-01 1.31198192e+00 9.90923285e-01
2.82708764e-01 -9.39234078e-01 -1.85040444e-01 7.65340388e-01
8.14761639e-01 -9.22412992e-01 -2.43539866e-02 -1.76203236e-01
-1.51748225e-01 1.36588120e+00 5.13844252e-01 -2.26566091e-01
8.62497747e-01 2.83330917e-01 -3.25090557e-01 -4.03863579e-01
-9.38898444e-01 6.76640213e-01 4.63072479e-01 3.02637875e-01
2.20306695e-01 2.22304210e-01 2.09416717e-01 6.20277345e-01
-9.25329387e-01 -1.58282399e-01 2.93136209e-01 4.99836743e-01
-2.75313795e-01 -5.68620861e-01 -1.25611961e-01 9.72925186e-01
-6.37450635e-01 -4.06488806e-01 3.93519342e-01 5.42445779e-01
7.07215071e-02 2.73718625e-01 3.33088875e-01 -8.09886932e-01
1.09415658e-01 -9.74254012e-02 5.67231357e-01 -1.52582265e-02
-4.93022591e-01 -1.95022762e-01 2.80567980e-03 -7.71876514e-01
-8.28079730e-02 -4.49124038e-01 -1.28372967e+00 -8.88368309e-01
-9.41761211e-02 2.97063356e-03 1.18157470e+00 8.03475976e-01
1.78847790e-01 6.05634451e-01 2.74729520e-01 -8.10950160e-01
-1.06541228e+00 -1.00967836e+00 -7.10765898e-01 1.10220432e-01
4.52070057e-01 -5.15056849e-01 -6.35753274e-01 -1.28425643e-01] | [8.525885581970215, 3.090400457382202] |
939c786a-6306-49c3-8012-8e67c3506aa8 | controlling-large-language-models-to-generate | 2302.05319 | null | https://arxiv.org/abs/2302.05319v2 | https://arxiv.org/pdf/2302.05319v2.pdf | Large Language Models for Code: Security Hardening and Adversarial Testing | Large language models (LMs) are increasingly pretrained on massive codebases and used to generate code. However, LMs lack awareness of security and are found to frequently produce unsafe code. This work studies the security of LMs along two important axes: (i) security hardening, which aims to enhance LMs' reliability in generating secure code, and (ii) adversarial testing, which seeks to evaluate LMs' security at an adversarial standpoint. We address both of these by formulating a new security task called controlled code generation. The task is parametric and takes as input a binary property to guide the LM to generate secure or unsafe code, while preserving the LM's capability of generating functionally correct code. We propose a novel learning-based approach called SVEN to solve this task. SVEN leverages property-specific continuous vectors to guide program generation towards the given property, without modifying the LM's weights. Our training procedure optimizes these continuous vectors by enforcing specialized loss terms on different regions of code, using a high-quality dataset carefully curated by us. Our extensive evaluation shows that SVEN is highly effective in achieving strong security control. For instance, a state-of-the-art CodeGen LM with 2.7B parameters generates secure code for 59.1% of the time. When we employ SVEN to perform security hardening (or adversarial testing) on this LM, the ratio is significantly boosted to 92.3% (or degraded to 36.8%). Importantly, SVEN closely matches the original LMs in functional correctness. | ['Martin Vechev', 'Jingxuan He'] | 2023-02-10 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 2.88576454e-01 7.16113895e-02 -3.32786798e-01 -3.07217538e-02
-1.32511353e+00 -1.09109008e+00 3.47264916e-01 2.36021370e-01
-1.11335970e-01 4.74824637e-01 -1.60272177e-02 -1.01348913e+00
3.14912438e-01 -1.01618683e+00 -1.08467972e+00 -2.82188803e-01
-1.78970084e-01 -2.07344934e-01 1.38957709e-01 -4.97452587e-01
3.04933906e-01 1.50532395e-01 -1.26895356e+00 4.49147642e-01
1.29153717e+00 5.50316632e-01 -2.67420113e-01 9.12585735e-01
1.26815885e-01 8.63278389e-01 -8.87037456e-01 -7.61268854e-01
2.14978978e-01 -6.60955906e-02 -8.20736825e-01 -5.38220882e-01
2.61728942e-01 -2.13188335e-01 9.57393721e-02 1.34807181e+00
2.12419540e-01 -2.80420512e-01 2.78893292e-01 -1.39457405e+00
-7.39961982e-01 1.11216903e+00 -4.85973805e-01 -2.99564004e-01
4.63526964e-01 2.65100151e-01 1.14298522e+00 -6.01957500e-01
2.68744469e-01 1.05393481e+00 6.41271830e-01 9.71210122e-01
-1.48391128e+00 -8.35146129e-01 -1.20099448e-01 -5.18891096e-01
-1.65392351e+00 -3.52428198e-01 5.98106623e-01 -6.69011652e-01
1.13551557e+00 4.84538078e-01 1.05008483e-02 1.15123475e+00
4.64930952e-01 4.06894684e-01 7.34982133e-01 -4.28148955e-01
3.06269228e-01 4.01429445e-01 -7.99629930e-03 6.72224224e-01
4.17922348e-01 5.42139411e-01 -1.61384732e-01 -7.39586115e-01
-8.42084549e-03 -4.31349814e-01 -2.82215983e-01 -1.57382429e-01
-9.48404133e-01 9.93644416e-01 2.31108010e-01 1.02003217e-02
3.84876311e-01 4.17839020e-01 6.01469517e-01 4.39276636e-01
5.79234622e-02 9.68313575e-01 -6.80670738e-01 -2.32123002e-01
-9.10500526e-01 4.00141567e-01 8.04485321e-01 8.89317453e-01
6.73334658e-01 3.26382548e-01 -1.40622154e-01 5.69523871e-01
3.86526465e-01 5.98213136e-01 2.70725220e-01 -4.97225702e-01
9.31949854e-01 6.48155570e-01 -2.33819544e-01 -9.04583037e-01
1.39473110e-01 -3.85847509e-01 -6.00441873e-01 6.31428480e-01
3.17621864e-02 -3.58345121e-01 -5.75215042e-01 2.02716875e+00
-7.53560215e-02 2.21587680e-02 3.75248730e-01 3.37303847e-01
2.44893968e-01 7.61725008e-01 4.22587395e-02 1.43748447e-01
1.00717008e+00 -7.38773525e-01 -6.61956593e-02 -2.53125817e-01
8.71187866e-01 -6.53360724e-01 1.50921118e+00 4.69916105e-01
-9.85121012e-01 -4.54556972e-01 -1.46271038e+00 3.25609744e-01
-2.87617505e-01 -3.94947529e-02 5.17821252e-01 1.34928489e+00
-1.32897198e+00 3.59117389e-01 -7.62610316e-01 5.46959877e-01
4.30790186e-01 4.54697460e-01 -2.80432016e-01 1.88217938e-01
-1.20923007e+00 4.92262006e-01 4.90642101e-01 -5.54145157e-01
-1.39930618e+00 -1.07155669e+00 -1.26180947e+00 2.32155293e-01
3.78869951e-01 -3.87866795e-01 1.24304473e+00 -7.60551989e-01
-1.46766663e+00 5.80396056e-01 2.64288276e-01 -4.15492684e-01
4.07608181e-01 -1.87802866e-01 -6.12300694e-01 -2.62534827e-01
-6.33746535e-02 2.30943546e-01 9.94900048e-01 -1.47280681e+00
-3.18405509e-01 2.79557228e-01 5.05405843e-01 -5.34396708e-01
-7.54033446e-01 2.52198845e-01 -3.09343904e-01 -9.66843665e-01
-7.52318978e-01 -1.00343227e+00 -1.21694297e-01 -1.94904193e-01
-5.84243059e-01 1.68690518e-01 6.01384461e-01 -6.92570150e-01
1.70196784e+00 -2.34398627e+00 1.55884027e-01 5.98327696e-01
2.18463525e-01 5.53038418e-01 -4.71812069e-01 3.30024749e-01
-2.64318794e-01 6.74804509e-01 -6.88649714e-01 -1.31676599e-01
1.75640598e-01 -1.54202029e-01 -7.76347876e-01 1.87659130e-01
5.48188090e-01 9.97574568e-01 -9.31392729e-01 -2.06713364e-01
-2.69236088e-01 4.61654037e-01 -1.29343534e+00 2.60017365e-01
-5.00276804e-01 -2.28727888e-02 -2.81638414e-01 6.90411806e-01
6.04049146e-01 -3.33358906e-02 2.14730844e-01 3.39941442e-01
1.43840641e-01 1.03093766e-01 -8.80602300e-01 1.59998357e+00
-8.92164946e-01 3.97851467e-01 -3.59351113e-02 -3.49969000e-01
8.26926887e-01 1.54675767e-01 -9.07650292e-02 -2.95808971e-01
-6.61174804e-02 2.31880441e-01 -9.78923216e-03 -2.74472207e-01
4.37665999e-01 7.50592202e-02 -6.98000789e-01 7.04803824e-01
-3.19316089e-01 -4.49455887e-01 -2.20566448e-02 4.50074106e-01
1.43695045e+00 2.85158120e-02 7.94417188e-02 -2.38794357e-01
1.06649971e+00 -2.87720948e-01 4.86505002e-01 6.85154140e-01
2.02459946e-01 3.29391152e-01 9.61035967e-01 -1.21703088e-01
-1.06142974e+00 -1.14265299e+00 3.86473611e-02 1.04277790e+00
-1.91365406e-01 -1.00563931e+00 -9.32923377e-01 -1.21314454e+00
5.03750779e-02 1.10671449e+00 -5.77272832e-01 -8.54440153e-01
-6.81209385e-01 -6.17802978e-01 1.16960239e+00 4.43261504e-01
2.54806101e-01 -8.39880407e-01 -3.79714072e-01 -1.29621834e-01
7.38492294e-04 -6.36278093e-01 -8.65645766e-01 -3.40727046e-02
-1.99817061e-01 -1.08338046e+00 -7.71231428e-02 -5.69748521e-01
1.01020706e+00 -2.93999553e-01 1.23493183e+00 7.35307932e-01
-2.92154223e-01 -3.39347236e-02 -4.75171715e-01 -1.53430045e-01
-1.37184274e+00 2.96526611e-01 1.33853564e-02 -2.39896148e-01
-2.68512875e-01 -4.90096539e-01 -9.78157371e-02 1.95003211e-01
-1.45117986e+00 -3.46360654e-01 4.98616040e-01 9.95434642e-01
2.05435053e-01 4.33620483e-01 3.84853870e-01 -1.10646927e+00
8.68688643e-01 -5.13664007e-01 -1.20768917e+00 4.48022574e-01
-8.44134152e-01 4.40968692e-01 1.20171976e+00 -4.74639475e-01
-8.23724031e-01 -1.54017046e-01 -5.15415311e-01 -1.97022542e-01
4.09499139e-01 5.09951293e-01 -4.37821150e-01 -5.70448041e-01
1.15493000e+00 2.07768485e-01 -1.73394054e-01 -9.66257509e-03
3.22841436e-01 4.85073924e-01 5.22482753e-01 -1.25354815e+00
1.61052883e+00 -5.14098257e-02 -2.14579538e-01 -2.01554932e-02
-3.91971678e-01 2.90494293e-01 2.37591546e-02 3.10465217e-01
4.75263149e-01 -7.27323174e-01 -7.85987735e-01 4.26886320e-01
-8.94016623e-01 -7.97567368e-01 -1.64916620e-01 -1.85226187e-01
-3.72169167e-01 5.24896860e-01 -6.07977331e-01 -6.59632742e-01
-6.14335775e-01 -1.54715276e+00 1.22911894e+00 -2.51368731e-01
-2.59273827e-01 -7.94442952e-01 2.41062790e-01 2.14948595e-01
5.50949812e-01 3.93153459e-01 1.40952194e+00 -4.45276886e-01
-4.55924958e-01 -4.07912791e-01 1.24358945e-01 8.03419113e-01
5.08643575e-02 3.53878200e-01 -9.17082906e-01 -7.04627633e-01
-1.51648998e-01 -5.62865138e-01 5.88558912e-01 -4.28717554e-01
1.40965712e+00 -6.45609021e-01 -3.07218246e-02 1.03537750e+00
1.56332099e+00 5.44630475e-02 6.79204226e-01 2.30284601e-01
4.57998276e-01 1.79469079e-01 3.90246630e-01 3.45421344e-01
6.48006052e-02 6.79491282e-01 6.04363859e-01 2.25827098e-01
1.11167349e-01 -5.48494041e-01 9.79428113e-01 5.52390099e-01
3.84420335e-01 -1.07969597e-01 -1.28540695e+00 1.95626438e-01
-1.34166098e+00 -7.38169193e-01 4.01684731e-01 2.26170015e+00
1.39541674e+00 4.38555032e-01 -1.31919265e-01 3.62539053e-01
2.89853543e-01 -5.98252099e-03 -2.45621249e-01 -6.07239604e-01
6.66925088e-02 4.84473348e-01 4.04491872e-01 7.40714133e-01
-1.04577982e+00 9.93160188e-01 5.87664413e+00 1.17789900e+00
-1.21519899e+00 7.06145838e-02 6.12641811e-01 1.01450242e-01
-9.79451180e-01 1.89356863e-01 -8.35472763e-01 6.14334702e-01
1.12758470e+00 -5.66375732e-01 8.19100976e-01 1.13263416e+00
-2.37832934e-01 3.18090051e-01 -9.69978929e-01 5.02915025e-01
1.02528334e-02 -1.40438092e+00 1.15086868e-01 -4.63574938e-02
9.04318810e-01 -3.81128699e-01 4.59444493e-01 8.01826477e-01
6.30467176e-01 -1.30501235e+00 9.93566573e-01 1.73663974e-01
1.39427841e+00 -1.18677437e+00 6.68887317e-01 3.88458371e-01
-1.19711268e+00 -2.74922669e-01 -6.70629367e-02 2.82400459e-01
-1.24777131e-01 3.79347384e-01 -8.14488292e-01 3.20527196e-01
3.71681035e-01 1.92732543e-01 -9.39687610e-01 5.18407941e-01
-6.30539358e-01 7.14777529e-01 1.18182071e-01 8.65925923e-02
5.37708439e-02 3.32399398e-01 3.97956491e-01 1.11880577e+00
2.95071155e-01 -4.34996516e-01 1.43260643e-01 1.32335031e+00
-3.74787003e-01 -1.13213196e-01 -5.26801229e-01 -1.97658077e-01
5.38248658e-01 1.19063032e+00 -2.57800698e-01 -1.30019367e-01
-3.28445584e-01 7.39957035e-01 4.17679816e-01 2.13033766e-01
-1.23816466e+00 -7.69132137e-01 9.54195082e-01 -7.20281824e-02
1.30641043e-01 -2.24097803e-01 -2.87502587e-01 -1.23686826e+00
1.78259596e-01 -1.54499805e+00 2.29985103e-01 -2.91691869e-01
-9.60644960e-01 1.01052809e+00 -1.11111388e-01 -1.01610398e+00
-3.79817635e-01 -4.87379700e-01 -6.76917195e-01 1.04061961e+00
-1.42295063e+00 -1.15847933e+00 5.45743480e-02 5.58832467e-01
2.34916449e-01 -5.81325591e-01 9.55999792e-01 4.30894881e-01
-5.44958353e-01 1.56510520e+00 -1.84103325e-01 1.94858998e-01
4.54439819e-01 -1.22992241e+00 9.69168842e-01 1.30174291e+00
-1.42206669e-01 1.14902842e+00 5.56025326e-01 -7.98959851e-01
-1.55160832e+00 -1.51218736e+00 7.68868983e-01 -8.42450857e-01
9.07003582e-01 -7.77512968e-01 -8.53136182e-01 4.70271051e-01
-2.56334394e-01 2.73631420e-02 7.89502859e-01 -2.08721951e-01
-1.08091617e+00 9.36962292e-03 -1.27910268e+00 7.47425020e-01
7.88705289e-01 -8.88071716e-01 -1.68655813e-01 2.53246158e-01
1.30709541e+00 -5.34581304e-01 -8.58044863e-01 3.46671283e-01
3.20514172e-01 -8.10991526e-01 1.07214928e+00 -8.00853610e-01
8.18307340e-01 -4.95883703e-01 -3.88853550e-01 -1.15866733e+00
-5.10217287e-02 -9.91052508e-01 -3.00140381e-01 1.30301762e+00
8.03362489e-01 -6.14062250e-01 7.14677155e-01 6.02856278e-01
-2.45652989e-01 -7.18585670e-01 -5.55366993e-01 -9.06067550e-01
5.20489752e-01 -7.20017076e-01 1.03709030e+00 8.84660363e-01
1.63577259e-01 -1.25746310e-01 -2.91050762e-01 3.56887996e-01
4.15961474e-01 -1.37841672e-01 8.26044977e-01 -6.08856857e-01
-7.59049714e-01 -5.80909669e-01 -2.33218551e-01 -3.89712602e-01
7.41613150e-01 -1.26414716e+00 6.33755550e-02 -5.30614853e-01
1.60076603e-01 -6.10566676e-01 -1.62215129e-01 8.28712761e-01
-4.00934756e-01 6.61593378e-02 1.76572844e-01 -1.66356668e-01
-2.50189662e-01 4.47211832e-01 5.94740331e-01 -4.32429552e-01
-1.11958273e-01 1.89256221e-01 -1.13200819e+00 4.12564963e-01
8.49198103e-01 -4.71245676e-01 -6.32851243e-01 -3.04367810e-01
8.63757074e-01 -1.57065442e-04 2.56272256e-01 -1.09960973e+00
4.32384461e-02 -1.63115546e-01 -2.86405593e-01 7.91492611e-02
-1.89387128e-01 -6.89011574e-01 2.14838102e-01 7.31553853e-01
-5.75318515e-01 2.79183745e-01 5.58033466e-01 4.13310707e-01
-1.67354837e-01 -3.17268193e-01 8.33399892e-01 2.83860505e-01
-4.17976588e-01 2.64461011e-01 -1.90689698e-01 3.06807905e-01
1.10344934e+00 2.84022480e-01 -5.23078024e-01 -1.43226504e-01
-1.81229800e-01 9.01586711e-02 8.71258914e-01 5.10448575e-01
7.31606424e-01 -1.29564679e+00 -7.87771523e-01 7.27197111e-01
4.02884334e-01 -3.60987663e-01 -2.45756563e-02 2.38454476e-01
-6.79682672e-01 1.91098407e-01 8.70319083e-02 -2.13582709e-01
-1.30355597e+00 9.37965810e-01 2.26947978e-01 -5.51480293e-01
-2.19467968e-01 1.19010031e+00 1.46335140e-01 -6.13059461e-01
2.73449253e-02 -2.73790300e-01 2.43679270e-01 -5.75356305e-01
6.16056740e-01 3.30866650e-02 1.26237690e-01 -3.27725202e-01
-3.86428386e-01 3.49222660e-01 -1.14891492e-01 -4.88774441e-02
1.08657563e+00 6.70735300e-01 -4.13448125e-01 -3.36313158e-01
1.36804676e+00 8.07392538e-01 -1.09393013e+00 7.12198839e-02
4.18616692e-03 -6.05030537e-01 -1.73520207e-01 -9.29361939e-01
-1.15955210e+00 6.63388014e-01 1.28908381e-01 1.52641103e-01
1.19432950e+00 -3.46284479e-01 8.55743527e-01 2.56063253e-01
7.12195694e-01 -6.54703677e-01 2.88848609e-01 6.19061232e-01
8.21146607e-01 -1.04475176e+00 -3.23032707e-01 -4.73945349e-01
-4.81337130e-01 9.57198977e-01 7.98703432e-01 8.94377679e-02
4.84281093e-01 9.24853623e-01 -1.15257025e-01 2.41682619e-01
-8.35287273e-01 5.79140782e-01 3.11415881e-01 5.88518083e-01
2.60620743e-01 1.01870395e-01 2.57038604e-02 8.87451947e-01
-5.56037724e-01 -3.00628603e-01 5.92313290e-01 9.87217009e-01
-5.62089146e-04 -1.53883994e+00 -4.68158215e-01 3.48257899e-01
-6.14911854e-01 -5.52250028e-01 -2.13026017e-01 4.09003675e-01
1.61796048e-01 1.02574289e+00 -5.61469793e-01 -8.90313804e-01
1.18107207e-01 -1.36172786e-01 3.29989381e-02 -8.98822606e-01
-1.05107510e+00 -5.05475819e-01 1.75456539e-01 -6.57125354e-01
5.16541481e-01 -3.38093728e-01 -1.11908877e+00 -5.43463528e-01
-2.39476070e-01 3.23113024e-01 5.49138963e-01 3.49209636e-01
3.51658106e-01 6.84477508e-01 9.10592079e-01 -3.71686518e-01
-1.06924832e+00 -1.73384085e-01 -2.65152961e-01 4.30701286e-01
5.33901632e-01 -2.56451458e-01 -5.48279285e-01 8.22877735e-02] | [7.047336101531982, 7.8436055183410645] |
4b22b193-f262-40d0-91f9-63fc339b58ee | neighborhood-mixture-model-for-knowledge-base | 1606.06461 | null | http://arxiv.org/abs/1606.06461v3 | http://arxiv.org/pdf/1606.06461v3.pdf | Neighborhood Mixture Model for Knowledge Base Completion | Knowledge bases are useful resources for many natural language processing
tasks, however, they are far from complete. In this paper, we define a novel
entity representation as a mixture of its neighborhood in the knowledge base
and apply this technique on TransE-a well-known embedding model for knowledge
base completion. Experimental results show that the neighborhood information
significantly helps to improve the results of the TransE model, leading to
better performance than obtained by other state-of-the-art embedding models on
three benchmark datasets for triple classification, entity prediction and
relation prediction tasks. | ['Mark Johnson', 'Lizhen Qu', 'Kairit Sirts', 'Dat Quoc Nguyen'] | 2016-06-21 | neighborhood-mixture-model-for-knowledge-base-1 | https://aclanthology.org/K16-1005 | https://aclanthology.org/K16-1005.pdf | conll-2016-8 | ['triple-classification'] | ['graphs'] | [-4.70123082e-01 1.59266785e-01 -7.56391644e-01 -9.37554613e-02
-2.67338306e-01 -3.08708400e-01 6.48561478e-01 5.26636064e-01
-6.53199196e-01 9.70210612e-01 6.46747530e-01 -2.36965373e-01
-4.43998963e-01 -1.09623981e+00 -6.05616450e-01 -2.12488547e-01
-1.36900872e-01 6.51488423e-01 3.51588219e-01 -5.66585124e-01
-1.75504580e-01 3.09006155e-01 -1.37367547e+00 4.06575680e-01
7.29707837e-01 7.29168773e-01 -3.03671926e-01 8.60733613e-02
-4.00358379e-01 9.34875906e-01 -1.69620082e-01 -9.24331248e-01
-1.92604065e-02 2.27056578e-01 -1.09658420e+00 -5.96153617e-01
1.22419111e-01 -6.67915419e-02 -8.95524681e-01 9.14757192e-01
3.78433377e-01 1.74356252e-01 8.58282208e-01 -1.17424643e+00
-1.40490377e+00 9.19898629e-01 -1.02860317e-01 2.75027484e-01
6.65705562e-01 -4.80288327e-01 1.47796690e+00 -1.44769216e+00
9.68699515e-01 1.18335474e+00 9.72654939e-01 2.47904435e-01
-1.21988189e+00 -3.60593796e-01 -8.52853805e-02 9.06406522e-01
-1.79761183e+00 -4.68128830e-01 5.98203838e-01 -3.80813271e-01
1.56537461e+00 1.39673457e-01 5.61807752e-01 9.12878931e-01
1.10197864e-01 8.10802639e-01 7.61192501e-01 -6.64431036e-01
