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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
43f85bf5-b0e5-40fc-b99a-27f6d614bf55 | 3d-petct-tumor-lesion-segmentation-via-gcn | 2302.12571 | null | https://arxiv.org/abs/2302.12571v1 | https://arxiv.org/pdf/2302.12571v1.pdf | 3D PETCT Tumor Lesion Segmentation via GCN Refinement | Whole-body PET/CT scan is an important tool for diagnosing various malignancies (e.g., malignant melanoma, lymphoma, or lung cancer), and accurate segmentation of tumors is a key part for subsequent treatment. In recent years, CNN-based segmentation methods have been extensively investigated. However, these methods often give inaccurate segmentation results, such as over-segmentation and under-segmentation. Therefore, to address such issues, we propose a post-processing method based on a graph convolutional neural network (GCN) to refine inaccurate segmentation parts and improve the overall segmentation accuracy. Firstly, nnUNet is used as an initial segmentation framework, and the uncertainty in the segmentation results is analyzed. Certainty and uncertainty nodes establish the nodes of a graph neural network. Each node and its 6 neighbors form an edge, and 32 nodes are randomly selected for uncertain nodes to form edges. The highly uncertain nodes are taken as the subsequent refinement targets. Secondly, the nnUNet result of the certainty nodes is used as label to form a semi-supervised graph network problem, and the uncertainty part is optimized through training the GCN network to improve the segmentation performance. This describes our proposed nnUNet-GCN segmentation framework. We perform tumor segmentation experiments on the PET/CT dataset in the MICCIA2022 autoPET challenge. Among them, 30 cases are randomly selected for testing, and the experimental results show that the false positive rate is effectively reduced with nnUNet-GCN refinement. In quantitative analysis, there is an improvement of 2.12 % on the average Dice score, 6.34 on 95 % Hausdorff Distance (HD95), and 1.72 on average symmetric surface distance (ASSD). The quantitative and qualitative evaluation results show that GCN post-processing methods can effectively improve tumor segmentation performance. | ['Yueyang Teng', 'YuDong Yao', 'Qingqing Fang', 'Hengzhi Xue'] | 2023-02-24 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [ 4.49578501e-02 3.17971230e-01 -2.99366087e-01 -3.78986239e-01
-6.91765666e-01 -6.42675385e-02 1.71125144e-01 2.89436072e-01
-5.70623219e-01 6.87998593e-01 -2.10152835e-01 -3.31062227e-01
-6.92904145e-02 -1.03310156e+00 -3.80042464e-01 -1.00226593e+00
6.65108114e-02 6.65830493e-01 5.87144792e-01 1.56113744e-01
-6.69178739e-02 4.52830642e-01 -7.19244421e-01 -3.95903587e-02
1.22027981e+00 1.23922217e+00 1.06502384e-01 2.12266684e-01
-2.78411627e-01 2.73713082e-01 -4.57836688e-01 -2.75720865e-01
1.86824296e-02 -3.47624630e-01 -8.16398919e-01 3.59454975e-02
-3.28323752e-01 -3.14136177e-01 -1.89432442e-01 1.48380888e+00
4.36146080e-01 5.89167736e-02 6.46516860e-01 -1.14491332e+00
-6.31558001e-02 7.14499056e-01 -7.71881998e-01 1.08972535e-01
-4.46166754e-01 1.58834249e-01 5.67103267e-01 -5.92324972e-01
3.91160101e-01 7.84324229e-01 7.24598229e-01 4.08845365e-01
-7.54348636e-01 -7.81747699e-01 -1.39099285e-01 -5.27641177e-03
-1.66632235e+00 1.35240376e-01 5.11394262e-01 -1.91836163e-01
3.84485990e-01 3.02358299e-01 9.28574741e-01 5.38010538e-01
6.72599912e-01 7.87805378e-01 6.51487827e-01 4.08404246e-02
2.08239704e-01 -9.46519747e-02 6.24983832e-02 7.07826614e-01
4.00363386e-01 -3.12083419e-02 3.97416770e-01 2.64954597e-01
1.04128993e+00 1.19096473e-01 -3.22355628e-01 -4.24017608e-02
-9.53043520e-01 8.82051408e-01 1.28493476e+00 4.17895645e-01
-3.59724611e-01 9.52067375e-02 4.53859299e-01 -3.47110093e-01
6.50735617e-01 7.14823306e-02 4.35281247e-02 1.63892657e-01
-1.04187584e+00 -1.76488832e-01 4.45014477e-01 6.17063284e-01
4.14060414e-01 -3.47384870e-01 -4.55411822e-01 7.63933480e-01
5.16200840e-01 1.96394339e-01 7.01357067e-01 -6.03789568e-01
1.83363706e-01 8.73943210e-01 -3.00941199e-01 -1.04892921e+00
-7.57555723e-01 -6.79330885e-01 -1.39558053e+00 -3.78313772e-02
4.62996453e-01 -8.62784907e-02 -1.35993063e+00 1.37555170e+00
5.20329952e-01 1.69295728e-01 -4.03121561e-01 1.12531853e+00
1.04273355e+00 4.38191354e-01 1.93670288e-01 -5.61711609e-01
1.25324118e+00 -1.06926823e+00 -8.08418572e-01 1.18680678e-01
7.70544827e-01 -4.37736809e-01 6.63379073e-01 1.79783463e-01
-9.33344126e-01 -2.77178407e-01 -7.24288166e-01 3.16558212e-01
-1.29433379e-01 3.31424065e-02 6.38142765e-01 5.37303925e-01
-1.00012827e+00 6.45553648e-01 -1.14764833e+00 -1.57670960e-01
8.81221533e-01 6.01962626e-01 -6.98633343e-02 -1.71987146e-01
-1.25234604e+00 5.91667593e-01 8.39473367e-01 4.06577975e-01
-7.32738376e-01 -6.57465458e-01 -7.46948004e-01 -1.19637623e-01
3.99264812e-01 -5.51046848e-01 1.14085603e+00 -7.60773480e-01
-1.34855354e+00 6.85096323e-01 1.51589155e-01 -3.28404158e-01
8.21415901e-01 6.45831227e-01 -3.89425457e-01 2.84851581e-01
2.40139887e-01 9.05062139e-01 3.56722891e-01 -9.75876868e-01
-6.64089620e-01 -4.54036444e-01 -3.75694752e-01 5.07032275e-01
-4.89306860e-02 -2.90524930e-01 -8.70864928e-01 -5.93217909e-01
6.42710567e-01 -8.22250187e-01 -5.95393956e-01 4.51324880e-02
-7.73830593e-01 -1.82741374e-01 7.98365057e-01 -7.31085837e-01
1.27998269e+00 -2.04042625e+00 -7.20213130e-02 8.04862678e-01
5.12576342e-01 1.95208937e-01 2.73614496e-01 -6.41412318e-01
-2.81927753e-02 4.76088673e-01 -6.93374693e-01 -1.41296804e-01
-4.26912904e-01 2.60215849e-01 5.77348650e-01 4.72621083e-01
-3.07419896e-02 1.19315493e+00 -8.85953903e-01 -9.50235665e-01
2.97000647e-01 3.86465251e-01 -1.56592876e-01 6.78512752e-02
-1.41529828e-01 5.90029418e-01 -7.37262011e-01 6.92600369e-01
8.06667149e-01 -3.77752692e-01 -8.44301134e-02 -2.84707934e-01
1.34349510e-01 -1.90990016e-01 -8.35244715e-01 1.35468805e+00
-2.88602877e-02 2.57651061e-01 8.08982849e-02 -7.55094528e-01
8.02727401e-01 1.68633372e-01 8.55177402e-01 -6.91715062e-01
6.81844413e-01 3.12556267e-01 3.20501745e-01 -3.60938311e-01
8.15855488e-02 -2.86199361e-01 1.68733478e-01 1.69499874e-01
-4.35615152e-01 -5.77493012e-01 9.59516764e-02 1.79825246e-01
9.27287817e-01 -3.92913014e-01 9.88795087e-02 -2.70580858e-01
5.71138442e-01 1.10954076e-01 7.70922780e-01 2.24432215e-01
-4.78599459e-01 9.75734353e-01 7.23461032e-01 -2.55154073e-01
-7.06593454e-01 -7.78366566e-01 -4.10618037e-01 2.13376120e-01
4.38197672e-01 -6.58832639e-02 -1.06499171e+00 -1.02339947e+00
-2.11502060e-01 7.17764437e-01 -7.07223177e-01 -2.91942298e-01
-3.51616144e-01 -1.09196091e+00 4.59945440e-01 6.86403811e-01
9.24112856e-01 -1.23307347e+00 -2.20959574e-01 3.26756209e-01
-2.54478961e-01 -8.81703973e-01 -4.21814799e-01 1.95370596e-02
-1.08122408e+00 -1.16023886e+00 -9.59487200e-01 -5.08067250e-01
1.01704812e+00 -3.79591323e-02 7.42571175e-01 4.60533321e-01
-1.79822564e-01 -2.25227416e-01 -8.61383155e-02 -2.69672096e-01
-4.07997727e-01 5.33814095e-02 -3.38256836e-01 -1.82393402e-01
1.51451960e-01 -1.88319072e-01 -7.24042654e-01 6.18349075e-01
-1.00470841e+00 1.97311863e-01 6.46234691e-01 8.71534824e-01
1.29203129e+00 5.43132007e-01 2.34182000e-01 -9.16684508e-01
5.11604130e-01 -3.50975901e-01 -5.95129013e-01 1.17584489e-01
-6.07232928e-01 -4.11378026e-01 5.43499410e-01 -2.31581017e-01
-7.96503305e-01 3.20196748e-02 -3.75993997e-01 -3.95521253e-01
-3.66961397e-03 8.02156270e-01 -1.78049952e-01 -2.97088176e-01
3.58927101e-01 -1.43272445e-01 2.90392935e-01 1.70297578e-01
-1.69956744e-01 5.55994034e-01 2.27092981e-01 -9.08373818e-02
3.93623292e-01 3.77789199e-01 2.21867889e-01 -3.87035161e-01
-6.69923425e-01 -2.43692383e-01 -3.79952252e-01 -3.22509557e-01
1.14815199e+00 -5.23427367e-01 -5.35697639e-01 6.77416265e-01
-7.80574262e-01 -4.76018429e-01 -3.30544114e-01 6.82992101e-01
-2.16332287e-01 3.35862398e-01 -7.03620791e-01 -3.03473681e-01
-7.04649687e-01 -1.69099975e+00 9.23583090e-01 8.43571901e-01
-6.77960366e-02 -1.11749208e+00 -3.28232527e-01 2.30580598e-01
3.28508496e-01 4.31161255e-01 7.88800716e-01 -7.38994241e-01
-3.66349310e-01 -2.91148335e-01 -6.06786847e-01 3.08234900e-01
1.91806197e-01 2.87974954e-01 -5.16962767e-01 -2.63042361e-01
-6.23925328e-02 -4.28043827e-02 7.94060290e-01 1.00389850e+00
1.75524485e+00 1.06939599e-01 -7.09143102e-01 8.90211582e-01
1.28052437e+00 3.82245749e-01 6.71647549e-01 1.55613750e-01
9.49096084e-01 2.72225231e-01 6.51229441e-01 1.65225074e-01
1.82564586e-01 2.39454195e-01 8.07399929e-01 -3.20308656e-01
-1.31502420e-01 1.34619130e-02 -3.55993301e-01 7.37865865e-01
-1.18603399e-02 -4.32334989e-01 -1.10525250e+00 4.05259430e-01
-1.70942688e+00 -3.31740201e-01 -4.66347516e-01 1.96982622e+00
7.65706480e-01 3.25879395e-01 -2.37530902e-01 1.26818232e-02
1.15421844e+00 -9.38058197e-02 -7.53210902e-01 1.81322936e-02
2.52627075e-01 2.58845389e-01 5.01051664e-01 3.61904144e-01
-1.06483519e+00 7.64251113e-01 5.00736523e+00 1.24846542e+00
-1.14080167e+00 -2.99626011e-02 1.32029319e+00 3.13626647e-01
-2.30489045e-01 -2.31982902e-01 -5.15830398e-01 6.04968011e-01
3.96842241e-01 -9.79645737e-03 8.91608447e-02 6.25384808e-01
3.42676699e-01 -4.77886736e-01 -6.99728310e-01 8.90507102e-01
-2.00745955e-01 -1.14956045e+00 -1.10590104e-02 2.07700014e-01
8.03396523e-01 -2.56094709e-02 -8.12141076e-02 1.69994846e-01
1.73352107e-01 -1.30385375e+00 1.45652041e-01 5.17834067e-01
9.08704877e-01 -1.04706168e+00 1.40777111e+00 4.68812287e-01
-1.14649963e+00 3.70964944e-01 -4.88307446e-01 6.32112205e-01
2.12671593e-01 1.03121126e+00 -1.00180888e+00 8.18563581e-01
4.83089030e-01 6.21700168e-01 -5.65838575e-01 1.38295031e+00
-2.68894613e-01 6.00533068e-01 -7.09088922e-01 -1.68215826e-01
3.85517627e-01 -3.32073092e-01 4.38849598e-01 7.89450169e-01
3.76114756e-01 4.51906711e-01 2.19038665e-01 9.66804862e-01
-3.16392750e-01 9.50217620e-02 1.61405832e-01 1.26524180e-01
2.37977073e-01 1.57903683e+00 -1.48846018e+00 -3.90632272e-01
1.80994645e-01 6.57424808e-01 4.36086729e-02 1.30675882e-01
-1.16871893e+00 -3.28684419e-01 -5.05489185e-02 -3.42896581e-03
-2.03651235e-01 3.22596461e-01 -5.74541569e-01 -7.92497754e-01
-2.80085146e-01 -5.43901801e-01 6.60460651e-01 -7.90066242e-01
-1.17474604e+00 6.31418467e-01 -1.31117418e-01 -9.63702202e-01
1.95418000e-01 -3.39746803e-01 -9.72491801e-01 8.53492618e-01
-1.13582420e+00 -7.26384819e-01 -8.06406438e-01 3.66826564e-01
3.74684840e-01 2.12546721e-01 4.79463160e-01 1.38370797e-01
-9.13579881e-01 6.25916600e-01 -2.42661715e-01 3.86541039e-01
2.88960665e-01 -1.13607442e+00 -6.49272352e-02 7.45579839e-01
-5.10617018e-01 1.63691461e-01 1.05613098e-01 -1.09268618e+00
-6.40896559e-01 -1.55849767e+00 3.43128383e-01 1.17099576e-01
4.27042335e-01 1.65184721e-01 -1.04441559e+00 3.57808083e-01
-1.92353398e-01 4.67582554e-01 3.86721909e-01 -5.75694740e-01
5.76808751e-01 5.21309040e-02 -1.57738256e+00 6.21858537e-01
6.18224084e-01 3.93081158e-02 -1.50640905e-01 4.62055087e-01
8.82320881e-01 -9.41537440e-01 -1.13049579e+00 9.14520502e-01
3.16760391e-01 -8.83691788e-01 7.20549285e-01 -3.31899454e-03
3.09644580e-01 -4.34917569e-01 3.23361129e-01 -1.33772850e+00
-2.71841168e-01 2.62754317e-02 3.12859625e-01 9.63047028e-01
5.12119055e-01 -6.07756317e-01 1.02710748e+00 8.60436618e-01
-5.32612503e-01 -1.29054952e+00 -9.72338021e-01 -4.60115731e-01
1.38772605e-03 -4.76863712e-01 6.95847094e-01 8.58310044e-01
-4.04356331e-01 -1.23224579e-01 4.85632867e-01 4.18740474e-02
5.76892436e-01 -3.49847525e-01 2.20523015e-01 -9.92014766e-01
1.80926993e-01 -8.65232229e-01 -4.00051862e-01 -7.01528847e-01
-1.81649029e-01 -1.02339578e+00 1.43416718e-01 -1.90573657e+00
1.80950239e-01 -6.59948766e-01 -4.09204662e-01 4.38872457e-01
-3.58969301e-01 1.71204656e-01 -2.71872401e-01 1.15413800e-01
-4.85467881e-01 6.01793110e-01 1.83202589e+00 -3.83850306e-01
-3.04613590e-01 4.12918597e-01 -3.64635825e-01 8.72431397e-01
9.38615263e-01 -4.17448908e-01 -3.31351697e-01 -9.53902677e-02
-1.27018020e-01 2.17355043e-01 2.38374531e-01 -9.56008613e-01
2.81156689e-01 -1.28850371e-01 7.22464025e-01 -8.85865271e-01
-8.14622343e-02 -9.33272958e-01 1.88705906e-01 6.84624195e-01
-3.59048471e-02 -3.80272627e-01 2.17768446e-01 3.85152668e-01
-1.62437424e-01 -2.38315433e-01 9.99665201e-01 -1.38150185e-01
-2.65329689e-01 9.81896520e-01 -8.91990587e-02 2.60674693e-02
1.29649854e+00 -3.55363131e-01 -9.59977433e-02 -4.55346368e-02
-8.23264778e-01 7.01046288e-01 2.26447314e-01 -1.77612782e-01
7.61465311e-01 -1.38885510e+00 -5.21417201e-01 1.23618208e-01
-5.17984331e-02 8.14161241e-01 4.13201600e-01 1.28420508e+00
-8.42098415e-01 1.52521329e-02 -9.23687592e-03 -9.44537103e-01
-9.02978122e-01 1.98884681e-01 8.19291174e-01 -6.41128421e-01
-5.11501491e-01 1.14781237e+00 3.05531412e-01 -2.70939082e-01
1.44597322e-01 -6.79598510e-01 -4.25821006e-01 -1.95549026e-01
1.70748234e-01 3.36189300e-01 2.82318115e-01 -5.11219442e-01
-3.78748834e-01 3.79670322e-01 -1.95152402e-01 2.12374385e-02
9.80478704e-01 7.24238902e-02 -4.60602969e-01 -3.05222590e-02
1.11353350e+00 -4.70260501e-01 -1.12848580e+00 -1.00831762e-01
-3.25365931e-01 -7.88486898e-02 4.13983762e-01 -8.18636298e-01
-1.58638191e+00 8.46180856e-01 6.67712212e-01 2.09648505e-01
1.20397651e+00 -6.39856979e-02 1.02999699e+00 -1.47045121e-01
8.85053305e-04 -9.99524832e-01 -1.84580728e-01 3.57720941e-01
6.79407954e-01 -1.19409025e+00 4.91643995e-02 -6.78794980e-01
-5.85094810e-01 1.16647565e+00 9.74527895e-01 -6.11850470e-02
6.59994721e-01 1.91153675e-01 -8.90671089e-02 -3.66497666e-01
-7.89320935e-03 -1.49819274e-02 2.72754490e-01 2.83118427e-01
1.04018420e-01 4.59802061e-01 -4.41393942e-01 9.45873439e-01
-1.43438995e-01 -7.85008371e-02 2.16681376e-01 5.34317732e-01
-4.34191883e-01 -6.34277284e-01 -2.55053163e-01 8.21495116e-01
-3.83739322e-01 9.10599716e-03 -2.34389380e-01 1.03604615e+00
1.84535444e-01 6.06071115e-01 1.31917819e-01 -4.35786813e-01
1.33418590e-01 -3.43472958e-01 1.79267794e-01 -5.55892467e-01
-6.01960421e-01 3.41758221e-01 -2.28795975e-01 -4.41626072e-01
-1.71632752e-01 -3.24926645e-01 -1.90883410e+00 -1.12005673e-01
-6.91726744e-01 3.39907140e-01 6.33650780e-01 1.04159737e+00
-2.10227087e-01 9.38538194e-01 6.29091442e-01 -4.30289537e-01
-2.42075831e-01 -8.76540244e-01 -5.72112679e-01 1.45350412e-01
-1.80545449e-01 -6.11260295e-01 -2.84212202e-01 -5.08210063e-01] | [14.588037490844727, -2.4886281490325928] |
b775f4c2-af67-4bf6-8a9f-ee66858e052c | consistent-classification-of-translation-1 | null | null | https://aclanthology.org/W17-0807 | https://aclanthology.org/W17-0807.pdf | Consistent Classification of Translation Revisions: A Case Study of English-Japanese Student Translations | Consistency is a crucial requirement in text annotation. It is especially important in educational applications, as lack of consistency directly affects learners{'} motivation and learning performance. This paper presents a quality assessment scheme for English-to-Japanese translations produced by learner translators at university. We constructed a revision typology and a decision tree manually through an application of the OntoNotes method, i.e., an iteration of assessing learners{'} translations and hypothesizing the conditions for consistent decision making, as well as re-organizing the typology. Intrinsic evaluation of the created scheme confirmed its potential contribution to the consistent classification of identified erroneous text spans, achieving visibly higher Cohen{'}s kappa values, up to 0.831, than previous work. This paper also describes an application of our scheme to an English-to-Japanese translation exercise course for undergraduate students at a university in Japan. | ['Atsushi Fujita', 'Anthony Hartley', 'Kyo Kageura', 'Kikuko Tanabe', 'Mayuka Yamamoto', 'Chiho Toyoshima'] | 2017-04-01 | null | null | null | ws-2017-4 | ['text-annotation'] | ['natural-language-processing'] | [ 1.22059703e-01 2.73370653e-01 -2.11643368e-01 -3.39272708e-01
-1.21575999e+00 -8.11593950e-01 2.65183568e-01 5.65953493e-01
-7.69157529e-01 1.04493475e+00 3.31011593e-01 -9.19808745e-01
-4.14133549e-01 -5.22781014e-01 -6.15488768e-01 -1.10889457e-01
9.87550974e-01 6.24238968e-01 2.77553797e-01 -3.55189294e-01
7.14158714e-01 2.83081263e-01 -1.65954006e+00 3.07080239e-01
1.67488742e+00 2.50707775e-01 3.93256009e-01 4.49766397e-01
-3.31976086e-01 5.02777278e-01 -8.35801065e-01 -8.38191152e-01
-3.14330518e-01 -5.75817943e-01 -1.19176316e+00 1.80691965e-02
6.97658181e-01 -7.04117119e-02 3.33309621e-01 1.34157538e+00
4.74802017e-01 2.24962354e-01 4.98408616e-01 -7.70892560e-01
-1.00865030e+00 1.04908907e+00 1.92093015e-01 1.71107680e-01
6.36505902e-01 -4.61482912e-01 8.53188455e-01 -1.08488905e+00
5.70318460e-01 6.63412273e-01 8.69098246e-01 4.38512921e-01
-1.05300558e+00 -7.91834593e-01 -3.86697017e-02 9.85468701e-02
-1.24754179e+00 -3.75234455e-01 2.88005233e-01 -7.70285606e-01
7.11712301e-01 3.55167568e-01 9.68341529e-01 8.63934457e-01
3.42568725e-01 2.40796342e-01 1.63557959e+00 -9.09772217e-01
9.21496525e-02 9.24258232e-01 1.94764793e-01 5.11454821e-01
6.64443195e-01 -3.98150682e-01 -6.80709243e-01 2.06638962e-01
4.49749470e-01 -4.35022146e-01 -1.76222697e-01 2.08133996e-01
-1.25733054e+00 3.93153876e-01 -3.04293185e-01 4.13456947e-01
-3.14702429e-02 -3.72267514e-01 3.36621970e-01 6.86741769e-01
3.02836776e-01 5.18934965e-01 -4.86651659e-01 -7.18967795e-01
-8.84840310e-01 1.78366110e-01 8.50056827e-01 1.42061949e+00
1.69689864e-01 -1.62522048e-01 2.08732449e-02 1.09957731e+00
4.87109154e-01 4.55258846e-01 9.23445344e-01 -7.83813655e-01
4.10988688e-01 7.63437808e-01 8.36211741e-02 -3.68144810e-01
-1.07793234e-01 -5.25650799e-01 -2.68591970e-01 1.77396744e-01
6.20840430e-01 -1.61014766e-01 -5.95968306e-01 1.35163963e+00
1.24151908e-01 -9.74352896e-01 1.91508621e-01 6.36641026e-01
1.06307054e+00 1.52529746e-01 2.58668423e-01 -6.24070406e-01
1.57063103e+00 -7.81646132e-01 -1.07072675e+00 4.05396104e-01
9.52431262e-01 -1.50152969e+00 1.53439033e+00 6.60284758e-01
-1.39312768e+00 -7.25832820e-01 -8.06429148e-01 -1.96291149e-01
-2.74269015e-01 2.68902838e-01 7.97141567e-02 1.18783593e+00
-1.09946597e+00 5.61817586e-01 -3.66355598e-01 -5.79995930e-01
-3.45190279e-02 3.89808148e-01 -3.42392892e-01 3.48172665e-01
-1.03908253e+00 1.03028774e+00 1.36112526e-01 -1.59877151e-01
-1.78091511e-01 -6.17218077e-01 -4.92704093e-01 -1.44785121e-01
6.09189123e-02 -4.55855161e-01 1.68348837e+00 -1.09558022e+00
-1.63692749e+00 9.88477945e-01 -7.93524683e-02 3.25790316e-01
5.91596901e-01 -2.13621423e-01 -5.88045537e-01 -2.17156589e-01
4.58906442e-01 2.55346596e-01 8.66376758e-02 -1.16478980e+00
-1.13063562e+00 -3.54633033e-01 -1.20023079e-01 6.40398800e-01
-7.66036928e-01 3.06712598e-01 -2.76496083e-01 -8.83201778e-01
6.04267657e-01 -1.06894588e+00 3.23837787e-01 -3.13334286e-01
9.41324309e-02 -6.08655751e-01 1.76309049e-01 -1.01250219e+00
1.73489368e+00 -1.63244009e+00 -1.30912781e-01 3.41725051e-01
2.17059478e-01 4.78327535e-02 3.80305588e-01 6.33552194e-01
1.10644676e-01 5.39986968e-01 4.41519618e-02 3.92019786e-02
4.58232537e-02 -2.85353791e-02 -1.84704233e-02 1.08896859e-01
-3.70335907e-01 3.92373770e-01 -1.12038600e+00 -7.06806898e-01
-2.44245768e-01 2.36099437e-02 -2.79364616e-01 -2.30171699e-02
3.25446248e-01 1.02498062e-01 -1.53881833e-01 5.60425162e-01
1.46323126e-02 1.85274139e-01 3.53043199e-01 3.77444774e-01
-7.43626356e-01 9.52780485e-01 -1.02684152e+00 1.51595020e+00
-5.24948180e-01 7.94739604e-01 -4.41955805e-01 -3.55502933e-01
1.19983613e+00 7.53986299e-01 1.91706240e-01 -7.80575454e-01
8.13622847e-02 7.58674145e-01 1.89780459e-01 -7.06526101e-01
1.09844327e+00 -2.86501914e-01 -8.86952784e-03 6.52350366e-01
1.56759039e-01 -3.24640006e-01 3.51321191e-01 1.56234652e-01
4.84067410e-01 3.70252103e-01 3.42769206e-01 -8.61740530e-01
3.38973582e-01 4.29501891e-01 3.86916250e-01 6.28466427e-01
-1.21727593e-01 3.35581452e-01 1.92088977e-01 -1.42462328e-01
-1.15750599e+00 -7.80805826e-01 -4.93697613e-01 9.92636919e-01
-1.60755470e-01 -5.57719707e-01 -9.43549514e-01 -8.32766533e-01
-3.61185789e-01 1.05551302e+00 -3.21579009e-01 -2.07805261e-02
-3.34762931e-01 -2.44082764e-01 5.33293843e-01 3.98724586e-01
2.78003633e-01 -8.92883778e-01 -5.16501069e-01 3.76684666e-01
-6.22257769e-01 -6.17889762e-01 -4.70074415e-01 2.52368450e-01
-1.01127994e+00 -6.48767829e-01 -4.21920866e-01 -1.22186589e+00
7.71746695e-01 3.15986961e-01 9.90851402e-01 3.57479542e-01
4.76370692e-01 4.43194628e-01 -5.80585778e-01 -6.77829862e-01
-8.49473894e-01 2.62357026e-01 2.96277165e-01 -1.00131679e+00
6.64936662e-01 -3.39804947e-01 -7.52807502e-03 4.67119277e-01
-8.42082024e-01 1.51942313e-01 6.48761153e-01 8.10581923e-01
4.66295600e-01 8.93429518e-02 9.16967034e-01 -1.08891821e+00
1.21205676e+00 -1.74969845e-02 -5.04681468e-01 6.32445931e-01
-1.31999969e+00 -2.40515038e-01 3.95401955e-01 -4.45463240e-01
-1.21016550e+00 -3.52089465e-01 -3.17477763e-01 3.40093762e-01
-1.56363979e-01 6.05950177e-01 -1.14850350e-01 -3.47203553e-01
8.43850791e-01 6.68440536e-02 -1.22787014e-01 -2.27293164e-01
-2.04146415e-01 1.04903555e+00 3.54836136e-01 -9.80897188e-01
6.96536720e-01 -5.11725605e-01 -4.67794865e-01 -7.42734313e-01
-5.96104205e-01 -4.72074568e-01 -9.96574938e-01 -5.29653728e-01
5.58799148e-01 -9.21640038e-01 -6.56567156e-01 1.20496646e-01
-8.84163618e-01 -2.68794417e-01 -3.21878999e-01 1.05364347e+00
-3.43730003e-01 2.25902766e-01 -5.54894924e-01 -5.21334052e-01
-1.16390184e-01 -1.32417202e+00 3.96458775e-01 3.94522995e-01
-8.09419334e-01 -1.15100527e+00 1.65912449e-01 8.96496356e-01
2.36244388e-02 -4.48209256e-01 1.06141579e+00 -9.25170302e-01
9.30016339e-02 7.62672722e-02 2.19547838e-01 1.65544495e-01
6.27059191e-02 1.58254638e-01 -7.74972975e-01 -3.73662934e-02
7.85118341e-02 -3.03922415e-01 9.08082649e-02 -1.81628257e-01
6.18563831e-01 -5.54115593e-01 1.33207738e-01 -7.18876645e-02
1.20356953e+00 5.10088742e-01 4.05769050e-01 7.85703003e-01
5.79312801e-01 8.28362286e-01 7.19839633e-01 -1.45088919e-02
6.31574750e-01 7.37910867e-01 -4.32960838e-01 3.57514054e-01
-3.77130061e-01 -3.54216874e-01 5.45167863e-01 1.84671342e+00
-2.51567036e-01 -7.82821327e-02 -1.12396753e+00 6.03378475e-01
-1.42455089e+00 -6.70111179e-01 -5.99373877e-01 2.44765115e+00
1.33437872e+00 2.47248128e-01 2.08510518e-01 2.45674118e-01
6.23524368e-01 -7.81372428e-01 2.91220456e-01 -9.61619198e-01
1.78767100e-01 1.96729481e-01 3.07376206e-01 7.52119958e-01
-3.47511470e-01 1.02377403e+00 6.26859188e+00 5.82462549e-01
-7.57052243e-01 8.84201080e-02 1.01159133e-01 3.67133647e-01
-8.65810871e-01 1.54088974e-01 -1.00313199e+00 3.20903301e-01
1.02751362e+00 -5.20770729e-01 5.94842993e-02 5.15374124e-01
2.49826863e-01 -8.17587003e-02 -7.51158595e-01 2.89541900e-01
2.05866322e-01 -9.58350003e-01 8.83428976e-02 1.10408366e-01
9.83692169e-01 -6.58907652e-01 -2.37169728e-01 3.34582090e-01
5.66788971e-01 -6.91690981e-01 1.18896604e+00 2.81372994e-01
8.35715055e-01 -6.22780263e-01 8.46827745e-01 1.75474867e-01
-6.04893863e-01 3.36794823e-01 -1.81300640e-01 -3.72601271e-01
-2.57874370e-01 3.16282392e-01 -1.21569192e+00 4.42646712e-01
6.99720979e-01 2.01406822e-01 -7.41386235e-01 9.75270748e-01