-1.40074447e-01 4.35847819e-01 6.63965523e-01 9.18408751e-01
6.20806396e-01 -5.78138232e-02 -6.55033588e-01 -2.25641638e-01
4.69268471e-01 -7.42542371e-02 -5.94779193e-01 -5.83621919e-01
-1.48864555e+00 7.31737256e-01 7.13854790e-01 4.70881969e-01
-7.62668610e-01 -1.74150497e-01 3.43573809e-01 4.71603006e-01
6.35551453e-01 3.68357033e-01 -4.02165473e-01 -6.33577034e-02
-2.79772252e-01 1.66461483e-01 1.22728062e+00 1.01160681e+00
6.57226503e-01 -2.81463087e-01 -3.47793847e-01 9.14551318e-01
1.94719836e-01 1.62483841e-01 3.31564873e-01 -2.24471137e-01
7.33605862e-01 1.12925041e+00 2.42396638e-01 -1.15594733e+00
-3.86307389e-01 -1.93575516e-01 -9.09114182e-01 -5.11168361e-01
7.93754086e-02 9.94075537e-02 -6.08749926e-01 1.49321651e+00
4.84547943e-01 4.48955536e-01 6.59607768e-01 5.80981374e-01
1.28610730e+00 7.83335805e-01 1.71389103e-01 2.78346743e-02
1.47705650e+00 -1.06264758e+00 -1.07813561e+00 -2.94521451e-05
8.03122103e-01 -4.54812378e-01 5.64209938e-01 -1.06448285e-01
-6.70663953e-01 -3.69599372e-01 -1.11152399e+00 -2.92503089e-01
-1.09451580e+00 1.98671073e-01 1.14996302e+00 6.52999043e-01
-8.90141129e-01 5.37421763e-01 -7.79412806e-01 -4.23814446e-01
4.00221735e-01 1.20888337e-01 -9.85196948e-01 -5.90954602e-01
-1.59902632e+00 1.41584885e+00 1.15770829e+00 1.36844516e-01
-1.68654338e-01 -1.05780542e+00 -1.14648569e+00 1.79210559e-01
4.21947747e-01 -1.01784480e+00 5.72461247e-01 9.21280608e-02
-1.08688724e+00 7.25939393e-01 -1.86762467e-01 -5.48817337e-01
7.18437806e-02 -3.71313155e-01 -8.26339185e-01 -2.57858515e-01
-1.06407762e-01 3.59915525e-01 1.75106958e-01 -1.09365737e+00
-5.22831976e-01 -2.90851533e-01 2.72613615e-01 2.12615252e-01
-8.36260259e-01 -2.14792460e-01 -7.18364716e-01 -4.66308802e-01
-1.27721757e-01 -8.29355061e-01 1.11645823e-02 -1.79870233e-01
-4.19252485e-01 -6.35428429e-01 6.74981773e-01 -8.29991817e-01
1.72295642e+00 -1.87832916e+00 3.79208505e-01 1.82568833e-01
3.35367829e-01 6.85851514e-01 -1.88123003e-01 8.02636266e-01
-1.85110226e-01 7.71183446e-02 1.30379170e-01 -2.20461175e-01
2.79530972e-01 4.74198073e-01 -3.81115735e-01 1.76119447e-01
1.82544202e-01 1.47986233e+00 -1.06213820e+00 -4.72108215e-01
2.28015497e-01 8.19765687e-01 -3.56862396e-01 -2.78761722e-02
4.28935476e-02 -3.70595217e-01 -4.51514184e-01 6.75011456e-01
5.16948521e-01 -1.40510648e-01 4.34156686e-01 -5.82313478e-01
3.17758560e-01 4.09278184e-01 -1.25111425e+00 1.49252200e+00
-4.71807420e-01 4.04395998e-01 -5.17749965e-01 -9.43991125e-01
7.14855731e-01 4.92917597e-01 3.07631493e-01 -2.88109452e-01
-1.42347023e-01 1.30781248e-01 -1.41691074e-01 -4.96101230e-01
9.07887578e-01 -1.30865678e-01 -4.48068455e-02 1.17520034e-01
4.13278371e-01 4.44444388e-01 4.56391305e-01 3.91021460e-01
1.14172053e+00 -1.59356028e-01 6.10917628e-01 -2.04651192e-01
5.37245333e-01 -1.58379246e-02 6.66630328e-01 5.46847224e-01
6.04385249e-02 3.10128536e-02 2.61896551e-01 -5.09189010e-01
-7.96905100e-01 -1.04142904e+00 -1.98873609e-01 8.12262237e-01
9.12230089e-02 -9.94024158e-01 -1.50902107e-01 -9.06470954e-01
5.05804062e-01 7.86694467e-01 -7.70778179e-01 -4.57667381e-01
-3.29578340e-01 -7.53179193e-01 6.07210994e-01 9.12159801e-01
4.16445971e-01 -1.04852664e+00 3.05327684e-01 4.05917913e-01
-2.77556658e-01 -1.36351371e+00 -2.60207176e-01 -3.55112664e-02
-8.11549664e-01 -1.25779450e+00 -3.67462218e-01 -8.90495718e-01
3.43575001e-01 2.38984600e-01 1.36577976e+00 6.04681186e-02
-1.45694003e-01 5.63335776e-01 -6.18449092e-01 -3.38668853e-01
-2.64460802e-01 2.31267244e-01 3.27566087e-01 -9.66453180e-02
9.33582842e-01 -4.98752862e-01 -1.41148835e-01 1.50284599e-04
-9.61686134e-01 -3.43647748e-01 4.68147993e-01 1.05120897e+00
6.10920250e-01 2.11529955e-01 8.76828492e-01 -1.14846551e+00
7.58413672e-01 -6.30464375e-01 -2.31562719e-01 7.77688444e-01
-7.61306882e-01 3.26684684e-01 4.85055357e-01 -1.37365356e-01
-9.53056753e-01 -2.72522539e-01 -1.23181418e-02 -5.08644164e-01
2.65286714e-01 1.20067239e+00 -2.03087777e-01 -1.01604424e-01
4.84029472e-01 1.80721253e-01 -3.85886639e-01 -6.20834649e-01
8.68574798e-01 5.99632144e-01 3.81066084e-01 -4.55934048e-01
7.63464570e-01 2.50941277e-01 -9.92866233e-02 -8.32067490e-01
-9.83627319e-01 -7.35075295e-01 -8.85173440e-01 3.21742266e-01
6.90248787e-01 -1.02632236e+00 -4.96148497e-01 -4.73083332e-02
-1.25601745e+00 4.22909737e-01 -2.75032997e-01 7.07384229e-01
-9.54579785e-02 3.46063882e-01 -5.41504920e-01 -4.12469864e-01
-4.72068369e-01 -5.50934911e-01 6.29948497e-01 9.26880836e-02
2.82625277e-02 -1.47395122e+00 4.81029689e-01 3.19045275e-01
4.05240744e-01 8.16646740e-02 1.11143482e+00 -7.92220652e-01
-5.75889885e-01 -4.64389175e-01 -2.37321585e-01 2.94405490e-01
1.53566003e-01 -2.79597253e-01 -6.85797691e-01 -1.88514099e-01
-6.89558506e-01 -2.59211987e-01 1.43710876e+00 -3.90203953e-01
6.72094285e-01 -2.62750894e-01 -7.95609772e-01 5.78530967e-01
1.62995434e+00 -3.47419888e-01 7.89890468e-01 3.82351965e-01
8.81115794e-01 3.04252416e-01 6.70322239e-01 3.74753535e-01
8.93309772e-01 7.13601232e-01 8.06808993e-02 2.64346808e-01
-2.69170076e-01 -5.87731302e-01 2.75613010e-01 1.19839585e+00
-3.36942941e-01 -6.84229359e-02 -1.03348839e+00 1.04186237e+00
-2.05277491e+00 -1.14538157e+00 -1.55857140e-02 1.81142867e+00
1.10996556e+00 -2.61106759e-01 -2.45652810e-01 1.53519427e-02
5.64378858e-01 2.82899439e-01 -1.02240577e-01 -1.35402054e-01
-2.55307257e-01 2.79925555e-01 4.34521019e-01 3.37016106e-01
-1.28660369e+00 1.24505043e+00 6.50516176e+00 8.55535626e-01
-4.38108355e-01 2.54361868e-01 -2.30380043e-01 4.39796507e-01
-4.36034709e-01 -1.07334577e-01 -1.09335923e+00 5.65367006e-02
8.76768470e-01 -5.17723739e-01 2.01132134e-01 8.56010556e-01
-6.91524863e-01 3.47741932e-01 -1.51651788e+00 1.12811053e+00
2.19954818e-01 -1.64680481e+00 2.59063929e-01 -1.44382164e-01
7.74440765e-01 2.98682097e-02 -2.32071862e-01 9.73954499e-01
4.72968519e-01 -1.02555382e+00 1.09398484e-01 7.47023702e-01
6.95891976e-01 -8.55421543e-01 1.00483537e+00 2.54573792e-01
-1.42452550e+00 5.41555323e-02 -7.96316266e-01 1.91178322e-01
1.78038672e-01 6.55668855e-01 -9.15604651e-01 1.25239456e+00
5.03984332e-01 9.17224228e-01 -7.61968315e-01 1.01408744e+00
-3.73884052e-01 4.29103464e-01 -2.57735431e-01 -4.04077433e-02
-9.86459702e-02 -1.17787980e-01 5.11386096e-01 1.45421731e+00
5.74829578e-02 2.14493304e-01 6.14005700e-02 6.64223015e-01
-6.90744221e-01 1.10541821e-01 -1.01581264e+00 -2.88880289e-01
8.33390117e-01 1.20103896e+00 -1.27513856e-01 -3.63710821e-01
-7.43283093e-01 8.00818861e-01 9.32614446e-01 3.42255026e-01
-7.13028371e-01 -7.28102684e-01 7.64717638e-01 -4.39318508e-01
6.74890637e-01 -1.27967194e-01 1.59113675e-01 -1.58906424e+00
3.12019497e-01 -4.82378393e-01 6.54427230e-01 -3.81033450e-01
-1.69037116e+00 5.03523886e-01 5.76088242e-02 -1.03364336e+00
-1.89403743e-01 -8.02044511e-01 -2.87529856e-01 7.65859544e-01
-1.91866422e+00 -1.43475688e+00 -9.71629247e-02 6.72886193e-01
-2.13941887e-01 -3.68514568e-01 1.37937546e+00 6.34984612e-01
-5.27289271e-01 8.38995755e-01 5.68063557e-01 5.96010625e-01
5.98877490e-01 -1.37508643e+00 3.73310626e-01 5.41669250e-01
7.50726759e-01 9.17041838e-01 2.21573964e-01 -5.58961153e-01
-1.87866807e+00 -1.22560942e+00 1.52956378e+00 -9.95098650e-01
9.21091020e-01 -7.46077746e-02 -1.15308964e+00 1.21206963e+00
1.70510978e-01 4.21571195e-01 1.01799524e+00 8.52703631e-01
-7.96963811e-01 -1.47131056e-01 -8.26872945e-01 4.12152976e-01
1.03055835e+00 -8.11159134e-01 -1.42414129e+00 2.11863205e-01
7.58809090e-01 -1.97734326e-01 -1.64827240e+00 5.48142314e-01
5.40651500e-01 -4.32326764e-01 1.44442534e+00 -1.05537260e+00
3.26032192e-01 -2.45313928e-01 -3.35054815e-01 -1.57312965e+00
-3.88577729e-01 -2.24869475e-01 -8.17739904e-01 1.53762436e+00
7.34218895e-01 -5.84688485e-01 7.32733428e-01 6.10562980e-01
2.16029789e-02 -8.04156959e-01 -1.04710770e+00 -1.06771934e+00
6.02430009e-05 -1.98745146e-01 6.80312037e-01 1.47890830e+00
4.66506749e-01 8.18458974e-01 -2.38395214e-01 2.44309068e-01
4.89467829e-01 2.98003107e-01 6.11369908e-01 -1.39212227e+00
-1.18474495e-02 -1.40121117e-01 -8.66243899e-01 -8.40672970e-01
5.98270416e-01 -1.46338356e+00 -6.39995575e-01 -2.00602388e+00
2.95056999e-01 -4.89250839e-01 -5.75309753e-01 7.14729607e-01
-4.91857767e-01 8.14309046e-02 2.10195780e-01 8.77447799e-02
-8.79662991e-01 1.15694892e+00 7.93157279e-01 -4.11655337e-01
-1.08910844e-01 -3.42934936e-01 -6.91676736e-01 4.33125705e-01
3.33255351e-01 -4.77788508e-01 -2.60141969e-01 -4.92744505e-01
4.26974356e-01 -8.11609179e-02 2.21798524e-01 -7.38023221e-01
5.15679657e-01 2.91078508e-01 1.93961650e-01 -5.45706511e-01
5.26246667e-01 -8.83953571e-01 7.15919212e-02 1.09504037e-01
-3.99506614e-02 6.83341548e-02 3.23422849e-01 9.20872867e-01
-7.23832130e-01 -2.15762541e-01 1.20695524e-01 2.95086622e-01
-1.21431053e+00 3.61851007e-01 2.13509724e-01 1.00619256e-01
9.34886515e-01 4.36722822e-02 -5.10790229e-01 2.73473449e-02
-8.50648224e-01 3.75614911e-01 7.99006373e-02 7.70371735e-01
8.88140023e-01 -2.02380252e+00 -7.77449071e-01 -5.64377382e-02
7.66407132e-01 -3.38211000e-01 6.12218566e-02 8.42845559e-01
-2.63621837e-01 8.27043831e-01 -6.33480549e-02 -8.46745819e-02
-1.16663718e+00 8.49993825e-01 2.75312830e-02 -6.96722567e-01
-6.22110844e-01 7.96067238e-01 -3.96090806e-01 -8.25132668e-01
2.28732079e-01 -2.53745764e-01 -6.91342175e-01 2.53301233e-01
6.34947598e-01 4.04826164e-01 3.10886890e-01 -6.56460464e-01
-6.49800420e-01 3.91789377e-01 -3.48888397e-01 2.68611491e-01
1.54619110e+00 1.60566732e-01 -3.99386376e-01 4.23557192e-01
1.25019264e+00 -4.74110357e-02 -3.27480316e-01 -8.95870984e-01
2.83042639e-01 -5.98303795e-01 1.71100393e-01 -5.86500466e-01
-7.09179461e-01 5.16640663e-01 1.90148622e-01 1.06794499e-01
6.85901463e-01 8.05940405e-02 7.37236083e-01 1.12406182e+00
5.32534003e-01 -8.89187455e-01 -4.22861248e-01 8.91343117e-01
8.54790151e-01 -1.24211967e+00 2.23252252e-01 -6.59046590e-01
-6.28040314e-01 9.38040733e-01 3.62363815e-01 8.18881318e-02
1.15658951e+00 -1.48627594e-01 -3.31219137e-01 -3.21532488e-01
-1.10709918e+00 -5.09539306e-01 7.03675151e-01 8.88287544e-01
5.21394670e-01 1.81694269e-01 -3.38883549e-01 9.96426582e-01
-7.37096602e-03 1.68121666e-01 1.79405823e-01 8.69804502e-01
-2.59107947e-01 -1.37226868e+00 3.61959040e-02 4.97694671e-01
-3.24976861e-01 -4.60309446e-01 -4.37877029e-01 1.06246316e+00
-1.16354235e-01 6.72589481e-01 -3.29876482e-01 -6.55268013e-01
6.99192226e-01 3.88011992e-01 4.44336265e-01 -7.56539464e-01
-4.32861418e-01 -8.63898814e-01 5.95373034e-01 -4.87828523e-01
-5.02164304e-01 -2.86966532e-01 -9.76993144e-01 -4.74249214e-01
-6.25421822e-01 3.24097186e-01 2.61507601e-01 7.19100118e-01
6.23459995e-01 3.70405734e-01 1.71668068e-01 -4.42887366e-01
-6.50812685e-01 -9.31626260e-01 -6.56047881e-01 5.44311106e-01
-5.52670546e-02 -9.72921312e-01 2.74363626e-02 -1.81350589e-01] | [8.860496520996094, 7.987196922302246] |
a4887d95-0479-430e-a6af-54fa466eb02c | 3d-photogrammetry-point-cloud-segmentation | null | null | https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29CP.1943-5487.0000929 | https://www.researchgate.net/profile/Meida-Chen/publication/344177868_3D_Photogrammetry_Point_Cloud_Segmentation_Using_a_Model_Ensembling_Framework/links/5f592c17a6fdcc116404704d/3D-Photogrammetry-Point-Cloud-Segmentation-Using-a-Model-Ensembling-Framework.pdf | 3D photogrammetry point cloud segmentation using a model ensembling framework | The US Army is paying increased attention to the development of rapid three-dimensional (3D) reconstruction using photogrammetry and unmanned aerial vehicle (UAV) technologies for creating virtual environments and simulations in areas of interest. The ability of the intelligence community, mission commanders, and front-line soldiers to understand their deployed physical environment in advance is critical in the planning and rehearsal phases of any military operation. In order to achieve various simulation capabilities such as destruction operations, route planning, and explosive-standoff distances computation among others, reconstructed 3D data needs to be properly attributed. In this paper, we introduce a model ensembling framework for segmenting a 3D photogrammetry point cloud into top-level terrain elements (i.e., ground, human-made objects, and vegetation). Preprocessing and postprocessing methods were designed to overcome the data segmentation challenges posed by photogrammetric data-quality issues. A large UAV-based photogrammetric database was created for validation purposes. The designed model ensembling framework was compared with existing point cloud segmentation algorithms, and it outperformed other algorithms and achieved the best F1-score. Because the ultimate goal of segmenting a photogrammetric-generated point cloud is to create realistic virtual environments for simulation. Qualitative results for creating virtual environments using the segmented data are also discussed in this paper. | ['Lucio Soibelman', 'Ryan McAlinden', 'Pratusha Bhuvana Prasad', 'Kyle McCullough', 'Andrew Feng', 'Meida Chen'] | 2020-11-01 | null | null | null | journal-of-computing-in-civil-engineering | ['point-cloud-segmentation'] | ['computer-vision'] | [ 1.75441042e-01 -3.19571495e-01 4.01841849e-01 -1.82097122e-01
-3.54836136e-01 -7.86687136e-01 4.74867851e-01 4.02614295e-01
-3.33778203e-01 5.04749954e-01 -3.62971991e-01 -7.64873445e-01
-4.73359764e-01 -1.16621637e+00 -2.37847731e-01 -1.86919838e-01
-4.47989196e-01 1.08892941e+00 2.35659219e-02 -9.01592076e-01
4.69123214e-01 1.43763387e+00 -1.37209344e+00 -4.01192516e-01
9.38328624e-01 8.32790196e-01 3.77182901e-01 6.72517300e-01
1.63353551e-02 -3.72972786e-01 -4.38487977e-01 -1.29602343e-01
8.30722094e-01 4.01677229e-02 -6.33329988e-01 3.22551131e-01
1.34533480e-01 -3.32909375e-01 1.79769784e-01 9.13188398e-01
3.43096226e-01 4.06098902e-01 5.73077023e-01 -1.09781075e+00
5.64968362e-02 -2.58061051e-01 -5.95166624e-01 7.25768507e-02
3.71918172e-01 1.48977995e-01 5.46077788e-01 -7.65066862e-01
5.87049425e-01 9.00576115e-01 8.02197754e-01 -2.18794987e-01
-9.16719735e-01 -3.02948534e-01 -2.09167749e-01 -1.61183506e-01
-1.52414167e+00 1.48281008e-01 6.40982926e-01 -7.55082905e-01
9.14692998e-01 6.48376286e-01 1.12903273e+00 1.30965039e-01
6.18646622e-01 -2.35966733e-03 8.64283741e-01 -4.23545301e-01
4.49551076e-01 -2.91861087e-01 -1.20938487e-01 4.89808232e-01
3.98197562e-01 3.65650892e-01 1.28925428e-01 -2.66519278e-01
1.23705840e+00 2.36329529e-02 -2.40585968e-01 -3.67171258e-01
-8.78621161e-01 6.17128909e-01 6.77137256e-01 4.50009704e-02
-9.54391181e-01 -1.81404024e-01 2.21312568e-01 -1.35950223e-01
3.54950279e-01 5.86511493e-01 -1.23274118e-01 -2.31610343e-01
-1.25567448e+00 3.36859971e-01 3.30171496e-01 8.94616544e-01
6.25976145e-01 4.68211025e-01 5.82304358e-01 3.09372604e-01
2.97439158e-01 5.80132008e-01 -5.38826808e-02 -8.22333038e-01
4.91614819e-01 9.42012727e-01 2.06027418e-01 -1.33118105e+00
-5.40548146e-01 -5.04714549e-01 -6.56814575e-01 5.41976035e-01
-2.96771020e-01 -7.17362985e-02 -1.22567308e+00 5.72217286e-01
5.83501577e-01 -2.07299992e-01 -4.86838818e-02 1.14909923e+00
2.93842375e-01 6.54852331e-01 9.97212529e-02 -1.22825091e-03
1.04127586e+00 -8.13586414e-02 -2.97723651e-01 -3.79989624e-01
3.24717939e-01 -6.54439926e-01 7.49913454e-01 2.01688379e-01
-7.87381828e-01 -5.49982727e-01 -1.08133471e+00 6.13510430e-01
-1.07119992e-01 -2.06851542e-01 7.37324953e-01 5.70347309e-01
-7.15160608e-01 7.47452557e-01 -7.96006978e-01 -2.70228952e-01
3.89213800e-01 2.67571986e-01 -4.35447454e-01 -2.76263785e-02
-9.73825276e-01 1.06931341e+00 3.34615201e-01 4.44944650e-01
-9.21635032e-01 -4.77082610e-01 -8.68637383e-01 -2.05632433e-01
-2.06272514e-03 -5.85178733e-01 7.99617290e-01 -4.83043641e-01
-7.68053174e-01 6.85539722e-01 4.25804317e-01 -4.32067007e-01
5.82726717e-01 1.25419006e-01 -3.73559952e-01 1.48236677e-01
9.70475450e-02 2.21119747e-01 4.04450685e-01 -1.73205328e+00
-6.24782205e-01 -8.33191037e-01 3.94536778e-02 6.51713371e-01
2.83564329e-01 -1.07783824e-01 -1.22092791e-01 -4.21861380e-01
6.78086162e-01 -9.52040792e-01 -5.63218772e-01 -3.87462288e-01
-1.85129166e-01 6.51289105e-01 8.83128703e-01 -1.00468040e+00
7.92289317e-01 -1.86809886e+00 3.61115150e-02 6.19609714e-01
-1.94378421e-01 3.35849375e-01 2.77745008e-01 8.02473485e-01
-1.05506495e-01 1.93673581e-01 -7.07923174e-01 5.49511313e-02
-2.59140939e-01 2.53560692e-01 -2.99946457e-01 5.79519928e-01
-3.01020086e-01 4.63239640e-01 -7.23177850e-01 -3.02568227e-01
6.71287239e-01 4.29186523e-01 2.51225997e-02 -2.18979985e-01
-1.24726342e-02 5.70177913e-01 -4.54722613e-01 1.17880952e+00
8.93249452e-01 7.12441146e-01 -2.08726302e-02 -3.11818011e-02
-5.66458941e-01 -3.96024734e-01 -1.01167512e+00 1.54868639e+00
-4.16703045e-01 4.41590011e-01 5.19074261e-01 -5.58009923e-01
1.47614634e+00 2.37316206e-01 7.51108766e-01 -5.62720120e-01
3.92381757e-01 2.39124179e-01 -1.95343122e-01 -2.93731242e-01
1.16551721e+00 -6.86781108e-01 -1.55660525e-01 -2.08566003e-02
-4.99288946e-01 -9.60139990e-01 -2.71593988e-01 -5.52647598e-02
4.63286191e-01 2.43934110e-01 2.32276365e-01 -2.99956977e-01
9.04701054e-02 1.17412221e+00 5.21251976e-01 -5.78205921e-02
-1.09119363e-01 4.85485524e-01 -2.14054361e-01 -5.85987210e-01
-1.19469750e+00 -9.76069570e-01 -2.94548899e-01 3.41847651e-02
5.42266846e-01 2.69617904e-02 -6.55448914e-01 1.22308906e-03
8.30438957e-02 8.96474898e-01 -1.54575929e-01 2.37719893e-01
-4.26980019e-01 -7.08362997e-01 2.21898809e-01 3.07111055e-01
5.43255866e-01 -5.35977781e-01 -1.19295037e+00 4.25352097e-01
-2.99500357e-02 -8.84671986e-01 2.59061843e-01 -2.61246920e-01
-1.26262987e+00 -1.19302142e+00 -3.51636916e-01 -4.83701020e-01
7.54645467e-01 5.60547769e-01 7.56520927e-01 1.39202490e-01
-5.21286011e-01 5.64983904e-01 -5.78879118e-01 -6.40929341e-01
-2.68734485e-01 -3.89643073e-01 1.17898606e-01 -3.83606881e-01
-1.33065432e-01 -5.56702018e-01 -4.16890383e-01 6.36830330e-01
-9.58924890e-01 1.80787772e-01 3.74363154e-01 2.82485336e-01
9.49229002e-01 6.53768063e-01 4.77287173e-02 -3.69141996e-01
7.33779013e-01 -3.01700860e-01 -7.10557818e-01 1.11021981e-01
-1.55485854e-01 -7.66737103e-01 3.35911155e-01 1.57248870e-01
-8.28867435e-01 3.00776005e-01 -2.18360290e-01 -3.86937886e-01
-3.46823424e-01 9.62319255e-01 -2.26635739e-01 -3.71285349e-01
7.36162961e-01 8.35284442e-02 1.01328984e-01 -2.66108423e-01
6.00868054e-02 8.16854894e-01 6.49090946e-01 -4.32085782e-01
1.11261702e+00 5.97870052e-01 3.49676102e-01 -1.17884266e+00
7.43668824e-02 -5.40121555e-01 -1.14741993e+00 -7.20701277e-01
5.80133975e-01 -6.31952107e-01 -2.61333361e-02 2.03449160e-01
-9.52430248e-01 -3.50175023e-01 -3.30752194e-01 4.95157629e-01
-5.25661826e-01 5.73846936e-01 1.77244827e-01 -1.00731206e+00
-4.56613272e-01 -1.35875487e+00 8.21800590e-01 1.44082420e-02
-1.86797634e-01 -8.55735481e-01 6.47977591e-02 3.97461414e-01
1.73281968e-01 1.14400589e+00 5.65619648e-01 -1.18660383e-01
-3.65656763e-01 -7.62669861e-01 1.66548714e-01 1.40583217e-01
9.57201049e-02 2.28482351e-01 -4.23512608e-01 -2.69488454e-01
8.78775939e-02 3.93765628e-01 1.60592750e-01 5.37677467e-01
5.51914155e-01 6.25699759e-02 -4.24133450e-01 5.76883614e-01
1.64473307e+00 5.97734272e-01 8.09291542e-01 5.96482992e-01
4.20805693e-01 7.72099912e-01 1.38294053e+00 5.76913178e-01
1.34755746e-01 7.41359591e-01 1.05107999e+00 -3.14852089e-01
2.52263576e-01 -1.64987296e-01 -2.56745279e-01 3.74466836e-01
-6.87336743e-01 7.91718960e-02 -1.51252854e+00 5.28961182e-01
-1.43692780e+00 -8.88878584e-01 -6.07566535e-01 2.53006506e+00
-8.24427158e-02 -6.44301400e-02 -6.57686964e-02 3.70121151e-01
6.39267564e-01 -1.41965300e-01 -3.27045798e-01 -4.76195306e-01
1.48139790e-01 2.15675279e-01 9.63838100e-01 6.27488911e-01
-9.61326718e-01 9.81914878e-01 5.44762325e+00 4.83773351e-01
-9.93869841e-01 -1.54550374e-01 -4.00675423e-02 3.71173352e-01
-3.19784820e-01 4.01226550e-01 -4.88838375e-01 -6.59355614e-03
6.84597611e-01 -2.72184402e-01 1.82602629e-01 7.78991580e-01
6.35329068e-01 -3.66963357e-01 -9.78688598e-02 7.83720672e-01
-9.08123776e-02 -1.34853137e+00 4.16750461e-02 4.10225451e-01
4.69350755e-01 -8.53278488e-02 -1.78360343e-01 -1.78620979e-01
1.50974721e-01 -8.55950713e-01 9.11494672e-01 5.32480955e-01
9.15198147e-01 -9.78637815e-01 7.07094252e-01 6.05920732e-01
-1.42910600e+00 -4.56786007e-02 -3.13343823e-01 -3.18224072e-01
7.73566842e-01 4.55915362e-01 -1.27110648e+00 1.16310775e+00
4.56874669e-01 5.86468391e-02 -2.05773994e-01 1.38447344e+00
4.38954942e-02 1.39484391e-01 -3.62860918e-01 4.30644721e-01
4.74435180e-01 -6.83560312e-01 7.78923750e-01 6.06964588e-01
4.76379395e-01 3.96326214e-01 2.98961878e-01 4.99081552e-01
5.66068292e-01 8.56504142e-02 -7.63436317e-01 -3.97852622e-03
6.52324080e-01 1.15412283e+00 -8.91550004e-01 3.10573447e-02
3.02025974e-01 7.25303590e-01 -4.65881795e-01 1.26229376e-01
-8.50448549e-01 -4.41394866e-01 6.50456846e-01 7.48306394e-01
-4.48504955e-01 -1.19073308e+00 -7.03268111e-01 -4.66944933e-01
-2.68564969e-01 -5.02992272e-01 -8.72408450e-02 -8.82589936e-01
-6.58098102e-01 9.46974993e-01 3.99242908e-01 -1.56406724e+00
-5.07935882e-02 -5.12659311e-01 -5.98571897e-01 1.17519450e+00
-1.12005794e+00 -1.41944659e+00 -7.32248366e-01 2.14485973e-01
7.11846828e-01 -2.26132974e-01 1.00766683e+00 -8.00886303e-02
-1.94117501e-01 -4.64836210e-01 -4.82753813e-02 -1.36576116e-01
-9.82340947e-02 -7.97339380e-01 7.27185547e-01 1.08551681e+00
-1.78709820e-01 3.68982911e-01 7.97391117e-01 -1.51315665e+00
-1.52490461e+00 -1.12420189e+00 2.03339517e-01 -1.14650227e-01
3.45727086e-01 8.10032636e-02 -6.46259069e-01 6.94162309e-01
-3.74210358e-01 -4.40542847e-01 6.21068716e-01 -2.28067771e-01
6.35272861e-01 2.62281060e-01 -1.44613290e+00 5.05932987e-01
8.60319197e-01 2.35895794e-02 -5.63481867e-01 3.30778927e-01
1.44209146e-01 -6.84622824e-01 -9.49860811e-01 6.64723217e-01
4.56118137e-01 -9.39541698e-01 1.16077268e+00 -2.43132055e-01
-2.61628255e-02 -3.90773296e-01 -2.70081222e-01 -1.54956019e+00
-1.48660213e-01 -3.38537484e-01 7.39296913e-01 6.59198463e-01
1.97925389e-01 -4.61408645e-01 8.54765773e-01 8.91776383e-01
-7.31923223e-01 -4.90263909e-01 -9.91222203e-01 -8.99714649e-01
-8.88072252e-02 -7.74852753e-01 7.72107422e-01 8.38565767e-01
-3.53572518e-01 -1.13644004e-01 -1.25582041e-02 7.59194136e-01
7.18085468e-01 1.29876375e-01 9.24417078e-01 -1.67486656e+00
3.60163122e-01 -1.17553562e-01 -8.63924742e-01 -3.14677149e-01
-1.94759220e-01 -7.15623200e-01 -2.25783944e-01 -2.14409804e+00
-5.81691384e-01 -1.04182875e+00 7.41452217e-01 1.72698990e-01
3.11780125e-01 5.57229994e-03 1.87011197e-01 5.31119406e-01
5.80976903e-01 5.19964516e-01 1.40231419e+00 8.91634524e-02
-4.85667080e-01 2.05124542e-01 -2.04455703e-01 6.31141722e-01
9.67974365e-01 -1.09544344e-01 -4.17235374e-01 -5.11557460e-01
5.17234318e-02 4.05688018e-01 5.88866949e-01 -1.26022458e+00
1.72554910e-01 -5.31778872e-01 2.52531946e-01 -1.02385986e+00
7.38064229e-01 -1.27477896e+00 8.59384000e-01 4.86433625e-01
5.72125733e-01 2.87908107e-01 5.42381644e-01 3.73835415e-01
-2.72464514e-01 -3.46946001e-01 5.65068305e-01 -3.08738470e-01
-1.10248888e+00 4.44421262e-01 -2.79957116e-01 -5.90359032e-01
1.56633830e+00 -9.10515726e-01 1.81352086e-02 -1.68214366e-01
-7.54889965e-01 1.94427550e-01 1.05206740e+00 -5.33061028e-02
1.14695573e+00 -9.12542403e-01 -7.32590437e-01 3.27653915e-01
-1.28905252e-01 3.97267163e-01 6.63900912e-01 4.35311556e-01
-1.57138801e+00 1.90602198e-01 -6.88212693e-01 -3.83216023e-01
-1.25617599e+00 2.72694081e-02 3.54108632e-01 1.19713984e-01
-6.16321087e-01 6.21891260e-01 -5.13511002e-01 -6.76068246e-01
-4.11289871e-01 -1.02066949e-01 -6.01861440e-02 -7.22374246e-02
3.02980751e-01 5.09502232e-01 4.16674614e-01 -1.21432173e+00
-3.28444332e-01 7.69937456e-01 6.74880922e-01 -3.63853097e-01
1.41551661e+00 -3.85075361e-02 1.07062206e-01 1.20514259e-02
4.62646991e-01 -1.03561051e-01 -1.16695106e+00 4.62553233e-01
-3.52039695e-01 -9.81649578e-01 5.17813206e-01 -6.00456059e-01
-7.23933160e-01 9.70575750e-01 4.75432247e-01 2.28447720e-01
1.05074894e+00 -4.55491751e-01 6.71147108e-01 -5.55615611e-02
1.11313665e+00 -9.63349700e-01 -6.74694598e-01 3.05105239e-01
1.17765772e+00 -7.93052912e-01 3.84845316e-01 -6.87679291e-01
-1.10633850e+00 1.20849240e+00 8.89913440e-02 -1.34038508e-01
4.28536147e-01 1.11708730e-01 2.66475398e-02 -6.29890382e-01
2.84772307e-01 -2.33731657e-01 1.33352563e-01 1.00149143e+00
-1.67362452e-01 4.79650944e-01 -7.63172582e-02 2.63557196e-01
-5.18877566e-01 -1.78260162e-01 5.49847543e-01 1.29252982e+00
-8.16885531e-01 -8.89275312e-01 -1.03043354e+00 3.17759335e-01
3.10435921e-01 8.08189362e-02 -3.67577016e-01 1.15340984e+00
-1.50423413e-02 7.79065490e-01 2.80782700e-01 -6.22106254e-01
7.92366207e-01 -4.45020378e-01 2.53186047e-01 -5.50866425e-01
-5.59057653e-01 -1.76411957e-01 2.81314403e-01 -1.22469574e-01
-1.22344434e-01 -6.78385794e-01 -1.49778402e+00 -4.81939763e-01
-2.22221568e-01 3.88356835e-01 1.56596661e+00 5.97072124e-01
2.37956136e-01 3.50728005e-01 7.01596141e-01 -1.37902606e+00
-2.74221599e-01 -6.97493494e-01 -9.12040353e-01 -6.30853176e-02
-4.02609974e-01 -7.97540665e-01 5.76229393e-02 -1.07283793e-01] | [8.37212085723877, -2.590467929840088] |
8ee070ee-2b82-4a03-ad78-70ba0522c973 | hrda-context-aware-high-resolution-domain | 2204.13132 | null | https://arxiv.org/abs/2204.13132v2 | https://arxiv.org/pdf/2204.13132v2.pdf | HRDA: Context-Aware High-Resolution Domain-Adaptive Semantic Segmentation | Unsupervised domain adaptation (UDA) aims to adapt a model trained on the source domain (e.g. synthetic data) to the target domain (e.g. real-world data) without requiring further annotations on the target domain. This work focuses on UDA for semantic segmentation as real-world pixel-wise annotations are particularly expensive to acquire. As UDA methods for semantic segmentation are usually GPU memory intensive, most previous methods operate only on downscaled images. We question this design as low-resolution predictions often fail to preserve fine details. The alternative of training with random crops of high-resolution images alleviates this problem but falls short in capturing long-range, domain-robust context information. Therefore, we propose HRDA, a multi-resolution training approach for UDA, that combines the strengths of small high-resolution crops to preserve fine segmentation details and large low-resolution crops to capture long-range context dependencies with a learned scale attention, while maintaining a manageable GPU memory footprint. HRDA enables adapting small objects and preserving fine segmentation details. It significantly improves the state-of-the-art performance by 5.5 mIoU for GTA-to-Cityscapes and 4.9 mIoU for Synthia-to-Cityscapes, resulting in unprecedented 73.8 and 65.8 mIoU, respectively. The implementation is available at https://github.com/lhoyer/HRDA. | ['Luc van Gool', 'Dengxin Dai', 'Lukas Hoyer'] | 2022-04-27 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 3.44257265e-01 8.61912034e-03 1.31831691e-01 -3.46249878e-01
-1.20494258e+00 -6.89197540e-01 4.41522866e-01 -3.04735489e-02
-3.75633895e-01 5.97854018e-01 -4.13842648e-02 -8.34141448e-02
4.45847332e-01 -1.20763040e+00 -9.99527991e-01 -6.32061720e-01
4.12825644e-01 7.02236593e-01 6.15177333e-01 -2.31614575e-01
-6.34068251e-02 4.05864924e-01 -1.48767519e+00 3.43566418e-01
1.15383911e+00 8.41293633e-01 7.45442569e-01 8.81459534e-01
-2.16161653e-01 1.88548416e-01 -4.34130549e-01 -1.48870677e-01
3.81056666e-01 -3.94426674e-01 -9.66268778e-01 2.49725416e-01
6.85091317e-01 -2.96480983e-01 -1.76376849e-02 9.22635257e-01
4.44725037e-01 2.27407038e-01 4.69308913e-01 -7.28734076e-01
-7.88379967e-01 2.56256670e-01 -7.77770638e-01 1.80110008e-01
-1.03167459e-01 1.82353258e-01 6.80953085e-01 -8.24814200e-01
7.67255127e-01 9.94376123e-01 7.57786334e-01 5.17584443e-01
-1.33756232e+00 -5.66723168e-01 2.73838937e-01 -2.41428331e-01
-1.29042506e+00 -2.62686342e-01 4.08001304e-01 -2.97928035e-01
1.12924218e+00 2.01152071e-01 5.41374743e-01 1.07264698e+00
5.68159521e-02 5.97971916e-01 1.21046710e+00 -2.44842663e-01
2.46796280e-01 -7.56161511e-02 -2.59429379e-03 4.56380755e-01
2.46034518e-01 4.27178107e-03 -3.74239355e-01 1.76490873e-01
1.06502473e+00 -1.20946765e-01 2.40675732e-02 -2.05622703e-01
-1.13971138e+00 6.70239568e-01 6.68014288e-01 1.70931667e-01
-4.17628944e-01 -7.60157332e-02 4.42713857e-01 3.84115875e-02
7.54103303e-01 4.40096676e-01 -8.15826237e-01 6.25262111e-02
-1.00886869e+00 1.19951941e-01 3.00004572e-01 1.15991092e+00
1.03265727e+00 1.27060458e-01 -7.88427889e-03 9.89164174e-01
-1.01582147e-01 6.41921341e-01 4.17835563e-01 -1.00946605e+00
3.28373939e-01 4.01562989e-01 9.38915312e-02 -5.95641136e-01
-4.13848788e-01 -3.60713303e-01 -8.21908295e-01 2.57802039e-01
7.52403378e-01 -2.80744713e-02 -1.55873358e+00 1.59269547e+00
5.44098318e-01 2.19987229e-01 1.43036604e-01 1.10889781e+00
6.96312845e-01 9.11884844e-01 3.69664520e-01 3.10098946e-01
1.35807145e+00 -1.29841053e+00 -1.35634854e-01 -7.28868425e-01
4.44785655e-01 -8.78462732e-01 1.66620028e+00 1.26534134e-01
-9.83846903e-01 -7.58069038e-01 -7.92482197e-01 -5.14113426e-01
-5.73672473e-01 2.80843489e-02 4.76941556e-01 3.72001529e-01
-1.00307536e+00 5.17537117e-01 -9.92995262e-01 -7.58547604e-01
6.06144488e-01 1.19669296e-01 -2.41239831e-01 -3.35087389e-01
-1.08448768e+00 5.15436351e-01 5.69964230e-01 -1.93084151e-01
-7.35565007e-01 -9.32347059e-01 -9.08801079e-01 -1.40338587e-02
3.69710177e-01 -6.94100440e-01 1.27501535e+00 -1.27172852e+00
-1.43336833e+00 1.13859224e+00 -3.26964930e-02 -5.24636567e-01
5.25747955e-01 -3.95541310e-01 -3.16445380e-01 1.19477041e-01
5.44943035e-01 1.05305219e+00 8.24157834e-01 -1.15997648e+00
-7.98929095e-01 -3.54515016e-01 -1.72914937e-01 3.27812433e-01
-3.18616740e-02 -2.61571735e-01 -7.25371480e-01 -9.62575972e-01
8.63159001e-02 -8.76845002e-01 -4.40814972e-01 -3.18679661e-02
-1.89345434e-01 3.35158587e-01 9.45296705e-01 -8.20526958e-01
8.01172912e-01 -2.04298830e+00 -8.90694335e-02 -9.84646156e-02
-7.98898637e-02 3.94414037e-01 -2.63939500e-01 -4.38538715e-02
9.83937755e-02 1.11459401e-02 -8.25560570e-01 -2.71579266e-01
-2.54027665e-01 3.52593541e-01 -2.17404678e-01 1.56063005e-01
4.01902974e-01 9.29022014e-01 -8.75356078e-01 -4.54052389e-01
3.80551875e-01 5.56165338e-01 -5.11936486e-01 1.82283163e-01
-5.17679274e-01 8.22254360e-01 -4.87969279e-01 7.24800587e-01
9.98158455e-01 -4.52422231e-01 1.60579458e-02 -6.84291124e-02
-2.79145896e-01 4.84072119e-02 -9.14894700e-01 2.04277778e+00
-6.71209812e-01 4.74754244e-01 1.42488435e-01 -7.33420551e-01
9.38508332e-01 -1.37367606e-01 2.87807673e-01 -1.02166486e+00
-5.76401018e-02 4.33094859e-01 -5.40887058e-01 -1.88501224e-01
6.66255474e-01 9.54584219e-03 -4.92610723e-01 1.43334642e-01
-4.42099385e-02 -4.03838962e-01 7.09018484e-02 2.49052607e-02
9.14732575e-01 4.99776781e-01 3.31516474e-01 -3.78955126e-01
2.38765001e-01 5.79949200e-01 7.35253692e-01 6.26379907e-01
-1.87037081e-01 1.25954425e+00 3.43821384e-02 -5.31178951e-01
-1.50012338e+00 -1.13652349e+00 -1.78005248e-01 1.31496501e+00
2.54712999e-01 -8.19323361e-02 -1.12909746e+00 -6.35726213e-01
-1.36032104e-01 6.95233703e-01 -6.03955269e-01 3.57182845e-02
-7.08800316e-01 -7.28172421e-01 3.75587374e-01 7.26675868e-01
1.01215672e+00 -1.21373844e+00 -7.27517128e-01 2.28665993e-01
-2.01994464e-01 -1.30221415e+00 -7.23753750e-01 1.21076658e-01
-7.99918532e-01 -7.65724421e-01 -1.02673519e+00 -8.62923265e-01
6.15958631e-01 2.71889836e-01 1.49068737e+00 -3.62039298e-01
-1.92757249e-01 5.85077442e-02 -3.35919052e-01 -1.23220690e-01
-3.46536934e-01 3.73404205e-01 -3.70655715e-01 -1.68028623e-01
2.27134272e-01 -5.11904359e-01 -6.22770250e-01 2.79222578e-01
-1.05163598e+00 4.81838375e-01 6.43242896e-01 8.67715597e-01
1.36491692e+00 -1.80982962e-01 4.30355668e-01 -1.39196098e+00
-8.48788302e-03 -4.06650245e-01 -7.76866913e-01 1.61147043e-01
-4.33984697e-01 -7.96505809e-02 8.56313765e-01 -4.47613746e-01
-1.38055623e+00 5.19461215e-01 -2.63713390e-01 -4.10117775e-01
-5.98904610e-01 1.60326004e-01 -2.53416568e-01 1.05798528e-01
8.39875281e-01 1.71722949e-01 -5.06207407e-01 -6.08622313e-01
5.41309953e-01 4.72407162e-01 8.83292735e-01 -6.92974687e-01
5.90121269e-01 6.14947915e-01 -3.23817998e-01 -8.00520480e-01
-1.02896655e+00 -5.28253317e-01 -8.12580943e-01 1.84046686e-01
1.12477553e+00 -1.15812826e+00 2.95453936e-01 6.82318211e-01
-7.85611928e-01 -1.04899096e+00 -5.49209893e-01 4.87839915e-02
-6.99188769e-01 8.05343837e-02 -5.36401868e-01 -1.56209141e-01
-5.34603834e-01 -9.71568108e-01 1.45961571e+00 5.36469340e-01
-1.25774667e-01 -7.50173151e-01 -4.20140363e-02 4.71007675e-01
4.12416041e-01 4.62370783e-01 5.67865551e-01 -2.48623997e-01
-7.25726128e-01 1.61084786e-01 -6.31403506e-01 2.36084133e-01
1.91359334e-02 -1.73582748e-01 -1.04406559e+00 -2.29327589e-01
-3.70412350e-01 -2.84295022e-01 1.01176608e+00 6.17213607e-01
1.30408788e+00 -5.75043559e-02 -2.37566620e-01 9.20959473e-01
1.53867996e+00 -7.97851756e-02 6.05538726e-01 5.51263213e-01
9.68132854e-01 3.16938102e-01 1.12115443e+00 2.87487239e-01
4.98844355e-01 6.59897268e-01 1.84131473e-01 -5.03137112e-01
-6.01802945e-01 -4.53065217e-01 7.87993670e-02 6.09197378e-01
9.95595679e-02 -2.44462758e-01 -1.21554029e+00 8.79194736e-01
-1.81311512e+00 -4.19219673e-01 -1.66584447e-01 2.19740343e+00
8.15612018e-01 1.45924360e-01 1.31420463e-01 -3.66747320e-01
7.39632785e-01 1.66469425e-01 -9.09252286e-01 -4.38205421e-01
-3.71596098e-01 3.89771163e-01 8.42029393e-01 6.09970748e-01
-1.32785606e+00 1.49110723e+00 4.94224405e+00 9.93900597e-01
-1.30686224e+00 3.71988237e-01 9.28678870e-01 -4.40044031e-02
-1.63518369e-01 -1.00286834e-01 -6.40887439e-01 3.94444346e-01
9.72547770e-01 1.82357773e-01 2.58813202e-01 9.34462190e-01
1.24872632e-01 -2.71300554e-01 -5.17687440e-01 7.96214283e-01
-2.36006498e-01 -1.25938296e+00 -3.57480906e-02 -1.66284710e-01
1.22914112e+00 4.26922768e-01 1.11853160e-01 3.53237361e-01
3.76905173e-01 -8.47308576e-01 6.66470706e-01 2.34304994e-01
1.24636471e+00 -7.65238762e-01 4.92402971e-01 2.24610507e-01
-1.33365083e+00 2.81252116e-01 -5.08071780e-01 8.94813165e-02
7.86738023e-02 6.28775895e-01 -7.47251689e-01 4.17822540e-01
1.21359682e+00 7.61238456e-01 -6.50824487e-01 6.77474260e-01
-9.01846886e-02 6.30848646e-01 -5.58711827e-01 4.95997518e-01
4.97885376e-01 -2.29915574e-01 3.47910583e-01 1.51027739e+00
4.60311681e-01 2.20336631e-01 2.15852275e-01 7.37194002e-01
-1.68286428e-01 2.09964602e-03 -5.26243925e-01 2.45616674e-01
3.66393775e-01 1.14500761e+00 -1.12704432e+00 -5.37921727e-01
-3.76893908e-01 1.45780110e+00 4.22340304e-01 4.19841260e-01
-9.34615374e-01 -2.59627253e-01 7.84918606e-01 3.28502148e-01