-5.95993876e-01 9.10314858e-01 2.47127101e-01 -4.71663594e-01
-5.10210358e-02 -4.27936584e-01 3.97539228e-01 1.28336477e+00
6.66494548e-01 2.64410466e-01 -4.34692092e-02 4.17713672e-01
-1.09638922e-01 6.79942667e-01 -3.92749518e-01 7.63118640e-02
9.74394262e-01 9.47729290e-01 -9.07035291e-01 -4.99323487e-01
-4.62347299e-01 8.53912532e-01 2.05712110e-01 2.83465326e-01
-3.96549493e-01 -5.69412172e-01 1.12157062e-01 9.16531906e-02
-1.94386154e-01 -1.36298046e-01 -9.14626062e-01 -8.24678779e-01
2.29259610e-01 -1.15913880e+00 1.81040376e-01 -4.99506623e-01
-9.02180076e-01 4.34448153e-01 -4.68386821e-02 -1.50538898e+00
2.92948753e-01 -4.27336782e-01 -3.30192745e-01 1.17434168e+00
-1.02872741e+00 -8.41596246e-01 -2.56197870e-01 1.54551685e-01
7.33547091e-01 -5.67011125e-02 9.05852079e-01 3.01109731e-01
-5.85311055e-01 9.62615490e-01 7.48696476e-02 -1.89710706e-01
1.12062776e+00 -1.58180475e+00 -4.76248078e-02 9.52069402e-01
6.38797060e-02 1.00880492e+00 7.06580341e-01 -1.03535664e+00
-1.09680295e+00 -7.68655300e-01 1.89220524e+00 -6.46769643e-01
7.93382645e-01 4.47853468e-02 -1.01127148e+00 5.43400645e-01
4.16704118e-01 -9.28671062e-01 1.31450915e+00 2.35604495e-01
4.64262255e-02 -6.49607927e-02 -8.88936698e-01 8.62380505e-01
1.04953897e+00 -5.30386090e-01 -1.17222571e+00 3.59851748e-01
6.20482922e-01 -7.01644957e-01 -1.39816427e+00 3.21698301e-02
8.82323027e-01 -6.75891161e-01 3.71243745e-01 -4.14017051e-01
6.68164790e-01 -3.22623581e-01 9.31120142e-02 -1.31735992e+00
-3.96189719e-01 -5.33056319e-01 7.29418337e-01 1.50693595e+00
8.53710294e-01 -3.91200721e-01 4.00300860e-01 7.99855411e-01
-8.81071806e-01 -5.31006455e-01 -8.05505872e-01 -5.57421207e-01
4.28838044e-01 -2.41632134e-01 3.00610095e-01 1.30935955e+00
5.51407874e-01 2.35764593e-01 7.75650665e-02 -2.68961675e-02
4.04566526e-01 -3.71629804e-01 5.75997829e-01 -1.44675672e+00
2.48898596e-01 -7.63790250e-01 -9.18614939e-02 -6.36766016e-01
3.10530560e-03 -1.02051985e+00 1.48021445e-01 -1.68184686e+00
-1.76609668e-03 -6.24323905e-01 2.34602895e-02 4.68953371e-01
-3.82355034e-01 2.45409101e-01 -9.11988020e-02 4.64339793e-01
-3.58572274e-01 1.19922698e-01 1.31310093e+00 3.24611217e-01
-4.72751111e-01 5.71231954e-02 -1.09260714e+00 5.85599780e-01
8.03080678e-01 -6.22174740e-01 -4.64783341e-01 -5.06169140e-01
6.54746532e-01 -5.72735779e-02 -1.02456152e-01 -9.24893498e-01
4.51693714e-01 -5.36346555e-01 4.65468615e-01 -2.01444775e-01
-4.64576006e-01 -8.43910575e-01 2.78969705e-01 4.12817270e-01
-6.08645439e-01 8.10450494e-01 3.05689424e-01 -2.21241564e-01
-2.04314858e-01 -6.85245633e-01 4.53537583e-01 -1.88873827e-01
-5.36237121e-01 -5.02047598e-01 -7.89703846e-01 -1.88875142e-02
9.23046350e-01 -5.21794081e-01 -4.05309796e-01 -1.48672879e-01
-5.06420195e-01 2.47444138e-02 7.72970259e-01 4.41078573e-01
5.13783753e-01 -1.31072211e+00 -5.94920576e-01 6.73942268e-02
3.45576882e-01 -2.41899878e-01 -4.38972525e-02 1.09623861e+00
-7.53850043e-01 4.19492960e-01 -4.27268237e-01 -2.28805959e-01
-1.51468611e+00 -1.32277876e-01 -1.45099014e-01 1.23360492e-02
-5.63795388e-01 6.54809773e-01 -5.05099714e-01 -5.54280877e-01
3.50001603e-01 -4.53616858e-01 -5.29039025e-01 2.38003701e-01
2.69929051e-01 6.76880896e-01 5.19459546e-01 -5.92844069e-01
1.63077260e-03 4.43070740e-01 -7.17042089e-02 -6.95009589e-01
6.89611554e-01 -4.52714920e-01 -1.61833718e-01 9.72853839e-01
4.91981894e-01 7.40521848e-01 -4.58948612e-01 -1.81515560e-01
3.09672028e-01 -4.42844212e-01 -1.12584397e-01 -1.30809283e+00
-7.78401271e-02 5.62852442e-01 3.49440455e-01 9.73390639e-02
8.23106706e-01 -4.52626139e-01 4.05245036e-01 5.40996850e-01
1.76453277e-01 -1.64165175e+00 -2.22444385e-01 6.45962179e-01
6.87296867e-01 -1.22333646e+00 7.36380294e-02 -3.07255298e-01
-7.18007565e-01 1.49109507e+00 6.00495994e-01 7.40059793e-01
2.13547215e-01 -1.35685235e-01 5.01792550e-01 9.02371630e-02
-6.01376235e-01 3.01451504e-01 6.47197306e-01 2.65294135e-01
1.24980342e+00 3.47847223e-01 -1.34770894e+00 6.51472747e-01
-7.65992880e-01 5.21386676e-02 9.95437384e-01 1.03823304e+00
-7.85434604e-01 -1.38258803e+00 -5.79278350e-01 3.70935798e-01
-5.10880888e-01 -2.57774621e-01 -6.39922798e-01 8.12292397e-01
1.93027854e-01 1.04811966e+00 -1.01125509e-01 -4.49961096e-01
4.80281323e-01 4.20499384e-01 4.57744956e-01 -8.37994516e-01
-1.20998907e+00 9.51419249e-02 3.62161726e-01 2.33829454e-01
-3.79320741e-01 -8.65980744e-01 -1.12460041e+00 -4.35806870e-01
-4.81429607e-01 7.71277905e-01 9.56366479e-01 1.03717959e+00
-1.58417732e-01 5.92314661e-01 2.59175241e-01 1.71740294e-01
-5.61495006e-01 -1.19998467e+00 -2.91313082e-01 3.48263949e-01
-1.54085204e-01 -2.91237295e-01 -1.30812213e-01 2.51000285e-01] | [11.263469696044922, 9.555848121643066] |
6baa326e-8f21-4dd3-bec7-7fa43bdf37e4 | revisiting-ipa-based-cross-lingual-text-to | 2110.07187 | null | https://arxiv.org/abs/2110.07187v2 | https://arxiv.org/pdf/2110.07187v2.pdf | Revisiting IPA-based Cross-lingual Text-to-speech | International Phonetic Alphabet (IPA) has been widely used in cross-lingual text-to-speech (TTS) to achieve cross-lingual voice cloning (CL VC). However, IPA itself has been understudied in cross-lingual TTS. In this paper, we report some empirical findings of building a cross-lingual TTS model using IPA as inputs. Experiments show that the way to process the IPA and suprasegmental sequence has a negligible impact on the CL VC performance. Furthermore, we find that using a dataset including one speaker per language to build an IPA-based TTS system would fail CL VC since the language-unique IPA and tone/stress symbols could leak the speaker information. In addition, we experiment with different combinations of speakers in the training dataset to further investigate the effect of the number of speakers on the CL VC performance. | ['Xinyuan Yu', 'Yang Zhang', 'Haoyue Zhan', 'Yue Lin', 'Haitong Zhang'] | 2021-10-14 | null | null | null | null | ['voice-cloning'] | ['speech'] | [-3.36371422e-01 -4.40943569e-01 -1.87971756e-01 -3.79581720e-01
-1.32538664e+00 -8.47354352e-01 2.91886747e-01 -4.07441914e-01
-2.66196996e-01 2.88896769e-01 4.12971973e-01 -9.33305144e-01
4.81282711e-01 -2.33886242e-01 -7.34294116e-01 -4.61654752e-01
3.15518051e-01 2.30106071e-01 1.78441092e-01 -3.96423507e-03
-1.71409845e-02 2.25775287e-01 -1.38636267e+00 2.85454154e-01
8.03636551e-01 5.83482563e-01 2.62723774e-01 6.51649952e-01
-3.42617244e-01 4.57758725e-01 -9.71897900e-01 -5.89023471e-01
5.57791770e-01 -7.14268804e-01 -5.79698384e-01 -3.30270380e-01
6.60942614e-01 1.77335441e-01 6.21212013e-02 6.90268219e-01
7.25229621e-01 5.83571866e-02 7.19285548e-01 -9.10174251e-01
-4.55397964e-01 1.13478613e+00 -2.38856435e-01 1.28957674e-01
3.54468793e-01 2.96151787e-01 1.20768178e+00 -8.06647718e-01
5.51318526e-01 1.63968205e+00 6.80322707e-01 4.38969254e-01
-1.36740172e+00 -9.47704256e-01 -1.11132540e-01 -8.64339843e-02
-1.51963985e+00 -8.49867821e-01 7.67527103e-01 -4.30288702e-01
6.89039171e-01 4.28213507e-01 4.25206542e-01 1.11127424e+00
1.11523382e-01 8.07979286e-01 1.46897101e+00 -8.37722182e-01
-8.26107562e-02 4.31400806e-01 -6.33910531e-05 2.73032606e-01
-2.96982825e-01 2.78704286e-01 -8.18613589e-01 8.00991356e-02
4.11265731e-01 -9.88804162e-01 -5.79749458e-02 3.84684503e-01
-7.32478142e-01 8.66978168e-01 -8.12477618e-02 5.67077637e-01
8.43046606e-02 2.30271406e-02 8.74822557e-01 5.64937234e-01
2.34202340e-01 4.64849919e-01 -7.52118051e-01 -3.39152396e-01
-9.40769196e-01 -1.30250975e-01 7.99144149e-01 7.54180074e-01
3.62194955e-01 5.89470923e-01 -1.00288540e-01 1.28404868e+00
3.21982175e-01 7.49075174e-01 7.61289537e-01 -9.05463874e-01
6.11107588e-01 -2.09579125e-01 -2.61103302e-01 -4.15040702e-01
1.77916661e-01 -3.11085671e-01 -8.47956538e-02 -2.24106357e-01
6.77774370e-01 -2.36573413e-01 -9.12651181e-01 1.95534348e+00
-1.35444179e-02 -1.09823301e-01 7.44881406e-02 4.96110916e-01
5.68930686e-01 9.05084729e-01 2.58523315e-01 -3.99690241e-01
1.36583090e+00 -8.06481719e-01 -9.46953535e-01 -1.38575315e-01
8.00948322e-01 -1.38491845e+00 2.00705171e+00 2.99178421e-01
-1.06338549e+00 -9.08997416e-01 -6.45341516e-01 -1.62524462e-01
-4.47981596e-01 2.18349993e-01 2.55364358e-01 1.25496864e+00
-7.67962635e-01 1.66759387e-01 -5.42848289e-01 -1.45551926e-02
-1.42045110e-01 1.20787203e-01 1.03951562e-02 1.56433240e-01
-1.51969230e+00 8.92169237e-01 1.09032519e-01 -2.43801519e-01
-6.38314724e-01 -7.67474353e-01 -7.39622116e-01 -2.15414166e-01
1.42661810e-01 5.20391390e-03 1.35193896e+00 -8.93559873e-01
-1.90675807e+00 7.03067183e-01 -5.58945715e-01 -1.19533733e-01
4.58306015e-01 -2.65341640e-01 -7.13587105e-01 -3.63319874e-01
3.08595747e-01 5.54906666e-01 7.37272561e-01 -1.35649586e+00
-6.69798970e-01 -2.69763112e-01 -8.12374055e-01 2.91581631e-01
6.06439114e-02 4.84274477e-01 -7.52513647e-01 -7.12181032e-01
-7.00665265e-02 -1.17288041e+00 3.64036471e-01 -8.99638474e-01
-6.38693929e-01 -3.95084172e-01 5.86300790e-01 -9.60262835e-01
1.40314913e+00 -2.54458547e+00 -2.91984081e-01 2.18918324e-01
-8.22607160e-01 1.83086798e-01 -2.28896160e-02 4.08896416e-01
2.39214301e-01 5.25348485e-01 1.10707907e-02 -7.05057263e-01
1.25919327e-01 4.86009359e-01 -3.44130218e-01 1.69648826e-01
-2.57053643e-01 7.71694958e-01 -4.04316097e-01 -6.59351945e-01
2.11095691e-01 3.81303400e-01 -4.47423309e-01 1.22538619e-01
7.05807582e-02 5.86759269e-01 7.88981542e-02 5.41694045e-01
3.93017709e-01 8.78462374e-01 2.13272005e-01 6.33618385e-02
-6.05697513e-01 1.09675491e+00 -8.30072939e-01 1.23574591e+00
-7.67961919e-01 7.09313333e-01 -2.92078331e-02 -2.79741049e-01
7.43347704e-01 5.53119302e-01 5.75183406e-02 -7.16372609e-01
1.72076032e-01 7.02839553e-01 6.52275801e-01 -2.10856944e-01
5.30133545e-01 -5.16245723e-01 -3.49299103e-01 3.08344603e-01
-4.84732948e-02 -5.28897107e-01 5.38017647e-03 -2.08632946e-01
3.99473429e-01 -1.39365390e-01 -6.93435073e-02 -4.12912995e-01
3.67349476e-01 -1.16865383e-02 6.27904356e-01 6.87207878e-01
-2.71701455e-01 4.46869940e-01 2.68284649e-01 3.36465299e-01
-8.46668899e-01 -1.25249577e+00 -4.96470600e-01 1.33231044e+00
-4.24197316e-01 -3.25633347e-01 -1.04749978e+00 -3.08435827e-01
-5.08767255e-02 1.21630943e+00 -8.12113881e-02 8.97232667e-02
-7.88706958e-01 -1.70567080e-01 1.29365635e+00 2.24579453e-01
1.31850675e-01 -1.38167274e+00 1.39229044e-01 2.90024608e-01
-3.80764842e-01 -1.29668999e+00 -1.03765976e+00 2.40973353e-01
-4.40717787e-01 -4.57756221e-01 -3.86308044e-01 -1.00756192e+00
-3.46370600e-02 -1.98334828e-03 6.95989132e-01 -4.60230649e-01
2.62268811e-01 1.41752467e-01 -2.39367962e-01 -5.95161378e-01
-1.15235496e+00 4.70518649e-01 3.98775101e-01 -9.90986452e-02
3.61182272e-01 -1.55758739e-01 9.80900079e-02 2.89324731e-01
-4.51196879e-01 -3.69617969e-01 1.53770700e-01 4.63531673e-01
5.65942109e-01 2.52075493e-01 6.66319966e-01 -9.63596821e-01
8.82670403e-01 -9.20975581e-02 -5.44141293e-01 5.03040180e-02
-2.82470584e-01 -1.01126775e-01 8.98039579e-01 -5.81444025e-01
-1.17804646e+00 6.23911843e-02 -6.58532083e-01 -4.07478064e-01
-2.39507645e-01 6.58900440e-01 -4.88338411e-01 1.49941906e-01
3.40172082e-01 2.02092752e-01 -1.33078143e-01 -8.24977815e-01
4.07403708e-01 9.93590951e-01 5.87257147e-01 -6.71683431e-01
6.33242607e-01 -2.19336465e-01 -5.89397013e-01 -1.11817527e+00
-6.66042209e-01 -4.22054738e-01 -6.31585598e-01 -1.91655010e-01
7.87849367e-01 -9.31257010e-01 -5.12523890e-01 4.93826747e-01
-9.76212144e-01 -5.03853858e-01 -2.00163022e-01 7.94453323e-01
-4.43390578e-01 1.42873555e-01 -7.54691601e-01 -7.75867641e-01
-7.27783516e-02 -1.47232211e+00 6.47930086e-01 -2.14873120e-01
-4.74592388e-01 -8.85401666e-01 1.26249731e-01 5.78115880e-01
2.42854059e-01 -3.63331825e-01 1.04731643e+00 -4.15545136e-01
-1.71729118e-01 1.94814220e-01 3.64582658e-01 7.39052236e-01
4.76464361e-01 3.53267580e-01 -1.31386065e+00 -8.24503899e-02
9.97943357e-02 2.28137756e-03 3.97889823e-01 6.68416321e-01
7.75489151e-01 -4.34608966e-01 9.89495367e-02 4.53200698e-01
1.12885821e+00 4.62048173e-01 5.86698115e-01 -1.68486703e-02
8.66733193e-01 8.41805935e-01 5.85522830e-01 -1.56929076e-01
4.44029391e-01 7.07702279e-01 -5.23289084e-01 6.62298724e-02
-5.17254114e-01 -5.46464026e-01 1.16360283e+00 1.87579346e+00
2.71242768e-01 -2.62793213e-01 -8.76364172e-01 7.57026613e-01
-9.58847880e-01 -6.77068532e-01 -3.79424691e-01 2.47468877e+00
1.40768683e+00 2.29733154e-01 2.41644368e-01 2.87415296e-01
6.85872436e-01 -4.28349301e-02 -5.40294163e-02 -1.20682275e+00
-2.53459573e-01 4.15129930e-01 6.35472298e-01 1.00084972e+00
-6.72065318e-01 1.75671732e+00 6.93687534e+00 1.24369335e+00
-1.52688789e+00 2.54082710e-01 3.24311376e-01 1.25108510e-01
-4.83999401e-01 -6.42581657e-02 -1.16094983e+00 6.36466861e-01
1.42925751e+00 1.14962526e-01 5.51414430e-01 5.30894935e-01
5.75022817e-01 -1.09090202e-01 -8.51829946e-01 6.86117649e-01
-1.73156232e-01 -6.22614563e-01 4.59961686e-03 1.16444968e-01
7.41029799e-01 2.15304315e-01 3.44049901e-01 4.82003212e-01
3.40636313e-01 -8.63503993e-01 1.11564362e+00 -2.34960243e-01
1.18197632e+00 -7.59015441e-01 4.13738668e-01 1.45961344e-01
-1.16878355e+00 2.11892903e-01 -9.69395190e-02 2.62999415e-01
4.44485366e-01 4.61646914e-02 -1.19214594e+00 2.96950102e-01
4.39581156e-01 1.45314317e-02 -5.74661195e-01 7.71025240e-01
-2.62665808e-01 1.56868696e+00 -4.49471116e-01 3.24348360e-01
3.29435796e-01 -2.00459994e-02 5.49350917e-01 1.50149333e+00
3.82413745e-01 -3.97958189e-01 8.80979300e-02 5.88642895e-01
6.92624748e-02 7.13192225e-01 -6.46366775e-01 -1.72044575e-01
8.77603471e-01 4.13684487e-01 -4.04091179e-01 -1.94818422e-01
-4.48694378e-01 8.23556662e-01 4.39942665e-02 4.12211865e-01
-6.76445723e-01 -1.25070989e-01 9.86778677e-01 1.16351090e-01
3.42048526e-01 -4.56224442e-01 -5.35489917e-01 -6.86595321e-01
-2.68233269e-01 -1.07560039e+00 5.55568784e-02 -3.79019737e-01
-1.33256197e+00 5.10578275e-01 -2.06353500e-01 -8.75058353e-01
-3.17562699e-01 -5.23171961e-01 -4.27966744e-01 1.21522236e+00
-1.45614123e+00 -1.00038433e+00 6.16792381e-01 5.90648055e-01
7.15326190e-01 -1.05691642e-01 8.24455738e-01 5.30808866e-01
-6.01527035e-01 1.13846445e+00 2.10080788e-01 3.44549805e-01
1.02092147e+00 -1.14521825e+00 5.55675030e-01 7.72491038e-01
3.89782190e-01 8.82410109e-01 6.80836201e-01 -7.69113064e-01
-1.13593388e+00 -1.07287943e+00 1.24890721e+00 -4.41472530e-01
5.07755458e-01 -5.20695984e-01 -9.13258970e-01 8.76924992e-01
3.67541671e-01 -4.88199770e-01 9.13191974e-01 2.57314086e-01
-3.18485528e-01 -1.02228388e-01 -8.22351515e-01 9.55168366e-01
7.06359148e-01 -1.12096405e+00 -5.79543829e-01 -2.69584656e-01
1.10361409e+00 -2.45544761e-01 -8.77875566e-01 4.83990833e-02
6.86478972e-01 -6.49666548e-01 6.81442976e-01 -3.34021866e-01
-1.85947523e-01 -3.13602239e-01 -4.89911914e-01 -1.59847665e+00
-1.64441064e-01 -7.14050055e-01 6.37114823e-01 1.80069435e+00
6.65594697e-01 -7.11491466e-01 1.17888853e-01 3.48665744e-01
-5.90992570e-01 -6.96934015e-02 -1.13223064e+00 -1.26540971e+00
5.38114488e-01 -9.58968937e-01 5.54219902e-01 9.51858580e-01
-2.95487434e-01 5.71724296e-01 -3.44942629e-01 2.49172777e-01
2.35164270e-01 -2.86447674e-01 6.99627817e-01 -9.16644096e-01
-1.86866269e-01 -3.94396067e-01 1.24292970e-01 -7.15377390e-01
3.50497067e-01 -9.69820857e-01 3.11748266e-01 -7.71041870e-01
-4.52068090e-01 -9.71761584e-01 -3.21027279e-01 4.19307262e-01
-8.34047645e-02 1.46597102e-01 5.80665052e-01 1.56870261e-01
2.16631398e-01 4.04043823e-01 1.32764316e+00 2.99149305e-01
-7.12675631e-01 2.31440678e-01 -4.29842293e-01 5.38658559e-01
1.05382919e+00 -8.16617608e-01 -2.19082162e-01 -3.50585908e-01
-5.40523708e-01 5.65513559e-02 -3.41151178e-01 -8.02333832e-01
-1.65769011e-02 -1.78321466e-01 -1.18231200e-01 -6.68707669e-01
4.91698742e-01 -5.77488601e-01 2.58946657e-01 3.07116657e-01
-3.03218275e-01 1.47338556e-02 5.44749439e-01 -2.33136341e-01
-3.46304476e-01 -2.96040177e-01 7.85609186e-01 -1.15707755e-01
-3.02177906e-01 -1.43326581e-01 -9.32927549e-01 3.94889057e-01
4.93821234e-01 -2.17884645e-01 1.22617602e-01 -1.70583770e-01
-3.95091772e-01 -2.57251322e-01 3.88345957e-01 6.13394558e-01
-7.65006617e-02 -1.25481582e+00 -5.95598876e-01 4.48698670e-01
-8.69281515e-02 -5.96246541e-01 -2.14918591e-02 5.40570855e-01
-3.27421695e-01 8.26548100e-01 -1.12614110e-02 -4.00460184e-01
-1.42534077e+00 2.21446037e-01 2.74902672e-01 1.15489952e-01
-1.50174156e-01 8.70610476e-01 2.17968285e-01 -8.80404353e-01
3.32123548e-01 -3.63282412e-01 1.95038185e-01 1.68384343e-01
-7.71494806e-02 2.84887940e-01 -7.06927553e-02 -1.18158901e+00
-3.66366357e-01 5.59695899e-01 1.70728322e-02 -8.06932926e-01
5.95992327e-01 -4.49409217e-01 1.98914409e-01 1.22083402e+00
1.37390602e+00 1.10957468e+00 -7.56942391e-01 -7.53800049e-02
1.93821713e-02 -4.33868766e-01 5.84846102e-02 -8.94067347e-01
-8.22361946e-01 1.05742025e+00 2.60875940e-01 1.70755565e-01
5.76765776e-01 -1.65267840e-01 1.05290544e+00 -2.36194938e-01
1.94679648e-01 -1.39242411e+00 -3.83117676e-01 8.30060363e-01
7.31334448e-01 -1.10731030e+00 -5.31185687e-01 -5.61625242e-01
-1.23996484e+00 6.62149370e-01 4.38023657e-01 3.90950859e-01
7.41977990e-01 2.70089597e-01 7.15790987e-01 1.38830706e-01
-5.05664647e-01 -1.83040902e-01 3.73668611e-01 3.81469995e-01
7.69156933e-01 6.76323950e-01 -3.52468520e-01 5.06578803e-01
-1.19287062e+00 -5.36223829e-01 3.06895018e-01 3.98691475e-01
-1.34865448e-01 -1.58508062e+00 -5.53815484e-01 1.39976904e-01
-9.49441493e-01 -5.90331435e-01 -5.05228162e-01 8.00096571e-01
2.52920777e-01 1.06647050e+00 2.98399366e-02 -4.53539699e-01
1.87489048e-01 6.95039392e-01 2.53163964e-01 -7.62553215e-01
-9.63185191e-01 5.37246108e-01 3.72046083e-01 -1.42546654e-01
1.16131596e-01 -1.06994677e+00 -1.28254259e+00 -4.86586094e-01
-5.65638661e-01 2.64907062e-01 8.54329348e-01 8.69042933e-01
6.95505440e-02 3.95355046e-01 7.69189000e-01 -2.57307440e-01
-3.38727087e-01 -1.09878063e+00 -7.51947105e-01 9.27740186e-02
2.28122368e-01 -3.90824348e-01 -7.33010173e-01 1.44450381e-01] | [14.614263534545898, 6.773756504058838] |
a2f2d8e5-c966-4d03-a11a-ea5e5d815c42 | rethinking-the-editing-of-generative | 2305.09454 | null | https://arxiv.org/abs/2305.09454v1 | https://arxiv.org/pdf/2305.09454v1.pdf | Rethinking the editing of generative adversarial networks: a method to estimate editing vectors based on dimension reduction | While Generative Adversarial Networks (GANs) have recently found applications in image editing, most previous GAN-based image editing methods require largescale datasets with semantic segmentation annotations for training, only provide high level control, or merely interpolate between different images. Previous researchers have proposed EditGAN for high-quality, high-precision semantic image editing with limited semantic annotations by finding `editing vectors'. However, it is noticed that there are many features that are not highly associated with semantics, and EditGAN may fail on them. Based on the orthogonality of latent space observed by EditGAN, we propose a method to estimate editing vectors that do not rely on semantic segmentation nor differentiable feature estimation network. Our method assumes that there is a correlation between the intensity distribution of features and the distribution of hidden vectors, and estimates the relationship between the above distributions by sampling the feature intensity of the image corresponding to several hidden vectors. We modified Linear Discriminant Analysis (LDA) to deal with both binary feature editing and continuous feature editing. We then found that this method has a good effect in processing features such as clothing type and texture, skin color and hair. | ['Xuyang Li', 'Qi Li', 'Zhenghong Yu', 'Haoran Jiang', 'Yuhan Cao'] | 2023-03-07 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 5.07243693e-01 5.07235453e-02 -1.09882340e-01 -5.94072282e-01
-3.67619306e-01 -5.40212691e-01 4.59099889e-01 -4.58305150e-01
-2.03798100e-01 6.36582196e-01 9.16357189e-02 3.24000269e-01
1.06211260e-01 -1.08116663e+00 -8.35752189e-01 -6.98411465e-01
2.64881432e-01 2.93350965e-01 4.33796458e-03 -2.21235141e-01
9.64541659e-02 2.57794946e-01 -1.39516056e+00 2.56878197e-01
1.05420732e+00 9.04989541e-01 9.25677791e-02 5.63612938e-01
-4.28253382e-01 6.04546666e-01 -8.21843863e-01 -5.31140149e-01
4.46825862e-01 -8.20133865e-01 -5.61202466e-01 3.93721789e-01
2.80768573e-01 -2.49590695e-01 -1.45724490e-01 1.48840868e+00
1.76928326e-01 -4.13826993e-03 9.65413094e-01 -1.52306283e+00
-1.20171773e+00 3.01468164e-01 -5.25066972e-01 -5.11509418e-01
3.03867817e-01 -1.17857186e-02 5.89312613e-01 -6.55691862e-01
7.55568027e-01 1.18772328e+00 8.01098764e-01 5.96732497e-01
-1.07034981e+00 -6.18375957e-01 -6.88501149e-02 2.98076347e-02
-1.28922153e+00 -2.15918154e-01 1.04714096e+00 -4.11460876e-01
2.81107545e-01 5.72601855e-01 9.84160304e-01 1.12486267e+00
1.85415864e-01 6.67762160e-01 1.33594275e+00 -6.60223365e-01
1.29621655e-01 3.29896897e-01 -4.91215795e-01 8.94310772e-01
-1.63687482e-01 -1.15612596e-01 -3.99558544e-01 -1.74744111e-02
1.17077315e+00 1.26967728e-01 -1.73431993e-01 -4.38062370e-01
-1.27113473e+00 1.13828027e+00 3.30693036e-01 1.85636416e-01
-3.42883140e-01 3.90116632e-01 1.85457900e-01 4.21075940e-01
6.04618132e-01 4.32044327e-01 -3.05731446e-01 -1.40481517e-02
-8.50771964e-01 -1.55637413e-01 5.34253180e-01 9.96716857e-01
1.04053617e+00 1.84182018e-01 -2.14495569e-01 9.63221788e-01
1.16193101e-01 6.55411363e-01 4.97217625e-01 -1.17491984e+00
7.78022707e-02 5.83450079e-01 -1.09226234e-01 -1.31220865e+00
1.40920579e-01 2.17721704e-02 -1.03416872e+00 2.55757898e-01
3.20447475e-01 -5.28458208e-02 -1.14086699e+00 1.81995738e+00
2.45084882e-01 -7.99179897e-02 -1.27876237e-01 8.72754276e-01
5.24334311e-01 5.75565696e-01 -1.29188523e-01 -1.68930694e-01
1.01423967e+00 -8.70590746e-01 -1.02765524e+00 3.37204672e-02
2.66815007e-01 -9.33414400e-01 1.33506536e+00 2.78100729e-01
-8.34685147e-01 -5.61442316e-01 -9.43438530e-01 -1.43557936e-01
-5.69939852e-01 2.46850461e-01 1.05341864e+00 7.33717382e-01
-9.95303035e-01 5.82571447e-01 -6.54289186e-01 -3.15482497e-01
3.04557562e-01 2.53561735e-01 -4.25664961e-01 3.37401666e-02
-1.11783397e+00 6.54437423e-01 2.26553500e-01 1.96815997e-01
-6.86169744e-01 -2.72030056e-01 -8.35428715e-01 -2.35974982e-01
3.14129889e-01 -7.08521068e-01 6.23782635e-01 -1.75534832e+00
-1.77885413e+00 8.10566962e-01 8.13999027e-03 4.93225306e-02
7.79167950e-01 -5.80192953e-02 -4.02639151e-01 7.92040750e-02
1.99437529e-01 9.86404657e-01 1.32337058e+00 -1.24331129e+00
-1.83463797e-01 -1.87346056e-01 6.28326088e-02 2.40625620e-01
-3.50440264e-01 -3.41868959e-02 -4.97255296e-01 -9.24512804e-01
2.45718434e-01 -8.91231358e-01 -1.85839057e-01 3.44187826e-01
-5.42186975e-01 1.19166844e-01 7.77889013e-01 -8.23133767e-01
6.46251738e-01 -2.28810310e+00 2.31974840e-01 4.20627803e-01
4.50255647e-02 -9.77725014e-02 -1.08850345e-01 1.57255113e-01
9.53512713e-02 1.93032026e-01 -5.59147239e-01 -8.05014968e-02
4.61708046e-02 4.26166028e-01 -1.89671293e-01 4.02750850e-01
9.44559649e-03 9.57973838e-01 -8.65216434e-01 -6.45864725e-01
2.48111188e-01 5.09153843e-01 -4.70000952e-01 2.28865474e-01
-3.22032094e-01 4.71891165e-01 -2.68612176e-01 6.14669800e-01
6.64839506e-01 9.87251624e-02 1.25932634e-01 -5.00062585e-01
2.11161897e-01 -4.32855129e-01 -1.17720723e+00 1.69905996e+00
-3.10414255e-01 6.90570772e-01 -7.89103657e-02 -9.68802631e-01
9.84438002e-01 9.62881446e-02 6.58626437e-01 -4.44092274e-01
1.49249226e-01 -1.90481078e-02 -3.17022413e-01 -3.89456511e-01
4.60952967e-01 1.68406554e-02 -1.47781134e-01 3.29155892e-01
1.27388328e-01 -4.35242087e-01 -8.12003314e-02 7.24763274e-02
6.97225988e-01 2.31960714e-01 1.68700352e-01 -1.05048455e-01
2.90869266e-01 -5.84440716e-02 5.39105475e-01 6.81860864e-01
1.02821685e-01 8.51494610e-01 6.17204487e-01 -1.90017462e-01
-1.29253185e+00 -1.06376719e+00 6.25127405e-02 9.39222038e-01
2.20467746e-01 -1.32599071e-01 -1.13416684e+00 -7.95439720e-01
-1.74679428e-01 6.37388289e-01 -8.91355395e-01 -3.61387283e-01
-1.54042065e-01 -6.99158728e-01 5.25872469e-01 4.07267332e-01
8.24503303e-01 -1.11827898e+00 -1.19355842e-01 -2.19533648e-02
-1.18783444e-01 -9.22249019e-01 -7.82299995e-01 -4.43518721e-02
-6.32345796e-01 -1.06774116e+00 -6.93557858e-01 -7.38324225e-01
1.16792631e+00 -5.97640648e-02 7.28368580e-01 3.20759080e-02
-3.89340192e-01 4.89188790e-01 -4.64079350e-01 -2.54061908e-01
-4.81453747e-01 -1.61639601e-01 -7.49516785e-02 2.24146023e-01
1.45358846e-01 -4.15918648e-01 -2.88097441e-01 4.23459709e-01
-1.05774689e+00 1.94752365e-01 5.44376433e-01 1.01599705e+00
5.53895116e-01 1.56750962e-01 3.40548158e-01 -1.19033146e+00
6.29483938e-01 -1.10316575e-01 -3.49897563e-01 4.25947547e-01
-5.82894325e-01 5.18582575e-03 5.02507269e-01 -5.68052590e-01
-1.02273905e+00 1.56110033e-01 -1.70053795e-01 -5.41448474e-01
-1.70017049e-01 1.96620643e-01 -4.49421823e-01 -2.53022283e-01
5.57172656e-01 3.24425101e-01 2.61056036e-01 -2.04429790e-01
7.10891366e-01 5.32775104e-01 4.31044817e-01 -3.78823251e-01
8.27270627e-01 5.67572534e-01 -2.07935557e-01 -7.46760070e-01
-6.88965857e-01 -2.92728655e-02 -8.39281678e-01 -2.66533345e-01
1.17153275e+00 -6.88858449e-01 -4.23124880e-01 6.68026567e-01
-9.99216735e-01 -2.61094600e-01 -6.48480415e-01 4.93941933e-01
-6.13485456e-01 5.15010715e-01 -6.21943355e-01 -5.28327405e-01
-1.31889388e-01 -1.07989919e+00 9.27212656e-01 5.76232225e-02
-1.89126462e-01 -1.09537160e+00 -1.94033176e-01 1.87748417e-01
5.26811361e-01 4.99846399e-01 9.90862072e-01 -6.80136457e-02
-5.30111790e-01 -2.52067119e-01 -1.84700772e-01 6.52081490e-01
4.79680806e-01 1.74121618e-01 -8.05812180e-01 -8.82408917e-02
3.94494422e-02 -8.18552747e-02 6.39106691e-01 4.75152224e-01
1.49137294e+00 -4.97459531e-01 -1.19632771e-02 9.40702319e-01
1.37711561e+00 2.40645558e-01 9.37866330e-01 1.15057528e-01
1.15131927e+00 4.50579673e-01 5.45486569e-01 1.15717620e-01
4.62316200e-02 6.54864788e-01 3.27025831e-01 -5.23026943e-01
-7.78802931e-02 -4.47825521e-01 4.95767087e-01 9.60278332e-01
-2.35861659e-01 -7.47201666e-02 -3.23967189e-01 5.00265360e-01
-1.64168406e+00 -7.20003545e-01 -1.52740963e-02 1.93945491e+00
1.04586983e+00 -5.43054491e-02 -1.59270808e-01 1.94916241e-02
8.55471551e-01 4.63418029e-02 -6.12756073e-01 -5.47264218e-01
-2.19307795e-01 2.94728369e-01 8.10578704e-01 3.78717870e-01
-9.15284157e-01 1.13793063e+00 7.07143974e+00 9.76944745e-01
-1.15417635e+00 2.51546443e-01 4.39436495e-01 2.97468632e-01
-6.48828983e-01 9.02140960e-02 -3.03340167e-01 5.91156662e-01
2.10386306e-01 1.50084108e-01 7.58467734e-01 9.15084183e-01
-2.24667117e-01 -4.24387790e-02 -7.97944605e-01 9.31325734e-01
2.67112553e-01 -1.01798677e+00 2.40616530e-01 1.27264678e-01
1.05877626e+00 -5.97974837e-01 2.41596892e-01 6.07270971e-02
4.19357955e-01 -1.08610177e+00 6.72980726e-01 7.95731306e-01
1.24200487e+00 -6.50299013e-01 7.33257830e-01 1.70204148e-01
-7.91757405e-01 4.05804485e-01 -5.76296151e-01 1.46654814e-01
-5.83575070e-02 7.25314081e-01 -6.80800915e-01 3.10332060e-01
4.10137534e-01 6.86141372e-01 -4.95601267e-01 4.20198798e-01
-5.77590764e-01 4.45273548e-01 -1.74482480e-01 1.91417530e-01
-1.04032084e-01 -5.46130121e-01 2.97301680e-01 8.33209872e-01
4.66481656e-01 -1.77580833e-01 8.80709365e-02 1.03554273e+00
-6.82549998e-02 2.52782643e-01 -6.12515509e-01 -2.27770150e-01
2.14928091e-01 1.14669287e+00 -9.52995420e-01 -3.50481808e-01
-2.07758412e-01 1.58567679e+00 -1.66077748e-01 3.97863686e-01
-1.07591403e+00 -4.72823083e-01 4.63086575e-01 -7.14794770e-02
2.34984845e-01 -3.22982848e-01 -4.08650994e-01 -1.29024172e+00
-1.53429598e-01 -8.32854152e-01 -7.70475119e-02 -9.57111001e-01
-1.36861122e+00 3.45149457e-01 -1.59166008e-01 -1.13036549e+00
-1.72255635e-01 -3.85906875e-01 -4.26583618e-01 7.58099616e-01
-1.02947986e+00 -1.46813738e+00 -4.43852991e-01 8.49775195e-01
4.99203742e-01 -2.95951575e-01 1.02793133e+00 2.31594354e-01
-2.59577423e-01 7.50123382e-01 1.73757866e-01 3.87600452e-01
7.19419122e-01 -1.22171545e+00 1.16218314e-01 6.80343509e-01
1.78457022e-01 4.91374135e-01 5.81364691e-01 -7.69709468e-01
-1.19862592e+00 -1.12438560e+00 4.33410108e-01 -2.70243704e-01
1.89622074e-01 -4.17520076e-01 -6.34527922e-01 9.03803766e-01
1.72630474e-01 -1.50825769e-01 5.44680834e-01 -6.13495037e-02
-7.93399196e-03 5.91798860e-04 -1.32734144e+00 4.28180695e-01
1.12964404e+00 -6.99778795e-01 -3.70542407e-01 3.81205767e-01
6.41261458e-01 -4.81518507e-01 -9.13210809e-01 3.22520077e-01
4.56913859e-01 -8.34320843e-01 8.37488770e-01 -4.45085287e-01
5.53366661e-01 -3.17569345e-01 -2.06083700e-01 -1.48814690e+00
-3.12776923e-01 -1.95621476e-01 3.37605864e-01 1.36718082e+00
1.24200188e-01 -6.16503477e-01 7.60596037e-01 4.29621547e-01
7.53476918e-02 -3.57889980e-01 -5.79545498e-01 -6.31208539e-01
-3.90754968e-01 -2.71966219e-01 7.48664737e-01 1.36747050e+00
-5.14637232e-01 -3.40644717e-02 -7.99511850e-01 -7.05512837e-02
6.85208499e-01 1.41004641e-02 9.61321115e-01 -9.33963835e-01
-2.05946490e-01 -1.27461389e-01 -6.38283372e-01 -7.76543140e-01
2.60619074e-01 -8.23291123e-01 9.19448435e-02 -1.37887168e+00
1.43759772e-01 -6.03737473e-01 -1.27768397e-01 6.27352118e-01
9.70194675e-03 5.53731263e-01 -8.30164063e-04 3.04182023e-01
-3.97149175e-01 6.72336578e-01 1.55505085e+00 -2.06303433e-01
-3.60258408e-02 -2.01789364e-01 -6.07538462e-01 9.01261806e-01
6.72290206e-01 -4.98068571e-01 -3.88310045e-01 -4.22430664e-01
1.71212599e-01 -1.13506027e-01 4.56471622e-01 -8.29403639e-01
-5.16872481e-02 -4.20992285e-01 7.69343972e-01 -2.21145868e-01
3.69100004e-01 -9.59270239e-01 5.40075064e-01 1.78190559e-01
-3.67018580e-01 -2.27460295e-01 -3.70016277e-01 4.65056330e-01
-4.29607719e-01 -2.30044946e-01 7.74860203e-01 -4.46301967e-01
-7.63864636e-01 2.05818847e-01 -2.85262346e-01 -1.81834847e-01
1.08303916e+00 -4.54067439e-01 7.15599060e-02 -5.86460292e-01
-8.32437634e-01 -2.40142986e-01 8.79376709e-01 3.23011577e-01
6.09710276e-01 -1.55879974e+00 -4.10366863e-01 5.58362782e-01
-1.51389375e-01 -3.50337272e-04 1.92923278e-01 5.30016899e-01
-6.67151153e-01 -1.41530320e-01 -6.33735597e-01 -7.11793482e-01
-1.07315183e+00 3.73811632e-01 2.58501917e-01 9.12571475e-02
-3.87637287e-01 8.55120003e-01 1.93507746e-01 -5.97634971e-01
-1.49165288e-01 -1.46959305e-01 -1.09646827e-01 8.86946023e-02
-7.10590463e-03 2.33399376e-01 -3.21194530e-01 -6.07684970e-01
6.61741719e-02 8.38058650e-01 1.45762727e-01 -1.03682861e-01
1.17512059e+00 -1.22896701e-01 -3.78298640e-01 5.04534960e-01
1.21368647e+00 2.94734478e-01 -1.28398180e+00 -2.71443930e-02
-3.83033365e-01 -8.80342424e-01 -5.39998077e-02 -5.63095510e-01
-1.41148901e+00 6.61932111e-01 7.88869381e-01 1.85937062e-01
1.26545489e+00 -1.39633849e-01 8.90000582e-01 -1.02082789e-01
3.94009739e-01 -1.38838243e+00 2.08923101e-01 2.58401245e-01
8.17111552e-01 -1.21373940e+00 3.03003415e-02 -7.07975149e-01
-8.70672405e-01 1.00106716e+00 6.10152245e-01 -2.28167161e-01
4.67564017e-01 5.85308075e-02 1.68708026e-01 -2.02074945e-01
-1.42802924e-01 -3.32604907e-03 2.66804993e-01 7.48843789e-01
1.27351612e-01 3.34160179e-01 -2.63996184e-01 2.00553477e-01
-3.67450774e-01 -6.54615015e-02 4.35070008e-01 6.44444525e-01
-1.52914405e-01 -1.24307442e+00 -3.87649238e-01 5.28067589e-01
-3.23607087e-01 -6.15651868e-02 -2.86189824e-01 7.22903073e-01
5.51676214e-01 4.38957423e-01 1.52790308e-01 -4.02400970e-01
1.87610075e-01 -1.65110044e-02 6.88638210e-01 -5.14479578e-01
-1.74098924e-01 6.15951754e-02 -2.42272943e-01 -5.51785409e-01
-5.52755415e-01 -5.66968679e-01 -9.57479596e-01 -2.65702307e-01
-3.92979831e-01 7.54971430e-02 1.06278765e+00 8.96711230e-01
2.05630343e-02 5.58289766e-01 6.77154422e-01 -5.16164243e-01
-2.26836354e-01 -9.25078809e-01 -9.12491262e-01 8.14855635e-01
-7.92252719e-02 -6.69918656e-01 -3.60005766e-01 5.25362790e-01] | [11.736138343811035, -0.45144644379615784] |
30ad301d-d49c-4d01-b2e0-ce97f1ded628 | a-dataset-for-building-code-mixed-goal | 1806.05997 | null | http://arxiv.org/abs/1806.05997v1 | http://arxiv.org/pdf/1806.05997v1.pdf | A Dataset for Building Code-Mixed Goal Oriented Conversation Systems | There is an increasing demand for goal-oriented conversation systems which
can assist users in various day-to-day activities such as booking tickets,
restaurant reservations, shopping, etc. Most of the existing datasets for
building such conversation systems focus on monolingual conversations and there
is hardly any work on multilingual and/or code-mixed conversations. Such
datasets and systems thus do not cater to the multilingual regions of the
world, such as India, where it is very common for people to speak more than one
language and seamlessly switch between them resulting in code-mixed
conversations. For example, a Hindi speaking user looking to book a restaurant
would typically ask, "Kya tum is restaurant mein ek table book karne mein meri
help karoge?" ("Can you help me in booking a table at this restaurant?"). To
facilitate the development of such code-mixed conversation models, we build a
goal-oriented dialog dataset containing code-mixed conversations. Specifically,
we take the text from the DSTC2 restaurant reservation dataset and create
code-mixed versions of it in Hindi-English, Bengali-English, Gujarati-English
and Tamil-English. We also establish initial baselines on this dataset using
existing state of the art models. This dataset along with our baseline
implementations is made publicly available for research purposes. | ['Suman Banerjee', 'Mitesh M. Khapra', 'Siddhartha Arora', 'Nikita Moghe'] | 2018-06-15 | a-dataset-for-building-code-mixed-goal-2 | https://aclanthology.org/C18-1319 | https://aclanthology.org/C18-1319.pdf | coling-2018-8 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-4.78752702e-01 -5.58755398e-02 -1.17605999e-01 -6.94611788e-01
-1.04426277e+00 -8.49968255e-01 7.24576890e-01 3.90477409e-03
-3.16605419e-01 1.01827466e+00 6.29723847e-01 -8.81130934e-01
2.37597436e-01 -5.64302206e-01 -1.61556646e-01 -3.21040779e-01
1.36852577e-01 1.04437816e+00 -2.35685706e-02 -1.22455537e+00
-4.14536819e-02 -1.05784893e-01 -9.86777782e-01 5.23497462e-01
7.63531983e-01 3.32149178e-01 4.44005340e-01 9.29950595e-01
-3.73437673e-01 1.06382442e+00 -4.21860635e-01 -6.70652568e-01
4.82116314e-03 -6.80910230e-01 -1.22677684e+00 1.29729137e-01
-2.55472828e-02 -3.99169058e-01 -1.96157366e-01 6.66214049e-01
4.00144875e-01 3.21888685e-01 5.16333699e-01 -1.28314710e+00
-5.04652679e-01 9.39355731e-01 -9.82954800e-02 -5.10757864e-02
7.50061631e-01 -1.52017936e-01 1.15189052e+00 -6.87075853e-01
6.93454444e-01 1.33682311e+00 4.13591474e-01 5.21126628e-01
-1.03538811e+00 -5.57385325e-01 8.12605023e-02 -8.27374682e-02
-1.05918431e+00 -6.70160651e-01 5.76048374e-01 -2.46025249e-01
1.03929818e+00 4.91370440e-01 3.00131470e-01 1.27036786e+00
3.97498421e-02 1.03668535e+00 1.08872747e+00 -2.27291331e-01
-1.23394258e-01 6.81551337e-01 7.43611977e-02 2.57898599e-01
-5.25556505e-01 -3.71426314e-01 -3.58951241e-01 -7.21376464e-02
1.47869393e-01 -1.13806993e-01 4.92768846e-02 7.00421929e-02
-1.42124677e+00 1.13566923e+00 2.79131174e-01 5.99407256e-01
-9.12354980e-03 -3.82646799e-01 6.08237565e-01 7.12069333e-01
4.19729590e-01 2.19100118e-02 -6.33480251e-01 -6.72813952e-01
-6.13197982e-01 7.26374686e-01 1.35743403e+00 1.47639239e+00
7.39927769e-01 -3.26264292e-01 2.61432439e-01 1.43972838e+00
3.65098119e-01 4.24832791e-01 2.74646550e-01 -8.52881372e-01
9.13542628e-01 5.46209574e-01 2.05934778e-01 -5.88422358e-01
-5.27066827e-01 1.27849087e-01 -7.52551198e-01 -4.57427442e-01
5.62144995e-01 -5.51401079e-01 -4.17741537e-01 1.33842456e+00
1.32992133e-01 -6.62415624e-01 5.05945802e-01 7.07582533e-01
1.19936585e+00 7.41683006e-01 -2.25183427e-01 4.81278002e-02
1.47010887e+00 -1.43322587e+00 -8.64433646e-01 -4.21513975e-01
7.72031724e-01 -1.21219051e+00 1.27083480e+00 3.05713471e-02
-9.82317507e-01 -4.01597679e-01 -5.28123379e-01 -3.48297834e-01
-7.27367401e-01 1.93500072e-02 6.85105264e-01 8.60262454e-01
-1.09660006e+00 -1.99154079e-01 -4.82782155e-01 -8.61609459e-01
-6.89704239e-01 5.21086529e-02 -4.25649047e-01 -2.57498890e-01
-1.58884251e+00 9.78283226e-01 1.20969549e-01 9.96159837e-02
-6.26178920e-01 -1.35043338e-01 -1.26110065e+00 -3.41680318e-01
1.81200653e-01 5.39150871e-02 1.83137989e+00 -5.79844654e-01
-1.55633950e+00 9.39460576e-01 -3.12283993e-01 -2.19026491e-01
6.32825613e-01 5.30626327e-02 -7.72433996e-01 -5.04217029e-01
4.67336476e-01 6.66514277e-01 1.61458086e-02 -1.08942676e+00
-9.36463594e-01 -2.64981449e-01 6.98867440e-01 5.38031042e-01
2.45839089e-01 3.11801821e-01 -4.12563354e-01 -3.38361800e-01
-5.61310723e-03 -1.25021172e+00 9.03046131e-03 -8.06263268e-01
-5.08932590e-01 -1.33737370e-01 8.63144457e-01 -8.32644701e-01
1.26985955e+00 -2.04216218e+00 -2.67764091e-01 -1.93773031e-01
-1.98015153e-01 -1.20165400e-01 1.85537949e-01 1.18256259e+00
3.88776422e-01 -9.27970931e-02 -1.27394691e-01 -4.66689676e-01
1.71673715e-01 5.95546365e-01 5.53666009e-03 1.37576377e-02
-2.44223312e-01 7.13581562e-01 -1.03595507e+00 -3.43243688e-01
3.70659411e-01 2.47125089e-01 -6.97316766e-01 3.11281830e-01
-2.04690889e-01 8.47003937e-01 -1.07052878e-01 8.00885677e-01
5.04831731e-01 1.10288382e-01 3.81903380e-01 3.37308258e-01
-4.14484441e-01 1.01759398e+00 -1.01959658e+00 1.64081252e+00
-9.43956792e-01 6.72767639e-01 5.10580003e-01 -6.50319397e-01
8.77324700e-01 5.45543849e-01 2.01893166e-01 -6.32517219e-01
1.18381433e-01 3.95471364e-01 1.26884148e-01 -3.49154264e-01
9.87747490e-01 7.77718029e-04 -7.15373635e-01 4.29780602e-01
-3.46876204e-01 -3.60397279e-01 5.12523293e-01 3.64890367e-01
7.55587935e-01 -1.62330419e-01 1.94426984e-01 -3.70989501e-01
9.17024851e-01 2.61066884e-01 1.09992698e-01 5.48400104e-01
-4.47241962e-01 4.45697606e-01 2.89135993e-01 -1.78467199e-01
-8.95454407e-01 -8.79816711e-01 -2.22946808e-01 1.76594162e+00
-7.06369206e-02 -4.25045818e-01 -4.86947268e-01 -4.52491254e-01
-3.31409693e-01 8.35674703e-01 -1.47097453e-01 4.54027593e-01
-4.96745199e-01 -3.49515855e-01 4.68833834e-01 4.52445783e-02
1.15440583e+00 -1.02721167e+00 2.15124995e-01 6.07354581e-01
-9.81078327e-01 -1.26580191e+00 -9.17315900e-01 2.93245852e-01
-2.65847236e-01 -6.41931474e-01 -7.21364677e-01 -1.17612767e+00
6.29037470e-02 4.32849377e-01 1.32763720e+00 -1.99923307e-01
2.31335700e-01 1.62123010e-01 -5.88669121e-01 -2.68466920e-01
-1.00402665e+00 5.27894080e-01 -1.05903663e-01 -2.18410179e-01
7.55822182e-01 -2.86787063e-01 -3.05341601e-01 6.51349068e-01
-3.71165514e-01 2.17113331e-01 7.91323259e-02 8.42485487e-01
-3.96329105e-01 -1.81541875e-01 8.22668374e-01 -1.10353220e+00
9.06580865e-01 -7.39028335e-01 -9.13631469e-02 1.29667309e-03
-8.60364661e-02 -2.20860347e-01 5.99462390e-01 -1.32771015e-01
-1.13715267e+00 -1.47806868e-01 -8.68585765e-01 6.38191998e-01
-3.20386052e-01 6.26283765e-01 -3.02455842e-01 3.63027006e-01
5.03962398e-01 2.99907357e-01 -1.56908572e-01 -3.95042747e-01
4.17938113e-01 1.62663829e+00 4.19223756e-01 -5.80073059e-01
2.90449053e-01 5.47698326e-02 -9.23163831e-01 -1.14825821e+00
-3.39528859e-01 -9.68046248e-01 -4.94515896e-01 -2.61843234e-01
9.37117815e-01 -1.10504150e+00 -8.71774435e-01 5.46504021e-01
-1.00484765e+00 -6.02271557e-01 6.90020561e-01 3.47634792e-01
-4.90460783e-01 6.40226714e-03 -1.02183962e+00 -8.75310481e-01
-7.92747065e-02 -1.47137260e+00 1.04868555e+00 2.39093341e-02
-6.36679411e-01 -1.29132700e+00 9.40435529e-02 8.66264224e-01
6.65161788e-01 -2.09889054e-01 9.72070813e-01 -8.39881480e-01
-1.50042549e-01 -3.40612754e-02 -1.17566653e-01 1.30571023e-01
5.55128872e-01 -2.13800654e-01 -7.95273364e-01 -2.41859630e-01
-2.14780316e-01 -6.71696305e-01 3.07164080e-02 2.07023695e-02
2.97482628e-02 -3.55721205e-01 -8.31607077e-03 -7.73426741e-02
9.02727187e-01 5.48806250e-01 4.18626010e-01 3.50846410e-01
4.90029693e-01 8.70854795e-01 6.74050272e-01 4.22320724e-01
1.44126999e+00 7.90898323e-01 -9.69360694e-02 1.15282513e-01
2.17115194e-01 -2.58820504e-01 4.97779876e-01 1.44642365e+00
3.70666087e-01 -1.25009820e-01 -1.04404652e+00 9.73810315e-01
-1.83469629e+00 -7.30835617e-01 -2.62292653e-01 1.94420874e+00
1.24872494e+00 1.10696675e-02 4.71431881e-01 -1.63338065e-01
6.03372812e-01 1.80295348e-01 -2.07596142e-02 -6.95531964e-01
2.63557345e-01 -4.46455687e-01 2.01184571e-01 1.05339479e+00
-1.07679689e+00 1.22195971e+00 5.40323305e+00 4.58234370e-01
-9.50189471e-01 2.24521995e-01 7.39776969e-01 2.68046498e-01
-2.06189215e-01 1.68483749e-01 -1.14255333e+00 5.67165792e-01
1.33101869e+00 -4.15346064e-02 7.68352628e-01 9.85493243e-01
3.28114957e-01 -2.75803149e-01 -1.20036829e+00 1.05223775e+00
-2.32945874e-01 -1.08213508e+00 -3.81134957e-01 -6.13983124e-02
5.67375243e-01 1.90335602e-01 -2.04221353e-01 1.05068076e+00
1.01907063e+00 -7.23419726e-01 6.43032789e-01 -2.37254918e-01
4.19074565e-01 -8.50717306e-01 7.62448251e-01 7.98326135e-01
-1.27845550e+00 1.58663526e-01 -1.17811663e-02 -1.71542197e-01
3.84581596e-01 9.38291661e-03 -1.05086315e+00 3.69432420e-01
8.09186101e-01 3.62252116e-01 -2.05032170e-01 3.20452631e-01
8.64613652e-02 3.23010236e-01 -1.86084196e-01 -3.32081586e-01
7.09846377e-01 -4.58220243e-01 2.82266617e-01 1.43071544e+00
1.38635427e-01 -5.81442080e-02 4.47429001e-01 4.20082062e-01
-1.35387897e-01 2.08929852e-01 -8.60420525e-01 -6.10321648e-02
4.02126521e-01 1.16267896e+00 -5.93522549e-01 -1.35584533e-01
-8.15335035e-01 1.04868019e+00 1.01083353e-01 2.44098350e-01
-6.51906788e-01 -5.44881225e-01 7.40604699e-01 6.96954876e-02
-3.08046848e-01 -5.01010120e-01 1.18749879e-01 -1.05030632e+00
-8.49496722e-02 -1.51719058e+00 2.64315575e-01 -5.93750536e-01
-1.19559336e+00 8.27859402e-01 8.14594328e-02 -9.01368678e-01
-9.42379773e-01 -2.59817451e-01 -2.94668347e-01 9.83230770e-01
-1.27414024e+00 -1.33696711e+00 -1.43804193e-01 7.51653314e-01
1.12798381e+00 -2.45224103e-01 9.52743232e-01 7.08319306e-01
-1.90843821e-01 5.54361939e-01 2.57541329e-01 5.44764876e-01
9.92705762e-01 -1.34559083e+00 6.79298818e-01 3.58063817e-01
-3.29157077e-02 7.78303027e-01 7.93582261e-01 -6.18189394e-01
-1.23347080e+00 -8.44306648e-01 1.61385870e+00 -6.77928567e-01
9.04535532e-01 -9.95303333e-01 -6.30503058e-01 9.43163335e-01
6.26067817e-01 -6.88304126e-01 6.50527358e-01 3.61688614e-01
6.56303987e-02 4.27542366e-02 -1.16417634e+00 8.35138083e-01
7.90383756e-01 -7.85761774e-01 -4.99954432e-01 5.68591833e-01
5.86930871e-01 -6.99775040e-01 -6.92869663e-01 -1.88879460e-01
4.51337337e-01 -1.17301309e+00 5.85119247e-01 -3.47627908e-01
2.18894899e-01 1.18172690e-01 -4.09574002e-01 -1.47598302e+00
1.70322597e-01 -9.11662936e-01 5.84597647e-01 1.49487424e+00
6.99787617e-01 -6.81819856e-01 6.30511940e-01 8.32095861e-01
-1.71494603e-01 -1.76854968e-01 -8.70750427e-01 -4.78726238e-01
3.91751111e-01 -3.94633919e-01 6.58870876e-01 9.51827645e-01
4.61701423e-01 6.48180008e-01 -5.25559843e-01 -1.77196324e-01
-2.31017149e-03 1.06814746e-02 1.16398287e+00 -7.26500273e-01
-2.52004296e-01 -1.47860482e-01 1.21876426e-01 -1.46323466e+00
9.34706628e-02 -8.67324531e-01 3.56222630e-01 -1.71980464e+00
-1.76810861e-01 -8.22181165e-01 5.71736693e-01 2.96745211e-01
4.34064679e-02 4.98775542e-02 1.57911718e-01 1.67597800e-01
-4.80958819e-01 1.98997110e-01 9.97631907e-01 -1.75150529e-01
-5.14896154e-01 4.46732938e-01 -8.44708622e-01 4.36827898e-01
6.77511513e-01 -1.62399352e-01 -5.69475591e-01 -3.20826173e-01
2.13097721e-01 6.37808621e-01 -2.96996951e-01 -7.50203907e-01
1.40659466e-01 -1.67873457e-01 -3.41078728e-01 -5.99375069e-01
4.16391939e-01 -8.65464211e-01 7.36495331e-02 1.24424696e-01
-3.84777874e-01 2.53591090e-01 7.43354857e-03 1.25228941e-01
-4.89086360e-01 -1.47989959e-01 6.68112993e-01 -4.90511447e-01
-6.76908374e-01 -1.03828229e-01 -1.05743015e+00 2.47608572e-01
8.47729206e-01 3.04785017e-02 -4.07786399e-01 -1.25333989e+00
-5.96676886e-01 7.91317165e-01 2.71738231e-01 8.47316682e-01
3.61058004e-02 -1.31425703e+00 -7.36856222e-01 2.44434237e-01
4.39824849e-01 -3.44384313e-01 1.80892915e-01 7.00937867e-01
-7.40949988e-01 9.05737996e-01 -1.14806600e-01 -3.84134918e-01
-1.14939201e+00 8.64023417e-02 1.53249323e-01 -3.04451883e-01
-2.14152455e-01 6.71640158e-01 -9.11434665e-02 -1.49321258e+00
3.03989053e-01 -3.09589356e-01 -2.09469646e-01 1.30008847e-01
3.94582599e-01 1.27078086e-01 1.97875798e-01 -9.71931458e-01
-4.28278834e-01 -1.21740133e-01 -3.00044149e-01 -4.87165481e-01
8.00184369e-01 -7.75761783e-01 -2.63498306e-01 8.70977819e-01
1.37586939e+00 1.69991195e-01 -6.02237344e-01 -1.63801327e-01
1.44781530e-01 -2.50913113e-01 -4.06342834e-01 -8.02261949e-01
-5.05939245e-01 7.59461880e-01 2.14807585e-01 5.89988589e-01
6.18338645e-01 1.51656121e-01 1.06434512e+00 7.24387586e-01
7.43226409e-01 -1.16071486e+00 -2.35746056e-01 1.29756069e+00
8.30492795e-01 -1.67443681e+00 -6.72358990e-01 -1.74188346e-01
-1.14486027e+00 8.14704120e-01 4.49664325e-01 4.00028259e-01
7.90057838e-01 2.72443175e-01 7.27061450e-01 7.25352690e-02
-7.76961982e-01 -4.69877273e-01 -2.02894688e-01 6.26441360e-01
1.13684475e+00 2.21846357e-01 -3.41785222e-01 4.08475518e-01
-8.24358165e-01 -3.95706892e-01 6.87380016e-01 1.11772990e+00
-2.65971303e-01 -1.52893686e+00 -2.41993994e-01 2.29923159e-01
-4.52872008e-01 -4.35408324e-01 -6.05895460e-01 1.00843203e+00
-2.51672119e-01 1.69918716e+00 2.67182128e-03 -4.10283387e-01
2.22241938e-01 4.24811095e-01 -1.60831064e-01 -7.52754986e-01
-1.01765919e+00 2.26855278e-01 7.64555573e-01 -4.01164219e-02
-4.44862545e-01 -5.42634487e-01 -1.20870447e+00 -1.02228045e+00
-9.33448970e-02 4.74779755e-01 7.16420770e-01 9.21154559e-01
-7.51179010e-02 1.24885872e-01 5.98746598e-01 -6.08294725e-01
-1.04596071e-01 -1.28829741e+00 -5.68535745e-01 2.96542913e-01
4.00017112e-01 -2.98083127e-01 -2.58275777e-01 2.23128963e-02] | [12.570096015930176, 8.221664428710938] |
46b9fc8a-7f75-4f17-98cb-52d048ed5cba | automated-top-view-registration-of-broadcast | 1703.01437 | null | http://arxiv.org/abs/1703.01437v1 | http://arxiv.org/pdf/1703.01437v1.pdf | Automated Top View Registration of Broadcast Football Videos | In this paper, we propose a novel method to register football broadcast video
frames on the static top view model of the playing surface. The proposed method
is fully automatic in contrast to the current state of the art which requires
manual initialization of point correspondences between the image and the static
model. Automatic registration using existing approaches has been difficult due
to the lack of sufficient point correspondences. We investigate an alternate
approach exploiting the edge information from the line markings on the field.