6.15210414e-01 -2.12512538e-01 -7.99797893e-01 -1.13639176e+00
-8.72759596e-02 -7.57210612e-01 4.99395728e-01 -8.38280320e-01
-1.12781417e+00 7.24639237e-01 -2.38751590e-01 -1.13352740e+00
-7.15994611e-02 -3.04024994e-01 -4.40423489e-01 9.99832034e-01
-1.66246414e+00 -1.60013962e+00 -5.43326676e-01 4.77571756e-01
1.09709537e+00 1.99572638e-01 8.79984140e-01 2.65438765e-01
-6.07919514e-01 4.93831545e-01 3.72359157e-01 -3.85673232e-02
8.51710320e-01 -1.30933619e+00 1.02369928e+00 1.06345451e+00
2.18457207e-02 -6.07431270e-02 5.61425567e-01 -7.00136006e-01
-7.14250624e-01 -1.68547130e+00 8.26014876e-01 -4.51701492e-01
3.38648349e-01 -4.81158793e-01 -1.36154640e+00 6.43661320e-01
-7.67803714e-02 4.13769454e-01 1.99611738e-01 -1.25869885e-01
-3.90355229e-01 -4.43605669e-02 -1.35763848e+00 6.06019795e-01
1.07697129e+00 -4.72969204e-01 -1.44571945e-01 2.31050521e-01
8.54471684e-01 -7.62931287e-01 -8.78642321e-01 3.78575444e-01
2.35242411e-01 -9.97203529e-01 1.08489597e+00 -2.44743556e-01
4.99159694e-01 -5.10312080e-01 -1.60136878e-01 -1.19661283e+00
-3.69516015e-01 -2.97776133e-01 1.24334544e-01 1.24441135e+00
4.79980379e-01 -5.98460197e-01 8.86835694e-01 5.74896038e-01
-2.89419144e-01 -4.88831401e-01 -7.47329891e-01 -6.75908387e-01
3.53373885e-01 -3.29272360e-01 6.80715084e-01 1.14488208e+00
-6.95349455e-01 3.82889867e-01 -1.73206374e-01 5.54461181e-01
5.45552492e-01 5.71250498e-01 6.71017826e-01 -8.83717895e-01
-2.25336000e-01 -2.90085256e-01 -4.35115993e-02 -9.85915005e-01
-4.38665934e-02 -6.74055696e-01 1.68992624e-01 -1.57164097e+00
6.28294051e-02 -7.34716654e-01 -1.02749422e-01 5.21520734e-01
-2.99858958e-01 6.95502579e-01 7.44636506e-02 2.13316679e-01
-5.61188519e-01 3.63944441e-01 1.36371756e+00 5.34392800e-03
-4.13206637e-01 -6.61176741e-02 -6.14912868e-01 7.47197032e-01
1.27131712e+00 -5.09011030e-01 -2.70041347e-01 -7.92917907e-01
-3.28896075e-01 -1.57531679e-01 3.17535698e-01 -1.11540747e+00
-1.00849062e-01 -1.76688865e-01 5.26804030e-01 -4.26641017e-01
2.22732261e-01 -5.58991432e-01 1.75861895e-01 7.02178776e-02
-1.00078896e-01 -2.52178282e-01 6.91739023e-01 5.01143992e-01
-1.82871729e-01 1.07350856e-01 1.22292674e+00 -2.75428981e-01
-1.32727134e+00 3.87942016e-01 -2.00875714e-01 3.79053593e-01
8.99107218e-01 -2.92650431e-01 -1.50098369e-01 4.43782434e-02
-8.01992357e-01 1.17409304e-01 9.99088168e-01 3.65615487e-01
2.81507790e-01 -1.00953019e+00 -7.35769510e-01 2.76815116e-01
1.61920384e-01 7.88208842e-01 6.27278388e-01 2.84194946e-01
-9.01947141e-01 2.20639646e-01 -4.81913537e-01 -6.77003264e-01
-1.11030459e+00 4.16340351e-01 2.97937870e-01 -4.14100170e-01
-9.22250271e-01 1.01688039e+00 6.83296025e-01 -7.08489001e-01
-3.32156509e-01 -4.41237122e-01 1.39186621e-01 -2.36217752e-01
2.92338133e-01 1.54142573e-01 2.29877606e-01 -7.29310095e-01
-1.39704242e-01 8.57216418e-01 -8.30622390e-02 7.80169740e-02
1.26299036e+00 -3.97665143e-01 3.25088620e-01 1.69336751e-01
9.85696197e-01 -1.61940441e-01 -1.94498014e+00 -4.33810204e-01
-1.39362186e-01 -5.21149933e-01 1.85980424e-01 -1.08025813e+00
-1.13056839e+00 9.08321083e-01 6.73973739e-01 -1.48125112e-01
1.58886218e+00 1.60538003e-01 1.13675344e+00 -3.55839287e-03
3.45777154e-01 -1.25365973e+00 -9.35549885e-02 5.55334568e-01
7.73121834e-01 -1.55168712e+00 -1.28759056e-01 -5.75441480e-01
-9.38363075e-01 7.88045943e-01 8.81351769e-01 -3.12488437e-01
2.82547444e-01 2.63645738e-01 2.86757439e-01 8.15610737e-02
-3.26767862e-01 -5.30235529e-01 1.14864439e-01 9.31735277e-01
1.54003575e-01 3.08775216e-01 2.45868862e-01 6.04531348e-01
-1.40475526e-01 -4.26133610e-02 3.64553183e-01 7.31541812e-01
-3.52147073e-01 -9.51891840e-01 -5.09307444e-01 4.04180348e-01
-3.03260237e-01 -2.06220195e-01 -1.59534961e-01 8.02823365e-01
2.32208535e-01 5.99917948e-01 4.50551152e-01 6.35850504e-02
5.56493163e-01 -1.38514712e-01 2.49747768e-01 -6.88490272e-01
-4.44320023e-01 3.21129113e-01 9.56331193e-03 -6.15198135e-01
-2.39918947e-01 -7.78422892e-01 -1.43577456e+00 -4.01994854e-01
2.04303697e-01 -3.79359394e-01 4.11574900e-01 5.69080353e-01
6.07249677e-01 5.96966088e-01 2.10522339e-01 -1.04109383e+00
1.10173047e-01 -7.86058307e-01 -4.91909862e-01 3.70872200e-01
2.89981157e-01 -2.97018051e-01 9.58218202e-02 3.36552083e-01] | [9.71345043182373, 1.2907793521881104] |
fa7d2493-766e-454f-9467-670c4cdb24df | overview-of-abusive-and-threatening-language | 2207.06710 | null | https://arxiv.org/abs/2207.06710v1 | https://arxiv.org/pdf/2207.06710v1.pdf | Overview of Abusive and Threatening Language Detection in Urdu at FIRE 2021 | With the growth of social media platform influence, the effect of their misuse becomes more and more impactful. The importance of automatic detection of threatening and abusive language can not be overestimated. However, most of the existing studies and state-of-the-art methods focus on English as the target language, with limited work on low- and medium-resource languages. In this paper, we present two shared tasks of abusive and threatening language detection for the Urdu language which has more than 170 million speakers worldwide. Both are posed as binary classification tasks where participating systems are required to classify tweets in Urdu into two classes, namely: (i) Abusive and Non-Abusive for the first task, and (ii) Threatening and Non-Threatening for the second. We present two manually annotated datasets containing tweets labelled as (i) Abusive and Non-Abusive, and (ii) Threatening and Non-Threatening. The abusive dataset contains 2400 annotated tweets in the train part and 1100 annotated tweets in the test part. The threatening dataset contains 6000 annotated tweets in the train part and 3950 annotated tweets in the test part. We also provide logistic regression and BERT-based baseline classifiers for both tasks. In this shared task, 21 teams from six countries registered for participation (India, Pakistan, China, Malaysia, United Arab Emirates, and Taiwan), 10 teams submitted their runs for Subtask A, which is Abusive Language Detection and 9 teams submitted their runs for Subtask B, which is Threatening Language detection, and seven teams submitted their technical reports. The best performing system achieved an F1-score value of 0.880 for Subtask A and 0.545 for Subtask B. For both subtasks, m-Bert based transformer model showed the best performance. | ['Alexander Gelbukh', 'Oxana Vitman', 'Hamza Imam Amjad', 'Sabur Butta', 'Andrey Labunets', 'Grigori Sidorov', 'Alisa Zhila', 'Maaz Amjad'] | 2022-07-14 | null | null | null | null | ['abusive-language'] | ['natural-language-processing'] | [-4.12143350e-01 -2.42854118e-01 -7.95930773e-02 -1.54882789e-01
-8.79712760e-01 -7.01582909e-01 8.01361859e-01 3.61955792e-01
-8.70528102e-01 1.00395477e+00 5.05303741e-02 -5.84252000e-01
5.90082943e-01 -6.21682465e-01 -1.27684444e-01 -5.59768438e-01
-1.55770462e-02 3.17816883e-01 4.48129065e-02 -6.46603286e-01
2.21463531e-01 4.00517255e-01 -1.05118704e+00 3.27335835e-01
8.08752120e-01 9.09794211e-01 -3.10020834e-01 7.51579225e-01
-1.49896517e-01 1.18741751e+00 -9.28310096e-01 -7.51249135e-01
9.09478143e-02 -3.23572159e-01 -8.53982508e-01 -1.69322774e-01
7.10653514e-02 -1.95562229e-01 -2.09901392e-01 1.00112402e+00
7.63722360e-01 -9.44999158e-02 6.14954591e-01 -1.28772187e+00
-4.68309134e-01 6.91622734e-01 -8.28252614e-01 6.73176587e-01
6.58270299e-01 7.31569082e-02 7.66931653e-01 -9.86857712e-01
4.03374076e-01 1.41159797e+00 4.87233251e-01 6.71493590e-01
-9.14108515e-01 -1.35769844e+00 1.68326139e-01 -1.72503427e-01
-1.49510431e+00 -3.36810082e-01 4.80720252e-01 -1.07969403e+00
9.50517058e-01 3.16177815e-01 5.56586087e-01 1.48058701e+00
3.79493982e-02 5.71962714e-01 1.36376715e+00 -8.31831545e-02
-1.54128805e-01 7.95028210e-01 4.13046986e-01 2.14092016e-01
2.09987268e-01 -2.35002831e-01 -2.94908613e-01 -4.79076982e-01
1.24796972e-01 -3.98067623e-01 -1.06750853e-01 8.62021267e-01
-8.76640260e-01 1.44464946e+00 3.96728292e-02 6.46026075e-01
-2.22191572e-01 -5.49509466e-01 7.63441801e-01 5.26056111e-01
9.80434239e-01 2.89024115e-01 -2.16366097e-01 -1.72903672e-01
-7.55872011e-01 2.72658855e-01 9.36341345e-01 5.91573358e-01
4.93838072e-01 7.54979178e-02 -1.52773663e-01 1.17188740e+00
2.33524263e-01 5.22764683e-01 7.22240567e-01 -1.80451982e-02
6.55869961e-01 7.05888152e-01 2.33075559e-01 -1.02256060e+00
-5.71996629e-01 -2.65569866e-01 -6.58218563e-01 -3.30565155e-01
3.38034600e-01 -7.25160122e-01 -5.39556801e-01 1.60634828e+00
4.32105422e-01 -8.33102167e-02 4.62829694e-02 6.93303287e-01
1.10317767e+00 9.14332330e-01 3.09945881e-01 -6.12857342e-01
1.46640527e+00 -7.16985524e-01 -6.82189941e-01 -1.72106773e-01
8.17715228e-01 -1.14446557e+00 1.15849423e+00 5.67170680e-01
-8.69844019e-01 -2.93559581e-01 -9.28555369e-01 7.95643553e-02
-5.29111803e-01 -3.37105870e-01 1.43784344e-01 1.08351147e+00
-8.87757003e-01 -1.94145009e-01 -7.29264691e-02 -4.35908645e-01
9.92585346e-02 6.60937130e-02 -3.88045639e-01 2.48968750e-01
-1.69694328e+00 1.03306818e+00 8.52125362e-02 -2.58652151e-01
-9.26501691e-01 -3.65617722e-01 -8.60815585e-01 -5.93386054e-01
-2.43517477e-03 3.44845086e-01 1.11463475e+00 -1.11750448e+00
-1.13549507e+00 1.61206722e+00 1.42473042e-01 -1.08104818e-01
8.70313346e-01 -3.18878710e-01 -8.59536707e-01 -2.04420060e-01
5.32841027e-01 3.03078830e-01 6.41220212e-01 -1.02542102e+00
-9.25465226e-01 -4.53208029e-01 -1.87882874e-02 6.51061013e-02
-5.92473269e-01 9.50604916e-01 4.82201055e-02 -6.26660466e-01
-3.45366120e-01 -7.22789109e-01 1.94729447e-01 -8.33312809e-01
-6.56771958e-01 -5.31854749e-01 1.06316257e+00 -1.17060149e+00
1.62673569e+00 -2.21910977e+00 -2.18562797e-01 5.71769569e-03
1.91172808e-01 3.15314233e-01 1.61890447e-01 4.35184807e-01
-1.34308442e-01 6.08129025e-01 1.55923143e-01 -4.43706065e-01
-2.00137705e-01 1.78997125e-02 -3.60574543e-01 6.64028525e-01
1.83769003e-01 2.41196081e-01 -9.29246783e-01 -5.04946828e-01
-1.95372537e-01 3.15197915e-01 -1.37292251e-01 5.18471837e-01
4.70446765e-01 6.64711297e-01 -2.58203834e-01 7.04537690e-01
7.60913491e-01 5.41941106e-01 -2.21144378e-01 3.78047258e-01
-4.57301646e-01 4.45312589e-01 -8.19044948e-01 4.69020903e-01
-5.12487054e-01 7.97424138e-01 4.97289002e-01 -5.82487345e-01
1.23950684e+00 4.67886388e-01 2.51876086e-01 -6.22627318e-01
4.90002036e-01 3.51407498e-01 9.76766273e-02 -6.41090095e-01
6.25116289e-01 -5.64666450e-01 -4.99445826e-01 6.10903442e-01
-3.85121435e-01 -1.08109765e-01 4.64261830e-01 3.51473212e-01
8.44630897e-01 -3.86665314e-01 6.15090668e-01 -2.44889855e-01
9.77763832e-01 -4.25257593e-01 4.80243474e-01 5.92634916e-01
-7.96615958e-01 3.77984196e-01 8.18686962e-01 -4.32731658e-01
-6.52442515e-01 -5.74229896e-01 -3.24638724e-01 1.64074326e+00
-1.20330974e-01 -9.21208709e-02 -7.65649498e-01 -8.09174836e-01
-1.01695232e-01 9.37589467e-01 -5.09496987e-01 4.51636240e-02
-5.27347744e-01 -1.00692022e+00 1.11832464e+00 -3.24324518e-02
6.60637200e-01 -1.16819739e+00 -2.69501299e-01 4.89590690e-02
-7.34844148e-01 -1.29328442e+00 -5.38456142e-01 -3.31207947e-03
-6.56597838e-02 -9.68332529e-01 -6.73816562e-01 -6.76111817e-01
1.89720586e-01 5.32273576e-02 9.19345498e-01 2.01419041e-01
-3.45411710e-02 -8.25029761e-02 -4.05388206e-01 -8.95195961e-01
-7.94051826e-01 1.02964930e-01 3.00069451e-01 3.02632511e-01
6.90192044e-01 4.06172387e-02 -5.85780628e-02 5.03510058e-01
-6.98265910e-01 -3.09693187e-01 -6.19843267e-02 3.83219421e-01
-1.08894475e-01 -4.51536000e-01 6.94675565e-01 -8.19204867e-01
8.44187617e-01 -1.06432152e+00 -2.03113079e-01 -2.25669667e-01
-7.32914582e-02 -1.03056169e+00 4.44768041e-01 -7.36202300e-01
-8.11272919e-01 -4.24195826e-01 -5.12231410e-01 1.09920934e-01
-2.49218673e-01 5.32000542e-01 -1.26031741e-01 3.44183147e-01
8.01012874e-01 -2.10445491e-03 -3.21683943e-01 -3.39398623e-01
-4.26786035e-01 1.60257697e+00 -7.98662379e-02 -2.36582711e-01
6.97720826e-01 1.68144703e-01 -7.26877689e-01 -1.21186769e+00
-1.07108963e+00 -8.85431647e-01 -5.09020686e-01 -4.44225729e-01
9.61253822e-01 -1.13101339e+00 -4.39331889e-01 8.78672659e-01
-1.07330024e+00 -3.45982790e-01 3.69357169e-01 3.26619148e-01
8.72438401e-02 1.86145440e-01 -8.76662016e-01 -1.24333668e+00
-5.38873851e-01 -1.07081044e+00 8.35696101e-01 -1.50343746e-01
-7.31519222e-01 -9.46445107e-01 1.92021698e-01 6.87140465e-01
2.53980756e-01 4.65303183e-01 7.56117046e-01 -1.41130292e+00
6.44281745e-01 -6.08323216e-01 -3.19725007e-01 4.04128641e-01
1.84025109e-01 6.26107827e-02 -1.15229237e+00 -4.51075375e-01
1.87750176e-01 -8.08301389e-01 4.48173910e-01 -1.78843543e-01
5.01830161e-01 -5.60418844e-01 -2.49014813e-02 -4.61710282e-02
8.43506455e-01 2.91005433e-01 5.53313851e-01 4.89869505e-01
6.09007061e-01 8.40874016e-01 7.17931688e-01 6.00233793e-01
4.86552536e-01 6.45227492e-01 3.33028436e-01 3.35394451e-03
3.31929356e-01 -1.20445661e-01 9.97300327e-01 6.97018325e-01
2.36229882e-01 -2.37132370e-01 -1.30613816e+00 6.03370547e-01
-1.46038783e+00 -9.57479477e-01 -6.01719975e-01 2.14810014e+00
1.06093621e+00 1.50495559e-01 7.55405903e-01 4.12046343e-01
9.57168221e-01 1.32189676e-01 -2.67410949e-02 -1.02773309e+00
-1.82272062e-01 -3.23918551e-01 -4.69701327e-02 8.34906518e-01
-1.42017841e+00 9.79040205e-01 5.34007978e+00 8.19404244e-01
-1.50369561e+00 5.76787829e-01 1.15794265e+00 -2.57286519e-01
1.29932910e-01 -5.26979864e-01 -1.21110225e+00 6.69194818e-01
1.15556312e+00 -5.12012020e-02 -1.68305766e-02 8.48428130e-01
3.69593203e-01 -2.13805780e-01 -7.30618417e-01 9.16801453e-01
4.68315899e-01 -4.06016827e-01 2.41098218e-02 2.09912956e-02
5.84030390e-01 1.99033886e-01 -9.75129008e-02 8.91947389e-01
7.25723058e-02 -1.26999569e+00 1.07247603e+00 -7.04568699e-02
8.71382773e-01 -9.10600662e-01 1.15258718e+00 8.24422598e-01
-7.09341705e-01 -1.21425360e-01 -1.87153876e-01 -6.48881197e-01
8.41865763e-02 5.97157657e-01 -5.65974832e-01 8.17647129e-02
8.68440449e-01 5.70930541e-01 -2.91064501e-01 4.68386889e-01
-4.39501196e-01 1.08728850e+00 -2.10453734e-01 -3.38206977e-01
5.89348853e-01 -1.59501299e-01 7.37019479e-01 1.78272760e+00
1.03045255e-01 -1.00666620e-02 4.53230888e-01 2.98603117e-01
-1.06751211e-01 6.47402823e-01 -7.21906245e-01 -1.73447784e-02
2.90348321e-01 1.47899175e+00 -6.01724207e-01 -2.97258943e-01
-4.38727349e-01 5.83866477e-01 4.25726622e-01 1.08900636e-01
-8.93594325e-01 -3.69165957e-01 4.73876119e-01 3.43602568e-01
-4.27349478e-01 9.07128304e-02 -4.18106526e-01 -9.76196527e-01
-3.86708260e-01 -1.19872046e+00 7.97947228e-01 -1.23142317e-01
-1.46403193e+00 8.23856831e-01 -6.68993639e-03 -7.95581400e-01
-2.17661709e-01 -5.59800982e-01 -5.35057604e-01 1.04709125e+00
-1.20479465e+00 -1.13208723e+00 -2.48822585e-01 6.09656572e-01
5.49896240e-01 -3.19049031e-01 6.13967240e-01 5.90595543e-01
-9.84066665e-01 6.87780976e-01 -4.59145039e-01 4.87908274e-01
8.24117005e-01 -8.57428551e-01 -1.79729939e-01 7.78130412e-01
-3.73366922e-01 3.61071199e-01 7.76067436e-01 -7.00359404e-01
-5.46380758e-01 -1.27100229e+00 1.62107837e+00 -5.00716507e-01
1.02573431e+00 -7.32890785e-01 -7.01820493e-01 6.95392609e-01
-2.88236588e-02 -3.24595988e-01 9.42969322e-01 1.74946785e-02
-4.50715035e-01 2.79811084e-01 -1.51777112e+00 5.54235697e-01
5.38996518e-01 -3.58485132e-01 -3.09244037e-01 5.46668291e-01
3.47411573e-01 -3.25466394e-01 -5.54386616e-01 -1.05002511e-03
4.28401589e-01 -9.52867925e-01 3.88400197e-01 -6.48085594e-01
6.24898672e-01 2.94259667e-01 -3.29743952e-01 -9.69218612e-01
-1.28353700e-01 -6.20657802e-01 3.84059280e-01 1.42389703e+00
5.51558971e-01 -9.69396830e-01 2.77333017e-02 3.12223136e-01
1.28352210e-01 -6.89844549e-01 -1.10726285e+00 -5.41805625e-01
6.85171187e-01 -6.26515269e-01 1.83724731e-01 1.37646925e+00
1.28894880e-01 6.61406577e-01 -4.49417114e-01 -1.19157873e-01
1.65076759e-02 -3.93898994e-01 9.12436306e-01 -9.27575350e-01
2.14317754e-01 -5.70384562e-01 -1.85343519e-01 -4.81277704e-01
3.22362661e-01 -7.13636756e-01 -2.95656919e-01 -1.11554289e+00
5.01203954e-01 -2.23217458e-01 3.87947597e-02 5.05693018e-01
-2.85598338e-01 8.67048144e-01 1.08041793e-01 1.93309441e-01
-2.58357555e-01 2.22829226e-02 6.12013698e-01 -1.18064955e-01
-3.01705897e-01 1.43121064e-01 -9.03607905e-01 8.97083044e-01
7.98171639e-01 -4.18317646e-01 3.06305941e-03 4.75810207e-02
3.56040388e-01 -3.07283580e-01 3.48718166e-02 -4.84775364e-01
-4.49775308e-01 -2.09966123e-01 1.03864521e-01 -4.54665691e-01
1.71405092e-01 -3.52223426e-01 -4.89229023e-01 6.95061326e-01
-1.84256002e-01 2.02890396e-01 3.07835966e-01 -1.47193447e-01
-3.07488084e-01 -3.17499608e-01 1.14056003e+00 -3.43093053e-02
-1.95880339e-01 2.49549344e-01 -9.87423420e-01 4.80970830e-01
1.35439491e+00 -2.51525551e-01 -2.77322501e-01 -8.13564360e-01
-5.69369256e-01 2.28284702e-01 3.51172239e-02 6.01725817e-01
4.20912474e-01 -9.47789192e-01 -1.41107714e+00 3.83307263e-02
1.47926211e-01 -5.81802011e-01 -6.34333640e-02 1.09576988e+00
-4.19621557e-01 2.80160218e-01 8.52999613e-02 -2.84890443e-01
-1.58353889e+00 2.69448191e-01 3.49047065e-01 -5.28226793e-01
1.25055552e-01 1.02311993e+00 2.19219297e-01 -7.15260386e-01
2.12557524e-01 4.43446249e-01 -7.49114394e-01 8.12963963e-01
8.98110747e-01 6.61434531e-01 5.07346541e-02 -1.56773543e+00
-4.39242214e-01 7.16419592e-02 -4.09947067e-01 -1.14356726e-01
8.51127923e-01 -1.37943298e-01 -4.78904903e-01 8.66683483e-01
1.25048089e+00 3.25566322e-01 -3.06943476e-01 2.54339892e-02