We formulate the registration problem as a nearest neighbour search over a
synthetically generated dictionary of edge map and homography pairs. The
synthetic dictionary generation allows us to exhaustively cover a wide variety
of camera angles and positions and reduce this problem to a minimal per-frame
edge map matching procedure. We show that the per-frame results can be improved
in videos using an optimization framework for temporal camera stabilization. We
demonstrate the efficacy of our approach by presenting extensive results on a
dataset collected from matches of football World Cup 2014. | ['Vineet Gandhi', 'C. V. Jawahar', 'Rahul Anand Sharma', 'Bharath Bhat'] | 2017-03-04 | null | null | null | null | ['bird-view-synthesis', 'homography-estimation'] | ['computer-vision', 'computer-vision'] | [ 3.19960654e-01 -3.23930621e-01 2.77762264e-02 -9.21006724e-02
-7.71774530e-01 -6.93594217e-01 6.00666344e-01 2.84500774e-02
-5.67238331e-01 3.67001176e-01 -8.77508372e-02 2.65994996e-01
-8.26954693e-02 -6.04751766e-01 -8.23671281e-01 -3.07305634e-01
6.64664358e-02 4.81846809e-01 7.87280262e-01 -5.50691605e-01
5.37935078e-01 5.40820301e-01 -1.61022985e+00 1.27549116e-02
2.09869713e-01 6.97719693e-01 7.94671401e-02 7.40846515e-01
4.21192139e-01 8.08546171e-02 -3.34762305e-01 -4.84375536e-01
8.02459538e-01 -4.75016326e-01 -6.70234501e-01 5.82134187e-01
8.81653428e-01 -1.62492901e-01 -3.66530299e-01 1.04265070e+00
5.42123795e-01 4.07680750e-01 8.78242850e-02 -1.20392776e+00
3.18925023e-01 -2.06437513e-01 -7.22498953e-01 9.59565043e-02
1.03143632e+00 -2.28104562e-01 6.42717004e-01 -8.00941288e-01
1.19348800e+00 8.04152131e-01 9.54929411e-01 1.21729180e-01
-1.30975902e+00 -4.21344906e-01 -1.46003872e-01 4.29668128e-01
-1.62519491e+00 -5.65319002e-01 9.01704609e-01 -5.17774940e-01
6.86884522e-01 2.28741989e-01 9.80709016e-01 5.20020306e-01
1.64062586e-02 3.35334204e-02 9.99686122e-01 -9.48708415e-01
-2.80332975e-02 -2.40926564e-01 3.35147940e-02 6.90189779e-01
1.82937577e-01 2.83346176e-01 -8.40042174e-01 -6.14504106e-02
1.17713678e+00 -2.12485611e-01 -3.04398000e-01 -1.00227582e+00
-1.34247804e+00 6.40770078e-01 -1.25077171e-02 2.44512215e-01
-2.89194942e-01 9.56132933e-02 1.88994646e-01 2.36372620e-01
2.83345163e-01 2.86993086e-01 4.12011705e-02 -4.09121901e-01
-1.18780625e+00 3.96012604e-01 8.43334377e-01 1.14678812e+00
9.32825327e-01 -1.80421710e-01 4.27708030e-01 5.65313101e-01
3.99344694e-03 2.12709025e-01 1.36525124e-01 -1.39070213e+00
4.43799138e-01 3.19025427e-01 2.46186018e-01 -1.54554498e+00
-2.90888786e-01 2.93454155e-02 -2.01445550e-01 2.44620055e-01
6.77800417e-01 8.10062140e-03 -4.52382356e-01 1.33898830e+00
5.27326643e-01 3.55517238e-01 -2.05730438e-01 8.56014371e-01
3.90054613e-01 3.73125106e-01 -6.98989511e-01 -2.94735193e-01
1.24033928e+00 -8.01660299e-01 -7.36732602e-01 -1.57738417e-01
3.13460141e-01 -1.26493883e+00 5.08692145e-01 3.76008987e-01
-1.41490579e+00 -5.64351737e-01 -1.10461688e+00 5.47362827e-02
8.75659510e-02 9.73273367e-02 1.48504570e-01 5.83047450e-01
-1.21772540e+00 5.85744083e-01 -9.18127060e-01 -5.57545066e-01
-3.51301581e-01 7.47848451e-01 -8.11973214e-01 2.41567716e-02
-1.03722119e+00 8.62154901e-01 3.87901992e-01 6.92684278e-02
-1.96758807e-01 -3.75999451e-01 -9.66885924e-01 -2.42999390e-01
5.45409441e-01 -6.23709977e-01 1.14075553e+00 -1.06458735e+00
-1.69500959e+00 1.11984253e+00 -2.57849336e-01 -3.23950678e-01
7.12250710e-01 -1.53911024e-01 -1.00522026e-01 4.44392979e-01
1.09055884e-01 3.52074027e-01 5.73693216e-01 -1.31520760e+00
-7.64419615e-01 -1.14980802e-01 3.53502154e-01 4.57427889e-01
3.79482843e-02 1.11728802e-01 -1.07673192e+00 -6.91866159e-01
5.18860638e-01 -1.44790220e+00 -3.23917627e-01 -1.56158745e-01
4.06302065e-02 5.89051127e-01 6.30746305e-01 -6.73344314e-01
1.26715124e+00 -2.05178905e+00 3.26344460e-01 6.16345644e-01
-2.70731270e-01 -1.93363614e-02 2.73717940e-01 5.78978062e-01
-1.89002290e-01 -5.40003479e-01 -2.55695917e-02 -1.50028944e-01
-3.00584346e-01 1.57270506e-01 1.25356521e-02 1.08868778e+00
-5.13874650e-01 2.94194549e-01 -7.72317469e-01 -6.46478713e-01
4.12934273e-01 5.92993200e-01 -7.01690197e-01 -9.81587544e-02
3.37688476e-01 4.43982750e-01 -4.24734093e-02 4.52239513e-01
6.43870354e-01 3.70299906e-01 1.81508914e-01 -3.29237431e-01
-3.99925619e-01 -1.07591311e-02 -1.95953000e+00 2.28154302e+00
-1.40025556e-01 6.83005393e-01 1.53384805e-01 -8.94532204e-01
9.25031424e-01 4.95972306e-01 9.28185880e-01 -3.88675630e-01
2.89203286e-01 2.11122587e-01 -9.69236344e-02 -2.09660232e-01
8.65258515e-01 -1.90652102e-01 -9.64376926e-02 2.54428566e-01
5.02838865e-02 -2.06379965e-01 4.62521136e-01 -8.74832645e-02
6.64267123e-01 5.48576117e-01 4.78037566e-01 -3.00977528e-01
6.54943764e-01 2.97053844e-01 5.25873184e-01 4.33121890e-01
-2.57163588e-02 1.08363867e+00 1.69031367e-01 -5.39649248e-01
-1.18642581e+00 -6.46941960e-01 -6.63223192e-02 4.26908225e-01
6.47735178e-01 -8.07134092e-01 -1.00382268e+00 -7.67634958e-02
-3.93248260e-01 5.60275801e-02 -3.74283880e-01 7.91013762e-02
-1.05328798e+00 -4.07260567e-01 1.51639536e-01 3.27528656e-01
4.36587572e-01 -5.10354996e-01 -7.91172862e-01 2.00184897e-01
-4.35922831e-01 -1.53100228e+00 -6.87670887e-01 -3.97469610e-01
-8.98021102e-01 -1.19285154e+00 -6.01891518e-01 -8.78899693e-01
7.65097082e-01 4.94657993e-01 9.92077351e-01 5.87448552e-02
-1.47120178e-01 7.39238858e-01 -3.80844474e-01 8.97633806e-02
-3.27488899e-01 -1.15819030e-01 2.87800103e-01 2.19016299e-01
3.25145304e-01 -4.91681933e-01 -4.79865730e-01 7.56581426e-01
-8.10347199e-01 1.20416813e-01 5.96738607e-02 6.89948559e-01
7.96600580e-01 -7.96954427e-03 -2.16298118e-01 -6.16474450e-01
1.82496756e-01 6.93688989e-02 -9.13001955e-01 1.13806546e-01
-1.38509542e-01 -2.78872043e-01 3.19582745e-02 -3.52555186e-01
-8.81746531e-01 7.74043977e-01 5.95982336e-02 -3.80551726e-01
1.43149737e-02 2.09923133e-01 -1.31851897e-01 -6.18029535e-01
5.39281487e-01 1.30292475e-02 7.15198517e-02 -2.86538333e-01
2.66104400e-01 2.92086363e-01 9.37425315e-01 -6.10424459e-01
9.37847614e-01 8.07772636e-01 3.38844806e-01 -9.01610196e-01
-3.70579243e-01 -1.03739953e+00 -1.34708118e+00 -6.00124478e-01
8.42234433e-01 -8.00203443e-01 -7.18697965e-01 2.03847483e-01
-1.30543196e+00 7.08213523e-02 -2.62052238e-01 9.37852442e-01
-1.12028277e+00 7.98544168e-01 -2.93490231e-01 -3.74421060e-01
1.24693535e-01 -1.36981940e+00 1.12299371e+00 6.10053763e-02
-2.89474487e-01 -9.76281285e-01 5.74400544e-01 2.56729543e-01
-1.93542227e-01 5.35217226e-01 1.90703776e-02 -1.12510517e-01
-6.51325107e-01 -5.79015195e-01 4.08226311e-01 -4.25169393e-02
8.02281201e-02 4.39801812e-02 -5.57863176e-01 -3.47810090e-01
9.78453159e-02 3.45582128e-01 3.10984939e-01 4.61252987e-01
4.36202526e-01 2.22114637e-01 -3.01031440e-01 8.43365371e-01
1.57210732e+00 1.45082459e-01 7.34887123e-01 1.02808523e+00
5.93731940e-01 5.87365866e-01 1.01242185e+00 3.73285681e-01
1.45217240e-01 1.53428280e+00 1.51748657e-01 -1.72533721e-01
-1.57152101e-01 -1.69999614e-01 3.75118732e-01 7.11176276e-01
-7.30989635e-01 1.03270032e-01 -8.42855096e-01 5.77950299e-01
-1.94538009e+00 -1.27730334e+00 -3.57198954e-01 2.51100516e+00
4.37909812e-01 -4.68235649e-02 3.05104464e-01 4.01783675e-01
1.10061753e+00 -5.40190935e-02 2.22175837e-01 -3.82519394e-01
3.26839164e-02 1.70643836e-01 8.06267142e-01 9.24438536e-01
-1.20601630e+00 8.69116783e-01 6.49932718e+00 5.60584843e-01
-8.51657331e-01 -1.51913762e-02 -2.27769762e-01 -1.48491422e-02
2.36131161e-01 5.07784665e-01 -8.64411652e-01 1.67689696e-01
5.66637576e-01 -1.36277393e-01 2.24558741e-01 6.37376964e-01
5.87092936e-02 -3.57714504e-01 -1.01881588e+00 1.28913832e+00
5.98437726e-01 -1.46493411e+00 -3.26781273e-01 2.86146402e-01
8.81687701e-01 -3.46279621e-01 -3.14810157e-01 -4.61425036e-01
-1.52583227e-01 -5.17994761e-01 7.75611579e-01 5.00147402e-01
7.87596703e-01 -7.98768044e-01 4.44906712e-01 1.59727275e-01
-1.42156613e+00 3.56529504e-01 -1.99115351e-01 -2.12299213e-01
6.53788328e-01 -8.79895166e-02 -5.71470320e-01 9.21850383e-01
6.48943007e-01 6.93058550e-01 -4.91981208e-01 1.35095024e+00
2.38244161e-01 2.01950334e-02 -5.28018653e-01 6.77148163e-01
6.11281432e-02 -7.60434449e-01 8.67897093e-01 1.02084553e+00
4.92707759e-01 2.12655321e-01 2.89210290e-01 -4.91462881e-03
3.70302469e-01 2.27683201e-01 -7.41498172e-01 6.15943789e-01
5.05150817e-02 1.15828705e+00 -1.13050687e+00 -1.40201002e-01
-6.86829865e-01 1.07869208e+00 -1.84770346e-01 7.85649717e-02
-7.21303165e-01 -2.02087641e-01 2.43648320e-01 6.48290336e-01
2.40347862e-01 -5.99106789e-01 5.46318628e-02 -1.34138787e+00
2.10510686e-01 -1.02726042e+00 3.42419356e-01 -5.77191830e-01
-5.52289069e-01 7.81438529e-01 5.30284107e-01 -1.89184713e+00
-5.56355834e-01 -3.48255068e-01 -2.92116821e-01 5.75024962e-01
-1.08007288e+00 -9.61239874e-01 -3.23800981e-01 8.42076540e-01
6.00334346e-01 -8.97866637e-02 6.47985280e-01 5.43108344e-01
-9.53945965e-02 3.46113324e-01 1.01222053e-01 -1.28918514e-02
8.68688583e-01 -9.56336617e-01 3.47422630e-01 1.16556370e+00
3.74584883e-01 5.12735844e-01 9.50261176e-01 -5.68029404e-01
-1.28803647e+00 -4.50674891e-01 8.55107427e-01 -5.01523733e-01
5.68197846e-01 -3.26640278e-01 -5.41075349e-01 9.49774325e-01
1.01966992e-01 8.81541171e-04 5.13057888e-01 -2.71361411e-01
2.69519240e-01 -1.16311744e-01 -7.68303871e-01 5.30539393e-01
9.39185679e-01 -2.28385553e-01 -6.23901844e-01 2.09288195e-01
-3.26036066e-02 -1.00996435e+00 -9.17396188e-01 2.52502322e-01
7.28392780e-01 -1.08363366e+00 1.27068341e+00 2.76515516e-03
-1.01091065e-01 -7.09212542e-01 -2.65542388e-01 -9.68035936e-01
1.14207398e-02 -1.06289196e+00 4.63940382e-01 1.10327554e+00
-1.48047864e-01 -1.77854359e-01 9.61510897e-01 6.26586378e-01
-1.08450390e-01 -1.63580999e-01 -1.21553135e+00 -7.43146241e-01
-5.03375053e-01 -3.06051612e-01 1.45888492e-01 8.30856681e-01
1.51937008e-01 -2.54777428e-02 -5.39425075e-01 2.42846295e-01
6.68159366e-01 2.97488689e-01 1.18449068e+00 -1.24150014e+00
-3.69642466e-01 -9.07091796e-02 -1.14784801e+00 -1.03304541e+00
5.21690026e-03 -5.18350601e-01 9.19324234e-02 -1.19470286e+00
-6.25364780e-02 -2.51044482e-01 4.08102751e-01 -8.17011893e-02
2.04676911e-01 6.93363965e-01 3.76825929e-01 3.49049300e-01
-4.20350373e-01 -3.86350937e-02 9.47620630e-01 4.40354466e-01
-2.34120965e-01 2.28080582e-02 -7.30080977e-02 1.13795948e+00
4.17908907e-01 -4.64080453e-01 -2.57146895e-01 -4.34739679e-01
1.89250156e-01 3.68626773e-01 4.25127536e-01 -1.20241857e+00
4.70434844e-01 -1.52610838e-01 9.88562703e-02 -5.06041408e-01
7.43813157e-01 -1.12444222e+00 7.64333487e-01 3.15282315e-01
1.11910582e-01 6.18970931e-01 1.14302061e-01 5.06674528e-01
-4.97998774e-01 -4.81132805e-01 6.48719609e-01 -1.70451477e-01
-9.24390018e-01 1.39549717e-01 -1.80163775e-02 -1.41446635e-01
1.39782655e+00 -8.93595099e-01 2.43068829e-01 -4.89521503e-01
-9.12012398e-01 -1.75015539e-01 1.17868984e+00 2.79855728e-01
4.46520537e-01 -1.37668920e+00 -4.34540957e-01 3.99857968e-01
-1.84889678e-02 -1.70196250e-01 2.18842581e-01 1.00950444e+00
-1.19165206e+00 2.35387310e-01 -4.63738233e-01 -9.54853654e-01
-1.67350984e+00 4.31670129e-01 4.33881879e-01 -1.08952373e-02
-7.72673309e-01 1.65174022e-01 -5.27027547e-02 2.90583018e-02
-4.41357121e-02 4.05762084e-02 -2.11476997e-01 -5.74963428e-02
3.27327043e-01 6.55589819e-01 3.01933914e-01 -1.51244152e+00
-3.59010875e-01 1.42340612e+00 3.00465077e-01 -6.25261426e-01
1.22443962e+00 -4.15834069e-01 9.90764424e-02 1.98408023e-01
1.09227264e+00 4.17313784e-01 -1.30024731e+00 1.16570201e-02
-1.84062317e-01 -9.94916320e-01 -3.72634642e-02 5.59045151e-02
-9.63483095e-01 5.32725990e-01 6.89747214e-01 6.25240849e-03
1.02273190e+00 -2.70524353e-01 6.98229790e-01 2.19115511e-01
4.77785766e-01 -1.20588672e+00 -2.83450663e-01 2.01654419e-01
7.23348260e-01 -9.08535123e-01 3.82954776e-01 -8.36615860e-01
-4.50845271e-01 1.39035141e+00 2.85394073e-01 -5.47635198e-01
5.39764404e-01 1.94834024e-01 1.13287017e-01 -1.99394897e-01
-7.60410503e-02 -2.39721417e-01 3.94831538e-01 6.29488468e-01
3.11546624e-01 -4.58586007e-01 -7.34382689e-01 -3.40690315e-02
-4.14118230e-01 7.64027834e-02 9.13359344e-01 1.27095592e+00
-2.63608217e-01 -1.65977550e+00 -8.53174984e-01 -3.62200886e-01
-5.52878559e-01 1.79887474e-01 -1.85068548e-01 1.15306056e+00
4.82991226e-02 8.78375649e-01 2.09893316e-01 -2.44190887e-01
7.64952064e-01 -1.59385756e-01 8.35620999e-01 -4.80784535e-01
-5.44974923e-01 5.37871003e-01 1.47846594e-01 -7.21818328e-01
-1.10725737e+00 -9.82925773e-01 -9.25510764e-01 -2.44397774e-01
-4.53455657e-01 1.80349559e-01 6.25054896e-01 7.58498609e-01
8.67021456e-02 -5.63951284e-02 5.12023151e-01 -1.25891376e+00
-4.43734787e-02 -3.41354519e-01 -5.07340968e-01 6.54809237e-01
6.77915439e-02 -9.78546441e-01 -7.27453828e-03 7.47522950e-01] | [7.971292495727539, -1.6502262353897095] |
1560807c-c883-46a8-bd45-fad08c9db5b5 | saliency-guided-mutual-learning-network-for | 2305.07180 | null | https://arxiv.org/abs/2305.07180v1 | https://arxiv.org/pdf/2305.07180v1.pdf | Saliency-Guided Mutual Learning Network for Few-shot Fine-grained Visual Recognition | Recognizing novel sub-categories with scarce samples is an essential and challenging research topic in computer vision. Existing literature focus on addressing this challenge through global-based or local-based representation approaches. The former employs global feature representations for recognization, which may lack fine-grained information. The latter captures local relationships with complex structures, possibly leading to high model complexity. To address the above challenges, this article proposes a novel framework called SGML-Net for few-shot fine-grained visual recognition. SGML-Net incorporates auxiliary information via saliency detection to guide discriminative representation learning, achieving high performance and low model complexity. Specifically, SGML-Net utilizes the saliency detection model to emphasize the key regions of each sub-category, providing a strong prior for representation learning. SGML-Net transfers such prior with two independent branches in a mutual learning paradigm. To achieve effective transfer, SGML-Net leverages the relationships among different regions, making the representation more informative and thus providing better guidance. The auxiliary branch is excluded upon the transfer's completion, ensuring low model complexity in deployment. The proposed approach is empirically evaluated on three widely-used benchmarks, demonstrating its superior performance. | ['Tong Zhang', 'Xinrong Gong', 'C. L. Philip Chen', 'Haiqi Liu'] | 2023-05-12 | null | null | null | null | ['fine-grained-visual-recognition', 'saliency-detection'] | ['computer-vision', 'computer-vision'] | [ 4.16626364e-01 -1.38722286e-01 -6.21476054e-01 -2.65778631e-01
-8.50559413e-01 -3.10312480e-01 6.29215837e-01 2.76037902e-01
2.23993305e-02 3.16174150e-01 4.79397506e-01 8.75500739e-02
-1.03420764e-01 -6.11861408e-01 -5.20421505e-01 -7.93973565e-01
2.41971642e-01 -1.60515234e-01 4.37351644e-01 1.12690642e-01
5.07936001e-01 4.63758171e-01 -1.66785538e+00 2.52322912e-01
1.00028598e+00 1.22571349e+00 4.79453743e-01 2.32935965e-01
-2.75652796e-01 9.39536691e-01 -2.81174898e-01 -4.65232544e-02
2.18596801e-01 -1.56368688e-01 -5.71657717e-01 1.52062461e-01
6.05925620e-01 -3.96626741e-01 -3.43016177e-01 1.08795989e+00
2.91574627e-01 3.58282596e-01 7.59337246e-01 -1.09333575e+00
-8.01254153e-01 2.53843129e-01 -7.46047139e-01 4.55352783e-01
1.82478353e-01 3.82658333e-01 1.32889307e+00 -1.10111117e+00
2.81431586e-01 1.34624934e+00 4.76160258e-01 3.38624448e-01
-1.09749889e+00 -6.64060712e-01 7.20496058e-01 4.13415730e-01
-1.42734039e+00 -5.51051021e-01 8.99867058e-01 -4.44984019e-01
7.86465228e-01 6.01823069e-02 3.36098313e-01 8.85079503e-01
-2.07908582e-02 1.20349634e+00 9.14891660e-01 -2.81146139e-01
2.04129517e-01 -2.56653856e-02 4.66050804e-01 8.02676737e-01
3.12087566e-01 1.24547571e-01 -7.42392123e-01 -2.37931609e-02
1.02458477e+00 6.36961520e-01 -2.94820696e-01 -5.96260786e-01
-1.08352947e+00 8.42201531e-01 8.22509110e-01 1.18775643e-01
-5.73619723e-01 -2.13509612e-02 3.75005633e-01 1.26256257e-01
3.74399930e-01 1.79225042e-01 -1.29608527e-01 -6.09965399e-02
-9.41206515e-01 -1.44093484e-01 3.06629717e-01 1.01502180e+00
1.03478992e+00 1.65242523e-01 -5.83431959e-01 1.05634749e+00
5.72958708e-01 2.98033416e-01 4.77816164e-01 -5.01572609e-01
5.71490228e-01 9.79890704e-01 -1.00559518e-01 -1.14164698e+00
2.66316924e-02 -8.37876856e-01 -8.06660950e-01 -4.02541794e-02
-3.22301574e-02 2.76848882e-01 -1.11626196e+00 1.61816096e+00
2.65959382e-01 6.19242191e-01 -1.32023349e-01 9.74076509e-01
8.92295301e-01 6.53618157e-01 2.63981014e-01 5.12153357e-02
1.19229949e+00 -1.39067864e+00 -2.52230912e-01 -4.06195432e-01
3.43462318e-01 -6.24313414e-01 1.01584268e+00 4.23045829e-02
-6.23088300e-01 -7.00986087e-01 -1.03082323e+00 -4.29891981e-02
-1.62271261e-01 3.07441890e-01 4.89337534e-01 2.44754791e-01
-9.19574797e-01 2.62766957e-01 -8.31887245e-01 -4.25635010e-01
7.77021945e-01 2.05086228e-02 -1.24459498e-01 -5.29624581e-01
-7.74829209e-01 6.17924809e-01 2.22919047e-01 7.76268020e-02
-1.11847115e+00 -7.44282544e-01 -1.02515674e+00 2.75389224e-01
3.65417898e-01 -4.87050414e-01 1.02253425e+00 -7.93299317e-01
-1.18985009e+00 5.59902012e-01 -3.71844858e-01 -3.20609719e-01
2.45676607e-01 -2.76355058e-01 -3.18783760e-01 2.15021968e-01
2.98172116e-01 6.55740798e-01 1.22293913e+00 -1.43904114e+00
-8.50272000e-01 -2.75727421e-01 1.49217606e-01 3.76141340e-01
-5.70389032e-01 -2.19217762e-01 -6.13557220e-01 -1.04662311e+00
1.77071109e-01 -6.75172448e-01 -1.96796387e-01 3.77502851e-02
-2.77729422e-01 -4.18195367e-01 9.48599935e-01 -4.78270918e-01
1.13420570e+00 -2.25276160e+00 2.29822174e-02 7.28285015e-02
3.93617541e-01 4.01646644e-01 -4.67493981e-01 2.81730533e-01
1.87976032e-01 -1.90107554e-01 -1.35848522e-01 -2.89649516e-01
-8.83861184e-02 1.60856009e-01 -5.01477003e-01 3.39685589e-01
6.03890061e-01 1.13441920e+00 -1.02449214e+00 -4.11175907e-01
3.86106938e-01 3.75856787e-01 -4.94258195e-01 2.82754660e-01
3.33754905e-02 1.39462709e-01 -7.55342126e-01 1.02448809e+00
6.72172725e-01 -4.79994953e-01 -1.23471454e-01 -4.50959861e-01
7.07264524e-03 2.68131420e-02 -1.13395691e+00 1.64998007e+00
-4.30737019e-01 4.00315940e-01 7.22520100e-03 -1.16304290e+00
1.27852988e+00 -2.03542456e-01 2.19894603e-01 -7.85305500e-01
1.46886660e-02 -3.81394364e-02 -1.10651061e-01 -2.62022704e-01
4.97331053e-01 -6.08661398e-03 1.72465034e-02 4.46872324e-01
1.99589580e-01 3.57556790e-01 -1.90774873e-01 3.10037702e-01
9.98927534e-01 1.32852525e-01 5.83625197e-01 -2.21186042e-01
4.34838086e-01 -3.01720351e-01 8.17242265e-01 9.19328809e-01
-5.78311622e-01 5.83203495e-01 -6.41380772e-02 -1.86111569e-01
-3.34554136e-01 -1.23946488e+00 1.75093099e-01 1.41300380e+00
6.44441366e-01 -4.80803162e-01 -2.82461941e-01 -8.89776349e-01
1.58106104e-01 5.60058951e-01 -7.63401091e-01 -6.06966913e-01
-2.93250144e-01 -2.56830126e-01 1.48325637e-01 7.38361537e-01
5.02449334e-01 -1.21574557e+00 -7.63010025e-01 1.07180506e-01
-3.72715965e-02 -6.95403397e-01 -5.53214371e-01 1.51256144e-01
-9.91124213e-01 -1.12621319e+00 -8.29581738e-01 -9.19291675e-01
8.27691317e-01 1.14808452e+00 8.86801839e-01 9.42061096e-02
-4.23962563e-01 5.66146851e-01 -6.47870600e-01 -2.77928919e-01
2.45026559e-01 5.95395379e-02 -1.76434234e-01 4.60211754e-01
6.21529698e-01 -5.46209157e-01 -7.73704350e-01 3.70318383e-01
-7.89461732e-01 6.87221531e-03 9.22790825e-01 1.05648065e+00
6.63128257e-01 -2.23314315e-01 7.68610477e-01 -5.48128664e-01
3.76446247e-01 -6.82924688e-01 -2.50786781e-01 3.21381569e-01
-4.21072751e-01 -8.81124102e-03 6.66010737e-01 -3.64752889e-01
-1.10570133e+00 -9.24907848e-02 2.73759484e-01 -8.45878243e-01
-2.00986430e-01 5.39237797e-01 -2.50930846e-01 -2.52583295e-01
3.21495622e-01 7.11236656e-01 -8.14522356e-02 -6.36948168e-01
3.02765906e-01 5.13729870e-01 4.82515454e-01 -6.19178951e-01
7.72333622e-01 3.65105212e-01 -4.52138305e-01 -7.05250502e-01
-1.22683775e+00 -9.38709080e-01 -6.22439563e-01 -1.77712068e-01
4.25308883e-01 -1.14946568e+00 -2.11092174e-01 3.61911297e-01
-7.53262162e-01 -1.99262556e-02 -3.68648380e-01 3.15754980e-01
-2.76770443e-01 3.77344728e-01 -4.01445985e-01 -7.46275008e-01
-3.82654548e-01 -1.07122147e+00 1.31694329e+00 6.81467414e-01
3.20932195e-02 -9.58058059e-01 -1.40750498e-01 2.90008456e-01
4.94233906e-01 -7.70798400e-02 6.77804649e-01 -5.97489536e-01
-7.99545527e-01 -1.38290972e-01 -7.64944613e-01 2.23961607e-01
4.74494249e-01 -2.60241181e-01 -9.89158154e-01 -5.37096679e-01
-1.47544146e-01 -5.73243737e-01 1.30157518e+00 2.01923460e-01
1.20878971e+00 -1.23986796e-01 -4.47303623e-01 5.63120246e-01
1.37741065e+00 4.23074886e-02 3.68949175e-01 3.08935463e-01
1.10017419e+00 4.73513037e-01 1.08663917e+00 5.68433762e-01
5.67081511e-01 6.66876554e-01 3.95208150e-01 -6.34770468e-02
-3.44942868e-01 -4.23110217e-01 3.33891094e-01 9.11615431e-01
2.42804810e-01 2.37129569e-01 -8.01996946e-01 6.73889160e-01
-2.10248780e+00 -9.13924932e-01 5.26896596e-01 1.91538370e+00
5.56089759e-01 -1.08112274e-02 2.38384828e-02 -1.72505006e-01
8.55339289e-01 5.53713620e-01 -9.48073804e-01 1.64761208e-02
9.02320817e-02 -3.30706611e-02 1.03718393e-01 7.72611722e-02
-1.17062283e+00 1.10802197e+00 5.33118486e+00 1.09205389e+00
-1.25348818e+00 1.89599209e-02 4.92606401e-01 -2.17262632e-03
-1.99423552e-01 -1.69424247e-03 -9.44847047e-01 4.21860129e-01
4.18477803e-01 -2.56552517e-01 6.74700886e-02 1.06440330e+00
6.66491538e-02 -2.33413037e-02 -9.64693606e-01 1.22737312e+00
3.27598631e-01 -1.34294415e+00 5.07895172e-01 -8.76448080e-02
7.34342098e-01 1.26844227e-01 9.18641537e-02 5.10116875e-01
2.55345941e-01 -7.89881229e-01 6.50889277e-01 7.21659601e-01
7.01767564e-01 -8.71379673e-01 5.44142306e-01 3.19561273e-01
-1.78055382e+00 -2.09131613e-01 -5.70411742e-01 4.71004285e-02
2.45355768e-04 5.16631544e-01 -5.90265095e-01 6.83926105e-01
4.84863222e-01 1.40626919e+00 -9.19911146e-01 1.40268385e+00
-1.38362154e-01 8.57584596e-01 -4.02120799e-02 2.07788453e-01
4.82375741e-01 -1.76922679e-01 5.26680171e-01 1.40664649e+00
4.69718091e-02 -7.30941966e-02 7.98729181e-01 7.27497578e-01
-1.03667371e-01 -3.73145081e-02 -4.47325140e-01 7.32159391e-02
6.50253296e-01 1.38112664e+00 -7.82557487e-01 -3.44985962e-01
-5.53317964e-01 9.08043861e-01 6.35680139e-01 3.52480859e-01
-6.14759922e-01 -4.47243512e-01 6.99790239e-01 -2.50398159e-01
6.64691806e-01 -4.85905521e-02 -3.56685072e-01 -1.25576508e+00
-5.29958121e-02 -9.14861202e-01 6.62007511e-01 -2.92158455e-01
-1.63602173e+00 3.88078630e-01 -1.70864269e-01 -1.41833627e+00
5.75988144e-02 -3.88851702e-01 -7.99156785e-01 8.61934185e-01
-1.89568019e+00 -1.56374824e+00 -6.09470487e-01 7.11487830e-01
9.94225860e-01 -4.07153219e-01 5.72008789e-01 1.62071839e-01
-7.64552832e-01 9.58983362e-01 -3.10578365e-02 -4.22674268e-02
7.01513469e-01 -1.03175068e+00 2.34915048e-01 1.04195499e+00
1.89000025e-01 9.47314560e-01 1.86370581e-01 -6.33563399e-01
-1.36679912e+00 -1.61995494e+00 4.18852359e-01 -1.70216233e-01
4.60235327e-01 -2.48948604e-01 -1.24666893e+00 3.88363570e-01
-2.46521384e-01 2.98724085e-01 7.85706222e-01 1.56860173e-01
-8.90800655e-01 -2.13986427e-01 -8.62946570e-01 4.73276675e-01
1.06560171e+00 -8.39193523e-01 -8.79569292e-01 1.55649846e-03
7.86760926e-01 1.31675214e-01 -5.73321164e-01 3.49535286e-01
4.93073016e-01 -8.02217782e-01 1.02100062e+00 -3.35827619e-01
4.25839067e-01 -5.44144034e-01 -2.74266332e-01 -1.25580072e+00
-7.79107869e-01 -1.98777974e-01 -5.57283401e-01 1.44164085e+00
2.30161077e-03 -6.05193734e-01 7.54582107e-01 2.40888506e-01
-2.34541640e-01 -9.26804245e-01 -6.14560723e-01 -7.83720434e-01
-3.39175820e-01 -2.36660287e-01 3.98140967e-01 9.89234030e-01
-3.26149702e-01 3.78246605e-01 -4.57832992e-01 9.13488120e-02
8.01097453e-01 5.95983028e-01 7.89305151e-01 -1.24279249e+00
-2.24316835e-01 -5.48246205e-01 -5.70496142e-01 -1.48273683e+00
1.60786301e-01 -9.90238965e-01 2.12756187e-01 -1.62865174e+00
6.03149951e-01 -5.09231865e-01 -9.92845714e-01 6.68076336e-01
-5.03958762e-01 3.36982608e-01 5.13082445e-01 5.82841098e-01
-1.07272780e+00 1.01924491e+00 1.08423674e+00 -3.51458371e-01
-1.74073696e-01 -7.72881731e-02 -1.04035711e+00 7.14440465e-01
5.85127771e-01 -1.98852032e-01 -6.60818517e-01 -2.87322938e-01
-6.68623865e-01 -4.73801225e-01 5.34689486e-01 -1.06649947e+00
4.60096538e-01 -3.78048390e-01 4.02281106e-01 -6.80683792e-01
2.11136207e-01 -6.76250041e-01 -3.78581226e-01 2.86643147e-01
-3.51658702e-01 -3.48212421e-01 9.59435180e-02 1.01950300e+00
-5.44910669e-01 -4.14493717e-02 8.46873164e-01 1.11394674e-01
-1.29908240e+00 6.60309315e-01 6.22204989e-02 -8.05296823e-02
1.05447793e+00 -4.36623156e-01 -1.55866995e-01 -2.41655424e-01
-3.99842203e-01 3.31120998e-01 5.00293374e-01 8.12576771e-01
1.02955055e+00 -1.45775962e+00 -4.81987953e-01 2.00349122e-01
6.57168984e-01 1.20435059e-02 5.23497403e-01 7.77611256e-01
-8.05065222e-03 4.05310482e-01 -3.27257693e-01 -8.53278220e-01
-1.15507329e+00 5.26780784e-01 -1.92876831e-02 -1.46012321e-01
-7.88886309e-01 9.36004519e-01 8.70883405e-01 -1.84286907e-01
2.46119112e-01 -3.89003642e-02 -4.46183652e-01 1.69967294e-01
8.44932556e-01 2.79869765e-01 -2.49893218e-01 -8.09770644e-01
-4.87275273e-01 6.43609285e-01 -6.51477039e-01 4.63176847e-01
1.35688412e+00 -4.69998121e-01 1.29812375e-01 4.68603313e-01
1.16540813e+00 -1.60166025e-01 -1.77127457e+00 -8.14926267e-01
1.10512696e-01 -7.56012619e-01 2.92262554e-01 -7.49750078e-01
-1.09644592e+00 9.94525552e-01 5.47736347e-01 -2.80598909e-01
1.27033699e+00 2.35188715e-02 7.69858479e-01 2.89621383e-01
5.54757774e-01 -8.54077160e-01 4.36605185e-01 6.75098956e-01
8.65345538e-01 -1.30529857e+00 -5.34466021e-02 -4.74765956e-01
-7.21792400e-01 8.41843605e-01 9.01836932e-01 -2.84651697e-01
4.52558041e-01 -2.80563682e-01 -6.32966161e-02 -9.88216847e-02
-7.03811407e-01 -3.17817152e-01 7.29425728e-01 7.29129195e-01
1.16641290e-01 6.24820031e-03 1.83209553e-01 8.87218297e-01
3.52873027e-01 -2.19417319e-01 3.68207172e-02 1.15861750e+00
-7.18923509e-01 -7.55630970e-01 -1.48924887e-01 8.35991144e-01
3.62314582e-02 -2.10846618e-01 -3.13891619e-01 3.39360654e-01
-5.66661917e-02 9.04850960e-01 -1.01162808e-03 -3.46744061e-01
2.02791363e-01 -1.82405382e-01 2.31171638e-01 -9.22512829e-01
-2.95046598e-01 1.11007914e-01 -2.38104388e-01 -7.58123040e-01
-4.09898043e-01 -6.61123335e-01 -9.04217482e-01 1.16801478e-01
-3.53773177e-01 4.79380041e-03 6.56544417e-02 1.03582883e+00
6.97084486e-01 5.80238521e-01 8.48518670e-01 -1.03183126e+00
-5.81475794e-01 -7.60447681e-01 -6.63954854e-01 3.39221120e-01
4.72063839e-01 -1.12441504e+00 -2.49874577e-01 1.85331609e-02] | [9.705199241638184, 1.950963020324707] |
c7175464-f7d3-4358-9ecb-b95e095bf46a | retrieval-augmented-chest-x-ray-report | 2305.03660 | null | https://arxiv.org/abs/2305.03660v1 | https://arxiv.org/pdf/2305.03660v1.pdf | Retrieval Augmented Chest X-Ray Report Generation using OpenAI GPT models | We propose Retrieval Augmented Generation (RAG) as an approach for automated radiology report writing that leverages multimodally aligned embeddings from a contrastively pretrained vision language model for retrieval of relevant candidate radiology text for an input radiology image and a general domain generative model like OpenAI text-davinci-003, gpt-3.5-turbo and gpt-4 for report generation using the relevant radiology text retrieved. This approach keeps hallucinated generations under check and provides capabilities to generate report content in the format we desire leveraging the instruction following capabilities of these generative models. Our approach achieves better clinical metrics with a BERTScore of 0.2865 ({\Delta}+ 25.88%) and Semb score of 0.4026 ({\Delta}+ 6.31%). Our approach can be broadly relevant for different clinical settings as it allows to augment the automated radiology report generation process with content relevant for that setting while also having the ability to inject user intents and requirements in the prompts as part of the report generation process to modulate the content and format of the generated reports as applicable for that clinical setting. | ['Tanuja Ganu', 'Ranjit Manuel', 'Gopinath Ganapathy', 'Mercy Ranjit'] | 2023-05-05 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [ 2.04562470e-01 9.39602017e-01 2.35131681e-02 -1.98379025e-01
-1.61137283e+00 -6.10220075e-01 6.26995146e-01 4.36087191e-01
-3.11086714e-01 5.32545865e-01 9.16514337e-01 -5.92371881e-01
-2.99625307e-01 -4.80107754e-01 -4.14742112e-01 -2.54338413e-01
-6.01560576e-03 7.54728019e-01 -2.12368906e-01 -3.39181662e-01
1.90207288e-01 1.70145467e-01 -8.51306558e-01 6.16491675e-01
5.51321268e-01 5.87070227e-01 4.79534090e-01 1.32503462e+00
1.36716723e-01 1.02365553e+00 -6.56524241e-01 -3.32715660e-01
1.18778251e-01 -4.76889700e-01 -7.22697198e-01 -1.00696525e-02
4.19395059e-01 -6.49277568e-01 -6.10413313e-01 5.47857821e-01
8.69135201e-01 -1.12553507e-01 9.70785379e-01 -5.95327199e-01
-1.08077765e+00 6.67748272e-01 -4.15104151e-01 4.63908374e-01
6.36888623e-01 4.42951739e-01 4.52952027e-01 -8.29577506e-01
9.19146538e-01 8.20700049e-01 1.55481070e-01 7.45288491e-01
-1.00739360e+00 -5.81844807e-01 -2.48108789e-01 -4.37743664e-01
-1.14384317e+00 -2.50962704e-01 2.91597545e-01 -5.46886325e-01
9.77580965e-01 5.05015373e-01 5.40085912e-01 1.16917217e+00
8.12452734e-01 4.18719471e-01 6.40403926e-01 -3.09651136e-01
1.01066902e-01 3.66447359e-01 3.33859324e-02 9.04413819e-01
2.24892810e-01 1.16705364e-02 -5.19585371e-01 -2.76809543e-01
9.63149309e-01 -1.61156878e-01 -3.72900367e-01 2.47494951e-01
-1.45457935e+00 1.06699479e+00 4.71547872e-01 -3.68145481e-02
-6.63173139e-01 1.27241656e-01 2.57916957e-01 2.17332747e-02
2.22811759e-01 8.21239054e-01 2.96694517e-01 2.24230886e-02
-1.07737529e+00 3.95349801e-01 8.04084539e-01 1.18739545e+00
9.21629742e-02 3.77524793e-02 -1.06734037e+00 7.28977442e-01
4.58548307e-01 7.04270124e-01 6.64040029e-01 -8.57704401e-01
3.59047025e-01 3.01116168e-01 1.70456897e-02 -5.99340200e-01
-3.68065029e-01 -5.90550363e-01 -3.14369977e-01 1.25314295e-02
-1.96505487e-01 -3.08919549e-01 -1.39058042e+00 1.50130570e+00
1.01619828e-02 -4.82916385e-01 3.22223455e-01 7.75689900e-01
1.49830878e+00 5.60543835e-01 3.67190421e-01 -9.07828361e-02
1.67798734e+00 -8.58705342e-01 -6.16833329e-01 -2.61434555e-01
5.47325730e-01 -1.18766844e+00 1.20566595e+00 1.56586945e-01
-1.49176526e+00 -3.16325516e-01 -1.12224305e+00 -7.73428753e-02
2.27429107e-01 3.03953707e-01 4.35744882e-01 3.22147489e-01
-1.58218026e+00 -1.97370932e-01 -7.12760091e-01 -4.26565677e-01
1.79419681e-01 3.91610533e-01 -3.42129946e-01 -2.57378817e-01
-6.85977340e-01 9.75951195e-01 3.15132290e-01 -3.47255260e-01
-1.06464243e+00 -1.13202286e+00 -8.99944246e-01 -2.23969817e-01
1.72900483e-01 -1.27977002e+00 1.74852240e+00 -2.08246484e-01
-1.02702713e+00 8.36119890e-01 3.09148848e-01 -3.40118706e-01
5.91547251e-01 1.17232308e-01 -3.92382562e-01 6.85668230e-01
3.30675513e-01 1.10229993e+00 6.86869919e-01 -1.19003391e+00
-1.88135803e-01 -1.12011038e-01 1.82479009e-01 5.48778772e-01
-9.14194509e-02 -1.87360987e-01 -4.82972801e-01 -8.46978068e-01
-9.37219411e-02 -1.09192598e+00 -5.20205677e-01 -7.15300348e-03
-5.20377278e-01 1.19314522e-01 3.94556910e-01 -8.96803558e-01
1.30810785e+00 -1.85785174e+00 -2.61751503e-01 2.16234326e-01
5.45456350e-01 -1.43600404e-01 -1.90532342e-01 6.63188577e-01
-1.79760873e-01 1.68288305e-01 -3.92954014e-02 -1.49695829e-01
-2.97956198e-01 -2.00673848e-01 -3.17332625e-01 -3.73258889e-02
4.44536269e-01 1.09920335e+00 -9.75569963e-01 -6.33621097e-01
-9.02250409e-03 5.69045246e-01 -8.65447283e-01 4.18031394e-01
-1.50832146e-01 4.62678403e-01 -5.18708408e-01 5.77504933e-01
1.93062454e-01 -5.70332885e-01 1.84411462e-02 -2.06923380e-01
1.07912958e-01 1.50368869e-01 -5.49139798e-01 1.88408053e+00
-6.49398088e-01 3.09379369e-01 -1.54254109e-01 2.57359058e-01
8.50335658e-01 4.90572006e-01 6.00237489e-01 -5.06263554e-01
-5.13378251e-03 4.35670139e-03 1.00356946e-02 -8.70105505e-01
9.41466093e-01 -3.58127832e-01 -3.12288672e-01 9.16388392e-01
1.40444905e-01 -5.44972897e-01 1.27087340e-01 8.77122223e-01
1.65714455e+00 -1.27014518e-01 3.00617725e-01 5.84812984e-02
-7.49478191e-02 3.28048259e-01 -4.77795243e-01 1.23035371e+00
1.27968341e-01 1.07615888e+00 1.32498503e-01 3.06324679e-02
-1.08945799e+00 -1.28815174e+00 -5.82802556e-02 7.15342224e-01
-3.56587619e-01 -3.50368530e-01 -3.56072009e-01 -7.44516551e-01
-4.14372265e-01 1.08684146e+00 -6.25438929e-01 -5.22691667e-01
-3.78520399e-01 -4.11139220e-01 4.70519036e-01 5.99946976e-01
-1.13902248e-01 -1.28701508e+00 -7.93805301e-01 3.36670995e-01
-3.67812186e-01 -9.73707318e-01 -9.42062140e-01 3.35598439e-02
-5.93247712e-01 -7.06800759e-01 -1.14835823e+00 -7.09282458e-01
9.50389624e-01 1.13124579e-01 1.09584975e+00 -2.28105001e-02
-6.54195249e-01 1.24156487e+00 -5.09497523e-01 -6.10829890e-01
-1.08826733e+00 -1.97244748e-01 -4.33426321e-01 -6.28970981e-01
2.54680235e-02 -3.67106237e-02 -9.38671649e-01 -3.93882006e-01
-1.44542778e+00 4.95122761e-01 9.81405079e-01 8.64240766e-01
4.93433446e-01 -9.69366491e-01 3.87395322e-01 -7.74135232e-01
1.22314954e+00 -8.02744210e-01 5.69941886e-02 2.92172223e-01
-7.80589461e-01 2.37168416e-01 4.06961627e-02 -3.96505266e-01
-9.35191393e-01 -2.65492886e-01 -5.11241406e-02 -4.58261549e-01
9.93119255e-02 7.04879522e-01 7.95198441e-01 1.90422609e-01
1.08435452e+00 4.08750802e-01 1.13687001e-01 1.30401388e-01
3.54361147e-01 7.25396037e-01 7.44082630e-01 -2.25556180e-01
7.55451500e-01 3.11992854e-01 -4.16121125e-01 -2.59738952e-01
-7.92859554e-01 -3.29658657e-01 6.75335228e-02 -2.72827238e-01
1.00081944e+00 -9.68357980e-01 -1.97440669e-01 -4.18478340e-01
-1.09674788e+00 6.51347861e-02 -4.72252995e-01 5.76187074e-01
-7.17395723e-01 9.42877755e-02 -7.71181941e-01 -6.57857478e-01
-1.06270325e+00 -1.49568987e+00 1.34052062e+00 4.05319005e-01
-7.52855659e-01 -7.46713877e-01 1.18735276e-01 3.83838952e-01
5.45707643e-01 1.78710446e-01 1.10756373e+00 -6.85801446e-01
-5.89662194e-01 -4.03814852e-01 -1.93625271e-01 -5.55531494e-02
2.12835401e-01 -1.79803744e-01 -6.97904944e-01 -2.68954843e-01
-1.97940186e-01 -2.26415232e-01 6.38368070e-01 3.64275724e-01
8.14819455e-01 -4.37769294e-01 -1.70384705e-01 3.18686068e-01
1.24461246e+00 3.32772702e-01 5.49414873e-01 1.50252029e-01
3.53083253e-01 3.74310553e-01 5.51432073e-01 6.36446834e-01
7.10215509e-01 2.87304848e-01 2.50967830e-01 -1.38208032e-01
-5.13046801e-01 -4.02955413e-01 2.34156266e-01 7.46432483e-01
2.68405944e-01 -2.28074297e-01 -1.05693221e+00 5.91810226e-01
-1.41437304e+00 -8.02368164e-01 1.43239349e-01 1.91361690e+00
1.03295612e+00 2.26866025e-02 -2.09060222e-01 -6.14489257e-01
3.17172289e-01 -9.18832347e-02 -3.33024293e-01 -6.16614342e-01
4.11340147e-01 6.75458014e-01 3.96201521e-01 3.61058533e-01
-6.11927450e-01 4.83643234e-01 6.48863983e+00 2.54143894e-01
-1.14329422e+00 1.41115934e-01 5.86020291e-01 -3.53791863e-01
-5.91966629e-01 -2.47439414e-01 -6.25855446e-01 2.19841242e-01
1.14892197e+00 -6.02722943e-01 -5.88260628e-02 5.45722961e-01
3.86639893e-01 -2.93650985e-01 -1.24546456e+00 9.24916327e-01
5.01469493e-01 -1.71883905e+00 5.05083323e-01 3.47489446e-01
5.52621722e-01 -2.45906025e-01 6.98994696e-01 5.01296341e-01
4.14561689e-01 -1.08023620e+00 6.69568121e-01 8.51204038e-01
1.18154085e+00 -4.61663425e-01 6.55631244e-01 -3.48501317e-02
-4.02596921e-01 1.25757396e-01 -5.02860025e-02 7.42695451e-01
2.44016990e-01 1.04595773e-01 -2.09085751e+00 5.22876322e-01
2.69708037e-01 -6.67451918e-02 -6.01380706e-01 9.88831162e-01
-3.72243440e-03 3.11699897e-01 -1.09446503e-01 1.40745882e-02
2.56973386e-01 5.01223564e-01 7.87436068e-01 1.34095216e+00
6.44987047e-01 2.30630249e-01 1.92518115e-01 7.88990974e-01
-7.21844360e-02 2.94399410e-01 -6.65909588e-01 -2.53405690e-01
2.94493377e-01 1.42439640e+00 -5.57938874e-01 -4.17877883e-01
-6.15631579e-04 7.89192677e-01 -1.60040498e-01 2.35058963e-01
-8.31343889e-01 -9.55637991e-02 -4.53179143e-02 4.76284832e-01
2.31093675e-01 -9.55226198e-02 -1.64855495e-01 -6.91631258e-01
-1.68547437e-01 -9.83488202e-01 6.05654597e-01 -1.36422300e+00
-1.04830527e+00 8.88939619e-01 1.75142378e-01 -1.54434633e+00
-8.43053520e-01 -2.28705928e-01 -4.03691113e-01 8.75507116e-01
-1.29862142e+00 -1.46446705e+00 -4.90552366e-01 3.91305029e-01
8.26595068e-01 -2.94221997e-01 1.16490626e+00 -5.32595031e-02
6.05683289e-02 8.01352203e-01 -5.68093121e-01 -1.53285200e-02
9.48170543e-01 -1.22650886e+00 -1.50343478e-01 4.99369293e-01
-9.60545540e-02 9.49834943e-01 8.01075935e-01 -8.79632175e-01
-1.36033320e+00 -1.07490528e+00 5.86134613e-01 -5.45767248e-01
6.38729155e-01 1.77837834e-01 -5.61777949e-01 5.61996579e-01
4.35478240e-01 -4.44949418e-01 1.04433763e+00 -5.26443481e-01
-1.41856194e-01 1.87147647e-01 -1.26939905e+00 9.54833329e-01
5.11392713e-01 -4.20107961e-01 -5.57285964e-01 6.64162815e-01
1.00812113e+00 -1.00749731e+00 -1.16085458e+00 1.13578483e-01
4.10704821e-01 -3.68327260e-01 7.07912982e-01 -5.32239079e-01
1.00558019e+00 -8.31229240e-02 -2.40056235e-02 -1.06571686e+00
-2.83277184e-01 -5.83571553e-01 -5.67060821e-02 5.55282414e-01
8.22273433e-01 -2.07546026e-01 3.59441906e-01 7.81073809e-01
-6.66117132e-01 -7.03704894e-01 -6.78021610e-01 9.05976724e-03
-5.17372340e-02 -4.67119515e-01 1.70850798e-01 5.53685009e-01
-3.47849950e-02 1.58664882e-01 1.44022837e-01 2.78714478e-01
2.91406602e-01 -2.59994745e-01 4.98841196e-01 -6.12453878e-01
-6.20350599e-01 -2.32070312e-01 -1.94893986e-01 -5.04412889e-01
-5.16909301e-01 -1.27011955e+00 -1.46094680e-01 -2.10930729e+00
6.20742500e-01 -3.49184304e-01 -2.93765754e-01 6.78828955e-01
-8.29121023e-02 3.07994127e-01 4.59360093e-01 2.91602194e-01
-3.33190531e-01 1.43385500e-01 1.50547171e+00 -3.64543527e-01
-2.75178373e-01 -1.84520081e-01 -1.48819554e+00 2.41736874e-01
3.83258253e-01 -5.28568506e-01 -6.94401205e-01 -3.42371076e-01
2.03675151e-01 5.13959110e-01 5.29784858e-01 -9.36795712e-01
3.38381708e-01 1.16674379e-02 4.67970937e-01 -7.81910539e-01
3.27964813e-01 -2.11934775e-01 2.38786355e-01 3.87828112e-01
-6.72519386e-01 3.43128562e-01 4.26011026e-01 4.38068599e-01
1.16380945e-01 -3.51895213e-01 4.12483692e-01 -5.77715456e-01
-2.89366066e-01 2.19906643e-01 -5.43658793e-01 5.75533696e-03
1.04144931e+00 -2.17039540e-01 -4.26755309e-01 -8.99819851e-01
-1.04017234e+00 2.26875961e-01 1.87071934e-01 6.49946213e-01
1.15313983e+00 -1.04227507e+00 -1.15638328e+00 2.57376600e-02
5.90702713e-01 1.10939052e-02 5.24801135e-01 8.42804372e-01
-5.83126605e-01 3.88064563e-01 6.78898245e-02 -5.22116840e-01
-1.08611143e+00 6.86267391e-02 1.90267086e-01 -7.26174653e-01
-8.04388583e-01 7.65471518e-01 3.11485887e-01 -4.07516509e-02
-1.57145530e-01 -1.85744643e-01 -1.53979827e-02 -2.23595753e-01
8.09978604e-01 -3.09393167e-01 3.64991814e-01 -1.30732924e-01
-8.28989446e-02 -1.87097207e-01 -7.56398499e-01 -6.91470385e-01
1.37084854e+00 1.86550841e-02 2.92867184e-01 4.96630408e-02
8.80239308e-01 1.73458979e-01 -7.63657093e-01 1.21414207e-01
-3.65997851e-01 1.87870026e-01 3.39065850e-01 -1.59166610e+00
-6.13285840e-01 3.21259499e-01 8.05002451e-01 -2.48767406e-01
8.83197546e-01 4.51781303e-01 6.01314604e-01 8.21754057e-03
1.79232270e-01 -7.68825114e-01 6.36027753e-01 2.12912694e-01
1.38503492e+00 -9.21340644e-01 8.04057121e-02 1.98197529e-01
-1.16119206e+00 1.17168427e+00 4.51980799e-01 7.32261762e-02
2.99431324e-01 5.64926386e-01 4.48122084e-01 -6.68129563e-01
-9.01992738e-01 7.64093548e-02 4.82403010e-01 6.96328759e-01
8.86369288e-01 1.64006099e-01 -3.85920793e-01 4.55414414e-01
-4.90424216e-01 2.31958516e-02 1.11162519e+00 1.18219674e+00
-1.54826581e-01 -7.96997726e-01 -4.59903538e-01 9.20011342e-01
-5.74114501e-01 -4.05835927e-01 -1.99532346e-03 6.31499767e-01
-4.39460367e-01 8.94744098e-01 -2.18826130e-01 -1.62856981e-01
4.03367996e-01 -6.14915155e-02 5.16159117e-01 -1.28731692e+00
-7.70317793e-01 3.84984136e-01 2.00356156e-01 -3.67431760e-01
9.31455269e-02 -5.97691834e-01 -1.43704712e+00 3.30354303e-01
2.89331656e-02 -5.28873317e-02 8.36761415e-01 5.07098794e-01
5.43699801e-01 9.22145963e-01 2.48791248e-01 -4.64807868e-01
-5.57697892e-01 -1.10610199e+00 -1.60537735e-01 1.81920305e-02
2.42909610e-01 -2.74135470e-01 1.32702067e-01 2.36515224e-01] | [15.050920486450195, -1.3866506814956665] |
d5686e23-26c7-4233-a2b2-6de49a7c1f21 | a-similarity-preserving-network-trained-on | null | null | http://papers.nips.cc/paper/9566-a-similarity-preserving-network-trained-on-transformed-images-recapitulates-salient-features-of-the-fly-motion-detection-circuit | http://papers.nips.cc/paper/9566-a-similarity-preserving-network-trained-on-transformed-images-recapitulates-salient-features-of-the-fly-motion-detection-circuit.pdf | A Similarity-preserving Network Trained on Transformed Images Recapitulates Salient Features of the Fly Motion Detection Circuit | Learning to detect content-independent transformations from data is one of the central problems in biological and artificial intelligence. An example of such problem is unsupervised learning of a visual motion detector from pairs of consecutive video frames. Rao and Ruderman formulated this problem in terms of learning infinitesimal transformation operators (Lie group generators) via minimizing image reconstruction error. Unfortunately, it is difficult to map their model onto a biologically plausible neural network (NN) with local learning rules. Here we propose a biologically plausible model of motion detection. We also adopt the transformation-operator approach but, instead of reconstruction-error minimization, start with a similarity-preserving objective function. An online algorithm that optimizes such an objective function naturally maps onto an NN with biologically plausible learning rules. The trained NN recapitulates major features of the well-studied motion detector in the fly. In particular, it is consistent with the experimental observation that local motion detectors combine information from at least three adjacent pixels, something that contradicts the celebrated Hassenstein-Reichardt model. | ['Yanis Bahroun', 'Anirvan Sengupta', 'Dmitri Chklovskii'] | 2019-12-01 | null | null | null | neurips-2019-12 | ['motion-detection'] | ['computer-vision'] | [ 5.25224626e-01 1.05514526e-01 -1.17913134e-01 -2.51519054e-01
3.14978715e-05 -4.50334221e-01 8.42671514e-01 -1.33044809e-01
-8.07359278e-01 5.04453361e-01 1.41259506e-01 -1.68394744e-02
-1.60168305e-01 -5.75058937e-01 -9.29370165e-01 -9.88227785e-01
9.10330340e-02 1.04291186e-01 3.91851008e-01 2.50622332e-02
4.53539997e-01 6.65025532e-01 -1.51711071e+00 1.43664107e-01
3.31980884e-01 4.92475033e-01 4.86639589e-01 1.00835466e+00
2.74173498e-01 1.03329480e+00 -3.51428017e-02 2.44512577e-02
4.45029616e-01 -9.66352940e-01 -9.49301541e-01 2.22459272e-01
3.67802590e-01 -5.22187017e-02 -3.55280876e-01 1.16250539e+00
3.49026523e-03 4.26959813e-01 1.04727435e+00 -1.05474353e+00
-6.69802010e-01 2.50709713e-01 -4.10619497e-01 3.16860199e-01
-2.87319887e-02 1.19907066e-01 8.51577878e-01 -7.73154378e-01
1.03308368e+00 1.17281604e+00 3.71720314e-01 8.56311858e-01
-1.60700309e+00 2.01388240e-01 -1.33679524e-01 2.50927836e-01
-1.28735709e+00 -5.48765719e-01 5.89185774e-01 -6.03224635e-01
7.56175041e-01 2.26504534e-01 6.40271723e-01 9.14203703e-01
4.71306145e-01 6.70651317e-01 7.52015710e-01 -7.92863131e-01
4.32146430e-01 -2.28419811e-01 -3.58725339e-02 9.33069348e-01
3.77234071e-01 1.84769377e-01 -5.77586114e-01 1.62254781e-01
1.32514298e+00 -5.79913296e-02 -3.70898485e-01 -8.56947362e-01
-1.57283247e+00 7.88723230e-01 4.92013991e-01 6.61502481e-01
-3.14914048e-01 3.59000951e-01 -5.70716411e-02 1.43978551e-01
-7.22886026e-02 3.53120148e-01 -1.18170045e-01 3.61758918e-01
-1.03974199e+00 -7.29721934e-02 7.13398337e-01 4.33997482e-01
9.11655605e-01 2.74596661e-01 2.38043040e-01 3.36454451e-01
5.83641469e-01 3.18072975e-01 7.53542900e-01 -1.47292662e+00
-2.52424508e-01 3.00941855e-01 -9.13638505e-04 -1.08388376e+00
-4.39284712e-01 -1.12281956e-01 -9.78362620e-01 7.64123261e-01
1.00403833e+00 1.18643641e-01 -5.42082012e-01 2.03541589e+00
1.80367976e-01 1.80022106e-01 9.59001258e-02 1.07875407e+00
2.81662256e-01 6.52739167e-01 1.40258614e-02 -5.23920357e-01
1.05098975e+00 -7.94652462e-01 -4.37363088e-01 -3.16142827e-01
6.01938784e-01 -3.29646796e-01 8.45451057e-01 3.32546115e-01