1.65451970e-02 -2.44978234e-01 -3.18759307e-02 -7.90964603e-01
-6.18399322e-01 9.02401090e-01 2.63336480e-01 6.58165157e-01
7.53744364e-01 -1.52532041e-01 7.68767416e-01 -2.47446060e-01
3.23970109e-01 -1.04428172e+00 1.81574911e-01 1.16129172e+00
1.27320278e+00 -1.28315949e+00 -2.67612666e-01 -4.70098913e-01
-8.27971101e-01 7.79750466e-01 8.35646868e-01 1.01193696e-01
6.13552034e-01 2.97447085e-01 6.13472402e-01 1.16244100e-01
-4.87288594e-01 -7.54937232e-02 1.73667341e-01 4.95015740e-01
8.83043587e-01 1.64945573e-01 -7.98614264e-01 9.48606551e-01
-6.00442052e-01 -6.02981746e-01 5.55007041e-01 6.50935411e-01
-3.20143521e-01 -6.01579487e-01 -8.08425188e-01 4.79528397e-01
-1.12905204e+00 -1.03665330e-01 -9.68891859e-01 1.17819905e+00
2.45333984e-01 1.43062365e+00 2.78260022e-01 -5.08803368e-01
2.96989262e-01 1.38827160e-01 -1.83219746e-01 -7.40913928e-01
-1.27824092e+00 3.46506052e-02 6.11225963e-01 -1.09379098e-01
6.45854026e-02 -7.15229213e-01 -1.00521159e+00 -6.38380587e-01
-2.76269168e-01 1.95519850e-01 4.38329905e-01 9.56607521e-01
-3.04137945e-01 -2.02862918e-01 8.74616742e-01 -7.45733500e-01
-6.54122531e-01 -1.57945728e+00 -5.73219061e-01 6.54985607e-01
3.43415588e-01 -4.82643276e-01 -7.10866570e-01 -4.54181254e-01] | [8.860923767089844, 10.59447956085205] |
4695e606-5f84-4cda-9839-3913b474c70b | introducing-representations-of-facial-affect | 2008.13369 | null | https://arxiv.org/abs/2008.13369v1 | https://arxiv.org/pdf/2008.13369v1.pdf | Introducing Representations of Facial Affect in Automated Multimodal Deception Detection | Automated deception detection systems can enhance health, justice, and security in society by helping humans detect deceivers in high-stakes situations across medical and legal domains, among others. This paper presents a novel analysis of the discriminative power of dimensional representations of facial affect for automated deception detection, along with interpretable features from visual, vocal, and verbal modalities. We used a video dataset of people communicating truthfully or deceptively in real-world, high-stakes courtroom situations. We leveraged recent advances in automated emotion recognition in-the-wild by implementing a state-of-the-art deep neural network trained on the Aff-Wild database to extract continuous representations of facial valence and facial arousal from speakers. We experimented with unimodal Support Vector Machines (SVM) and SVM-based multimodal fusion methods to identify effective features, modalities, and modeling approaches for detecting deception. Unimodal models trained on facial affect achieved an AUC of 80%, and facial affect contributed towards the highest-performing multimodal approach (adaptive boosting) that achieved an AUC of 91% when tested on speakers who were not part of training sets. This approach achieved a higher AUC than existing automated machine learning approaches that used interpretable visual, vocal, and verbal features to detect deception in this dataset, but did not use facial affect. Across all videos, deceptive and truthful speakers exhibited significant differences in facial valence and facial arousal, contributing computational support to existing psychological theories on affect and deception. The demonstrated importance of facial affect in our models informs and motivates the future development of automated, affect-aware machine learning approaches for modeling and detecting deception and other social behaviors in-the-wild. | ['Maja J. Matarić', 'Leena Mathur'] | 2020-08-31 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [ 1.41370252e-01 1.06614962e-01 -8.25895667e-02 -8.80287111e-01
-8.18968654e-01 -7.19023943e-01 5.42851508e-01 7.63914511e-02
-4.90285575e-01 4.10484523e-01 4.28027898e-01 -1.26031220e-01
1.71959534e-01 -7.05807284e-02 -1.93955116e-02 -4.50941026e-01
-1.62822172e-01 -1.43829435e-01 -8.24549198e-01 -1.31624460e-01
2.86004722e-01 7.76111364e-01 -1.46033955e+00 8.14086378e-01
3.53627354e-01 1.38031936e+00 -1.16964114e+00 8.41064990e-01
4.49380845e-01 1.08570099e+00 -8.67035329e-01 -1.30356693e+00
-8.14622268e-02 -6.33678555e-01 -5.54600298e-01 1.31273996e-02
8.87182117e-01 -8.41789424e-01 -2.49493793e-01 7.35240519e-01
5.30051708e-01 -1.43928438e-01 8.56326997e-01 -1.50505364e+00
-6.67256773e-01 -1.24679334e-01 -4.77461606e-01 3.73953640e-01
7.54309714e-01 3.68139476e-01 8.40322435e-01 -7.14435637e-01
5.11209905e-01 1.50314069e+00 7.48547792e-01 1.10669994e+00
-9.84299719e-01 -9.12988842e-01 -2.12123901e-01 4.02980000e-01
-8.82110834e-01 -9.86105144e-01 8.25460196e-01 -6.05751455e-01
1.00745106e+00 4.70606863e-01 1.00879407e+00 1.70528579e+00
3.42036843e-01 7.17621148e-01 1.16004014e+00 7.34178871e-02
1.17908977e-01 1.99929446e-01 1.30368307e-01 8.77835095e-01
-1.02433965e-01 3.88097554e-01 -9.01413202e-01 -1.01274765e+00
1.31490380e-01 -1.33949623e-01 -1.51322052e-01 2.70503640e-01
-6.29360735e-01 1.10838640e+00 1.57386407e-01 1.35528579e-01
-4.63465929e-01 1.65663183e-01 8.51805925e-01 4.08687681e-01
6.57523036e-01 5.00996888e-01 -8.44026804e-02 -6.16931140e-01
-9.44547236e-01 4.88784276e-02 7.11409032e-01 -1.83676973e-01
1.05103761e-01 3.50767612e-01 -6.90965876e-02 9.53948438e-01
1.22535795e-01 5.91165185e-01 4.46257621e-01 -1.14246356e+00
-2.31607305e-03 5.32670557e-01 -1.23606555e-01 -1.53174520e+00
-6.46170855e-01 6.75068870e-02 -4.20656919e-01 4.39346433e-01
1.17205270e-01 -3.32917929e-01 -7.83272803e-01 1.63115871e+00
2.99814165e-01 -2.47874752e-01 -4.94545922e-02 1.08322334e+00
8.67397666e-01 1.87966481e-01 4.51050162e-01 -2.70308584e-01
1.37770998e+00 -2.52850056e-01 -7.70468950e-01 -3.16329658e-01
5.80499351e-01 -3.83364230e-01 6.32786512e-01 6.93361878e-01
-1.05215120e+00 -2.42162049e-02 -9.18343604e-01 8.71791411e-03
-1.83164194e-01 -2.48029511e-02 7.89038837e-01 1.41001952e+00
-7.68819451e-01 4.22713757e-01 -4.92248625e-01 -1.32902861e-01
1.04958761e+00 1.92179516e-01 -8.48698735e-01 8.30110684e-02
-1.03212571e+00 1.34230506e+00 -4.54199523e-01 3.65383595e-01
-9.51888263e-01 -3.59484315e-01 -1.10953784e+00 -1.43186808e-01
-2.78282553e-01 -3.46336365e-01 1.04302180e+00 -2.05428004e+00
-1.31498086e+00 1.33053076e+00 -1.81374580e-01 -2.14707449e-01
2.88486451e-01 -1.02158049e-02 -8.17343473e-01 7.46284306e-01
-3.60770047e-01 6.73162043e-01 1.45341146e+00 -1.01065075e+00
-4.57685534e-03 -9.10838246e-01 -2.99779773e-01 1.57687202e-01
-4.82713968e-01 5.45027018e-01 8.71462464e-01 -1.67110100e-01
-3.50989074e-01 -6.08852923e-01 4.91786510e-01 2.47440487e-01
4.62557115e-02 -4.57583964e-02 8.72689009e-01 -1.14031982e+00
8.18770587e-01 -2.25362039e+00 -6.05051145e-02 2.47821286e-01
5.49810529e-01 3.80996913e-01 -1.31375387e-01 1.36093542e-01
-2.67731726e-01 2.34367266e-01 1.13743264e-02 -5.85285604e-01
-4.27704528e-02 4.84437197e-02 -1.17752895e-01 9.70088124e-01
2.71678239e-01 1.03523433e+00 -7.38100410e-01 -5.85937381e-01
1.96285576e-01 5.93716264e-01 -3.95989239e-01 1.77470282e-01
5.61888695e-01 1.95026863e-02 -2.34191045e-02 1.20787299e+00
4.37721848e-01 5.73438466e-01 5.16116433e-02 -1.10931285e-02
4.31170374e-01 -1.50514171e-01 -1.12535879e-01 1.00957918e+00
-1.32825494e-01 1.00854266e+00 7.05952525e-01 -7.87984490e-01
1.01332223e+00 4.29225415e-01 2.87207276e-01 -6.49987996e-01
6.75030828e-01 1.03268273e-01 5.54028116e-02 -9.17038023e-01
2.16610938e-01 -9.60626006e-01 -2.72065461e-01 2.85494447e-01
1.78341120e-01 -1.55772701e-01 -6.17097795e-01 1.25202268e-01
1.15695763e+00 -3.22578639e-01 2.78144628e-01 2.74249673e-01
2.92828888e-01 -1.82204992e-01 2.68118531e-01 3.82879943e-01
-1.04702961e+00 4.23953831e-01 8.35077107e-01 -6.60596907e-01
-4.74555850e-01 -8.07232797e-01 -1.20473914e-01 1.33099747e+00
-3.69627506e-01 -9.36210679e-04 -6.83982551e-01 -8.70619178e-01
1.86561435e-01 9.97195601e-01 -1.08651400e+00 -7.66868114e-01
3.44344266e-02 -6.93058968e-01 1.16478145e+00 2.87605405e-01
-1.29575469e-02 -1.12143207e+00 -8.55798125e-01 -2.40216047e-01
-2.69519269e-01 -8.96899283e-01 -2.13670090e-01 -2.72426337e-01
-3.01936835e-01 -1.26887798e+00 -2.79683828e-01 -1.13598481e-01
3.09273779e-01 -2.01219290e-01 7.87133396e-01 2.96631660e-02
-6.49381876e-01 1.02566683e+00 -3.06098253e-01 -6.30314708e-01
-6.24024928e-01 -7.09064245e-01 2.15754524e-01 4.81400788e-01
5.29315114e-01 -3.54508460e-01 -5.13101280e-01 -8.07768665e-03
-8.29759896e-01 -5.44813573e-01 2.81555057e-01 8.13756287e-01
-3.53503883e-01 -9.98175025e-01 9.23957646e-01 -1.36496052e-01
1.18503356e+00 -7.00999200e-01 2.15047583e-01 1.20942909e-02
-2.41579637e-01 -6.35488510e-01 3.03126067e-01 -4.81213987e-01
-1.02004945e+00 -2.09282607e-01 -2.49730453e-01 -5.23307323e-01
-3.69114459e-01 2.97852278e-01 2.43130341e-01 -3.34015816e-01
8.85459602e-01 -6.61543831e-02 4.96578932e-01 1.66831687e-01
8.94445777e-02 1.26053107e+00 4.72895145e-01 -1.98000446e-01
1.39955431e-01 5.53134143e-01 2.91758776e-02 -9.47730124e-01
-7.76221454e-01 -2.78506935e-01 -3.71355563e-01 -7.22619534e-01
7.16283381e-01 -6.64525986e-01 -1.38503325e+00 6.07838511e-01
-1.09928215e+00 2.25955173e-01 1.81601703e-01 3.92333508e-01
-4.68514621e-01 5.76889634e-01 -6.43276393e-01 -1.51577926e+00
-6.95420086e-01 -6.05117917e-01 1.15168822e+00 -2.41261631e-01
-9.04519916e-01 -7.39965796e-01 1.56512022e-01 1.08404875e+00
2.98607469e-01 8.05634677e-01 7.22060144e-01 -8.78226399e-01
6.39372289e-01 -7.05313921e-01 -1.23646580e-01 6.53770208e-01
2.06265994e-03 5.66401929e-02 -1.37858033e+00 -2.96501052e-02
1.62964985e-01 -1.25902319e+00 8.00299823e-01 1.23568162e-01
9.32842910e-01 -5.24991870e-01 1.30526990e-01 3.05215597e-01
5.91093838e-01 1.95156649e-01 6.26131833e-01 -2.99424499e-01
3.45599741e-01 8.26883197e-01 2.20436871e-01 5.90116858e-01
2.62127250e-01 4.65274274e-01 8.12950850e-01 -9.02010649e-02
3.61054659e-01 -4.03407812e-02 8.09654117e-01 -1.31332532e-01
1.56251062e-02 -1.67189166e-02 -7.43202746e-01 4.16685194e-01
-1.41121721e+00 -1.49408078e+00 7.62047470e-02 1.81527102e+00
7.28099704e-01 -4.30468976e-01 4.76643562e-01 1.20592125e-01
4.83603865e-01 1.88529596e-01 -6.51182950e-01 -1.57572293e+00
-1.00965224e-01 6.46967441e-02 -2.41840333e-01 4.22507912e-01
-1.02900064e+00 5.20646513e-01 6.84982586e+00 4.53478307e-01
-1.38136292e+00 5.35849743e-02 1.15584993e+00 -7.86999345e-01
-1.75191686e-01 -7.27683306e-01 -3.48000675e-02 3.02317291e-01
1.39572024e+00 1.81852654e-01 5.68449259e-01 8.46562624e-01
3.27918977e-01 -9.64685306e-02 -1.12697184e+00 1.17285109e+00
9.94668841e-01 -8.03875804e-01 -2.16500908e-01 -4.05860879e-02
1.15121402e-01 -2.84604222e-01 2.89939135e-01 1.92644998e-01
-2.83470631e-01 -1.65680933e+00 6.89603031e-01 6.22862279e-01
9.53115284e-01 -8.08027864e-01 8.01928461e-01 9.82722491e-02
-3.36271785e-02 -3.67007464e-01 1.08135268e-01 -3.62442881e-01
3.24572250e-02 1.53649971e-01 -8.93176973e-01 -9.55698639e-02
5.27342200e-01 4.86851633e-01 -2.96201825e-01 3.34551722e-01
3.35502848e-02 6.19559526e-01 -9.98526588e-02 -3.29989493e-01
2.03093797e-01 3.78518075e-01 4.85556990e-01 1.47193170e+00
-4.49488265e-03 2.56916374e-01 -4.65379983e-01 7.61727452e-01
-6.71356767e-02 9.90422890e-02 -5.59791446e-01 -3.16627353e-01
-9.56726633e-03 1.60937965e+00 -4.23950441e-02 -1.96318462e-01
-2.29092479e-01 1.09449911e+00 3.22427928e-01 3.72262783e-02
-8.69497776e-01 -4.12592990e-03 9.58003461e-01 -3.22284326e-02
-2.69991875e-01 1.80599406e-01 -3.56461525e-01 -1.14704406e+00
-8.65115747e-02 -1.15929604e+00 5.62020361e-01 -8.93029213e-01
-1.49486685e+00 5.38105786e-01 -8.94170180e-02 -7.22268581e-01
-4.23482478e-01 -7.25558698e-01 -5.55248141e-01 6.17240012e-01
-9.94894564e-01 -1.43569279e+00 -2.97849327e-01 7.28093445e-01
-1.49699971e-02 -8.21088478e-02 1.17565274e+00 -2.16483384e-01
-4.27619874e-01 7.58504331e-01 -4.28738415e-01 2.76088268e-01
7.88456082e-01 -7.39285529e-01 -4.18718070e-01 3.58968228e-01
-1.48097113e-01 3.85039210e-01 4.37090904e-01 -4.51013774e-01
-1.26440501e+00 -5.73973179e-01 6.74584329e-01 -5.65339386e-01
5.89858174e-01 -3.54581684e-01 -5.43642879e-01 6.20501220e-01
1.73218846e-01 -1.26725778e-01 1.36138380e+00 2.43359491e-01
-8.62560034e-01 4.51032780e-02 -1.98751819e+00 2.95545220e-01
7.31012344e-01 -8.24676454e-01 -7.17359543e-01 1.21702202e-01
-1.48805186e-01 -5.52321114e-02 -7.36617565e-01 1.33048415e-01
1.24535584e+00 -1.29451299e+00 7.01224029e-01 -1.16246176e+00
7.88072586e-01 5.18478394e-01 -7.66290128e-02 -1.37199020e+00
3.27555859e-03 -4.34077948e-01 -7.16311112e-02 8.46171319e-01
2.16248453e-01 -6.98002040e-01 5.44251084e-01 1.21961355e+00
8.78698900e-02 -8.61532211e-01 -1.50181687e+00 -1.54746681e-01
1.65396824e-01 -5.76144934e-01 8.09359103e-02 1.04458559e+00
8.13911259e-01 1.65751681e-01 -6.84782267e-01 -2.63868898e-01
4.11845118e-01 -1.30417019e-01 3.59991461e-01 -1.08156490e+00
1.37693688e-01 -4.42376763e-01 -1.11202288e+00 1.54675737e-01
7.11414218e-01 -7.46091545e-01 -4.77067560e-01 -8.81052613e-01
5.21026373e-01 4.74277318e-01 5.40673062e-02 9.93749976e-01
8.44572484e-03 5.76245189e-01 1.43116593e-01 -2.45108753e-01
-2.96843946e-01 6.69153571e-01 8.99731874e-01 -1.55009195e-01
1.12217516e-01 -2.49948457e-01 -1.14593613e+00 8.10116947e-01
7.45358109e-01 -2.37398058e-01 7.42268376e-03 9.09499899e-02
1.29257217e-01 4.33990300e-01 9.85494196e-01 -4.41377342e-01
-2.77588725e-01 -2.22194478e-01 1.00135660e+00 1.32038966e-01
1.19721329e+00 -6.39616251e-01 -3.03080350e-01 3.38414043e-01
-3.72682661e-01 -1.18123114e-01 4.46014225e-01 2.22747535e-01
-6.59161508e-02 -1.20347880e-01 9.82930064e-01 -1.07399389e-01
-1.83747083e-01 -1.92571208e-01 -7.96966970e-01 -1.66800067e-01
1.00671685e+00 -2.99032062e-01 -6.38004661e-01 -1.19386196e+00
-9.07922387e-01 -2.20526427e-01 1.78879336e-01 4.44683433e-01
1.02802896e+00 -1.06748712e+00 -9.11102474e-01 1.75092757e-01
1.14868112e-01 -1.32695985e+00 4.02747065e-01 1.01045907e+00
-2.44378317e-02 7.70432130e-02 -5.54204941e-01 -2.29731828e-01
-1.88840604e+00 2.70034611e-01 8.34126830e-01 4.17293638e-01
2.86761522e-01 8.28610241e-01 -2.39199311e-01 -2.69579828e-01
-8.48744214e-02 1.81708902e-01 -3.83502096e-02 4.24965292e-01
7.77363598e-01 5.75701714e-01 8.92751068e-02 -1.17973471e+00
-6.43652022e-01 1.71134081e-02 4.94270548e-02 -2.37802491e-01
1.18945968e+00 1.84445456e-01 -1.94720253e-01 2.97345579e-01
1.52649403e+00 -2.12475657e-02 -5.61008513e-01 6.21868372e-01
-5.95233858e-01 -5.79409957e-01 2.57822067e-01 -1.41544175e+00
-9.08491373e-01 9.96877253e-01 5.59581280e-01 1.40386716e-01
1.14699244e+00 -3.00424267e-02 5.35682082e-01 1.37147769e-01
-2.90291458e-02 -1.04970503e+00 2.52941728e-01 -5.86034693e-02
1.26793993e+00 -1.27744281e+00 9.39565748e-02 6.29390730e-03
-1.17333555e+00 1.25269020e+00 5.03440857e-01 2.31433123e-01
3.13772917e-01 -1.43967345e-02 3.74135107e-01 -5.48618853e-01
-6.71131790e-01 2.90615618e-01 5.36971748e-01 7.51854002e-01
3.76256287e-01 2.95617640e-01 -3.89877319e-01 9.05636072e-01
-3.33978266e-01 -9.34336782e-02 4.50144827e-01 5.96835673e-01
-2.59728998e-01 -5.73064327e-01 -5.05800664e-01 8.18237126e-01
-8.46677721e-01 4.12459597e-02 -1.45239151e+00 4.52464014e-01
1.39320418e-01 1.40717828e+00 -3.76931159e-03 -5.45444608e-01
1.50569856e-01 4.81687754e-01 5.64449847e-01 -1.15126543e-01
-1.10888934e+00 -4.04508173e-01 7.21301258e-01 -8.72425497e-01
-3.10414612e-01 -1.14805710e+00 -8.46741974e-01 -5.41838109e-01
-1.01431645e-01 -3.28169852e-01 7.09436178e-01 9.04755175e-01
4.96633857e-01 -2.82781839e-01 6.82791650e-01 -9.22894299e-01
-5.39924264e-01 -1.11558175e+00 -5.41064024e-01 7.34709382e-01
7.55837679e-01 -4.79187727e-01 -8.62170696e-01 -2.79816061e-01] | [13.312671661376953, 2.060544967651367] |
0d01968d-7c9b-4a15-b828-9f2f4b9edeea | depth-cooperated-trimodal-network-for-video | 2202.06060 | null | https://arxiv.org/abs/2202.06060v2 | https://arxiv.org/pdf/2202.06060v2.pdf | Depth-Cooperated Trimodal Network for Video Salient Object Detection | Depth can provide useful geographical cues for salient object detection (SOD), and has been proven helpful in recent RGB-D SOD methods. However, existing video salient object detection (VSOD) methods only utilize spatiotemporal information and seldom exploit depth information for detection. In this paper, we propose a depth-cooperated trimodal network, called DCTNet for VSOD, which is a pioneering work to incorporate depth information to assist VSOD. To this end, we first generate depth from RGB frames, and then propose an approach to treat the three modalities unequally. Specifically, a multi-modal attention module (MAM) is designed to model multi-modal long-range dependencies between the main modality (RGB) and the two auxiliary modalities (depth, optical flow). We also introduce a refinement fusion module (RFM) to suppress noises in each modality and select useful information dynamically for further feature refinement. Lastly, a progressive fusion strategy is adopted after the refined features to achieve final cross-modal fusion. Experiments on five benchmark datasets demonstrate the superiority of our depth-cooperated model against 12 state-of-the-art methods, and the necessity of depth is also validated. | ['Qijun Zhao', 'Keren Fu', 'Dingyao Min', 'Yukang Lu'] | 2022-02-12 | null | null | null | null | ['video-salient-object-detection'] | ['computer-vision'] | [ 4.11912240e-02 -1.77421451e-01 -2.75639236e-01 -1.36504889e-01