-1.07251346e+00 -6.36844754e-01 -1.26719534e+00 -2.57855326e-01
-1.72571123e-01 4.13062535e-02 5.64137220e-01 1.90299347e-01
-1.21984303e+00 8.84890795e-01 -1.11081660e+00 -1.00117910e+00
-1.12440693e-03 3.58367175e-01 -5.15956998e-01 4.40148354e-01
-6.48832142e-01 1.01803625e+00 5.90413451e-01 9.13354084e-02
-1.02278435e+00 -3.54913622e-02 -6.97560132e-01 -2.14703307e-01
9.64765400e-02 -1.19932663e+00 1.12085140e+00 -1.50757921e+00
-1.57900536e+00 1.25640392e+00 -4.97706264e-01 -7.76105642e-01
4.24353659e-01 1.58672109e-01 -8.41416121e-02 3.63650650e-01
-1.87679470e-01 1.09205437e+00 1.04521418e+00 -1.08940995e+00
-3.47656399e-01 -2.18660817e-01 -1.21260799e-01 1.66849896e-01
-1.81265563e-01 -7.93603137e-02 -2.93168146e-02 -6.93405747e-01
3.20769966e-01 -7.96836317e-01 -3.66576850e-01 5.77034771e-01
-1.00011103e-01 3.14736776e-02 3.93808246e-01 -4.89373893e-01
8.34660649e-01 -2.02864003e+00 5.25145710e-01 -2.30518728e-02
2.93107957e-01 1.62441295e-03 -2.15514809e-01 1.55421048e-01
-1.83915734e-01 -1.96450204e-01 -4.91672486e-01 1.63654730e-01
-1.51762977e-01 2.98218936e-01 -1.89767286e-01 8.19965243e-01
1.99951768e-01 9.66110229e-01 -8.56830478e-01 -5.06476760e-01
3.22259933e-01 3.13347220e-01 -5.56152701e-01 -1.91830508e-02
-2.94546783e-01 6.90004230e-01 -2.08099693e-01 1.72895044e-01
1.86703846e-01 -3.18924993e-01 1.77218318e-01 -1.17853113e-01
-3.00164431e-01 -1.86676338e-01 -9.83561099e-01 1.88422930e+00
9.86253619e-02 1.08059990e+00 3.22989412e-02 -1.54724979e+00
7.97697842e-01 2.30108976e-01 5.60844064e-01 -3.22649121e-01
3.12198669e-01 1.21017620e-01 1.41712338e-01 -5.75045288e-01
1.24352835e-01 -2.17433855e-01 3.55998576e-01 4.93392736e-01
3.01627755e-01 -3.56467371e-03 2.47527361e-01 -4.50432971e-02
1.08680677e+00 3.99318546e-01 8.35788310e-01 -4.48063970e-01
6.38890743e-01 -8.05756673e-02 5.51617980e-01 8.19594622e-01
-3.45400900e-01 7.05934405e-01 1.66870549e-01 -6.79311931e-01
-1.23060012e+00 -1.18516481e+00 1.93944871e-02 7.92173207e-01
1.55204132e-01 2.18913555e-02 -1.03371847e+00 -1.48795888e-01
-3.57368320e-01 4.74724859e-01 -7.58903563e-01 -2.62001365e-01
-6.07252717e-01 -5.48027396e-01 2.95024365e-01 1.47547901e-01
4.15854007e-01 -1.29996622e+00 -1.17430568e+00 1.72341064e-01
-5.59819415e-02 -9.37059104e-01 -2.57448852e-01 2.65828282e-01
-1.07584858e+00 -9.23926950e-01 -8.13326061e-01 -9.86825466e-01
8.23526919e-01 3.60122085e-01 7.38230586e-01 -9.32392031e-02
-3.96539807e-01 3.61279756e-01 3.94913927e-02 -1.56261295e-01
-6.05874598e-01 -3.21639061e-01 4.25092816e-01 1.66602120e-01
3.78639936e-01 -6.07104838e-01 -5.08260310e-01 3.27043712e-01
-1.06090212e+00 1.49929330e-01 5.40058315e-01 7.40537643e-01
7.52935290e-01 -2.31046706e-01 1.31599262e-01 -3.16085190e-01
3.33443396e-02 -1.43185243e-01 -6.87943459e-01 2.93489963e-01
-2.46868312e-01 5.29293716e-01 5.76677501e-01 -6.17549360e-01
-9.19080555e-01 7.98646927e-01 2.10125417e-01 -2.55555719e-01
-3.84348929e-01 2.46307477e-01 -1.28114596e-01 -3.07171494e-01
9.94523287e-01 5.17924607e-01 -1.31173711e-02 -5.87929934e-02
5.45480251e-01 1.25137553e-01 1.00483358e+00 -1.84429139e-01
8.01344991e-01 9.08894658e-01 3.53533894e-01 -1.25945735e+00
-4.54210311e-01 -3.19737703e-01 -9.92587268e-01 -3.52910578e-01
1.20541978e+00 -4.83070880e-01 -6.74058557e-01 4.11520541e-01
-1.34763122e+00 -3.28901201e-01 -5.37281692e-01 8.85249078e-01
-1.34631431e+00 5.00064313e-01 -4.82395202e-01 -8.11454237e-01
1.68037742e-01 -6.77377999e-01 6.00570142e-01 2.74867594e-01
-4.49733436e-01 -1.16318858e+00 3.75553548e-01 -7.75334165e-02
8.66692364e-02 3.34094703e-01 7.63039827e-01 -3.23889047e-01
-7.13194847e-01 -2.38109217e-03 8.03229287e-02 1.39421031e-01
6.88048378e-02 2.74865687e-01 -1.01170099e+00 -5.42938076e-02
5.95839202e-01 -1.13634460e-01 1.14904761e+00 9.21233237e-01
7.48156190e-01 -2.04494119e-01 -2.73304135e-01 6.81557417e-01
1.44031715e+00 2.96998620e-01 5.22660255e-01 3.69045049e-01
5.42443454e-01 8.57366443e-01 1.01002358e-01 1.93526633e-02
-7.90699124e-02 5.19744396e-01 3.26305687e-01 1.01381429e-01
-1.36976346e-01 -3.02240551e-01 6.35604858e-01 7.55685747e-01
-3.87018263e-01 -4.98939157e-02 -7.34446228e-01 3.55823547e-01
-2.18294787e+00 -1.30472612e+00 -2.49826405e-02 2.48755717e+00
6.16096258e-01 1.44977704e-01 1.80434808e-01 6.75703287e-02
7.95664608e-01 7.02754930e-02 -5.07688403e-01 -3.11279148e-01
-5.75287580e-01 -2.85329074e-01 4.93732184e-01 7.10870683e-01
-1.11777890e+00 8.78468931e-01 6.74348164e+00 4.43882942e-01
-1.02986038e+00 -7.00838789e-02 4.82268095e-01 1.36293203e-01
5.50809428e-02 2.09833860e-01 -5.49941003e-01 9.74368230e-02
9.72620666e-01 -4.50651139e-01 6.14889324e-01 5.06070077e-01
6.41953528e-01 -3.32687587e-01 -1.48700511e+00 1.02876103e+00
1.69828281e-01 -1.41081583e+00 1.95486546e-01 5.34604453e-02
7.22204864e-01 -1.12957835e-01 -1.08850129e-01 -4.20154572e-01
2.83689409e-01 -8.43581438e-01 7.64197946e-01 9.92642701e-01
1.55834809e-01 -2.78388351e-01 1.12857558e-01 7.30943561e-01
-1.05124092e+00 -1.08969603e-02 -6.86028659e-01 -2.88332522e-01
1.96123704e-01 2.75150657e-01 -4.00920898e-01 -7.09716231e-03
2.98895538e-01 8.66852701e-01 -4.37767684e-01 1.22257972e+00
-1.51967615e-01 2.26182431e-01 -3.43025804e-01 -1.35067450e-02
1.01165436e-01 -3.94801468e-01 6.87578976e-01 1.01728427e+00
2.78571427e-01 7.66822100e-02 -2.96386570e-01 1.12690163e+00
6.58723712e-02 1.20703951e-02 -1.01844835e+00 -8.47993717e-02
-2.96409847e-03 1.10407639e+00 -1.18114877e+00 -2.88666844e-01
-3.38994950e-01 1.09736753e+00 3.12697023e-01 4.65019733e-01
-4.53861654e-01 9.24686864e-02 3.62422466e-01 -5.26637712e-04
3.29405040e-01 -4.45932329e-01 -1.80627957e-01 -1.21304917e+00
-2.62682855e-01 -3.80435765e-01 1.51967574e-02 -8.64250720e-01
-9.32070374e-01 2.85428673e-01 4.07028245e-03 -1.42728877e+00
-5.22268772e-01 -8.91245961e-01 -6.37089789e-01 4.26484346e-01
-9.42184806e-01 -8.71454358e-01 9.59361903e-03 5.50698161e-01
5.80112338e-01 -1.61105841e-02 7.09768653e-01 -2.89128453e-01
-3.90378743e-01 1.03538699e-01 2.52683908e-01 4.82289381e-02
4.28969055e-01 -1.17149413e+00 3.30757201e-01 1.22761905e+00
8.09393704e-01 5.73938310e-01 9.52253163e-01 -3.49459976e-01
-1.18388009e+00 -9.00748312e-01 9.62773442e-01 -4.58933860e-01
7.73120642e-01 -2.22313032e-01 -9.42524910e-01 6.65012181e-01
2.73040771e-01 1.94508046e-01 2.06346527e-01 -7.70463645e-01
-1.44695878e-01 5.82223684e-02 -8.08600843e-01 8.56260300e-01
1.15997660e+00 -4.41104114e-01 -7.54851043e-01 3.50427866e-01
2.71596819e-01 1.33632630e-01 -3.54659855e-01 1.01738214e-01
7.23398209e-01 -1.03481328e+00 1.01359022e+00 -6.40104771e-01
2.84793198e-01 -6.54525220e-01 -8.45532268e-02 -1.09964550e+00
-5.74961662e-01 -6.65648997e-01 -7.64200985e-02 6.68067873e-01
1.99401885e-01 -3.87731582e-01 7.37831056e-01 4.06248778e-01
2.48199224e-01 -1.05211549e-01 -8.75797629e-01 -1.01794434e+00
9.38868821e-02 -1.36571765e-01 -4.07424957e-01 9.26879883e-01
1.75262854e-01 2.74858594e-01 -4.01392370e-01 -8.34563896e-02
9.20552075e-01 -1.31065652e-01 6.61116719e-01 -1.16544032e+00
-5.49132168e-01 -6.08710229e-01 -7.95957863e-01 -1.18122804e+00
2.71191925e-01 -8.38839471e-01 3.43612373e-01 -1.26814127e+00
2.72417426e-01 4.83690768e-01 -2.01513678e-01 2.67657399e-01
2.61345059e-01 2.03785390e-01 2.16399446e-01 4.98718917e-01
-5.10988832e-01 2.19697818e-01 1.14905643e+00 -3.28257568e-02
-1.51789367e-01 -1.35792702e-01 -2.53998399e-01 1.29186869e+00
7.92113423e-01 -5.28461397e-01 -4.30915982e-01 -5.08409500e-01
1.22009307e-01 -2.62497608e-02 6.84987903e-01 -1.12753665e+00
7.33697295e-01 -4.27966624e-01 4.34161097e-01 -2.01387227e-01
1.09357461e-01 -7.09340811e-01 3.11223149e-01 8.73960912e-01
-5.67213118e-01 1.95838325e-02 -1.72125533e-01 7.18594372e-01
-1.43909693e-01 -5.51148951e-01 1.12647378e+00 -4.10542250e-01
-1.10320091e+00 -7.69982487e-02 -1.02352798e+00 -1.51370034e-01
9.90109622e-01 -6.55374050e-01 -2.38338619e-01 -3.15643251e-01
-1.00479794e+00 -4.41286206e-01 6.99223876e-01 1.59145132e-01
6.85255945e-01 -1.25276911e+00 -6.08218968e-01 1.78258061e-01
3.28854434e-02 -3.83248538e-01 -1.76577732e-01 8.76858175e-01
-8.86010766e-01 4.85923409e-01 -5.92745900e-01 -6.76298738e-01
-1.08385682e+00 7.82834709e-01 6.13012791e-01 2.20663309e-01
-3.84864211e-01 6.78898335e-01 6.43771470e-01 1.16781384e-01
-4.23840471e-02 -5.31765044e-01 -1.34741202e-01 -3.48920345e-01
6.56381130e-01 2.69848287e-01 -4.38985765e-01 -8.77441525e-01
-2.45502725e-01 8.07490885e-01 3.00878078e-01 -3.79097044e-01
1.21457291e+00 -2.39313483e-01 -1.96151555e-01 7.53038824e-01
1.17021465e+00 -3.37109059e-01 -1.38959050e+00 -3.18072408e-01
3.88017386e-01 -1.21907704e-01 -2.24263817e-01 -1.11452363e-01
-7.95950711e-01 9.89674687e-01 5.57347000e-01 1.32009208e-01
1.11170161e+00 -7.64609948e-02 2.50680894e-01 8.40818703e-01
3.78128946e-01 -1.01221192e+00 1.51224121e-01 3.72738302e-01
8.82095754e-01 -1.26785660e+00 -1.48388982e-01 -1.00331947e-01
-2.89528489e-01 1.30344498e+00 3.68601948e-01 -5.93962073e-01
5.58551610e-01 7.86429718e-02 -8.99685621e-02 1.32848740e-01
-8.71912479e-01 -2.58345068e-01 3.52795005e-01 7.57041693e-01
4.06651825e-01 -2.33942568e-01 -5.23611784e-01 -3.87436822e-02
1.21875502e-01 2.61879414e-01 5.09779751e-01 8.76826167e-01
-8.77597690e-01 -8.10902119e-01 -3.27761441e-01 -7.96330944e-02
-8.98550302e-02 9.11011472e-02 -6.03524566e-01 6.44899249e-01
9.91073437e-03 6.40993357e-01 8.01489204e-02 -1.94926381e-01
-1.43638909e-01 -1.30454767e-02 8.25710833e-01 -4.66915935e-01
9.28718410e-03 2.16008753e-01 -5.22008181e-01 -6.20655954e-01
-1.09980178e+00 -8.27302158e-01 -1.39220214e+00 -8.09016153e-02
1.22946009e-01 -8.00735354e-02 4.80788589e-01 1.07432926e+00
-1.51541606e-01 4.20932956e-02 3.87375027e-01 -1.00388694e+00
-3.07728529e-01 -5.48782110e-01 -7.31394231e-01 5.39138377e-01
4.28300649e-01 -3.83726835e-01 -5.78249931e-01 8.80047500e-01] | [8.945442199707031, -0.3936833143234253] |
02e7e4c1-a561-4595-9657-7d9514f522c6 | deep-bv-a-fully-automated-system-for-brain | 1811.03601 | null | http://arxiv.org/abs/1811.03601v1 | http://arxiv.org/pdf/1811.03601v1.pdf | Deep BV: A Fully Automated System for Brain Ventricle Localization and Segmentation in 3D Ultrasound Images of Embryonic Mice | Volumetric analysis of brain ventricle (BV) structure is a key tool in the
study of central nervous system development in embryonic mice. High-frequency
ultrasound (HFU) is the only non-invasive, real-time modality available for
rapid volumetric imaging of embryos in utero. However, manual segmentation of
the BV from HFU volumes is tedious, time-consuming, and requires specialized
expertise. In this paper, we propose a novel deep learning based BV
segmentation system for whole-body HFU images of mouse embryos. Our fully
automated system consists of two modules: localization and segmentation. It
first applies a volumetric convolutional neural network on a 3D sliding window
over the entire volume to identify a 3D bounding box containing the entire BV.
It then employs a fully convolutional network to segment the detected bounding
box into BV and background. The system achieves a Dice Similarity Coefficient
(DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set
surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of
25%. | ['Jeffrey Ketterling', 'Orlando Aristizabal', 'Jack Langerman', 'Yao Wang', 'Nitin Nair', 'Jonathan Mamou', 'Ziming Qiu', 'Daniel H. Turnbull'] | 2018-11-05 | null | null | null | null | ['brain-ventricle-localization-and-segmentation'] | ['medical'] | [-4.12648842e-02 2.05950871e-01 4.28637594e-01 -1.91840410e-01
-5.13885260e-01 -6.36729658e-01 1.17561929e-01 3.31179887e-01
-6.60314441e-01 4.83101428e-01 -7.54756927e-01 -2.98302114e-01
4.58378851e-01 -7.95862615e-01 -6.87260807e-01 -6.95752919e-01
-1.91345453e-01 7.32472360e-01 5.99966466e-01 1.99372128e-01
2.29697391e-01 7.40791857e-01 -1.02845180e+00 -5.25175035e-02
4.73340750e-01 1.15376604e+00 4.00832593e-01 8.23491395e-01
-3.09119731e-01 1.90285534e-01 -5.65957546e-01 -3.33712809e-02
2.28098840e-01 -5.52339852e-01 -9.24288452e-01 -1.06997423e-01
2.51631588e-01 -7.42340207e-01 6.37764186e-02 7.76935458e-01
7.24147618e-01 -2.84066588e-01 8.83877456e-01 -6.63676560e-01
-8.10234472e-02 3.19633335e-01 -7.89038777e-01 7.52661526e-01
1.08775005e-01 -2.84612507e-01 4.39491749e-01 -8.49153936e-01
9.24275398e-01 5.38686156e-01 6.01768315e-01 7.09291995e-01
-1.31109321e+00 -8.28364015e-01 -5.39030671e-01 -4.60731357e-01
-1.27569926e+00 -1.79624230e-01 3.37106138e-01 -1.15620553e+00
7.11988807e-01 -1.38566092e-01 9.75221932e-01 1.23008288e-01
4.69267637e-01 5.13604701e-01 7.50383973e-01 -1.76607043e-01
4.12719965e-01 -2.88462847e-01 -1.94673538e-02 1.13384545e+00
2.44271997e-02 -3.34218979e-01 -1.18983187e-01 -9.98825505e-02
1.24579453e+00 -2.80593961e-01 -2.92990267e-01 -4.51764435e-01
-9.65621293e-01 8.16321373e-01 7.19096288e-02 4.59897667e-01
-1.67118073e-01 3.95827234e-01 4.86057967e-01 -8.44987035e-02
7.14690149e-01 2.00333983e-01 -3.17669779e-01 2.44592559e-02
-1.24092042e+00 2.05618382e-01 3.97743136e-01 5.56778133e-01
4.22631592e-01 7.79049750e-03 -2.05736719e-02 7.28444278e-01
4.49133158e-01 3.45252842e-01 4.56146300e-01 -9.72468376e-01
1.34633273e-01 5.17303646e-01 -4.15588111e-01 -4.25049156e-01
-7.52245784e-01 -1.26842007e-01 -6.42741323e-01 3.92410904e-01
5.94930351e-01 -5.89158356e-01 -1.07114065e+00 1.32134843e+00
7.98440635e-01 -3.25404964e-02 -4.18261766e-01 9.40228283e-01
1.39933896e+00 4.20454532e-01 6.28761947e-03 -1.85470074e-01
1.17754912e+00 -3.02717745e-01 -3.65100861e-01 1.58854768e-01
8.67512643e-01 -5.02280116e-01 3.37130398e-01 -2.72112656e-02
-1.02101767e+00 8.02389979e-02 -1.04483879e+00 -4.78398763e-02
-1.96653023e-01 2.12198541e-01 4.99529064e-01 6.62577331e-01
-1.24346447e+00 6.54325783e-01 -1.10842001e+00 -3.29868466e-01
1.04055333e+00 6.16057336e-01 -6.69936538e-01 4.36441004e-01
-5.36725581e-01 5.82874179e-01 2.06263334e-01 -1.93577304e-01
-9.77496028e-01 -1.16404366e+00 -9.74601269e-01 -6.84231520e-02
-6.77165464e-02 -4.00547415e-01 1.12038112e+00 -2.20020667e-01
-1.45578849e+00 1.33101189e+00 2.65284181e-02 -4.37040001e-01
7.09365964e-01 8.45636800e-02 4.40043926e-01 6.59918129e-01
2.46730059e-01 7.32767582e-01 5.94034314e-01 -1.02866387e+00
-3.44085932e-01 -5.81412435e-01 -4.53495562e-01 -2.94742167e-01
2.43195668e-01 1.97465762e-01 -2.89214641e-01 -3.59465539e-01
5.43942809e-01 -6.10715091e-01 -5.24513386e-02 3.69537592e-01
3.20853554e-02 -3.88185307e-02 8.55686069e-01 -1.03134501e+00
7.19286382e-01 -1.77470338e+00 -5.04925661e-02 3.02831560e-01
7.57175446e-01 3.84658486e-01 2.38666087e-01 -1.44524306e-01
1.12009421e-01 1.48392841e-02 -5.69311202e-01 -3.46256793e-01
-5.85269928e-01 -2.56835669e-01 2.74949968e-01 9.63708878e-01
-4.94253673e-02 8.22841465e-01 -8.93236101e-01 -8.97550166e-01
2.33148560e-01 5.46354711e-01 -3.68815392e-01 2.44218990e-01
1.82594225e-01 7.84687996e-01 -3.05153757e-01 7.69649684e-01
7.90835738e-01 -1.76197886e-02 -2.99880821e-02 2.15002820e-01
-1.29712954e-01 -5.71044445e-01 -5.13722122e-01 1.66664958e+00
-1.67226583e-01 1.07273555e+00 3.25673014e-01 -1.08235776e+00
9.43725526e-01 3.78897995e-01 8.10735822e-01 -4.38664019e-01
6.46062195e-01 3.21534842e-01 -6.94622621e-02 -6.18546009e-01
-2.53297269e-01 -4.08884943e-01 3.79932940e-01 4.81478482e-01
6.19020164e-01 -3.80830079e-01 4.00223255e-01 1.42699052e-02
1.16244316e+00 2.31877610e-01 1.03902951e-01 -5.99543393e-01
4.70394433e-01 -3.96046013e-01 5.76739371e-01 6.46664202e-02
-5.87669194e-01 1.22221267e+00 8.19204867e-01 -4.84360486e-01
-1.18396032e+00 -9.04991448e-01 -5.84570408e-01 6.42723083e-01
-4.23607305e-02 8.41971040e-02 -1.26865840e+00 -6.21262550e-01
1.29042909e-01 1.34270638e-01 -8.73496890e-01 4.14351881e-01
-5.48358977e-01 -7.40446746e-01 5.39585114e-01 3.88445467e-01
4.91578341e-01 -8.92863810e-01 -1.19939935e+00 1.92251116e-01
1.97426770e-02 -1.20432460e+00 -3.25906016e-02 2.37253327e-02
-8.40298116e-01 -1.18042648e+00 -1.39570940e+00 -8.29731286e-01
9.06830788e-01 -3.42870563e-01 8.91533792e-01 2.72319049e-01
-7.92354286e-01 2.45647989e-02 -1.46783128e-01 -4.55945909e-01
-3.76112133e-01 -3.59123617e-01 -2.60969847e-01 -4.25834358e-01
-2.78392970e-03 -4.65372831e-01 -7.89107680e-01 7.94187784e-02
-6.43584788e-01 -1.21569119e-01 1.37971908e-01 5.60355067e-01
8.24131668e-01 -4.55130607e-01 4.77610737e-01 -9.17027056e-01
2.05784097e-01 -3.65920484e-01 -1.03740275e+00 3.33459564e-02
-6.69144839e-02 -5.22727549e-01 4.04798090e-01 -2.12627277e-01
-6.51060104e-01 2.80673087e-01 -3.59698296e-01 -3.26290935e-01
-2.30897829e-01 2.22250670e-01 4.24686879e-01 -5.92773080e-01
3.96474361e-01 2.02067867e-01 2.11623743e-01 -2.48407394e-01
-9.58529115e-02 4.29007053e-01 4.29282039e-01 -1.56119123e-01
3.91875595e-01 6.22739255e-01 5.11595607e-01 -9.51690733e-01
-2.52043724e-01 -5.11376441e-01 -1.00803006e+00 -6.20948970e-01
1.22499824e+00 -5.43018818e-01 -6.20560348e-01 4.00463402e-01
-1.15064323e+00 -7.10509241e-01 3.05103153e-01 4.06532139e-01
-5.41189790e-01 5.69879889e-01 -8.41837287e-01 -6.48825109e-01
-7.55112529e-01 -1.20672798e+00 1.03761375e+00 2.39616603e-01
-3.02570730e-01 -8.28980327e-01 2.44699582e-01 5.47834277e-01
1.87246770e-01 8.54426682e-01 8.39921772e-01 -5.09264290e-01
-2.64211800e-02 -5.04383862e-01 -2.86543995e-01 2.54949808e-01
-1.44615546e-01 1.71591431e-01 -9.91898656e-01 -1.38028726e-01
-1.46995306e-01 -2.46163025e-01 8.85338724e-01 8.40584338e-01
1.07114708e+00 4.20634687e-01 -3.70532781e-01 8.09662044e-01
1.49680769e+00 5.21410346e-01 1.41063705e-01 2.03028306e-01
5.43170393e-01 6.35803103e-01 5.25832117e-01 6.26060188e-01
-4.02007103e-02 1.21880852e-01 4.11556870e-01 -3.52822363e-01
5.76492809e-02 4.64611590e-01 -3.13125849e-01 1.75295040e-01
-3.50392520e-01 1.75293684e-02 -1.33627474e+00 7.21095502e-01
-1.34706688e+00 -5.51456869e-01 -5.21977805e-02 2.15510702e+00
7.48499751e-01 -2.44242381e-02 1.71610281e-01 1.84723251e-02
7.39898682e-01 -3.65368307e-01 -2.65379488e-01 -4.85279173e-01
2.97321171e-01 3.58549386e-01 4.47104305e-01 2.83382624e-01
-1.33777595e+00 7.53466725e-01 6.31045008e+00 4.62338299e-01
-1.19296491e+00 5.47181629e-02 9.60050702e-01 -1.19610786e-01
2.52055019e-01 -5.41961312e-01 -4.39715862e-01 4.90993828e-01
5.02820969e-01 5.32657504e-02 5.39778359e-02 6.89112186e-01
-8.33251327e-02 -4.96107161e-01 -1.09387803e+00 8.70927036e-01
1.17623890e-02 -1.46695137e+00 -3.67710680e-01 1.20625973e-01
6.46229804e-01 1.89333797e-01 -2.03910634e-01 -3.00850961e-02
-4.08092201e-01 -1.02207696e+00 5.34691513e-01 2.40663156e-01
1.07131779e+00 -8.98647487e-01 1.27765083e+00 4.00720119e-01
-1.04708850e+00 3.51881832e-01 -4.91967112e-01 2.24023208e-01
7.37180188e-02 6.77631497e-01 -1.11884224e+00 -4.69090156e-02
7.99288332e-01 -5.55516360e-03 -2.37390891e-01 1.22660482e+00
2.52109081e-01 5.03887951e-01 -3.53166252e-01 1.69978723e-01
1.13911271e-01 -3.82677346e-01 4.05218273e-01 1.30611837e+00
4.33052450e-01 3.42315912e-01 -3.49190652e-01 1.07539356e+00
-1.29041061e-01 3.72077703e-01 -6.09644353e-01 -7.86786824e-02
1.13786891e-01 1.47017515e+00 -1.56932664e+00 -3.15144271e-01
-3.05837423e-01 6.62659645e-01 3.42796654e-01 -1.42869009e-02
-7.26886392e-01 -4.57849175e-01 1.48020074e-01 2.99525678e-01
4.11650449e-01 -7.58178756e-02 -3.37519705e-01 -6.04188085e-01
-3.18747520e-01 7.54412040e-02 1.12197295e-01 -3.92044395e-01
-5.21827340e-01 4.16017950e-01 -1.10689342e-01 -8.37551236e-01
-6.31570071e-02 -4.34770018e-01 -8.39972675e-01 8.24962497e-01
-1.13982081e+00 -7.67410100e-01 -3.02197963e-01 -1.15053244e-01
3.60898554e-01 -1.43884897e-01 7.97873318e-01 1.78286076e-01
-6.09135091e-01 3.36787641e-01 2.33455095e-02 5.29789507e-01
2.38821983e-01 -1.40329552e+00 1.83891281e-01 6.89299703e-01
-4.21082407e-01 5.61024964e-01 6.46136940e-01 -7.80754089e-01
-8.40971410e-01 -1.06121051e+00 6.12706363e-01 -1.05440326e-01
2.70268053e-01 -3.08475107e-01 -8.36733520e-01 5.77462792e-01
-4.62697484e-02 5.68410218e-01 1.00366592e+00 -6.61024451e-01
2.00148180e-01 3.72181743e-01 -1.61834872e+00 3.11847001e-01
4.45400625e-01 6.56310171e-02 -3.06409359e-01 2.75419354e-01
3.52745384e-01 -7.28188932e-01 -1.23303926e+00 5.93153536e-01
7.58135617e-01 -9.68846738e-01 9.05869305e-01 -9.27810296e-02
8.00515234e-01 -1.04949556e-01 3.56009632e-01 -8.92058790e-01
1.57964695e-02 -3.92112970e-01 3.77249229e-03 1.03249025e+00
3.77820820e-01 -4.39155579e-01 1.12647927e+00 5.68157732e-01
-2.21812829e-01 -1.21805644e+00 -1.18636096e+00 -4.58590508e-01
3.37664127e-01 -1.96631432e-01 2.52626419e-01 6.17873013e-01
1.42916188e-01 -3.64174061e-02 4.55085605e-01 -1.38249069e-01
7.92187154e-01 1.27636552e-01 3.51192951e-01 -1.21061647e+00
3.41892779e-01 -6.12753570e-01 -7.50336647e-01 -3.90936226e-01
4.96474579e-02 -8.94896150e-01 1.99760765e-01 -1.73579180e+00
2.63327748e-01 -1.06469309e-02 7.42439693e-03 2.66259253e-01
2.74689030e-02 5.98728716e-01 -1.60488695e-01 -1.13932490e-01
-3.60726655e-01 2.89077342e-01 1.32528281e+00 1.13167062e-01
-2.29695663e-01 -9.49222688e-03 1.14889704e-01 9.98444855e-01
1.02885139e+00 -3.71326119e-01 -1.61696240e-01 -1.80471018e-01
-2.41390571e-01 2.85389364e-01 8.76910463e-02 -1.08674288e+00
4.65525053e-02 3.57504159e-01 9.22693551e-01 -8.83389354e-01
7.24524856e-02 -6.38758242e-01 -1.88137516e-01 6.92240357e-01
-3.61117385e-02 -4.93544400e-01 2.63764322e-01 1.33121923e-01
6.61667809e-02 -4.66477364e-01 1.18480778e+00 -4.27052081e-01
-2.59083688e-01 4.27988946e-01 -8.18931460e-01 9.58305076e-02
1.34754014e+00 -3.26466590e-01 1.64566096e-02 1.42085835e-01
-6.26650453e-01 2.37823457e-01 3.85652125e-01 -4.67454106e-01
8.48750234e-01 -7.78087735e-01 -6.40706182e-01 2.52249062e-01
-1.87794328e-01 5.47529519e-01 3.29594374e-01 1.00411165e+00
-1.64041626e+00 4.01733935e-01 -3.56974542e-01 -9.07582700e-01
-1.66703534e+00 6.03495575e-02 5.02895534e-01 1.50873587e-01
-8.21517527e-01 1.40427375e+00 3.47793341e-01 -4.56229895e-01
1.02811260e-02 -4.48087841e-01 -6.52286828e-01 1.18125625e-01
4.40187484e-01 4.69322354e-01 3.92476916e-02 -9.60124016e-01
-3.89657885e-01 1.03621233e+00 2.99216896e-01 1.09006859e-01
1.35979712e+00 4.79799733e-02 -3.85283321e-01 2.37678781e-01
1.43902612e+00 -4.84584481e-01 -1.05030513e+00 1.50104061e-01
-2.76112884e-01 -3.28565896e-01 3.41335773e-01 -2.83777237e-01
-1.49471664e+00 1.01121068e+00 6.47954881e-01 1.25542432e-01
7.12451160e-01 1.31452248e-01 1.01848090e+00 -2.50325110e-02
2.83983797e-01 -1.08847272e+00 -2.15198681e-01 3.89755160e-01
7.59869397e-01 -1.32622778e+00 2.35965714e-01 -6.50393665e-01
-2.45206982e-01 1.27611864e+00 6.74116135e-01 -1.63430676e-01
8.38571250e-01 4.26024705e-01 3.57784778e-01 -5.13083756e-01
-3.06585461e-01 4.57281470e-02 2.43331671e-01 6.22246802e-01
6.96222901e-01 -1.54666737e-01 -5.28880537e-01 4.21152085e-01
-1.29288405e-01 2.62578595e-02 4.42010790e-01 1.13598788e+00
-7.56018579e-01 -2.56725401e-01 -1.74479410e-01 6.31091774e-01
-1.11962128e+00 1.68124631e-01 -1.93929642e-01 6.81694806e-01
3.47804755e-01 4.38573122e-01 2.26539239e-01 1.00110695e-01
-1.35673016e-01 4.86857112e-04 6.43457532e-01 -7.78833091e-01
-4.93819028e-01 4.58458602e-01 -3.08680773e-01 -6.52665555e-01
-1.99924842e-01 -5.15592992e-01 -1.91848123e+00 6.35769740e-02
-2.39166349e-01 -1.89279571e-01 1.01769137e+00 1.03115046e+00
-1.65933475e-01 5.67081571e-01 3.83569419e-01 -1.01868021e+00
2.58946002e-01 -7.96106756e-01 -9.47014809e-01 -5.82830086e-02
4.06393677e-01 -6.97476387e-01 -3.59767109e-01 1.82931259e-01] | [14.331670761108398, -2.6734108924865723] |
29e2f54e-571a-4919-bd7b-c4760bae2415 | maximal-multiverse-learning-for-promoting | null | null | https://aclanthology.org/2021.eacl-main.14 | https://aclanthology.org/2021.eacl-main.14.pdf | Maximal Multiverse Learning for Promoting Cross-Task Generalization of Fine-Tuned Language Models | Language modeling with BERT consists of two phases of (i) unsupervised pre-training on unlabeled text, and (ii) fine-tuning for a specific supervised task. We present a method that leverages the second phase to its fullest, by applying an extensive number of parallel classifier heads, which are enforced to be orthogonal, while adaptively eliminating the weaker heads during training. We conduct an extensive inter- and intra-dataset evaluation, showing that our method improves the generalization ability of BERT, sometimes leading to a +9{\%} gain in accuracy. These results highlight the importance of a proper fine-tuning procedure, especially for relatively smaller-sized datasets. Our code is attached as supplementary. | ['Lior Wolf', 'Itzik Malkiel'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['unsupervised-pre-training'] | ['methodology'] | [ 2.49103814e-01 1.36790037e-01 -3.47919852e-01 -6.19230747e-01
-9.21556175e-01 -7.87618697e-01 7.24381387e-01 2.75463432e-01
-7.17391193e-01 7.18538523e-01 -4.70629835e-04 -6.71039581e-01
-1.98688242e-03 -4.75949705e-01 -5.02850235e-01 -5.48159778e-01
-1.15163058e-01 6.11829042e-01 3.09302062e-01 -2.37121172e-02
9.37769040e-02 4.58500862e-01 -1.22101176e+00 1.08240947e-01
7.66436517e-01 9.50203955e-01 -1.09876975e-01 4.80293751e-01
-4.20300737e-02 7.67509401e-01 -4.29839343e-01 -6.38384819e-01
2.86381721e-01 -3.80184166e-02 -7.95783699e-01 3.09921473e-01
-2.72530317e-03 -7.09358752e-02 7.01936707e-03 7.97994316e-01
2.47810796e-01 3.31859291e-01 7.35727847e-01 -1.00579476e+00
-2.23013833e-01 7.05004811e-01 -6.29182637e-01 3.20665866e-01
-1.09475091e-01 1.78440958e-01 1.20288384e+00 -1.03299260e+00
1.58169061e-01 8.76562119e-01 7.67419815e-01 2.99728811e-01
-1.44352341e+00 -8.86268318e-01 2.63501912e-01 -2.97814220e-01
-1.57843173e+00 -5.96684515e-01 7.30035543e-01 -4.66318935e-01
9.39376712e-01 1.31956428e-01 1.60540730e-01 8.25212836e-01
-1.60551026e-01 7.72428095e-01 1.12656724e+00 -7.39043057e-01
3.45855445e-01 4.59435165e-01 5.64266801e-01 4.38467056e-01
1.66612938e-01 4.10674550e-02 -2.67994523e-01 -3.26279432e-01
3.48193169e-01 -2.45863080e-01 2.47661714e-02 -4.08392668e-01
-8.77901196e-01 1.02360654e+00 1.68785930e-01 2.91775584e-01
-7.70157874e-02 -2.49717608e-01 4.12837535e-01 1.89604461e-01
5.70681930e-01 5.54694772e-01 -7.34395683e-01 -1.15239486e-01
-1.10873449e+00 1.04802512e-01 8.76932204e-01 9.20051754e-01
9.75615919e-01 -2.26488426e-01 -1.29674137e-01 1.08802652e+00
1.06006339e-01 2.05675676e-01 6.54875815e-01 -5.66456437e-01
5.81531823e-01 6.37824237e-01 4.34381105e-02 -5.15513837e-01
-5.59716821e-01 -5.68177283e-01 -6.09436452e-01 -1.96734086e-01
4.75120097e-01 -4.05810386e-01 -8.60674322e-01 1.89613283e+00
3.07825685e-01 -4.02869172e-02 -2.10553840e-01 4.79751468e-01
4.02857780e-01 2.23499656e-01 4.14889842e-01 -2.82683611e-01
1.19669580e+00 -1.07580268e+00 -5.52027285e-01 -3.63475621e-01
9.54704106e-01 -6.28914356e-01 1.34700954e+00 3.43861848e-01
-8.14301312e-01 -3.82278383e-01 -9.02206540e-01 -1.16750054e-01
-3.23218912e-01 3.05668026e-01 8.09386075e-01 9.34480727e-01
-8.60108912e-01 4.20043468e-01 -8.54851484e-01 -7.69903883e-02
3.50972652e-01 7.68169403e-01 -3.43133003e-01 6.86661303e-02
-1.21066070e+00 5.97292840e-01 5.90752006e-01 -1.00990564e-01
-4.33255553e-01 -5.15484095e-01 -8.05995941e-01 7.33567625e-02
6.26334012e-01 -3.68884861e-01 1.40657425e+00 -8.49512994e-01
-1.40921426e+00 8.59401047e-01 -2.14289293e-01 -5.38006842e-01
6.20265424e-01 -2.07769677e-01 -3.14423621e-01 -1.66973770e-01
-1.01224057e-01 6.86335921e-01 6.36225700e-01 -1.22765267e+00
-5.73468328e-01 -3.90696585e-01 -3.29851592e-03 1.63150415e-01
-7.85963118e-01 8.34169313e-02 -7.95349002e-01 -8.12218487e-01
-1.06969185e-01 -1.05120087e+00 -2.42982209e-01 -4.39029068e-01
-4.18527693e-01 -2.92461276e-01 5.91853738e-01 -3.50441515e-01
1.59832203e+00 -2.18580151e+00 -2.92920709e-01 4.83348429e-01
3.01748127e-01 4.04219508e-01 -3.80839519e-02 3.51124048e-01
-6.97943494e-02 2.20353067e-01 -3.18719298e-01 -7.13366449e-01
-1.99555263e-01 1.76266119e-01 -2.98036724e-01 2.78706640e-01
3.48458230e-01 8.24533761e-01 -7.11230099e-01 -4.54461485e-01
-9.04121026e-02 7.26768449e-02 -6.00385010e-01 -1.12124672e-02
-2.13933542e-01 3.98509234e-01 -3.93639356e-01 3.89952242e-01
4.72290367e-01 -5.80788910e-01 4.87039000e-01 1.53673097e-01
-4.55838256e-02 5.39072216e-01 -1.31576753e+00 1.12944841e+00
-2.79570788e-01 3.19292694e-01 5.17509766e-02 -8.37242126e-01
7.93163002e-01 1.15058117e-01 5.57797372e-01 -2.82916397e-01
1.81231886e-01 1.82676747e-01 4.48514260e-02 -1.43212467e-01
4.41110760e-01 -1.56445473e-01 1.47237172e-02 5.83656371e-01
4.47564572e-02 9.47250351e-02 3.35986525e-01 1.95769027e-01
9.75292027e-01 -2.16800719e-01 5.03957748e-01 -5.45189500e-01
5.58760405e-01 -9.74599570e-02 5.01538992e-01 9.63815451e-01
-2.58302987e-01 4.49963331e-01 4.92201239e-01 -1.36266559e-01
-8.67578208e-01 -7.69209862e-01 -3.03959042e-01 1.28763509e+00
-1.95277795e-01 -4.73216057e-01 -7.66939700e-01 -1.10016751e+00
1.95889339e-01 7.96235025e-01 -8.12483549e-01 -1.02485515e-01
-5.67110300e-01 -1.00820816e+00 6.77214801e-01 7.51553655e-01
2.97547311e-01 -8.86312962e-01 -2.13670984e-01 -2.69565210e-02
1.11073047e-01 -1.17990220e+00 -4.94805634e-01 8.47705841e-01
-8.66655469e-01 -9.85854685e-01 -4.22563016e-01 -7.47794092e-01
7.22070456e-01 1.63302451e-01 8.58298182e-01 2.56166816e-01
2.74116307e-01 1.82309042e-04 -3.51190269e-01 -3.84965360e-01
-3.42805535e-01 4.52133000e-01 6.46394640e-02 -1.21478535e-01
5.39649487e-01 -4.59591329e-01 -2.24469587e-01 4.66742665e-01
-9.10131574e-01 7.99094737e-02 6.55558109e-01 1.10845697e+00
4.27760750e-01 1.34041697e-01 4.17814344e-01 -1.37745821e+00
5.19941449e-01 -4.44232285e-01 -6.44117951e-01 1.99984401e-01
-8.54157388e-01 1.91897944e-01 7.78300643e-01 -5.54101646e-01
-8.38069677e-01 1.86308727e-01 -1.87779739e-01 -2.21692085e-01
-1.86991915e-01 7.11817861e-01 -1.57261029e-01 -1.40104005e-02
7.01965213e-01 -1.34127215e-01 -1.17790721e-01 -5.98986983e-01
3.46827388e-01 8.41656864e-01 3.76647830e-01 -6.60778642e-01
9.15636301e-01 3.11807454e-01 -3.37413371e-01 -7.31573105e-01
-1.02216744e+00 -7.02867627e-01 -9.14565384e-01 2.29832098e-01
3.52196932e-01 -8.37870955e-01 -4.01987106e-01 2.59831220e-01
-4.98845100e-01 -7.38046527e-01 -2.14845315e-01 6.19273424e-01
-5.07001989e-02 5.05417407e-01 -6.34741306e-01 -6.59446359e-01
-3.03831518e-01 -1.03493702e+00 9.76756454e-01 -6.44509643e-02
-2.93874592e-01 -1.04286623e+00 5.77709451e-02 4.31428850e-01
1.67820036e-01 -4.96697545e-01 9.15043175e-01 -1.34229183e+00
3.97747234e-02 -3.13617468e-01 -2.52702743e-01 3.68675143e-01
1.81579128e-01 -3.43709327e-02 -1.12743235e+00 -4.11466449e-01
-1.24331862e-01 -4.84254003e-01 9.79982972e-01 1.40383154e-01
1.33221066e+00 -1.40742779e-01 -3.56472969e-01 4.79234517e-01
1.06690347e+00 6.60175979e-02 1.87188223e-01 3.82517129e-01
5.95198750e-01 4.74558353e-01 5.48003197e-01 3.94588500e-01
3.33353907e-01 6.98375404e-01 -6.16644472e-02 -1.87142089e-01
2.30750382e-01 -2.08718210e-01 9.99175906e-02 5.89690566e-01
1.77940875e-01 -1.75170004e-01 -1.11174560e+00 2.70495415e-01
-1.66577005e+00 -6.36105955e-01 1.81046292e-01 2.42128730e+00
1.11180103e+00 5.81767201e-01 4.65411842e-01 3.78881574e-01
5.70608020e-01 -1.23893984e-01 -4.84178692e-01 -7.73006678e-02
2.00584429e-04 2.85685569e-01 6.01231217e-01 6.56285942e-01
-1.26089835e+00 1.27799761e+00 7.03120136e+00 9.13437426e-01
-1.31949353e+00 4.07323353e-02 8.70525062e-01 -4.91406135e-02
-2.78671592e-01 2.06877708e-01 -1.00200212e+00 3.38593125e-01
9.38993633e-01 -5.10884225e-02 4.11978364e-01 8.45519781e-01
2.21015960e-01 -3.01982332e-02 -1.03159332e+00 5.50442338e-01
-1.38506323e-01 -1.00720263e+00 -7.65767545e-02 8.64845440e-02
5.92829883e-01 2.38335550e-01 -2.63459861e-01 5.50443769e-01
5.16936600e-01 -8.52678716e-01 7.15319037e-01 -8.68770257e-02
6.95522904e-01 -7.87693441e-01 6.10683382e-01 6.93081439e-01
-1.04824197e+00 -1.79545030e-01 -5.61224706e-02 -3.64872417e-03
-1.19090602e-01 9.24949169e-01 -8.22580755e-01 3.96824896e-01
4.28987920e-01 4.82772082e-01 -8.17996740e-01 7.45833457e-01
-2.94414163e-01 9.93023396e-01 -3.57989162e-01 -5.76201491e-02
1.62707150e-01 -5.20146936e-02 2.19010979e-01 1.57780075e+00
-2.59759009e-01 4.98628244e-02 4.52794552e-01 6.25471056e-01
-2.03536749e-01 2.26324484e-01 -2.18008623e-01 -8.34060684e-02
6.63761020e-01 1.29638863e+00 -8.74330819e-01 -4.43954051e-01
-4.37677681e-01 5.58368981e-01 7.40602672e-01 4.57526892e-01
-8.51056635e-01 -2.51264155e-01 1.33613810e-01 2.83383187e-02
4.12550628e-01 -3.90420556e-01 -6.49317563e-01 -1.10575831e+00
-1.03836261e-01 -9.55353439e-01 6.35817051e-01 -2.19665751e-01
-1.18220973e+00 5.80225348e-01 -5.02628973e-03 -9.62470710e-01
-3.00918490e-01 -6.73461258e-01 -4.48918641e-01 6.27395928e-01
-1.44266224e+00 -1.08150470e+00 -1.45510405e-01 4.65641260e-01
2.66659081e-01 -2.00909423e-03 7.72115707e-01 2.97048777e-01
-9.83334363e-01 1.03988063e+00 1.00191765e-01 2.15148389e-01
8.50570261e-01 -1.25115442e+00 2.11105809e-01 7.72728682e-01
1.58386797e-01 9.53518271e-01 5.29688001e-01 -5.58277845e-01
-9.29736853e-01 -1.06636858e+00 1.07215905e+00 -3.29123497e-01
9.25496221e-01 -7.06249595e-01 -9.72186565e-01 9.59165633e-01
-2.42830932e-01 -9.64777991e-02 8.69951189e-01 6.82643294e-01
-4.27520037e-01 -7.44282007e-02 -9.11160707e-01 5.85162878e-01
8.56113076e-01 -3.20028424e-01 -3.89737248e-01 4.12013561e-01
5.18944979e-01 -2.79117227e-01 -8.32152247e-01 5.06041527e-01
4.26193863e-01 -7.22542942e-01 6.74988687e-01 -5.88077545e-01
2.68103480e-01 -7.26677664e-03 -9.18346196e-02 -1.06433725e+00
-2.98792154e-01 -6.61491156e-01 -3.98831740e-02 1.33349800e+00
7.13476539e-01 -7.98958302e-01 7.95707583e-01 7.95322061e-01
2.61743981e-02 -9.40711915e-01 -6.12008691e-01 -7.08236814e-01
3.01798820e-01 -6.79557502e-01 3.89292777e-01 1.02873063e+00
-1.75903533e-02 6.13531232e-01 -1.93533555e-01 8.48377571e-02
1.74716070e-01 -8.30784664e-02 8.65093589e-01 -1.17054260e+00
-4.32146728e-01 -5.54992139e-01 4.51729484e-02 -1.15646648e+00
2.62023121e-01 -7.91320384e-01 2.16032341e-02 -7.83214092e-01
4.25959200e-01 -8.64948273e-01 -4.77506101e-01 8.88307929e-01
-6.02445483e-01 2.70118445e-01 -1.31041586e-01 5.29590905e-01
-5.78896940e-01 4.73702610e-01 8.54283214e-01 2.00904310e-01
-4.78942007e-01 1.66655794e-01 -9.81754482e-01 7.26770103e-01
8.07515800e-01 -4.93346900e-01 -4.26709175e-01 -2.22373694e-01
-2.60869302e-02 -3.90277714e-01 1.49443045e-01 -8.54387879e-01
9.58191752e-02 -1.85323939e-01 3.61230135e-01 -2.70489812e-01
1.17449157e-01 -5.84888995e-01 -2.09673122e-01 3.02259237e-01
-6.05964661e-01 3.18471417e-02 4.08031851e-01 4.34107423e-01
-1.11711718e-01 -2.38025606e-01 9.40634787e-01 1.85321376e-01
-5.24578452e-01 1.64731339e-01 -1.32605538e-01 1.16373725e-01
7.83927619e-01 -1.20382875e-01 -2.48667020e-02 -3.00215691e-01
-5.72150111e-01 3.49837542e-01 4.13596183e-01 2.26700112e-01
-1.93046138e-01 -8.16151142e-01 -4.19935286e-01 4.21474546e-01
1.97161540e-01 -5.04665263e-02 -1.61500916e-01 9.74248767e-01
-1.03547916e-01 5.52043140e-01 3.65139067e-01 -6.39136672e-01
-1.18123031e+00 6.54984295e-01 3.38077635e-01 -8.07801127e-01
-5.82251012e-01 7.74508774e-01 2.24748492e-01 -6.14902258e-01
3.69804323e-01 -1.67945489e-01 -1.75116464e-01 -4.57419343e-02
3.54811549e-01 1.95859820e-01 3.46229017e-01 -4.54047561e-01
-4.64684993e-01 9.49961767e-02 -5.07268131e-01 -2.64015913e-01
1.11851525e+00 -1.44887477e-01 1.96179539e-01 3.72051150e-01
1.03624904e+00 4.84377861e-01 -1.33200359e+00 -3.63627493e-01
3.18662941e-01 -2.11122498e-01 1.13368869e-01 -9.41642284e-01
-8.47939730e-01 4.95225251e-01 1.18456580e-01 1.61750868e-01
1.19652021e+00 -1.58370078e-01 4.87781823e-01 5.42997777e-01
1.82585150e-01 -1.04639077e+00 -8.63595083e-02 7.24126518e-01
4.46958125e-01 -1.33743441e+00 5.88179529e-02 -5.42472780e-01
-7.62256265e-01 8.63124788e-01 4.59744453e-01 1.14034370e-01
7.61291265e-01 5.68825781e-01 2.27013394e-01 -6.75998628e-02
-7.46665597e-01 -7.61370361e-02 3.52330685e-01 1.32818213e-02
7.39716172e-01 -5.28933071e-02 -4.12167490e-01 7.90706098e-01
-3.17311555e-01 1.02611355e-01 -1.49849411e-02 1.06446588e+00
-3.01277041e-01 -1.29444122e+00 -2.37115711e-01 6.23122215e-01
-4.65146750e-01 -2.28619725e-01 -4.88627970e-01 1.02487159e+00
-7.81162605e-02 1.17333102e+00 -4.75759692e-02 -4.81768548e-01
1.76156476e-01 3.51350486e-01 1.76974148e-01 -7.64686167e-01
-7.17278361e-01 3.68019164e-01 1.33316413e-01 -2.95260757e-01
-1.75855830e-01 -6.80931628e-01 -1.26264310e+00 -1.63327396e-01
-6.89472914e-01 2.76769817e-01 3.32299620e-01 1.24470723e+00
2.69293964e-01 3.14467728e-01 7.07702696e-01 -6.86716497e-01
-9.98593569e-01 -1.17164898e+00 -6.03489220e-01 5.26948273e-01
1.55452654e-01 -7.71932065e-01 -5.19494474e-01 -7.60649815e-02] | [9.463397979736328, 3.643970251083374] |
5bc9aa61-6855-454f-9366-711769dc6f34 | from-images-to-sentences-through-scene | 1511.03292 | null | http://arxiv.org/abs/1511.03292v1 | http://arxiv.org/pdf/1511.03292v1.pdf | From Images to Sentences through Scene Description Graphs using Commonsense Reasoning and Knowledge | In this paper we propose the construction of linguistic descriptions of
images. This is achieved through the extraction of scene description graphs
(SDGs) from visual scenes using an automatically constructed knowledge base.
SDGs are constructed using both vision and reasoning. Specifically, commonsense
reasoning is applied on (a) detections obtained from existing perception
methods on given images, (b) a "commonsense" knowledge base constructed using
natural language processing of image annotations and (c) lexical ontological
knowledge from resources such as WordNet. Amazon Mechanical Turk(AMT)-based
evaluations on Flickr8k, Flickr30k and MS-COCO datasets show that in most
cases, sentences auto-constructed from SDGs obtained by our method give a more
relevant and thorough description of an image than a recent state-of-the-art
image caption based approach. Our Image-Sentence Alignment Evaluation results
are also comparable to that of the recent state-of-the art approaches. | ['Somak Aditya', 'Cornelia Fermuller', 'Chitta Baral', 'Yiannis Aloimonos', 'Yezhou Yang'] | 2015-11-10 | null | null | null | null | ['image-sentence-alignment'] | ['natural-language-processing'] | [ 3.16103190e-01 2.50484616e-01 2.14474201e-01 -6.40759408e-01
-6.71007633e-01 -7.64572620e-01 1.10807729e+00 3.70763630e-01
-7.00135589e-01 6.43307686e-01 4.50663894e-01 -1.22478753e-01
-8.37948397e-02 -5.00829697e-01 -8.67404819e-01 -1.89573228e-01
4.45594758e-01 3.52674037e-01 3.87097061e-01 -4.09323066e-01
5.59297442e-01 4.91159528e-01 -1.74815702e+00 6.17600441e-01
3.00604731e-01 1.05256653e+00 4.60660160e-01 7.00455666e-01
-2.11814135e-01 1.51039112e+00 -1.98690981e-01 -6.95252478e-01
-6.18673535e-03 -4.07367975e-01 -1.28347075e+00 5.11429787e-01
9.64830875e-01 -9.50970575e-02 -1.89724758e-01 1.49058008e+00
2.09907100e-01 1.50621414e-01 5.60248256e-01 -1.31211174e+00
-1.31842172e+00 4.61867750e-01 -1.50105432e-01 9.00672525e-02
8.05571735e-01 2.31519520e-01 7.96989977e-01 -8.37374330e-01
1.19891024e+00 1.20157599e+00 2.61933953e-01 6.04422450e-01
-1.38269293e+00 -1.95464835e-01 -2.08472490e-01 5.45047045e-01
-1.71273458e+00 -4.91346061e-01 8.24126542e-01 -6.10653102e-01
1.24622595e+00 2.94851959e-01 6.21367455e-01 9.20986056e-01
-1.20565683e-01 4.70843762e-01 1.35460234e+00 -8.66634727e-01
3.46633822e-01 6.85755372e-01 1.08111855e-02 6.82420969e-01
3.45769487e-02 -3.08580130e-01 -5.41510463e-01 1.27030611e-01
7.48052299e-01 -4.38508093e-01 -1.68864414e-01 -5.59420705e-01
-1.43947721e+00 5.88673532e-01 6.45640910e-01 4.84055459e-01
-6.91570103e-01 3.73536885e-01 5.63885689e-01 -9.50562879e-02
1.29784688e-01 4.08914626e-01 8.60354863e-03 1.33803889e-01
-8.64193618e-01 2.70428210e-01 7.15385854e-01 1.30691981e+00
1.03314471e+00 -2.61831075e-01 -1.82923913e-01 5.14154673e-01
2.87190437e-01 8.10667753e-01 4.40935224e-01 -1.18746245e+00
1.93810299e-01 6.08489931e-01 3.76057893e-01 -1.46423197e+00
-8.46997835e-03 2.23203763e-01 -4.79835629e-01 -6.78272173e-02
2.10394651e-01 4.93945748e-01 -9.35455978e-01 1.51118803e+00
1.24894641e-01 -1.24040782e-01 4.46526408e-01 1.04048932e+00
1.16422117e+00 3.65817666e-01 1.18802592e-01 6.70614606e-03
1.67084670e+00 -8.45061183e-01 -7.28813589e-01 -3.76620531e-01
1.68776959e-01 -7.63221979e-01 1.13448274e+00 1.68532819e-01
-8.60450268e-01 -7.24010587e-01 -8.80316019e-01 -3.91226381e-01
-8.41899216e-01 1.43368199e-01 1.76107273e-01 3.21244329e-01
-1.46190774e+00 8.52450356e-02 -1.84870020e-01 -9.15884733e-01
3.13927680e-01 -1.37096137e-01 -6.19229436e-01 -3.23942959e-01
-9.60531473e-01 1.15511322e+00 8.14587057e-01 -1.91571742e-01
-1.19986928e+00 -1.81364432e-01 -1.23671377e+00 -3.83421242e-01
4.91206765e-01 -6.85678601e-01 1.05710030e+00 -1.38966823e+00
-1.05887139e+00 1.77983499e+00 1.80401411e-02 -7.69243538e-01
1.86216980e-01 -1.86214283e-01 -2.91901618e-01 7.60924518e-01
3.53749454e-01 1.12655628e+00 6.93083286e-01 -1.57335889e+00
-4.33633089e-01 -2.07650810e-01 5.51447213e-01 1.55566499e-01
-4.99235606e-03 3.01090062e-01 -6.44753516e-01 -2.63470292e-01
-1.87073752e-01 -9.42492485e-01 -1.17830470e-01 5.88052906e-03
-5.89222014e-01 -1.13752402e-01 4.49780494e-01 -9.29381311e-01
8.41818213e-01 -1.95415533e+00 -1.56571101e-02 3.17898244e-02
8.50336552e-02 2.88741291e-01 -2.35651582e-01 5.33375382e-01
-9.18405503e-02 1.63704529e-01 4.38322872e-02 -1.82203069e-01
2.09926039e-01 4.60754633e-01 -4.88682151e-01 2.64748096e-01
1.60047233e-01 9.33982313e-01 -1.32921970e+00 -9.97616410e-01
7.42746115e-01 3.32553327e-01 -1.18771501e-01 2.19726786e-01
-3.43622416e-01 1.77784711e-01 -3.20123762e-01 3.71800989e-01
3.61501485e-01 -2.02346489e-01 6.91130757e-02 -5.33322334e-01
-1.43613189e-01 -8.96287486e-02 -7.57747829e-01 2.01751447e+00
-3.31323206e-01 8.94885182e-01 -5.77034414e-01 -8.38261187e-01
9.05187726e-01 2.22369105e-01 -5.39145898e-03 -9.54515457e-01
1.54400706e-01 2.71471497e-03 -4.89327013e-01 -1.02136528e+00
7.36138046e-01 -9.38194916e-02 -1.42782748e-01 6.13086335e-02
4.11019981e-01 -6.50955439e-01 5.48875630e-01 5.57687521e-01
8.73219728e-01 3.89606357e-01 7.08735049e-01 -3.06845605e-01
9.27029967e-01 5.89202344e-01 6.82722777e-02 8.13689232e-01
-4.05064136e-01 7.41140783e-01 2.68686026e-01 -4.82097387e-01
-1.53853643e+00 -1.00072396e+00 1.05317511e-01 6.87797487e-01
3.24625909e-01 -5.83037615e-01 -1.11351168e+00 -3.82863581e-01
-2.92048544e-01 1.17293489e+00 -8.20584953e-01 1.98745102e-01
-1.62683558e-02 6.83123292e-03 5.71154892e-01 3.51447821e-01
8.57169867e-01 -1.44551480e+00 -9.14010644e-01 -8.45533609e-02
-4.93347019e-01 -2.14725780e+00 -3.03409368e-01 -4.09752727e-01
-9.11108404e-02 -1.18111086e+00 -3.69262695e-01 -8.01885009e-01
7.27819085e-01 2.54651397e-01 1.40284026e+00 1.03997532e-02
-5.54406643e-01 1.07502615e+00 -6.59424424e-01 -5.95833719e-01
-6.94415390e-01 -7.75792539e-01 1.30398124e-02 1.66283339e-01
7.59662509e-01 -1.76303536e-01 -4.02423948e-01 -1.04102055e-02
-1.21050048e+00 4.04915869e-01 4.22231078e-01 3.14475894e-01
7.99778402e-01 1.19048923e-01 2.13927791e-01 -4.59028453e-01
4.99733627e-01 3.91881634e-03 -5.12148857e-01 5.48066795e-01
-3.63420069e-01 1.04815908e-01 2.97760576e-01 -3.07122052e-01
-1.09848678e+00 4.04258966e-01 3.98242056e-01 -7.12445438e-01
-6.46434188e-01 3.47168475e-01 1.15899794e-01 -3.50793377e-02
9.92867947e-01 4.59500939e-01 -6.86135367e-02 8.75361413e-02
7.24948227e-01 7.40100980e-01 1.06340158e+00 -3.88115168e-01
6.81527853e-01 8.13762963e-01 -2.04581209e-03 -9.96902525e-01
-9.25336301e-01 -7.34341741e-01 -1.05116928e+00 -5.13952553e-01
1.49415445e+00 -9.05137181e-01 -3.22937340e-01 7.21974224e-02
-1.47872233e+00 -5.82447136e-03 -4.00683463e-01 3.89574200e-01
-9.97408509e-01 4.37508345e-01 -2.23031953e-01 -9.71414924e-01
-3.52558255e-01 -8.85566771e-01 1.23061645e+00 2.29626253e-01
-1.94554806e-01 -9.01857674e-01 8.87118932e-03 7.96968460e-01
1.75024822e-01 5.04557669e-01 5.62538385e-01 -6.04219437e-01
-3.99348319e-01 -2.09799573e-01 -7.64120936e-01 7.18656659e-01
-6.67691901e-02 6.24930970e-02 -1.07472456e+00 2.59992748e-01
-2.95381606e-01 -5.59861004e-01 4.73616451e-01 1.02119759e-01
5.95116079e-01 -3.34396601e-01 -2.83854902e-02 -8.15493837e-02
2.06657672e+00 -7.12613575e-03 1.01602352e+00 5.86306572e-01
6.38197303e-01 8.73795331e-01 7.97574401e-01 2.27380365e-01
6.77984118e-01 7.07517564e-01 6.83869958e-01 7.31469132e-03