-5.53277612e-01 -4.41005304e-02 5.58050334e-01 -1.57845214e-01
-4.43901449e-01 4.72798139e-01 6.54074907e-01 1.58180386e-01
5.24522252e-02 -6.90108716e-01 -3.07976961e-01 -9.94332910e-01
1.94549784e-01 -3.50892544e-01 7.49894023e-01 -2.46049061e-01
2.75140077e-01 2.69238055e-01 -1.65226710e+00 1.39493585e-01
9.36628461e-01 1.32360733e+00 6.73379540e-01 3.21517557e-01
-7.39313141e-02 1.10264003e+00 -2.27310970e-01 1.80085246e-02
1.42669842e-01 -4.24127132e-01 -6.98397279e-01 2.96643645e-01
2.80215740e-01 -6.87638998e-01 -7.47238934e-01 1.07434058e+00
6.67456627e-01 4.21906769e-01 3.16064984e-01 -1.41974080e+00
-6.52708769e-01 3.81165117e-01 -8.93551350e-01 5.89294314e-01
4.03726131e-01 4.05534983e-01 7.93974161e-01 -9.46627021e-01
5.11053741e-01 1.37037110e+00 2.07880601e-01 6.04745746e-01
-7.35792398e-01 -4.52665061e-01 6.20087683e-01 4.67143536e-01
-1.16374302e+00 -3.25260967e-01 1.23215926e+00 -2.36322701e-01
5.64849019e-01 9.32757780e-02 8.10954630e-01 8.61631095e-01
6.75930232e-02 1.31519210e+00 7.64128208e-01 -9.36578363e-02
4.89870831e-02 -1.14164993e-01 -2.53730386e-01 8.43165994e-01
1.14044741e-01 7.00521395e-02 -7.45126009e-01 2.02053860e-01
8.76304924e-01 3.45917344e-01 -6.22590125e-01 -4.63068664e-01
-1.59589875e+00 6.65689051e-01 9.51690018e-01 2.95245469e-01
-5.85709810e-01 1.47321016e-01 4.42418665e-01 -1.48923606e-01
4.60965395e-01 -4.30764742e-02 -1.60038069e-01 2.21559126e-03
-6.73385382e-01 -3.36324461e-02 1.17509112e-01 8.27106357e-01
8.21818650e-01 3.30002941e-02 -5.52019894e-01 6.71046436e-01
5.64034343e-01 4.88473386e-01 6.20020390e-01 -9.78004515e-01
5.85033000e-01 8.43885362e-01 1.73527375e-01 -1.20307577e+00
-5.18818259e-01 -1.74109802e-01 -9.60419357e-01 -1.32338451e-02
5.00997268e-02 -2.16317456e-02 -9.46444213e-01 1.67991495e+00
6.49619102e-01 4.21139508e-01 1.10948689e-01 1.63262200e+00
1.30824697e+00 7.15512931e-01 2.50155509e-01 -3.12752724e-01
1.12939942e+00 -1.07120597e+00 -6.80001676e-01 -2.73776531e-01
3.35020632e-01 -5.60471773e-01 8.57020915e-01 1.24625817e-01
-1.17365170e+00 -7.24746704e-01 -9.06075716e-01 -4.68073547e-01
-1.82753876e-01 -1.20787583e-02 7.69745648e-01 1.15120806e-01
-8.81879151e-01 5.79994731e-02 -8.76564741e-01 -2.45628491e-01
4.50680375e-01 2.76747882e-01 -2.79500812e-01 -2.42480978e-01
-1.37627268e+00 6.95920408e-01 5.62430620e-01 5.12681544e-01
-9.79499698e-01 -4.14017886e-01 -1.11497509e+00 -2.08500981e-01
4.54008996e-01 -9.82903302e-01 9.42366421e-01 -9.05080199e-01
-1.31663644e+00 6.29814088e-01 -4.15873379e-01 -1.01153195e-01
3.96198362e-01 -1.06360324e-01 -3.50381821e-01 6.81431711e-01
2.37305835e-01 1.00788879e+00 8.50609422e-01 -1.29060233e+00
-1.15998340e+00 -3.34480464e-01 3.91441077e-01 6.97393060e-01
-5.59353650e-01 -1.89668685e-01 -8.00912917e-01 -5.55599332e-01
6.13580287e-01 -5.00897765e-01 -2.55957961e-01 2.22950503e-01
-4.62438047e-01 -2.82947183e-01 9.93879318e-01 -6.23887420e-01
1.23143780e+00 -2.17888331e+00 6.17758632e-01 -1.25577077e-01
4.86216456e-01 5.37079684e-02 -8.38735402e-02 -8.09405372e-02
2.67518729e-01 -3.32810730e-01 -2.31070474e-01 -6.44083738e-01
-1.33243054e-01 1.92136332e-01 -1.05705500e-01 5.99593282e-01
2.62569815e-01 9.96680379e-01 -1.24373257e+00 -7.65593767e-01
6.98514283e-01 6.83586478e-01 -3.72139990e-01 1.99717358e-01
-1.37366459e-01 4.85628068e-01 -7.40235388e-01 1.06436777e+00
8.20394874e-01 -3.24815154e-01 -4.35063004e-01 -4.80074465e-01
-3.07440192e-01 5.25614619e-02 -1.18273616e+00 1.94479322e+00
-2.73545891e-01 4.10651892e-01 -2.22782721e-03 -7.02807307e-01
7.52306879e-01 -2.68560704e-02 5.99406779e-01 -9.74344850e-01
2.87679821e-01 1.84562251e-01 -3.10467958e-01 -6.87491417e-01
6.99650884e-01 7.39047676e-02 8.65236968e-02 1.04186408e-01
-8.90109763e-02 5.19820750e-02 1.54459514e-02 2.89768666e-01
6.67637289e-01 1.46894887e-01 7.44553730e-02 1.50575470e-02
9.80570316e-01 -1.49556175e-01 9.39419985e-01 4.01289940e-01
-6.41851068e-01 9.14976478e-01 2.84208566e-01 -3.53447139e-01
-5.81249774e-01 -8.61684620e-01 2.17683971e-01 9.23437953e-01
1.20176661e+00 -1.77084103e-01 -3.25988591e-01 -6.90274775e-01
-2.67893169e-02 2.12131158e-01 -9.03328121e-01 -4.43374813e-01
-4.52324003e-01 -5.98010778e-01 1.20371677e-01 6.42813802e-01
1.08261514e+00 -1.05553722e+00 -8.93566608e-01 2.16034696e-01
-5.98631263e-01 -1.10599947e+00 -6.80965126e-01 -1.19113408e-01
-7.65793145e-01 -8.90241206e-01 -1.06770432e+00 -8.86647463e-01
4.68448341e-01 9.88221705e-01 5.99991083e-01 1.57415718e-01
1.00907132e-01 3.49482626e-01 -4.78848606e-01 -6.04645088e-02
2.74383813e-01 1.11512735e-01 -1.49422903e-02 2.65948862e-01
3.19779634e-01 -3.87350202e-01 -1.14851081e+00 2.40515321e-01
-1.07816207e+00 3.79944116e-01 6.72756016e-01 7.99367964e-01
5.89087963e-01 -1.80445150e-01 3.60773593e-01 -2.20515013e-01
1.87670350e-01 -5.85015953e-01 -3.15430462e-01 1.54485792e-01
-1.41367331e-01 -1.73842102e-01 3.61606181e-01 -3.46361905e-01
-1.19304013e+00 9.91890430e-02 1.20583326e-02 -8.55190635e-01
4.27196659e-02 4.39008981e-01 -4.70636845e-01 -1.42969295e-01
1.05385117e-01 5.74279189e-01 -1.38680533e-01 -3.14754754e-01
2.69352406e-01 5.98047435e-01 6.13850594e-01 -2.76415914e-01
7.00511336e-01 8.70855391e-01 -8.40443671e-02 -5.66688299e-01
-9.73854303e-01 -6.23552144e-01 -5.42665899e-01 -3.05272430e-01
9.95699346e-01 -1.23264575e+00 -7.60879159e-01 7.55125523e-01
-1.12781084e+00 -1.96310595e-01 -1.13752685e-01 6.30517244e-01
-3.73250633e-01 5.37014961e-01 -6.02941692e-01 -7.86228240e-01
-3.07664096e-01 -1.18655264e+00 1.50168920e+00 7.62991309e-01
4.96519208e-01 -8.78449321e-01 -2.67351538e-01 3.22182864e-01
2.84695357e-01 3.03217858e-01 2.50587821e-01 1.32011339e-01
-9.54598546e-01 2.21435219e-01 -6.82821989e-01 1.19646072e-01
3.27702254e-01 -2.30049074e-01 -9.08377469e-01 -2.11921185e-01
-1.78069457e-01 -7.40022063e-02 1.31281877e+00 5.13813317e-01
1.02388418e+00 4.59354706e-02 -3.87256891e-01 8.32965493e-01
1.21657205e+00 2.77190097e-02 5.71459293e-01 6.30830646e-01
1.24308777e+00 4.38581795e-01 1.05474377e+00 5.66999674e-01
8.03203404e-01 5.58843017e-01 7.66582668e-01 -3.32416832e-01
-1.95258230e-01 -2.43188173e-01 3.71458113e-01 5.60979486e-01
-1.22723520e-01 -1.60609573e-01 -5.13619006e-01 6.47547364e-01
-1.96310186e+00 -1.00087166e+00 -6.25955611e-02 1.80838013e+00
5.80406785e-01 1.40168130e-01 2.88468301e-01 1.09017089e-01
7.57327437e-01 4.97054160e-01 -6.90854073e-01 3.36053073e-01
-5.13878584e-01 -4.94937509e-01 3.42548281e-01 3.07706773e-01
-1.38707304e+00 8.75793993e-01 4.98419237e+00 7.85064101e-01
-1.35751736e+00 1.16858982e-01 5.33344686e-01 -2.17176691e-01
-4.91619021e-01 -8.48241374e-02 -5.87505460e-01 6.80675328e-01
3.37876044e-02 -4.94308909e-03 1.03955872e-01 8.03730428e-01
4.68976855e-01 -4.02823359e-01 -5.61632812e-01 1.23527646e+00
1.95554078e-01 -1.26038063e+00 6.45987540e-02 -1.57748729e-01
7.57520914e-01 5.45392325e-03 4.67917360e-02 8.00061151e-02
-1.78649202e-01 -4.37703431e-01 9.53597307e-01 6.99004889e-01
4.74790812e-01 -8.01449358e-01 7.80348361e-01 8.86628777e-02
-1.68578017e+00 -4.13023680e-01 -4.50409204e-01 8.86452943e-02
3.80860388e-01 5.97630799e-01 -4.37922180e-02 7.56579995e-01
8.99371386e-01 1.38694215e+00 -6.57854497e-01 1.27667582e+00
-2.49472722e-01 1.90631542e-02 -2.94178605e-01 -3.43059492e-03
5.02663493e-01 1.11459848e-02 7.25596070e-01 9.67776000e-01
1.98911682e-01 3.35493237e-01 2.19853565e-01 6.64942145e-01
2.13846698e-01 -1.48051023e-01 -1.93619475e-01 4.23254132e-01
4.15253252e-01 1.34971070e+00 -6.90005839e-01 -3.82127494e-01
-6.10279322e-01 1.26646686e+00 5.78450151e-02 4.11972880e-01
-8.76789153e-01 -3.12372923e-01 6.85847282e-01 -1.77075759e-01
5.38865507e-01 -1.23076409e-01 -1.81353539e-02 -1.44940770e+00
1.14180610e-01 -4.62872148e-01 6.65265501e-01 -1.13478816e+00
-1.15343773e+00 4.25662696e-01 -1.34015903e-01 -1.74315178e+00
1.82154268e-01 -2.22793430e-01 -5.48032343e-01 9.32335854e-01
-2.08711195e+00 -1.38325465e+00 -8.39555681e-01 8.75171840e-01
5.25144041e-01 2.28171721e-01 -2.41670646e-02 3.91163647e-01
-8.23632121e-01 2.87200868e-01 -2.67798930e-01 8.73603523e-02
5.30138850e-01 -1.01763844e+00 -1.53111191e-02 1.13401318e+00
-3.21864933e-01 4.84236628e-01 4.00582343e-01 -5.75471878e-01
-1.55376303e+00 -1.27298558e+00 5.18034637e-01 8.64214357e-03
4.27979439e-01 -3.34083810e-02 -8.05735111e-01 4.56957400e-01
-2.58265380e-02 3.67900699e-01 9.25507843e-02 -4.40483093e-01
-9.31084231e-02 -2.07075879e-01 -9.01572108e-01 5.70831954e-01
1.16650617e+00 -5.26234090e-01 -5.86681843e-01 -5.23604788e-02
1.08934605e+00 -6.20217860e-01 -6.75731361e-01 5.05522370e-01
4.18394685e-01 -1.14627993e+00 1.04341102e+00 -6.72341138e-02
5.65426767e-01 -9.01368022e-01 -2.18271300e-01 -1.03243470e+00
-2.53154218e-01 -4.59657758e-01 -5.92785299e-01 1.23706877e+00
-1.95736796e-01 -4.58596706e-01 7.08365798e-01 4.08738583e-01
-4.46697265e-01 -7.67971814e-01 -9.69194472e-01 -3.37582737e-01
-5.74268997e-01 -2.84699857e-01 5.04733562e-01 9.24626946e-01
-4.46646400e-02 2.04498067e-01 -5.44688106e-01 2.86631167e-01
5.89752376e-01 4.42146748e-01 6.48795009e-01 -1.02512860e+00
-3.39170210e-02 -6.02440953e-01 -6.32663131e-01 -1.60174704e+00
-1.97330341e-01 -4.18229908e-01 1.51534334e-01 -1.75383139e+00
3.74892533e-01 -3.28025550e-01 -6.03140414e-01 4.02041018e-01
-6.86562121e-01 4.05949891e-01 2.76029021e-01 2.25842461e-01
-1.00776875e+00 1.18524802e+00 1.66548848e+00 -1.16114743e-01
-4.05491710e-01 -3.14525634e-01 -8.84696126e-01 6.15169883e-01
4.73643959e-01 -1.12539694e-01 -4.74386781e-01 -5.02797484e-01
-1.63849264e-01 1.73536286e-01 7.25937128e-01 -1.06809044e+00
4.50671285e-01 -2.86912024e-01 5.90020180e-01 -9.86453891e-01
5.54741919e-01 -7.12582648e-01 -3.98151696e-01 3.64589989e-01
-8.34237114e-02 -1.25743374e-01 7.20093027e-02 7.45651960e-01
-4.96061236e-01 2.89041400e-01 6.58814192e-01 -4.36914973e-02
-1.43715680e+00 7.25298882e-01 -6.88656271e-02 -3.65703739e-02
1.05552387e+00 -3.92694831e-01 -3.38322103e-01 -2.71582425e-01
-4.89205241e-01 6.63907886e-01 4.37743396e-01 6.57977700e-01
1.04617155e+00 -1.45708406e+00 -5.34669638e-01 1.79909647e-01
2.83477724e-01 2.84868717e-01 7.88501143e-01 1.22579801e+00
-3.44894022e-01 1.54552698e-01 -1.44348577e-01 -8.77606332e-01
-1.00596404e+00 6.69630766e-01 4.58990335e-01 2.61064857e-01
-7.33013749e-01 1.04577327e+00 4.52236533e-01 9.01361704e-02
3.48069459e-01 -3.64621341e-01 -4.79443908e-01 2.85558194e-01
7.41909623e-01 2.65094906e-01 -2.95252025e-01 -9.08700645e-01
-5.47080398e-01 7.77674198e-01 8.80519822e-02 1.80360861e-02
1.21060491e+00 -7.79949307e-01 -3.65192816e-02 4.62058961e-01
1.12250197e+00 -3.53346109e-01 -1.70639324e+00 -5.15311360e-01
-4.81976092e-01 -7.24305034e-01 3.72826397e-01 -2.48163894e-01
-1.51550317e+00 1.00366771e+00 6.11784756e-01 8.56591612e-02
1.60360610e+00 1.42231034e-02 1.07827401e+00 4.28916067e-02
2.15651661e-01 -9.26737428e-01 2.92292506e-01 3.21725905e-01
6.21337235e-01 -1.50634265e+00 -2.82794964e-02 -5.07465422e-01
-7.93202102e-01 9.60916460e-01 9.83081341e-01 8.72079059e-02
4.62844282e-01 -2.97921687e-01 -6.27423599e-02 -7.00246915e-02
-5.17602324e-01 -7.33478904e-01 3.81329119e-01 5.73520839e-01
8.71840678e-03 -3.41439247e-01 -1.19023092e-01 4.92538124e-01
4.93131489e-01 1.10801652e-01 3.71441841e-01 9.87689435e-01
-5.44653714e-01 -5.12720168e-01 -2.89734900e-01 1.65997326e-01
-3.58005799e-02 -1.90279007e-01 -1.06272228e-01 7.72455037e-01
4.01854813e-01 9.62644815e-01 4.06684875e-02 -6.37467325e-01
2.65313864e-01 -6.59530520e-01 2.35732064e-01 -3.35718930e-01
-3.37522477e-01 1.98148012e-01 -2.73655474e-01 -7.49037921e-01
-1.00624394e+00 -7.85380781e-01 -1.32764661e+00 -3.02680165e-01
-2.53311187e-01 -2.98979110e-03 1.27133265e-01 9.22810853e-01
2.79970735e-01 6.15179300e-01 8.09076428e-01 -1.27784157e+00
9.29689780e-02 -7.38633752e-01 -5.53524256e-01 3.22517455e-01
7.77801275e-01 -1.07332540e+00 -5.48581779e-01 -9.27597657e-02] | [9.583089828491211, -0.633668839931488] |
81d383a7-d824-41ff-a1e6-34c70a66abb1 | occlusion-handling-in-generic-object | 2101.08845 | null | https://arxiv.org/abs/2101.08845v1 | https://arxiv.org/pdf/2101.08845v1.pdf | Occlusion Handling in Generic Object Detection: A Review | The significant power of deep learning networks has led to enormous development in object detection. Over the last few years, object detector frameworks have achieved tremendous success in both accuracy and efficiency. However, their ability is far from that of human beings due to several factors, occlusion being one of them. Since occlusion can happen in various locations, scale, and ratio, it is very difficult to handle. In this paper, we address the challenges in occlusion handling in generic object detection in both outdoor and indoor scenes, then we refer to the recent works that have been carried out to overcome these challenges. Finally, we discuss some possible future directions of research. | ['Zoltán Vámossy', 'Sándor Szénási', 'Kaziwa Saleh'] | 2021-01-21 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [-8.59382227e-02 -4.83888119e-01 3.47929373e-02 -3.34967941e-01
-6.46904409e-02 -2.82680929e-01 3.55685920e-01 4.21426184e-02
-4.19848412e-01 5.40909827e-01 -6.25272617e-02 -7.93175921e-02
2.63536274e-02 -7.79331386e-01 -3.14216256e-01 -5.03330290e-01
-9.43822972e-03 9.82071459e-03 8.17746639e-01 9.56701562e-02
4.77278866e-02 1.02654791e+00 -1.69570112e+00 -1.73579410e-01
5.06892145e-01 9.56997752e-01 1.99524194e-01 5.80368280e-01
1.75513607e-02 5.68179548e-01 -7.44253337e-01 -3.64744723e-01
3.80752504e-01 -2.89463904e-02 -3.34617376e-01 2.92171270e-01
6.30731761e-01 -5.67641795e-01 -5.55585265e-01 9.97058392e-01
7.01423824e-01 3.57302204e-02 4.67547476e-01 -1.26943874e+00
-4.83345032e-01 2.01425236e-02 -7.45489836e-01 4.01173890e-01
2.33654246e-01 -5.39173298e-02 5.74267149e-01 -1.01398098e+00
4.01147097e-01 1.49321377e+00 6.66695058e-01 2.43712768e-01
-8.22797775e-01 -6.87291980e-01 3.42671663e-01 2.61526167e-01
-1.33242273e+00 -3.36192161e-01 7.55173326e-01 -3.24271679e-01
9.77156997e-01 8.01346302e-02 7.31241882e-01 8.12321842e-01
2.64067471e-01 9.47450459e-01 9.06889677e-01 -5.21917105e-01
1.39577299e-01 2.27520704e-01 -5.04105585e-04 3.91799659e-01
5.27882338e-01 4.16823849e-02 -1.13118336e-01 2.70664934e-02
8.88356090e-01 7.12440684e-02 4.55524772e-03 -6.65848613e-01
-8.58705044e-01 7.79796183e-01 7.04844177e-01 2.32706904e-01
-1.48848712e-01 8.31969827e-02 3.53404552e-01 -2.94688474e-02
2.61093229e-01 2.29455739e-01 -2.65734732e-01 1.55277699e-01
-6.64705515e-01 3.64108592e-01 7.89694309e-01 1.03640342e+00
3.16565126e-01 1.26334786e-01 2.13017195e-01 7.10835874e-01
4.27200913e-01 2.91340500e-01 8.08290169e-02 -6.32218599e-01
2.85217434e-01 5.12093306e-01 1.63889632e-01 -1.30055928e+00
-6.86732888e-01 -5.99265635e-01 -7.48613477e-01 4.93085265e-01
4.63427931e-01 1.39824348e-02 -8.21960449e-01 1.27506840e+00
6.02957010e-01 -1.27298594e-01 -2.43726179e-01 9.05205309e-01
1.03725958e+00 3.46058637e-01 2.89488465e-01 1.63966984e-01
1.36535883e+00 -9.49576497e-01 -8.60916018e-01 -5.41446209e-01
1.33401796e-01 -1.22655439e+00 5.99633217e-01 4.04963166e-01
-8.73570263e-01 -7.83866823e-01 -1.28922689e+00 -2.58766234e-01
-4.97043699e-01 6.25813901e-02 9.53215122e-01 8.16164017e-01
-7.11471438e-01 2.12718964e-01 -8.49799216e-01 -8.02934527e-01
6.94024861e-01 5.55556655e-01 -7.92118162e-02 -1.70115456e-01
-8.25213730e-01 1.11803794e+00 4.41577286e-01 2.81864256e-01
-5.00051200e-01 -1.32588208e-01 -5.05192339e-01 -1.39429137e-01
4.53344703e-01 -5.08503079e-01 1.06268167e+00 -5.29444277e-01
-1.06614912e+00 7.61116683e-01 -1.02187678e-01 -4.34913933e-01
7.62856662e-01 -5.93750894e-01 -5.03019571e-01 -2.63586968e-01
-1.09179921e-01 7.30823696e-01 4.87444699e-01 -1.04268456e+00
-9.27063942e-01 -3.69441926e-01 4.97056127e-01 2.24220097e-01
-2.63404131e-01 6.53244376e-01 -5.78918874e-01 -5.13921082e-01
3.17134142e-01 -8.67182910e-01 -3.02675724e-01 5.66339135e-01
-2.88056970e-01 -3.73047352e-01 1.29212165e+00 -2.63975769e-01
9.84380662e-01 -2.16983938e+00 -1.34923592e-01 -3.59362721e-01
2.28057966e-01 6.06415033e-01 1.65913716e-01 4.62609887e-01
1.59704670e-01 -4.51977365e-02 8.91761929e-02 -4.76569623e-01
-9.91187021e-02 1.57467306e-01 -3.36024672e-01 6.39281690e-01
1.78825021e-01 6.90698683e-01 -7.19127297e-01 -6.09831095e-01
6.75614953e-01 7.89238811e-01 -2.61593461e-01 6.63622916e-02
-1.65663578e-03 2.71603912e-01 -5.84598601e-01 1.04049325e+00
8.70307207e-01 1.16954215e-01 -1.06101528e-01 -1.76225424e-01
-3.88894856e-01 1.74152181e-01 -1.53731406e+00 1.20970237e+00
-2.89912254e-01 1.01326025e+00 1.05607413e-01 -9.14800107e-01
7.37961948e-01 2.97150493e-01 3.85544270e-01 -4.95383620e-01
3.79177153e-01 2.47958601e-01 2.26945013e-01 -6.27452850e-01
6.15138590e-01 1.41169727e-01 2.36509979e-01 1.56891972e-01
-3.45063567e-01 -2.15465635e-01 3.82515937e-01 -1.41659781e-01
8.25971007e-01 1.38516560e-01 7.23409235e-01 3.41052338e-02
4.69724625e-01 -1.16192825e-01 6.35928154e-01 8.93436611e-01
-7.64606535e-01 7.28984952e-01 3.93016562e-02 -9.71144974e-01
-8.51039886e-01 -1.02788544e+00 -4.78631794e-01 8.50847721e-01
3.64300579e-01 -4.98677641e-02 -4.83267218e-01 -3.49171579e-01
3.45241129e-02 8.91302973e-02 -2.83591211e-01 1.02985002e-01
-8.76948178e-01 -8.81342530e-01 3.16775262e-01 1.01335001e+00
8.36113572e-01 -1.14852500e+00 -1.10409510e+00 4.53338236e-01
2.26431526e-02 -1.54410625e+00 9.61819440e-02 8.76225010e-02
-9.96312320e-01 -9.39528346e-01 -4.86672133e-01 -8.54979694e-01
4.96769100e-01 8.80084038e-01 1.11379409e+00 2.49990776e-01
-8.94364834e-01 8.81675929e-02 -2.83884615e-01 -1.00551534e+00
6.18930049e-02 -9.23808385e-03 2.30123714e-01 -3.90795410e-01
6.96559846e-01 -4.13737297e-01 -7.42672563e-01 5.29214799e-01
-9.03814733e-01 -1.59309417e-01 4.91950691e-01 3.07328194e-01
4.87055987e-01 3.51128101e-01 2.42162913e-01 -3.42817515e-01
1.98674396e-01 -1.87597200e-01 -8.83694887e-01 2.26328708e-02
-6.32446557e-02 -6.25740945e-01 3.17013115e-01 -4.91824180e-01
-8.53206575e-01 2.52190828e-01 -1.95706859e-01 1.19869843e-01
-5.37059605e-01 3.80692109e-02 -2.27212191e-01 -4.46781814e-01
4.31820124e-01 -3.48112792e-01 -5.78885198e-01 -5.61066747e-01
1.01495340e-01 7.19936967e-01 3.33701879e-01 -2.18912706e-01
8.59964848e-01 7.73426712e-01 -1.98856350e-02 -1.13456476e+00
-1.10634458e+00 -4.75659966e-01 -9.79542196e-01 -3.46266568e-01
8.67092192e-01 -8.22430134e-01 -5.18210471e-01 6.31265163e-01
-1.31281352e+00 2.41920993e-01 -1.30466953e-01 5.69671988e-01
-4.39418815e-02 3.28557342e-01 -5.18694639e-01 -1.04227233e+00
-2.25943830e-02 -1.25891972e+00 9.21224296e-01 6.48624539e-01
-7.11495802e-02 -8.95302594e-01 -2.92589635e-01 8.34160596e-02
6.12685442e-01 3.82856071e-01 3.98223877e-01 -1.68085143e-01
-9.90233541e-01 -6.64548755e-01 -6.02092326e-01 1.04083687e-01
1.94811702e-01 2.01641113e-01 -9.98201132e-01 -4.51843858e-01
3.27270508e-01 -2.58521676e-01 6.76229894e-01 6.75162375e-01
1.05110645e+00 3.22186112e-01 -6.62699819e-01 5.26217997e-01
1.41348016e+00 4.22080725e-01 5.63378632e-01 7.05610216e-01
5.82820356e-01 5.37311018e-01 7.78787255e-01 2.92142004e-01
1.94667935e-01 9.99141514e-01 6.82638586e-01 -4.49795306e-01
-4.23211157e-01 1.09930150e-01 -2.00997666e-01 3.05975109e-01
-2.20984444e-01 -4.33744073e-01 -7.53744006e-01 4.35837388e-01
-1.66604638e+00 -7.32681811e-01 -2.99781263e-01 2.00174785e+00
2.03716859e-01 3.73892903e-01 8.83790106e-02 2.94360697e-01
6.49253428e-01 2.15604767e-01 -5.45932055e-01 -1.76392913e-01
-2.93626785e-01 -1.44559726e-01 3.95979047e-01 2.45733306e-01
-1.57549739e+00 9.07224238e-01 7.34596062e+00 4.68559921e-01
-1.14973152e+00 2.49345265e-02 3.05455476e-01 4.83953804e-02
4.58695918e-01 -3.47172581e-02 -1.14354551e+00 2.32078120e-01
2.16294289e-01 2.09182620e-01 -5.89653142e-02 1.06624317e+00
-5.38514368e-02 -2.99987048e-01 -9.41203654e-01 8.73436034e-01
1.06044471e-01 -6.43165231e-01 -3.84254456e-01 3.18398476e-02
6.16405487e-01 1.37225211e-01 1.92356110e-01 1.65282935e-01
7.03685656e-02 -8.31093848e-01 7.06069112e-01 -4.34107557e-02
2.05111921e-01 -5.68897784e-01 8.66483867e-01 3.44664156e-01
-1.44011033e+00 -1.39000982e-01 -8.68294060e-01 -4.34461772e-01
2.69071996e-01 7.09305167e-01 -6.23546720e-01 2.18510300e-01
9.22055721e-01 3.86055976e-01 -5.43099463e-01 1.73325086e+00
-3.82541150e-01 1.94685444e-01 -7.43040860e-01 -1.36437759e-01
1.41543210e-01 -3.08101159e-02 4.54795212e-01 1.19648540e+00
1.20941214e-01 6.01229705e-02 4.42729294e-01 6.12550497e-01
2.09113389e-01 2.49373205e-02 -5.11333704e-01 3.43621850e-01
4.90067095e-01 1.29510427e+00 -1.17645025e+00 -1.27024129e-01
-7.96011031e-01 6.87600493e-01 2.24808216e-01 3.05917382e-01
-9.76515174e-01 -2.80852735e-01 6.99449539e-01 1.02829598e-01
3.32603693e-01 -6.25328243e-01 -1.49903089e-01 -9.44783092e-01
2.08625644e-01 -5.46746552e-01 2.35028937e-01 -5.51842391e-01
-1.05830288e+00 6.36082172e-01 3.93621847e-02 -1.20684218e+00
3.12408090e-01 -8.38296413e-01 -4.78906840e-01 5.16245067e-01
-1.46370435e+00 -1.06137133e+00 -4.99499649e-01 3.47668491e-02
8.73723984e-01 1.29640251e-01 6.96438432e-01 6.82760894e-01
-6.67082608e-01 3.17400753e-01 -1.18918382e-02 1.92949802e-01
5.89554310e-01 -8.70491087e-01 6.44444466e-01 1.06181550e+00
2.37326741e-01 5.38777828e-01 9.77008939e-01 -3.80085349e-01
-1.23668766e+00 -8.29992771e-01 8.52718174e-01 -2.52286643e-01
4.38841820e-01 -5.23909926e-01 -7.82393932e-01 7.00001895e-01
-1.97519716e-02 3.47843945e-01 3.45524400e-01 1.55585974e-01
-2.53706843e-01 -1.13469176e-01 -1.10481870e+00 5.87080121e-01
9.54801798e-01 -2.10175395e-01 -4.80695218e-01 4.49304044e-01
4.41265345e-01 -8.22460234e-01 -5.17613530e-01 6.31156147e-01
8.68309855e-01 -1.27565897e+00 1.28832233e+00 -2.57582724e-01
-3.14161927e-02 -4.81170774e-01 -1.54473737e-01 -6.76102281e-01
-4.89812464e-01 -6.20786250e-02 -1.25353947e-01 1.18333304e+00
-2.10074663e-01 -6.61550224e-01 9.24885929e-01 6.72722042e-01
3.31930853e-02 -5.80941617e-01 -8.85418296e-01 -9.05021012e-01
-2.65674025e-01 -5.08976400e-01 4.94353831e-01 3.22917789e-01
-6.52991772e-01 1.29517108e-01 -4.45153058e-01 2.51925468e-01
6.47681355e-01 1.99546367e-01 8.76571894e-01 -1.39332592e+00
-2.85419170e-02 -3.89004678e-01 -7.87625253e-01 -1.24855089e+00
-2.78221339e-01 -1.85214490e-01 5.75783476e-03 -1.66271973e+00
2.40899742e-01 -4.64371026e-01 -2.08157793e-01 2.54201621e-01
-2.61985332e-01 7.29762614e-01 3.31298023e-01 2.43421391e-01
-6.27116680e-01 3.05390120e-01 1.01192343e+00 -1.08198628e-01
-5.59716485e-02 1.57923028e-01 -3.33308548e-01 1.10272861e+00
8.66429508e-01 -5.02362847e-01 -4.35939664e-03 -7.42629230e-01
-4.20729928e-02 -4.93331134e-01 3.86553437e-01 -1.55242634e+00
1.89838156e-01 4.57792282e-02 7.11484969e-01 -1.08131063e+00
6.59229338e-01 -1.02682889e+00 4.45741601e-02 5.62278569e-01
1.63367212e-01 3.93885523e-02 3.18099827e-01 4.36965287e-01
-2.28683561e-01 -2.45262951e-01 9.31502461e-01 -8.36425200e-02
-8.44559371e-01 2.15099454e-01 -4.01881248e-01 -4.52582479e-01
1.23400176e+00 -4.65025604e-01 -8.31659213e-02 -1.39456585e-01
-5.18071711e-01 7.40304291e-02 3.61023307e-01 7.06677675e-01
4.47825730e-01 -1.09659708e+00 -4.26168442e-01 3.83939147e-02
-1.24195009e-01 -2.82253399e-02 1.15831658e-01 6.82364047e-01
-6.79423928e-01 6.73681021e-01 -2.58271575e-01 -7.52667606e-01
-1.48416555e+00 7.65360057e-01 1.68428659e-01 2.17973981e-02
-7.48675942e-01 9.36258674e-01 3.15604359e-01 3.26080248e-02
7.35900402e-01 -1.61043867e-01 -2.76209474e-01 -8.68373290e-02
7.06281841e-01 5.92650175e-01 6.97556362e-02 -6.60156488e-01
-5.81240177e-01 9.18003142e-01 -2.64348060e-01 3.84028018e-01
1.11246264e+00 -2.08117172e-01 5.05144931e-02 1.99620828e-01
8.80787313e-01 -2.87979916e-02 -1.23656189e+00 -2.04466298e-01
-1.01239584e-01 -6.69699728e-01 3.99730466e-02 -3.61437261e-01
-1.12481785e+00 1.08118951e+00 9.24251258e-01 3.95843446e-01
1.06557012e+00 -1.31416589e-01 6.34309769e-01 3.78002495e-01
5.83869755e-01 -1.01245224e+00 -9.91046950e-02 5.26522875e-01
7.74039268e-01 -1.66984427e+00 5.05566001e-01 -8.01472008e-01
5.66492714e-02 1.15945303e+00 7.85740316e-01 -3.14088345e-01
7.91680217e-01 4.56111729e-01 1.58293933e-01 -1.80504918e-01
-1.43003926e-01 -3.38626027e-01 1.85950324e-01 7.56045401e-01
7.58438408e-01 -1.76685989e-01 -3.75892669e-01 -1.97992519e-01
-3.23576406e-02 -5.65456413e-02 1.71222165e-01 1.15933263e+00
-6.84735715e-01 -1.15242791e+00 -7.85172403e-01 2.53340960e-01
-6.37770772e-01 3.40023339e-01 -2.38870740e-01 1.08151972e+00
3.72991562e-01 1.21234536e+00 -6.86003715e-02 -2.96455063e-02
5.71603358e-01 -4.83664453e-01 5.56054950e-01 -5.65981984e-01
-2.52235532e-01 2.28384156e-02 -1.65375814e-01 -5.72646737e-01
-5.48744023e-01 -6.48700893e-01 -9.94078398e-01 -2.16671657e-02
-5.59445262e-01 -3.76806051e-01 8.21663499e-01 9.10479844e-01
-2.45809965e-02 6.16764188e-01 2.30969846e-01 -1.18960476e+00
-2.25443959e-01 -9.67928112e-01 -4.66742694e-01 5.92716364e-03
2.53624529e-01 -1.05326521e+00 -2.91739762e-01 -2.37503901e-01] | [8.64079475402832, -0.4951813220977783] |
d688c7d7-fe30-4af4-b199-f76ea0f8bbf9 | basis-pursuit-denoising-via-recurrent-neural | 2305.14209 | null | https://arxiv.org/abs/2305.14209v1 | https://arxiv.org/pdf/2305.14209v1.pdf | Basis Pursuit Denoising via Recurrent Neural Network Applied to Super-resolving SAR Tomography | Finding sparse solutions of underdetermined linear systems commonly requires the solving of L1 regularized least squares minimization problem, which is also known as the basis pursuit denoising (BPDN). They are computationally expensive since they cannot be solved analytically. An emerging technique known as deep unrolling provided a good combination of the descriptive ability of neural networks, explainable, and computational efficiency for BPDN. Many unrolled neural networks for BPDN, e.g. learned iterative shrinkage thresholding algorithm and its variants, employ shrinkage functions to prune elements with small magnitude. Through experiments on synthetic aperture radar tomography (TomoSAR), we discover the shrinkage step leads to unavoidable information loss in the dynamics of networks and degrades the performance of the model. We propose a recurrent neural network (RNN) with novel sparse minimal gated units (SMGUs) to solve the information loss issue. The proposed RNN architecture with SMGUs benefits from incorporating historical information into optimization, and thus effectively preserves full information in the final output. Taking TomoSAR inversion as an example, extensive simulations demonstrated that the proposed RNN outperforms the state-of-the-art deep learning-based algorithm in terms of super-resolution power as well as generalization ability. It achieved a 10% to 20% higher double scatterers detection rate and is less sensitive to phase and amplitude ratio differences between scatterers. Test on real TerraSAR-X spotlight images also shows a high-quality 3-D reconstruction of the test site. | ['Xiao Xiang Zhu', 'Yilei Shi', 'Peter Jung', 'Yuanyuan Wang', 'Kun Qian'] | 2023-05-23 | null | null | null | null | ['unrolling'] | ['computer-vision'] | [ 1.60136506e-01 -8.55508670e-02 5.00215709e-01 -4.11039561e-01
-7.80444503e-01 6.54166713e-02 8.37960839e-02 -7.82184184e-01
-2.02373743e-01 8.07489455e-01 1.91572219e-01 -1.63507119e-01
-4.46270645e-01 -6.83411777e-01 -7.77247548e-01 -1.37004173e+00
-2.57774681e-01 4.12533909e-01 -3.62396479e-01 -3.98255438e-01
-6.26576785e-03 5.47103822e-01 -1.13800085e+00 8.45459814e-04
8.93398046e-01 1.13504815e+00 3.54373842e-01 -3.27239409e-02
2.89486587e-01 8.52369130e-01 -2.51664817e-01 2.97231704e-01
5.21763504e-01 -4.39097881e-01 -4.77502085e-02 1.74900949e-01
4.40263003e-01 -5.26654780e-01 -7.31749356e-01 1.09303796e+00
6.30718112e-01 2.35707000e-01 5.15774846e-01 -5.84766984e-01
-6.81745768e-01 5.79320908e-01 -9.14231598e-01 4.10624117e-01
-4.40325201e-01 9.14092138e-02 7.15217054e-01 -1.19974828e+00
3.24116111e-01 9.10097659e-01 1.07013261e+00 1.77561924e-01
-1.01600647e+00 -6.58046365e-01 -1.33179784e-01 1.49982452e-01
-1.62106431e+00 -6.71503127e-01 8.51282895e-01 -1.42513692e-01
9.48346734e-01 2.83127338e-01 5.80834746e-01 6.96007848e-01
2.27394879e-01 3.98499310e-01 1.03488803e+00 -3.01188886e-01
5.93583621e-02 -2.19548240e-01 1.54279023e-01 8.35182786e-01
5.63623309e-01 3.13162327e-01 -4.36936438e-01 -1.39441900e-02
1.09835362e+00 1.57879204e-01 -8.47268879e-01 -6.21775351e-03
-1.10024548e+00 9.68202889e-01 8.34603608e-01 2.81490326e-01
-7.96277404e-01 2.14931831e-01 -1.04573354e-01 3.47276628e-01
6.45702124e-01 4.35757488e-01 -1.09457843e-01 6.14062548e-01
-1.24935651e+00 -1.04729459e-02 5.83643496e-01 4.69096571e-01
9.19339955e-01 1.16365600e+00 1.36343213e-02 8.45488071e-01
5.95477700e-01 1.00881934e+00 5.37070572e-01 -8.51165175e-01
3.49312901e-01 2.17748642e-01 3.36668342e-02 -1.33359516e+00
-5.84305286e-01 -1.11225247e+00 -1.54138768e+00 2.41421133e-01
-1.06262282e-01 -2.86162198e-01 -1.28203499e+00 1.64132738e+00
1.65813714e-01 3.95870656e-01 1.91075861e-01 1.49888110e+00
8.44221115e-01 1.13011706e+00 -4.58793670e-01 -6.45256817e-01
9.46159124e-01 -6.60950541e-01 -8.97438347e-01 -7.40753949e-01
2.47710884e-01 -6.43351018e-01 3.88128698e-01 5.01014173e-01
-8.29303503e-01 -3.23298275e-01 -1.22739220e+00 1.33554965e-01
4.53712046e-01 2.14095458e-01 8.55746031e-01 3.04757953e-01
-9.56975639e-01 7.06168056e-01 -9.67355847e-01 -1.59859750e-02
3.77536625e-01 1.31281555e-01 -1.74017265e-01 -1.83211625e-01
-1.17981076e+00 7.90610492e-01 -3.67934108e-02 9.00994658e-01
-1.42935193e+00 -6.87920868e-01 -8.14898074e-01 1.77809030e-01
1.97875410e-01 -6.11038446e-01 7.59962916e-01 -8.77600729e-01
-1.23688877e+00 5.78309298e-01 4.00635265e-02 -6.35715902e-01
3.46910767e-02 -3.26017857e-01 -4.45937097e-01 1.14530697e-01
8.21713209e-02 7.81726167e-02 1.24029922e+00 -1.17972374e+00
-1.68320477e-01 -4.76949811e-01 -5.99611461e-01 1.81191891e-01
-1.35664232e-02 -3.89605671e-01 1.55369014e-01 -7.62813628e-01
1.03727245e+00 -7.48381436e-01 -5.26391983e-01 -1.69516891e-01
-1.54450119e-01 3.26741576e-01 6.79494143e-01 -1.02548754e+00
1.13769114e+00 -2.22084045e+00 7.34511688e-02 2.89803296e-01
1.91520572e-01 7.97756910e-02 -1.26306549e-01 2.49777138e-01
-3.47066253e-01 -3.65882427e-01 -6.34794414e-01 -4.05258946e-02
-3.49941283e-01 1.94832295e-01 -6.31137133e-01 9.84285533e-01
1.62178092e-02 5.93756676e-01 -3.52350205e-01 -6.27602935e-02
3.13443020e-02 6.90356493e-01 -4.15838599e-01 1.58542097e-01
-8.62276256e-02 5.27345479e-01 -5.23119450e-01 8.86166811e-01
9.37910140e-01 -4.27084535e-01 -5.02940714e-02 -6.21983171e-01
-2.44943500e-01 -1.14880830e-01 -1.32660913e+00 1.47644889e+00
-5.20968318e-01 7.48644710e-01 5.80707133e-01 -1.38530934e+00
1.36206007e+00 3.50432783e-01 3.32500309e-01 -7.60259449e-01
9.67274532e-02 5.03948271e-01 -2.46472210e-02 -7.42529035e-01
2.93126315e-01 -5.59910417e-01 3.67966443e-01 2.95988113e-01
-1.53241947e-01 -5.99530786e-02 -4.05545086e-01 5.46533316e-02
9.78156447e-01 -9.09760743e-02 2.90773690e-01 -3.73145670e-01
3.03761512e-01 3.85246463e-02 9.60839808e-01 8.65192056e-01
2.79527783e-01 8.36008668e-01 -1.60508394e-01 -7.76834607e-01
-1.04818809e+00 -7.66164720e-01 -3.36493284e-01 6.51893318e-01
-8.50166231e-02 4.02077794e-01 -2.01616824e-01 2.00054750e-01
-1.81060910e-01 6.32601142e-01 -3.51005167e-01 -5.67639284e-02
-8.92308176e-01 -1.39057922e+00 3.85271758e-01 1.22251429e-01
9.87202764e-01 -8.72736871e-01 -5.24588525e-01 2.41863132e-01
-4.25985605e-01 -1.01122534e+00 4.69896384e-02 3.05041343e-01
-1.27754593e+00 -6.15838230e-01 -7.55570292e-01 -7.18185663e-01
7.18330085e-01 5.24073422e-01 9.36896205e-01 -9.88738611e-02
-2.78268635e-01 -4.14544456e-02 -2.24296495e-01 -8.96898434e-02
3.95585671e-02 -3.42731178e-01 2.74300665e-01 2.69184649e-01
-2.62467861e-02 -1.17809474e+00 -6.52522385e-01 1.64251715e-01
-7.90347755e-01 -5.86862043e-02 1.06385732e+00 1.08576000e+00
7.64660299e-01 1.27591163e-01 4.97424632e-01 -7.19202459e-01
3.81786138e-01 -5.15209258e-01 -7.25681722e-01 6.72858581e-02
-6.71521306e-01 2.83580814e-02 5.24245441e-01 -4.81367521e-02
-1.29746437e+00 -1.22683033e-01 1.97225232e-02 -6.70105457e-01
3.54971468e-01 9.44007277e-01 1.53616354e-01 -5.44056833e-01
8.35381567e-01 7.27138638e-01 2.65970491e-02 -5.37857115e-01
-5.86866634e-03 4.33149844e-01 5.18873155e-01 -1.73776776e-01
9.31842864e-01 7.73632765e-01 2.01905325e-01 -1.22192204e+00
-1.13920450e+00 -2.39275843e-01 2.23925505e-02 -1.55978784e-01
4.30531532e-01 -1.35787892e+00 -5.11178315e-01 5.31859577e-01
-9.45283473e-01 -5.76805323e-02 -1.01680234e-01 8.60048473e-01
-1.77753240e-01 4.44291323e-01 -6.26226366e-01 -8.47039044e-01
-7.59313583e-01 -6.30745173e-01 6.27021611e-01 2.16633111e-01
4.06195492e-01 -6.54056072e-01 -1.48728434e-02 3.46202135e-01
8.07355344e-01 3.15915853e-01 7.71490037e-01 -2.27737144e-01
-8.92018914e-01 -1.24189273e-01 -2.78890729e-01 6.66934848e-01
-1.97655573e-01 -4.47867006e-01 -8.44875574e-01 -6.58624470e-01
1.00082815e+00 -2.38112748e-01 1.40083957e+00 1.07242370e+00
6.67108119e-01 -6.06534243e-01 -1.23158537e-01 1.29764962e+00
1.81169546e+00 1.76560786e-02 8.41847956e-01 2.49001533e-01
7.51377344e-01 3.43038678e-01 3.81700218e-01 7.02695549e-01
-2.31287256e-01 1.12377822e-01 7.31860042e-01 -8.23109001e-02
-1.17524844e-02 3.18947226e-01 3.80244613e-01 1.27834415e+00
-4.11203414e-01 -4.25619222e-02 -8.12591970e-01 5.92568755e-01
-1.73439717e+00 -1.01539874e+00 -1.82052538e-01 1.80465794e+00
4.84794199e-01 -1.45587176e-01 -9.02862966e-01 -1.14028931e-01
5.56668699e-01 5.18395245e-01 -6.95901990e-01 1.10788807e-01
-4.26554054e-01 2.29202926e-01 7.62955725e-01 5.33398092e-01
-9.02686417e-01 7.01879382e-01 5.47934008e+00 8.40613008e-01
-1.26554573e+00 2.44004756e-01 4.72726941e-01 -1.61529765e-01
-2.67539352e-01 -8.58942270e-02 -6.82962358e-01 1.48817271e-01
5.91503918e-01 1.60578355e-01 6.09144330e-01 7.55095422e-01
5.52937329e-01 -6.03910759e-02 -6.08047009e-01 1.18955171e+00
2.28910342e-01 -1.48231721e+00 1.67684019e-01 -1.59353703e-01
9.38488841e-01 4.65821683e-01 8.22639912e-02 2.64687538e-01
-5.13445120e-03 -1.07795715e+00 3.88451397e-01 6.96355522e-01
5.02862871e-01 -6.29877985e-01 8.84132683e-01 3.68846595e-01
-7.66984761e-01 -1.40452772e-01 -8.24753463e-01 -2.64666021e-01
4.73708004e-01 1.30329859e+00 -5.19494891e-01 5.70946693e-01
7.41474807e-01 9.10946906e-01 -5.89521565e-02 9.32689905e-01
-3.31773996e-01 9.38206851e-01 -6.27133906e-01 4.03040200e-01
6.64632440e-01 -7.46580601e-01 1.06997192e+00 8.24444056e-01
8.48973751e-01 6.47983789e-01 1.91306211e-02 9.89994645e-01
3.23244557e-02 -2.76668578e-01 -5.31613111e-01 1.90624610e-01
2.99814552e-01 1.44921935e+00 -4.36153203e-01 -2.12350097e-02
-1.91925332e-01 4.13256794e-01 -1.28845543e-01 6.99269235e-01
-7.49449730e-01 -1.12392813e-01 4.19683158e-01 2.07062528e-01
4.97275084e-01 -2.03669578e-01 -2.43171409e-01 -1.11143351e+00
9.37569421e-03 -9.95743155e-01 1.40303940e-01 -1.03588712e+00
-1.23716772e+00 9.05525208e-01 -2.96881825e-01 -1.33263814e+00
1.30134359e-01 -4.84759629e-01 -4.60750371e-01 9.32484448e-01
-1.66624343e+00 -1.15331650e+00 -6.96412444e-01 5.30616283e-01
2.66408235e-01 -3.66999269e-01 5.27881324e-01 5.19194722e-01
-6.26802623e-01 -5.18619865e-02 4.04953092e-01 -2.06249170e-02
3.40631366e-01 -6.21878266e-01 -7.30965585e-02 1.12207389e+00
9.95754823e-03 4.54718977e-01 1.02148676e+00 -6.04338586e-01
-1.49552429e+00 -1.10313594e+00 4.02160585e-01 4.15991187e-01
4.84514296e-01 2.46539727e-01 -1.13297999e+00 9.83830929e-01
1.53044596e-01 1.30372271e-01 2.15263441e-01 -2.07330510e-01
-5.14427125e-02 -5.15204787e-01 -9.51763391e-01 1.92573532e-01
7.23038554e-01 -1.79551959e-01 -7.04340160e-01 6.49129033e-01
6.21485353e-01 -4.68004674e-01 -4.85970169e-01 8.44056606e-01
2.20798224e-01 -1.13204002e+00 9.14219797e-01 -2.57815331e-01
4.66582239e-01 -3.31727386e-01 -5.49721956e-01 -1.24393177e+00
-7.21022069e-01 -5.79101205e-01 -6.73394767e-04 7.80241013e-01
3.09712201e-01 -8.46551597e-01 7.93286860e-01 1.28821015e-01
-5.75670540e-01 -7.24432290e-01 -1.17726469e+00 -6.67491198e-01
-4.32096571e-01 -1.30493745e-01 8.03564265e-02 1.15443623e+00
-6.76460803e-01 4.26674306e-01 -8.96438062e-01 8.09472799e-01
1.14592361e+00 4.20791388e-01 1.93039715e-01 -1.09285426e+00
-3.84938687e-01 2.49882080e-02 -8.04562345e-02 -1.14609098e+00
-1.76952571e-01 -9.42186832e-01 1.80959865e-01 -1.50901866e+00
-4.23167422e-02 -4.27488536e-01 -2.77109683e-01 3.70411903e-01
3.26376200e-01 3.38708401e-01 -1.56701639e-01 5.85264087e-01
-1.36006519e-01 8.29944968e-01 1.28547215e+00 -1.43332809e-01
-7.66189843e-02 1.03536285e-02 -4.82610404e-01 8.86168003e-01
5.86943746e-01 -8.54803503e-01 -1.17348284e-01 -1.08144009e+00
5.09341598e-01 4.72588778e-01 5.05955875e-01 -1.24623203e+00
4.85845089e-01 4.15867791e-02 3.67548227e-01 -6.46570802e-01
4.21308070e-01 -7.96865940e-01 5.33546746e-01 4.92865235e-01
9.91160572e-02 -2.61958361e-01 6.82342425e-03 6.23937130e-01
-6.22198045e-01 -3.75455767e-01 1.07266688e+00 -4.94383603e-01
-7.24513710e-01 3.22760969e-01 -1.69454575e-01 -1.96624041e-01
5.57157338e-01 -1.02646664e-01 -1.36072978e-01 -4.43426251e-01
-6.85546517e-01 1.26900345e-01 -2.01309174e-01 -3.40953976e-01
9.10774946e-01 -1.18343592e+00 -1.18451929e+00 4.78837729e-01
-4.66153532e-01 1.96651682e-01 6.47809207e-01 1.16340983e+00
-8.52729201e-01 3.28398108e-01 -1.89971447e-01 -6.89146042e-01
-6.66877389e-01 -5.21866307e-02 6.68848634e-01 -3.19161415e-01
-8.94538105e-01 1.15636218e+00 1.89885318e-01 -3.66231024e-01
2.50479300e-02 -1.98356256e-01 -5.24505340e-02 1.58234015e-01
3.57955098e-01 4.22214359e-01 2.44238526e-01 -4.36482191e-01
-3.32978964e-01 7.40190387e-01 1.32978093e-02 2.21863747e-01
1.95804048e+00 -1.35015726e-01 -6.04434907e-01 7.61782229e-02
1.04436123e+00 -3.19362164e-01 -1.12868202e+00 -5.11222005e-01
-4.83641475e-01 -2.37013757e-01 7.38301516e-01 -5.84348083e-01
-1.63367522e+00 7.65403509e-01 6.55602276e-01 -3.12467739e-02
1.24638700e+00 -4.96642649e-01 1.00944698e+00 9.19727981e-01
3.06140244e-01 -8.11060846e-01 -9.02338922e-02 6.22619450e-01
1.17461014e+00 -1.33876228e+00 4.86429900e-01 -2.39758957e-02
-4.27420437e-01 1.11310172e+00 3.40795726e-01 -5.67769945e-01
8.33108842e-01 1.67693153e-01 1.37497291e-01 -5.99066317e-01
-5.37456930e-01 3.02815914e-01 1.82743073e-02 1.06142528e-01
6.54905364e-02 -1.08912148e-01 -1.88715562e-01 6.76400244e-01
-4.51140068e-02 -2.87450310e-02 5.60214579e-01 6.05453908e-01
-9.11805391e-01 -3.05037145e-02 -6.74655199e-01 5.46958447e-01
-3.49104404e-01 -4.00455028e-01 4.82593894e-01 6.86499476e-01
-1.81168765e-01 6.02933586e-01 -1.55881010e-02 -1.98616628e-02
2.15242341e-01 -3.86684895e-01 2.54640579e-01 -4.29960042e-01
-4.14717346e-01 3.58625203e-01 -7.92579427e-02 -5.47894895e-01
-4.99338090e-01 -4.17272061e-01 -1.03278172e+00 -2.14718461e-01
-4.82956499e-01 2.40521193e-01 4.53884810e-01 8.21197987e-01
1.85097829e-01 6.81699753e-01 7.41971672e-01 -1.01612091e+00
-9.73688185e-01 -1.20499027e+00 -1.14584696e+00 -1.89631283e-02
6.81645989e-01 -4.15208578e-01 -9.30867732e-01 -1.33654550e-01] | [10.609759330749512, -2.1363468170166016] |
5d4d9d08-4b99-4e6d-b14f-c357722b8265 | co-creating-a-globally-interpretable-model | 2306.13381 | null | https://arxiv.org/abs/2306.13381v1 | https://arxiv.org/pdf/2306.13381v1.pdf | Co-creating a globally interpretable model with human input | We consider an aggregated human-AI collaboration aimed at generating a joint interpretable model. The model takes the form of Boolean decision rules, where human input is provided in the form of logical conditions or as partial templates. This focus on the combined construction of a model offers a different perspective on joint decision making. Previous efforts have typically focused on aggregating outcomes rather than decisions logic. We demonstrate the proposed approach through two examples and highlight the usefulness and challenges of the approach. | ['Rahul Nair'] | 2023-06-23 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 5.21484196e-01 1.04467487e+00 -1.70484483e-01 -5.47054946e-01
-3.69380832e-01 -5.97399771e-01 1.12411535e+00 2.34904423e-01
1.94535866e-01 1.07124817e+00 2.12578058e-01 -7.07628846e-01
-3.91356856e-01 -1.00141120e+00 -3.19751710e-01 -1.92195967e-01
1.74093589e-01 6.20389104e-01 -3.47012579e-02 -7.59006515e-02
4.13244404e-02 3.10846716e-01 -1.47369373e+00 6.38734519e-01
8.52773964e-01 1.02727664e+00 -4.05895203e-01 6.00242376e-01
-1.32895827e-01 1.49893320e+00 -8.33827436e-01 -6.13935590e-01
4.61499423e-01 -5.31173408e-01 -9.56132889e-01 1.82934552e-01
-1.88863859e-01 -3.70383739e-01 3.28464419e-01 8.68870318e-01
-4.72880378e-02 -2.30056077e-01 7.78825045e-01 -1.91601241e+00
-4.66645002e-01 1.06863809e+00 4.28950489e-01 -5.81380248e-01
1.02430391e+00 5.04362404e-01 1.18778265e+00 -5.35867453e-01
7.38458574e-01 1.58499694e+00 2.81291813e-01 4.37537372e-01
-1.50951624e+00 -1.82954490e-01 2.53714323e-01 3.03116232e-01
-1.41088498e+00 -4.50924516e-01 6.40434563e-01 -5.79982102e-01
1.16028666e+00 7.08321750e-01 8.93580496e-01 8.24793577e-01
3.91408652e-01 6.98956430e-01 1.47239161e+00 -8.36198568e-01
5.85329592e-01 2.96560168e-01 2.74824739e-01 3.12347770e-01
7.32555926e-01 6.03467450e-02 -4.38025296e-01 -5.73869169e-01
6.33825064e-01 -1.27567947e-01 3.20083231e-01 -4.40577269e-02
-1.29135120e+00 3.99711043e-01 3.14999670e-02 3.13462734e-01
-7.01106489e-01 2.60190368e-01 -3.39635350e-02 5.22889376e-01
-2.17096955e-02 7.76507914e-01 -3.47713828e-01 1.09750949e-01
-5.54575980e-01 8.01831543e-01 1.29513025e+00 9.48943436e-01
4.98634011e-01 2.21049003e-02 -6.04344726e-01 -2.09666088e-01
6.65477395e-01 3.19570094e-01 -4.54106539e-01 -1.26366878e+00
2.05200791e-01 1.19142377e+00 8.90146196e-01 -8.13908398e-01
8.61577019e-02 1.35142028e-01 -4.10870612e-01 6.82821631e-01
3.30072582e-01 -5.76516390e-01 -7.51219988e-01 1.46787345e+00
1.63958192e-01 -4.85021889e-01 3.41709912e-01 4.71759647e-01
6.10307813e-01 5.11464059e-01 3.61757874e-01 -2.71476388e-01
1.00744379e+00 -5.13457179e-01 -1.28385139e+00 1.69281512e-01
3.20125014e-01 -2.69459456e-01 4.58516985e-01 8.44555199e-01
-1.43912637e+00 -1.39276162e-01 -1.13268948e+00 2.17055589e-01
-4.28649426e-01 8.35839435e-02 6.80968463e-01 4.71351087e-01
-1.14265335e+00 1.16416290e-01 -4.57245082e-01 -1.40371900e-02
-8.43722094e-03 4.69306707e-01 -4.73950878e-02 -5.91133779e-04
-1.33852470e+00 1.13820314e+00 3.93122137e-01 2.33782351e-01
-7.10772634e-01 -1.67658985e-01 -8.23471725e-01 2.18142092e-01
7.61587024e-01 -9.79011774e-01 1.38111866e+00 -1.03210866e+00
-1.52689278e+00 6.08847141e-01 -1.60237253e-01 -6.33826971e-01
9.40372407e-01 6.47048280e-02 -4.72530812e-01 -2.13623524e-01
-1.59229606e-01 4.86102521e-01 3.24074954e-01 -1.32819200e+00
-6.72031581e-01 -2.25623325e-01 6.16579175e-01 -7.35672563e-02
3.77130687e-01 5.42520046e-01 2.22584561e-01 -2.95282632e-01
-1.16851345e-01 -6.64667070e-01 -4.79590803e-01 -1.36448383e-01
-6.31934822e-01 -4.31592107e-01 3.82879227e-01 -3.94084841e-01
1.87037897e+00 -1.42082596e+00 3.75607461e-01 6.32128775e-01
2.12064043e-01 -1.34295508e-01 4.30710971e-01 8.92905772e-01
1.85678378e-02 8.07286620e-01 -1.61182493e-01 -9.77674406e-03
5.85060716e-01 2.33494774e-01 -2.65009850e-01 -2.14130998e-01
6.67823792e-01 9.15197134e-01 -6.96775675e-01 -6.15017235e-01
1.13187164e-01 -6.18741177e-02 -2.52074689e-01 3.64667147e-01
-8.55496049e-01 1.60396248e-01 -7.43324280e-01 1.05389583e+00
3.03742588e-01 1.82392403e-01 9.49445069e-01 7.80517682e-02
-2.54567891e-01 4.49328452e-01 -1.53880775e+00 1.06384087e+00
-2.18891427e-01 2.00707883e-01 1.41834304e-01 -5.43062568e-01
8.48152161e-01 8.69176149e-01 1.90604895e-01 -1.10614352e-01
7.18708634e-02 1.09971359e-01 3.48909229e-01 -5.21052480e-01
1.26513526e-01 -2.14246333e-01 -5.27856231e-01 7.18181252e-01
-2.17626870e-01 -5.45330584e-01 3.18815619e-01 -1.08703040e-02
1.18876433e+00 3.16880971e-01 1.12737763e+00 -1.25595912e-01
5.68935752e-01 4.32101935e-01 8.09944332e-01 8.16586137e-01
7.07415342e-02 -1.56395193e-02 1.18266332e+00 -6.77444637e-01
-8.80333722e-01 -7.88825095e-01 7.99013451e-02 3.80061775e-01
-1.56659380e-01 -7.52409279e-01 -5.53741455e-01 -7.97442079e-01
1.12546481e-01 1.21360719e+00 -6.25402749e-01 2.17048645e-01
-3.85757923e-01 -1.61888450e-01 6.08379483e-01 4.49189037e-01
2.06882894e-01 -1.24018550e+00 -9.62514222e-01 2.68945873e-01
5.51493987e-02 -8.72099698e-01 4.89264160e-01 1.33019447e-01
-5.31291902e-01 -1.00933647e+00 1.96988031e-01 -7.91253373e-02
8.15051913e-01 -4.97080684e-01 1.15515721e+00 2.25517914e-01
1.98371243e-02 4.12723660e-01 -3.17288399e-01 -9.10294354e-01
-7.55958140e-01 -5.77445805e-01 -2.35139057e-01 1.75047383e-01
2.88108349e-01 -4.17616755e-01 -4.91595082e-02 2.65801400e-01
-1.04036641e+00 5.29607773e-01 5.35182655e-01 6.12494051e-01
1.01563394e-01 4.87108566e-02 6.49066091e-01 -1.13921869e+00
1.26158094e+00 -5.01052737e-01 -4.68733102e-01 9.31086838e-01
-7.92817354e-01 2.16014564e-01 5.18454373e-01 -2.98215859e-02
-1.40496230e+00 1.07441813e-01 5.30924141e-01 2.76586879e-02
-6.40421271e-01 7.69901395e-01 -5.28906345e-01 3.42551708e-01
5.09612441e-01 -4.23718035e-01 -1.43705383e-01 -1.66789978e-03
2.72247910e-01 6.61550462e-01 1.97193906e-01 -1.12684047e+00
4.39007074e-01 4.69270721e-02 2.60475665e-01 2.60642380e-01
1.06572412e-01 5.27387619e-01 -5.93545735e-01 -6.07274413e-01
7.72816360e-01 -4.27063942e-01 -7.92337835e-01 7.09581897e-02
-1.32698250e+00 -2.98818439e-01 -6.18668973e-01 3.94194387e-02
-6.12248123e-01 -2.27045015e-01 -1.68215558e-01 -1.58569205e+00
-1.25261545e-01 -1.28090286e+00 6.87909245e-01 -1.59185484e-01
-9.87825572e-01 -9.26652133e-01 -2.69161224e-01 1.44540593e-01
2.21548826e-01 7.04149604e-01 1.23632455e+00 -9.73375261e-01
-6.93101287e-01 -5.89799523e-01 1.29002631e-01 4.68272306e-02
2.90012300e-01 5.21835685e-01 -6.54507279e-01 4.77791518e-01
-2.52099901e-01 -2.78939456e-01 -2.91479141e-01 -1.13215238e-01
7.41989434e-01 -6.05937362e-01 -2.18933299e-01 -4.07928854e-01
1.22934318e+00 7.06152320e-01 6.50310576e-01 8.25465098e-02
-6.11745343e-02 8.81875873e-01 8.95170093e-01 7.13248789e-01
4.98644531e-01 7.85046220e-01 2.19227076e-01 1.54527560e-01
1.47216603e-01 -8.74382332e-02 1.34402469e-01 -1.03048906e-01
-5.71000040e-01 -2.85634935e-01 -1.48867571e+00 4.05210584e-01
-2.46179652e+00 -1.05014908e+00 -2.08406989e-02 1.80611479e+00
8.01460564e-01 2.36332536e-01 -4.92466278e-02 4.07798111e-01
4.82865900e-01 -2.45205253e-01 -1.41127750e-01 -1.07186604e+00
1.08363777e-01 2.43603483e-01 -1.66770950e-01 7.61614680e-01
-5.58617115e-01 6.06223166e-01 7.99634838e+00 3.30292195e-01
-6.94617152e-01 -2.60157853e-01 5.10172486e-01 -4.71412465e-02
-8.19815516e-01 5.69497108e-01 -4.43661809e-01 9.92797315e-02
6.51845455e-01 -4.88912612e-01 3.29483688e-01 5.78809857e-01
4.44449812e-01 -4.02664065e-01 -1.56312001e+00 3.27765107e-01
-2.70372033e-01 -1.31067717e+00 3.58983338e-01 1.16730034e-01
6.90638006e-01 -1.09026968e+00 -3.69670540e-01 -2.75313929e-02
8.75521064e-01 -1.24053168e+00 1.29595447e+00 1.07630122e+00
4.76798415e-01 -5.01743197e-01 7.10732579e-01 3.91991943e-01
-7.10090637e-01 -4.32600081e-01 4.98336554e-01 -8.48418415e-01
2.20345065e-01 3.92103493e-01 -1.26565564e+00 9.48561013e-01
2.02583477e-01 3.31163287e-01 -2.52519161e-01 6.75203979e-01
-6.18800282e-01 3.84181172e-01 -3.74150246e-01 -2.05793798e-01
-6.29466623e-02 -3.51904303e-01 5.04816294e-01 1.19257486e+00
1.75542906e-01 3.35030973e-01 2.22557649e-01 1.33770919e+00
5.25001824e-01 -3.54330570e-01 -8.80559027e-01 -6.88290223e-02
5.72873652e-01 9.34267700e-01 -2.94138253e-01 -7.74189532e-01
-2.75578558e-01 8.41369852e-02 3.21730673e-02 4.65663999e-01
-5.33566058e-01 -1.41364321e-01 4.09164041e-01 1.49739295e-01
-1.37131587e-01 -1.40763968e-01 -8.91653776e-01 -1.01807606e+00
2.04817683e-01 -1.22217858e+00 2.28838831e-01 -1.01657367e+00
-8.08749735e-01 4.95641649e-01 8.46944571e-01 -1.02392924e+00
-8.74100864e-01 -3.34355652e-01 -1.05445778e+00 1.07290387e+00
-5.94502270e-01 -1.38024116e+00 -2.00268388e-01 1.99529380e-01
1.52991444e-01 6.61730617e-02 1.00555229e+00 -2.95348227e-01
-6.14569306e-01 1.03935622e-01 -9.59724009e-01 -4.92537856e-01
2.29342684e-01 -1.18804371e+00 9.39691067e-03 1.00294757e+00
-4.69609380e-01 9.56733763e-01 1.02262402e+00 -9.09300625e-01
-1.49875200e+00 -6.36711657e-01 1.29036224e+00 -5.58554828e-01
4.82813478e-01 -2.82308728e-01 -3.66212785e-01 1.00117743e+00
7.34669566e-01 -4.29742038e-01 6.32717133e-01 -1.72978356e-01
1.55578017e-01 -1.48962289e-01 -1.60843766e+00 8.31606328e-01
9.42046344e-01 -3.32497865e-01 -1.00785697e+00 -7.51243383e-02
4.94472146e-01 1.23950690e-01 -8.94721091e-01 5.65064132e-01
5.44357598e-01 -8.64025235e-01 4.40331638e-01 -1.14846361e+00
3.74445021e-01 -7.88330078e-01 -2.17645943e-01 -1.04632068e+00
-2.20398590e-01 -8.90931606e-01 -1.19811632e-01 1.15397286e+00
7.71862447e-01 -5.93537688e-01 1.81836143e-01 1.77635992e+00
3.77542228e-02 -8.99571955e-01 -7.36690879e-01 -1.95273876e-01
-5.01671493e-01 -7.17273414e-01 9.68850970e-01 6.86208129e-01
5.95500171e-01 1.36781354e-02 -1.94483548e-01 -4.75074202e-02
4.28860158e-01 1.69074893e-01 7.12106824e-01 -1.40627134e+00
-2.70955086e-01 -3.33474964e-01 -4.04308766e-01 -1.35194629e-01
1.75406411e-01 -5.27451336e-01 -7.74721578e-02 -1.86306345e+00
-1.53999627e-01 -4.32865739e-01 -1.80520937e-01 9.73391294e-01
1.87621012e-01 -4.70136017e-01 4.15311635e-01 -4.35893349e-02
-5.63921750e-01 -2.11533550e-02 1.07153523e+00 -1.02265507e-01
-2.22479656e-01 -4.11224365e-02 -1.00158012e+00 8.15464139e-01
6.07312024e-01 -2.87404269e-01 -4.32253897e-01 -1.03265382e-01
8.79669487e-01 5.18343151e-01 3.87042463e-01 -7.61742115e-01
4.37720001e-01 -9.88436341e-01 -2.50292905e-02 -9.72229019e-02
1.88321009e-01 -1.03966093e+00 1.12892234e+00 3.54483038e-01
-6.47826076e-01 1.99888244e-01 -2.33328134e-01 1.22836538e-01
-4.59122986e-01 -2.84137666e-01 -8.07644054e-02 -1.36323422e-01
-4.39228714e-01 -5.58485150e-01 -8.55670810e-01 -8.29610348e-01
1.41121709e+00 -2.44712591e-01 -1.01465158e-01 -4.12224650e-01
-1.08039999e+00 4.88574445e-01 4.43340421e-01 -1.77909136e-02
6.77988291e-01 -1.22364628e+00 -5.34884691e-01 7.88688287e-02
-1.15447016e-02 1.10223547e-01 -4.48165238e-01 7.14038730e-01
-3.13538879e-01 6.51515722e-01 -5.79726756e-01 -8.10589939e-02
-1.05780196e+00 3.31639290e-01 3.74842644e-01 -5.18631399e-01
9.38798636e-02 8.05015191e-02 -1.36033610e-01 -2.84754395e-01
6.71328381e-02 -7.87789762e-01 -3.48621726e-01 7.16765225e-02
3.63711357e-01 4.86803591e-01 -1.43540382e-01 -3.64305414e-02
-4.17623580e-01 1.07583411e-01 4.16839629e-01 -6.67645037e-01
1.09681773e+00 4.11714837e-02 -5.42532206e-01 6.55694664e-01
-8.38134885e-02 1.09780245e-02 -7.34091103e-01 -9.09149274e-02
1.25248417e-01 -1.88836351e-01 -2.79932797e-01 -1.30001771e+00
-1.86444968e-01 3.21084201e-01 -2.16411471e-01 7.62454689e-01
1.05723536e+00 -1.45516753e-01 -6.62418008e-02 6.76130116e-01
7.02922165e-01 -9.06323075e-01 -6.07588053e-01 1.85782820e-01
1.46352732e+00 -7.45423079e-01 2.14954570e-01 -6.27551258e-01
-9.08668280e-01 1.25602806e+00 6.44807458e-01 1.08200833e-01
4.02595878e-01 6.54033363e-01 -7.83672109e-02 -1.21647045e-02
-1.58303916e+00 -1.42874420e-01 1.03090785e-01 4.27996695e-01
4.17389005e-01 5.59224904e-01 -7.11393416e-01 1.05309498e+00
3.38148363e-02 7.23127007e-01 6.14428103e-01 1.53555071e+00
-9.09679532e-02 -1.40970457e+00 -6.53320730e-01 4.40869063e-01
-4.15487021e-01 2.04822436e-01 -1.20400584e+00 6.11419678e-01
4.38627690e-01 1.42040777e+00 -2.80662268e-01 -5.61187565e-01
5.43207347e-01 4.76163119e-01 5.05051374e-01 -6.77473664e-01
-1.12821698e+00 -2.72050470e-01 8.38040173e-01 -5.07177293e-01
-5.27516007e-01 -8.33983839e-01 -1.06708241e+00 -1.34035915e-01
2.14798469e-02 1.67194560e-01 4.21952009e-01 1.04680181e+00
2.41681948e-01 6.11178279e-01 3.48214954e-01 -5.47008574e-01
-5.55550933e-01 -8.42559516e-01 -1.63136676e-01 2.85478920e-01
4.53748405e-02 -6.10658586e-01 -2.22598732e-01 1.47017837e-01] | [9.001288414001465, 6.625453948974609] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.