-3.40642780e-01 -2.52055138e-01 3.45262945e-01 4.40267980e-01
-1.63059175e-01 -5.88810220e-02 -1.28570020e+00 1.00114679e+00
-2.18536448e+00 -1.15599835e+00 -2.52906650e-01 1.90513909e+00
8.19078565e-01 -1.10689387e-01 -8.01468790e-02 -2.70986408e-01
7.22753942e-01 6.36284845e-03 -4.04687710e-02 -5.96516848e-01
-2.79309362e-01 -1.80747226e-01 5.58255792e-01 2.57370651e-01
-1.12460458e+00 1.26207352e+00 5.68441439e+00 7.92259753e-01
-6.90568268e-01 1.97327286e-01 2.58870393e-01 2.58844078e-01
-5.27010225e-02 2.26533160e-01 -3.59228849e-01 7.52794594e-02
6.77502811e-01 -2.20537201e-01 6.00965917e-01 8.56746256e-01
2.38568380e-01 -4.78112519e-01 -1.20480132e+00 1.54253316e+00
9.34346914e-01 -1.46713316e+00 3.12308878e-01 -1.65827662e-01
7.83977866e-01 2.02376693e-01 -4.32736754e-01 -3.54608446e-02
3.71338427e-01 -9.10608530e-01 1.10788178e+00 9.54028010e-01
6.99057817e-01 -3.77190888e-01 1.02541459e+00 1.86611652e-01
-9.36652243e-01 2.49279812e-01 -6.46059394e-01 9.88488570e-02
2.22002879e-01 4.38441902e-01 -1.00604093e+00 7.79412746e-01
9.77498293e-01 5.75274825e-01 -1.23920059e+00 6.91547692e-01
-5.88565290e-01 1.88433185e-01 1.09565057e-01 -2.10048914e-01
4.34137195e-01 -6.74111098e-02 6.11484826e-01 1.43157005e+00
-1.21766932e-01 1.48079559e-01 2.38453150e-01 1.24321675e+00
1.40005186e-01 2.28043690e-01 -9.59813356e-01 -2.97688752e-01
1.24053091e-01 1.39655674e+00 -8.46849263e-01 -7.41187274e-01
-2.41809770e-01 1.12237012e+00 2.63467699e-01 2.65188485e-01
-8.33500803e-01 -2.62756735e-01 1.73226923e-01 -6.05760030e-02
1.38224721e-01 -1.24899186e-01 2.80303299e-01 -1.13385296e+00
3.24014761e-02 -6.22217119e-01 7.39512295e-02 -2.15871048e+00
-1.28623712e+00 9.49368298e-01 4.68051404e-01 -1.10001135e+00
-4.56499517e-01 -7.23655641e-01 1.18172597e-02 5.77734649e-01
-1.46435261e+00 -1.54535067e+00 -7.34208703e-01 5.71230948e-01
4.85455543e-01 3.72915007e-02 9.05849814e-01 -7.91803226e-02
8.34683329e-03 -2.75462627e-01 -4.67339426e-01 3.96188796e-01
5.57405233e-01 -1.30812085e+00 1.35059282e-01 9.04929698e-01
6.30436480e-01 4.40080702e-01 1.05709791e+00 -5.58425248e-01
-1.03523254e+00 -1.04902434e+00 9.34728622e-01 -8.58692944e-01
8.25587928e-01 -1.39162168e-01 -6.99805677e-01 6.08438790e-01
6.87878072e-01 1.16603717e-01 4.87774611e-01 -5.81932962e-01
-5.87370634e-01 4.45343778e-02 -1.06699681e+00 6.66117847e-01
9.49541390e-01 -1.03422487e+00 -1.11963606e+00 5.44790030e-01
6.74111485e-01 -2.13634297e-01 -6.71303213e-01 1.03313841e-01
5.20662785e-01 -7.76848733e-01 1.02653456e+00 -6.09583974e-01
6.74678206e-01 -5.45136333e-01 -7.74367034e-01 -8.72300446e-01
-3.17440033e-02 -1.84318811e-01 6.07465923e-01 1.32278204e+00
5.07067293e-02 4.08599563e-02 1.69815108e-01 7.32834816e-01
1.92301452e-01 -1.00834453e-02 -6.16482317e-01 -7.32958138e-01
-5.11042714e-01 -6.85745180e-01 2.25234941e-01 1.05528045e+00
-5.83680868e-02 5.98125756e-01 -1.78925857e-01 1.12381347e-01
7.47080982e-01 -1.46401137e-01 9.57297087e-01 -8.47065747e-01
1.19820677e-01 -8.03242698e-02 -1.16596329e+00 -2.88521320e-01
3.69702041e-01 -8.17946732e-01 2.54474610e-01 -1.97261727e+00
7.23900259e-01 3.08373511e-01 -9.84928012e-02 6.92973673e-01
3.55854124e-01 7.12219059e-01 4.74587262e-01 1.03082061e-01
-1.18159628e+00 1.98697791e-01 1.10956979e+00 -1.37667447e-01
2.08955109e-01 -1.19173110e+00 -4.90677267e-01 9.85816181e-01
5.58290362e-01 -3.62766862e-01 -3.61005187e-01 -1.98273018e-01
5.45046449e-01 -1.83848664e-01 1.09208107e+00 -1.06409347e+00
3.16432565e-01 -3.58633459e-01 3.11783031e-02 -5.69529057e-01
3.13489944e-01 -8.60296488e-01 2.36477390e-01 1.78907290e-01
-4.36696172e-01 -4.39226441e-02 2.66238689e-01 6.64241254e-01
-2.46370420e-01 -4.60698694e-01 5.16993165e-01 -4.51381356e-01
-1.50836480e+00 -4.45282310e-01 -2.38118201e-01 6.09558891e-04
1.13258076e+00 -4.18676168e-01 -4.98994887e-01 -6.69268370e-01
-6.92012131e-01 -2.77229883e-02 6.93820596e-01 5.01357734e-01
9.19239223e-01 -1.37133920e+00 -6.53557122e-01 -3.61898869e-01
7.53518879e-01 -1.50227249e-01 1.86765805e-01 7.56103575e-01
-9.49636877e-01 6.64234102e-01 -5.05224466e-01 -6.21130407e-01
-1.47001731e+00 1.04741907e+00 1.79939628e-01 7.74873346e-02
-5.90709150e-01 3.78412843e-01 1.49018511e-01 -1.88520268e-01
-2.15477094e-01 -2.76210129e-01 -4.32756096e-01 -1.62882149e-01
6.02314591e-01 -4.18866202e-02 -2.29207069e-01 -1.43561041e+00
-6.93624139e-01 6.31802022e-01 2.91927725e-01 -4.98189539e-01
1.06006157e+00 -5.54515064e-01 -4.08769876e-01 4.89304066e-01
9.84685302e-01 -1.92007661e-01 -7.26200521e-01 -3.47953409e-01
7.61750117e-02 -3.67913485e-01 -1.09552620e-02 -8.68900895e-01
-5.00378847e-01 6.88604951e-01 5.19052923e-01 3.53446871e-01
1.17147827e+00 3.48941356e-01 1.56247586e-01 5.61709702e-01
7.31579423e-01 -1.25959897e+00 3.88158768e-01 3.09835762e-01
1.35638559e+00 -1.46025884e+00 1.65274277e-01 -5.26652455e-01
-1.13165462e+00 1.20260143e+00 2.57274389e-01 -2.77157098e-01
4.27637577e-01 -3.61331195e-01 8.59811455e-02 -5.60032487e-01
-5.38843036e-01 -8.56274426e-01 6.82406962e-01 8.13557744e-01
2.82885749e-02 3.42081301e-02 -1.17412627e-01 4.43587303e-01
-7.08632991e-02 2.55701840e-01 8.43144894e-01 7.71080256e-01
-4.11982834e-01 -6.26042008e-01 -3.65277708e-01 -1.89561546e-01
-2.44125441e-01 -3.37559700e-01 -8.31666708e-01 9.89142597e-01
2.74883240e-01 1.20720398e+00 1.54501408e-01 -2.80638725e-01
2.56467521e-01 -5.88492267e-02 6.71305060e-01 -8.68546724e-01
-2.80234039e-01 -3.47287565e-01 4.07985091e-01 -7.57740319e-01
-1.28034675e+00 -4.49806541e-01 -1.28053868e+00 2.40308434e-01
-1.37625605e-01 -1.27891973e-01 1.02115571e+00 1.17032015e+00
2.59957939e-01 3.61701101e-01 1.68646290e-03 -8.39330792e-01
-3.66146415e-02 -7.71731496e-01 -5.21743357e-01 1.13587546e+00
-2.52588335e-02 -3.54288936e-01 -2.99162477e-01 9.13900495e-01] | [10.811554908752441, 1.266785740852356] |
cfd332bc-64f8-472b-bfe9-fa779565854d | wildfire-detection-via-transfer-learning-a | 2306.12276 | null | https://arxiv.org/abs/2306.12276v1 | https://arxiv.org/pdf/2306.12276v1.pdf | Wildfire Detection Via Transfer Learning: A Survey | This paper surveys different publicly available neural network models used for detecting wildfires using regular visible-range cameras which are placed on hilltops or forest lookout towers. The neural network models are pre-trained on ImageNet-1K and fine-tuned on a custom wildfire dataset. The performance of these models is evaluated on a diverse set of wildfire images, and the survey provides useful information for those interested in using transfer learning for wildfire detection. Swin Transformer-tiny has the highest AUC value but ConvNext-tiny detects all the wildfire events and has the lowest false alarm rate in our dataset. | ['A. Enis Cetin', 'Hongyi Pan', 'Tianxiao Ye', 'Yifei Zhao', 'Emadeldeen Hamdan', 'Ziliang Hong'] | 2023-06-21 | null | null | null | null | ['transfer-learning'] | ['miscellaneous'] | [ 2.98351049e-01 -5.64857602e-01 -1.23552263e-01 -3.45940083e-01
-2.17500359e-01 -6.79867744e-01 4.60557520e-01 -1.87169522e-01
-8.75875235e-01 4.67023313e-01 2.04988331e-01 -6.13251328e-01
-2.38968194e-01 -1.15138686e+00 -3.74162376e-01 -6.47810817e-01
-7.65187621e-01 -6.67672306e-02 3.96263063e-01 -3.69852781e-01
-1.38075247e-01 7.80094087e-01 -1.30217028e+00 2.42390350e-01
1.31544828e-01 9.00637925e-01 4.86026518e-02 1.17184973e+00
6.03434980e-01 1.00422227e+00 -6.96521401e-01 1.75939769e-01
8.35087001e-01 1.43140271e-01 -3.53831828e-01 -6.12657309e-01
9.50504363e-01 -1.01360631e+00 -5.60470104e-01 9.52591658e-01
4.97498572e-01 3.31969485e-02 5.74809074e-01 -9.70388055e-01
-4.01437849e-01 4.91497278e-01 -5.85388780e-01 1.15837812e+00
-2.76082516e-01 6.96520567e-01 6.60164714e-01 -6.86289310e-01
1.65476382e-01 1.17306888e+00 1.37463474e+00 1.34736255e-01
-1.01599824e+00 -1.07917809e+00 -1.48029670e-01 -3.50021087e-02
-1.45142710e+00 -2.82105684e-01 1.90606505e-01 -4.73336786e-01
1.50407457e+00 2.30690464e-02 6.80677414e-01 1.27906919e+00
3.91187608e-01 8.17088410e-02 1.27098155e+00 -2.72113923e-03
1.37214214e-01 -6.97201908e-01 3.14572603e-01 7.48903394e-01
5.97122312e-01 1.07103646e+00 -1.82652026e-01 -2.69600213e-01
1.17437935e+00 6.88153028e-01 -4.22862619e-01 2.97368407e-01
-1.05983615e+00 1.01043773e+00 1.32247794e+00 1.57196239e-01
-6.21482372e-01 2.44813919e-01 3.47575366e-01 3.89357328e-01
5.47731042e-01 3.48712146e-01 -5.99491894e-01 4.42617893e-01
-1.18570650e+00 2.68858463e-01 4.36079025e-01 4.32904601e-01
6.53220534e-01 2.28924245e-01 5.18096723e-02 7.33516395e-01
3.29999775e-01 1.19518614e+00 -1.74697056e-01 -8.42291236e-01
2.19031528e-01 3.56946230e-01 1.34976789e-01 -9.39503968e-01
-7.18010426e-01 -4.48389709e-01 -1.14668274e+00 9.46801186e-01
2.33730510e-01 -5.42550206e-01 -1.50265193e+00 1.02655494e+00
-3.73738676e-01 4.01847780e-01 2.85298396e-02 1.03715825e+00
9.42626953e-01 8.42883229e-01 5.24184823e-01 2.31414795e-01
1.12270355e+00 -7.11277246e-01 -2.07499027e-01 -8.72034609e-01
-3.73341702e-02 -2.00546846e-01 6.92731440e-01 -4.42695469e-02
2.27252282e-02 -4.83965456e-01 -1.19081831e+00 4.67672408e-01
-6.90826595e-01 1.62361294e-01 6.67162418e-01 3.16197008e-01
-1.13414693e+00 4.46392059e-01 -7.47556984e-01 -8.63160491e-01
6.48168087e-01 -5.30632734e-02 -2.76454896e-01 -1.25342131e-01
-1.05464053e+00 9.94985700e-01 4.04639035e-01 8.14301908e-01
-1.46249604e+00 -5.77464342e-01 -6.22483075e-01 9.27460194e-02
-1.68128684e-02 -6.70517325e-01 1.05131197e+00 -5.99738002e-01
-7.31956542e-01 9.13474917e-01 5.75812817e-01 -8.71637940e-01
2.53378808e-01 -1.38826743e-01 -5.99073172e-01 3.24191228e-02
2.70771563e-01 7.82464147e-01 9.57359433e-01 -1.14469028e+00
-1.11199665e+00 -3.56505126e-01 4.86302316e-01 -1.04197115e-01
1.91822529e-01 3.63215894e-01 7.24480689e-01 -6.89364552e-01
-1.80087894e-01 -7.36002922e-01 -4.99369085e-01 1.18075520e-01
-1.43647939e-01 2.54908442e-01 1.01854324e+00 -4.54376131e-01
8.46771896e-01 -2.04131031e+00 -6.86859369e-01 4.86347117e-02
2.35205501e-01 7.91490018e-01 -2.94518888e-01 3.23945880e-01
-6.44203275e-02 2.95144081e-01 -4.51346457e-01 7.85224319e-01
-5.54256082e-01 4.06692177e-01 -5.64743519e-01 4.39832956e-01
1.46568134e-01 5.09303749e-01 -1.00988030e+00 -8.07804540e-02
5.05355299e-01 4.54605758e-01 -2.33967882e-02 3.70047688e-01
-6.06340170e-02 2.02737272e-01 -2.02060148e-01 8.25881958e-01
8.39113593e-01 1.23255081e-01 -3.74097764e-01 -7.63643384e-02
-4.33631539e-01 -2.35445067e-01 -7.60251343e-01 9.02269006e-01
-3.57826531e-01 7.17231214e-01 9.16200504e-02 -4.44061160e-01
9.95874166e-01 1.56382412e-01 6.85471147e-02 -4.78121996e-01
4.34290394e-02 -7.27178752e-02 -3.08257878e-01 -6.76606059e-01
1.89724833e-01 -3.66152853e-01 5.38745150e-02 2.44048685e-01
-2.90389135e-02 3.91287565e-01 -1.26770437e-01 -3.02176297e-01
1.73839056e+00 -2.15702504e-01 2.52058029e-01 -1.55681878e-01
-8.37114453e-02 6.55004025e-01 2.91458696e-01 1.18789279e+00
-6.45517468e-01 5.22737145e-01 -3.37352633e-01 -1.41179883e+00
-6.70539141e-01 -1.56156492e+00 -1.85111240e-01 1.32288980e+00
-9.36587062e-03 4.34565842e-02 -8.20299760e-02 -8.28808188e-01
-4.66524586e-02 4.57800448e-01 -7.92281508e-01 -9.05009732e-02
-2.39538819e-01 -1.00086975e+00 1.20609403e+00 8.91861200e-01
1.16877556e+00 -1.51505363e+00 -1.35414076e+00 -2.49374136e-02
-1.24823213e-01 -9.15735126e-01 1.49634451e-01 8.96648288e-01
-9.84179616e-01 -1.43515646e+00 -4.10581619e-01 -7.60711253e-01
1.54334083e-01 8.89942884e-01 1.22903275e+00 4.17741202e-02
-4.57763553e-01 1.31431311e-01 -4.08222109e-01 -6.05324805e-01
6.85961843e-02 1.57499924e-01 -3.23204935e-01 -5.19809127e-01
8.08834016e-01 -8.19269538e-01 -6.16509199e-01 -4.77671549e-02
-7.84164786e-01 -5.98387003e-01 7.83986211e-01 6.66961908e-01
1.35360703e-01 3.00193220e-01 -6.02625087e-02 -3.92344415e-01
4.90231425e-01 -5.93904376e-01 -7.50483751e-01 -7.21059740e-02
-3.27352524e-01 -1.39744699e-01 5.31148434e-01 -1.62796319e-01
-9.43076551e-01 3.69446218e-01 1.46178275e-01 -5.72770655e-01
-7.27133334e-01 5.65822780e-01 5.61996698e-01 4.57356535e-02
1.34107399e+00 -2.17377052e-01 -5.59606075e-01 -4.73648548e-01
6.05687238e-02 7.36322522e-01 1.10314703e+00 2.43679509e-01
1.25840282e+00 6.64546192e-01 -1.60121977e-01 -1.19923687e+00
-1.09779572e+00 -7.98553467e-01 -7.40996361e-01 -3.14172417e-01
9.30166662e-01 -1.16592765e+00 -3.63645494e-01 7.77500272e-01
-1.09308136e+00 -5.38393259e-01 -8.81556794e-02 4.74234521e-01
1.47728413e-01 -3.32277834e-01 -4.06242520e-01 -8.10067356e-01
-8.70056510e-01 -3.35255384e-01 1.12515473e+00 4.25410748e-01
3.12073529e-01 -6.78057790e-01 7.18399167e-01 -2.77030915e-02
8.41003418e-01 5.62314212e-01 3.37664545e-01 -7.67122507e-02
-3.45102847e-01 -2.59443402e-01 -6.63514972e-01 3.49849820e-01
1.07856467e-01 2.92182267e-01 -1.19935548e+00 -1.08988866e-01
-4.48389292e-01 -2.38257796e-01 1.78760862e+00 8.14673960e-01
6.54710352e-01 -2.47878164e-01 -3.29268634e-01 9.63887870e-01
1.81931567e+00 4.09737825e-02 6.85679317e-01 7.51277924e-01
4.91309643e-01 -7.82741904e-02 2.03943938e-01 3.60650897e-01
1.60079777e-01 9.08003822e-02 1.17276597e+00 -5.13930202e-01
2.20207330e-02 -1.70886308e-01 5.75652003e-01 -4.64046896e-01
-6.97052777e-01 -2.61492599e-02 -1.15952873e+00 5.49883485e-01
-1.78876042e+00 -1.84236073e+00 -2.32022315e-01 1.95879889e+00
1.55632138e-01 -3.94448340e-02 1.62108868e-01 -1.89557225e-01
7.39555597e-01 7.68285990e-01 -2.93695390e-01 -1.82720095e-01
-2.56232142e-01 6.36478901e-01 1.32821965e+00 3.99912894e-01
-2.01717019e+00 1.17406642e+00 7.93928623e+00 -4.06897366e-02
-1.30246186e+00 1.07102305e-01 1.95442364e-02 -8.93666297e-02
6.68561220e-01 8.74188021e-02 -5.93105078e-01 -1.03538640e-01
1.04706383e+00 4.29843247e-01 6.46830440e-01 8.68814945e-01
5.05553007e-01 -3.16034794e-01 -3.15736175e-01 5.32241702e-01
-3.21435004e-01 -1.15519595e+00 -1.67693287e-01 -8.19778144e-02
3.34578186e-01 1.12761497e+00 -3.90140772e-01 5.34644604e-01
9.95923638e-01 -1.09669852e+00 3.25777292e-01 3.42460275e-01
5.78498125e-01 -4.75372970e-01 9.41148341e-01 2.25968599e-01
-1.24167264e+00 -3.89827490e-01 -7.74354517e-01 -6.68829679e-01
-1.29621968e-01 5.60198426e-01 -8.11698496e-01 -1.10825762e-01
1.46385527e+00 8.92686367e-01 -8.13317060e-01 1.21604228e+00
-5.30368209e-01 1.00564027e+00 -6.47150338e-01 3.11659425e-01
6.54317141e-01 7.69352727e-03 5.20338535e-01 1.50872612e+00
4.54658009e-02 8.55144709e-02 4.79602635e-01 7.11684167e-01
8.71974975e-02 -6.89226389e-01 -1.38030601e+00 3.05500329e-01
3.66622537e-01 1.46404028e+00 -5.00843525e-01 -1.58936396e-01
-1.46320507e-01 5.74060440e-01 -8.06688983e-03 4.43406731e-01
-9.14377809e-01 -3.78468931e-01 8.60900044e-01 1.60296604e-01
4.47376758e-01 -2.96982944e-01 6.40525222e-02 -8.63269210e-01
-3.81804854e-01 -4.39006239e-01 7.91287065e-01 -1.19262719e+00
-1.45171952e+00 7.78125286e-01 2.27713645e-01 -1.09941149e+00
5.27008697e-02 -8.56934845e-01 -1.05985975e+00 6.69658184e-01
-1.72324371e+00 -1.52005315e+00 -1.01188421e+00 6.64166868e-01
2.67381310e-01 -1.23462915e-01 1.14572597e+00 7.04746619e-02
-4.88294065e-01 -1.44640476e-01 -1.71350420e-01 6.90214157e-01
5.37918746e-01 -1.05864120e+00 6.19977355e-01 1.28766012e+00
7.12031946e-02 2.46336654e-01 4.02579665e-01 -6.09886825e-01
-8.21041107e-01 -1.92623019e+00 6.12911463e-01 -1.84863925e-01
5.81262887e-01 2.30195880e-01 -5.78833938e-01 1.15546560e+00
2.03415215e-01 4.56728101e-01 2.40661919e-01 5.67172319e-02
-9.44360495e-01 -3.28585207e-01 -1.31729496e+00 1.96546733e-01
1.04453206e+00 -4.27925706e-01 -6.64445043e-01 4.34350133e-01
1.27604827e-01 3.92803550e-03 -5.26495993e-01 6.20980680e-01
7.20111489e-01 -1.08278286e+00 1.15458488e+00 -6.22558713e-01
4.89807963e-01 -4.93442774e-01 -3.36500049e-01 -1.49203324e+00
-9.74336982e-01 1.96207583e-01 3.01058561e-01 5.76531172e-01
3.64986897e-01 -5.35131216e-01 3.98786217e-01 -3.58708560e-01
5.37148044e-02 1.55105501e-01 -7.64630914e-01 -9.90644693e-01
-1.98783338e-01 -3.13696176e-01 1.57792062e-01 7.47115970e-01
-5.44107318e-01 5.73503911e-01 -3.94065171e-01 9.92057860e-01
8.42385769e-01 -7.77192041e-02 5.63852072e-01 -1.53522527e+00
1.70366168e-01 -2.84124017e-01 -5.58450937e-01 -3.97962570e-01
-1.57573551e-01 -6.82763755e-01 3.21188450e-01 -1.52197146e+00
1.00468777e-01 -1.67265281e-01 -3.88192713e-01 1.38058102e+00
4.69313040e-02 6.85203314e-01 -7.25636333e-02 2.95252502e-01
-1.28607288e-01 1.50176585e-01 5.05186558e-01 -6.56755805e-01
-1.47397742e-01 1.58843428e-01 -2.99673259e-01 9.23387766e-01
1.01012552e+00 -7.50407219e-01 9.01088957e-03 -7.60820210e-01
-1.32054478e-01 -2.28622496e-01 1.16467166e+00 -1.58083510e+00
2.02867948e-02 -2.99313009e-01 5.44064581e-01 -6.27605379e-01
-8.96656513e-02 -1.04743159e+00 1.58867002e-01 7.44845390e-01
-2.48009227e-02 9.20425728e-02 2.89640814e-01 5.72133064e-01
1.75726071e-01 5.94819523e-02 1.05669558e+00 -3.95343274e-01
-1.22920561e+00 2.99034894e-01 -8.56915236e-01 -1.07811019e-01
6.78772271e-01 -9.31087583e-02 -8.34054708e-01 -2.23970652e-01
-3.80131423e-01 -5.60768768e-02 6.18851557e-02 6.67664826e-01
4.97838527e-01 -9.16372418e-01 -9.44417000e-01 3.55176806e-01
4.57119673e-01 -2.35670358e-01 -1.98690772e-01 3.43799770e-01
-9.83173430e-01 1.56400412e-01 -7.19292521e-01 -7.13808417e-01
-1.13310039e+00 4.18192267e-01 8.69051933e-01 -2.65452534e-01
-9.11379457e-01 6.77905440e-01 -3.87783974e-01 -6.30094111e-01
1.01088092e-01 -7.23518372e-01 -2.49556735e-01 -2.47399509e-01
7.50262022e-01 4.01210904e-01 1.05859071e-01 -5.17885804e-01
-5.90999961e-01 3.72338831e-01 3.80978346e-01 1.49817556e-01
1.75728297e+00 4.48717654e-01 8.80776644e-02 2.84774750e-01
6.38022423e-01 -7.30451524e-01 -1.23829818e+00 -1.18798211e-01
-2.27871090e-01 -3.94640028e-01 9.12083745e-01 -1.17069364e+00
-1.50779080e+00 7.37794816e-01 1.44222558e+00 -4.71619368e-02
1.21244502e+00 -3.31633866e-01 5.88546038e-01 8.80498052e-01
4.60093528e-01 -5.59965968e-01 -4.13919985e-01 8.92114758e-01
8.59299839e-01 -1.26195717e+00 -1.22781090e-01 1.98045805e-01
-2.01461434e-01 1.17548013e+00 5.57074308e-01 -5.27393878e-01
9.81325805e-01 6.41395986e-01 6.87624574e-01 -6.72960520e-01
-3.64937842e-01 -6.96090579e-01 -4.77691948e-01 1.07129788e+00
1.98053531e-02 4.19603497e-01 3.19662333e-01 7.30387419e-02
1.22158624e-01 3.30094397e-01 3.01649839e-01 1.17662275e+00
-9.68748450e-01 -1.69442326e-01 -6.82752430e-01 7.49704242e-01
-9.45089757e-02 -3.69726598e-01 -7.87347615e-01 7.12174714e-01
2.87873566e-01 1.16305625e+00 3.95757072e-02 -6.90234780e-01
6.19119287e-01 -2.51145869e-01 1.23707399e-01 -4.06909615e-01
-1.16033006e+00 -3.22130620e-01 2.90721096e-03 -4.45908189e-01
-7.86004663e-01 -1.88919723e-01 -7.09669590e-01 -5.20259619e-01
-1.96959853e-01 -3.57464820e-01 3.28856826e-01 7.57611394e-01
1.51603371e-01 2.13504821e-01 6.74726188e-01 -1.26969779e+00
-5.29582739e-01 -1.42227149e+00 -7.64806747e-01 -8.79938900e-02
5.88217735e-01 -5.51614881e-01 -6.28019452e-01 2.26449203e-02] | [9.261537551879883, -1.300600290298462] |
4c3cd0e4-ccc7-4e7c-81c7-8fa836e918e1 | deep-hyperedges-a-framework-for-transductive | 1910.02633 | null | https://arxiv.org/abs/1910.02633v1 | https://arxiv.org/pdf/1910.02633v1.pdf | Deep Hyperedges: a Framework for Transductive and Inductive Learning on Hypergraphs | From social networks to protein complexes to disease genomes to visual data, hypergraphs are everywhere. However, the scope of research studying deep learning on hypergraphs is still quite sparse and nascent, as there has not yet existed an effective, unified framework for using hyperedge and vertex embeddings jointly in the hypergraph context, despite a large body of prior work that has shown the utility of deep learning over graphs and sets. Building upon these recent advances, we propose \textit{Deep Hyperedges} (DHE), a modular framework that jointly uses contextual and permutation-invariant vertex membership properties of hyperedges in hypergraphs to perform classification and regression in transductive and inductive learning settings. In our experiments, we use a novel random walk procedure and show that our model achieves and, in most cases, surpasses state-of-the-art performance on benchmark datasets. Additionally, we study our framework's performance on a variety of diverse, non-standard hypergraph datasets and propose several avenues of future work to further enhance DHE. | ['Josh Payne'] | 2019-10-07 | null | null | null | null | ['hypergraph-embedding', 'hyperedge-classification'] | ['graphs', 'graphs'] | [ 3.43694955e-01 3.39852840e-01 -2.26128444e-01 -1.87806070e-01
-2.32412100e-01 -7.81203449e-01 7.60400593e-01 1.28179476e-01
-2.45191120e-02 6.85910821e-01 1.81346387e-01 -6.44096196e-01
-4.06068683e-01 -1.03803360e+00 -7.90686369e-01 -7.23028898e-01
-4.24457282e-01 8.25164080e-01 8.60527828e-02 -1.95620686e-01
-1.19417667e-01 3.48253340e-01 -1.17223561e+00 1.33106604e-01
5.59584796e-01 3.51229310e-01 -3.10351729e-01 7.66415358e-01
-1.60040595e-02 6.07386112e-01 -1.57586366e-01 -7.38379419e-01
3.13621879e-01 -3.91575277e-01 -1.03634834e+00 2.05858916e-01
8.06118488e-01 1.40800858e-02 -8.10205758e-01 9.35653806e-01
6.02994442e-01 -2.24062562e-01 6.81649506e-01 -1.49131870e+00
-1.40408516e+00 7.35945582e-01 -6.52437210e-01 2.19469205e-01
2.36952677e-01 5.89575052e-01 1.65239584e+00 -6.13458931e-01
1.04612470e+00 1.40181220e+00 7.90740848e-01 2.89419621e-01
-1.71277070e+00 -4.31980699e-01 7.36692175e-02 9.93058681e-02
-1.08769059e+00 4.23394181e-02 5.04624009e-01 -4.66684192e-01
1.21410406e+00 5.62550798e-02 8.93341660e-01 1.23511362e+00
2.63751578e-02 5.99699974e-01 1.08620071e+00 -3.71260136e-01
-1.00943752e-01 -3.09691727e-01 2.80755162e-01 1.15542150e+00
6.10995293e-01 5.65220751e-02 -2.45813400e-01 -1.44784838e-01
8.55678797e-01 -1.84497640e-01 -3.64172339e-01 -1.02265990e+00
-1.25563562e+00 1.07062876e+00 8.59210968e-01 2.78709710e-01
-1.00656569e-01 4.74229336e-01 4.33585435e-01 4.19709504e-01
5.16438007e-01 5.49548507e-01 -2.54491717e-01 1.74601763e-01
-3.56600553e-01 5.46263717e-02 1.02514517e+00 9.22525525e-01
7.39236236e-01 -1.64128646e-01 -1.48242190e-02 6.41103029e-01
1.46561831e-01 2.81970859e-01 -4.46550995e-01 -3.82429957e-01
4.44970071e-01 9.90015030e-01 -4.20161217e-01 -1.25054729e+00
-6.87352717e-01 -3.29305887e-01 -9.45477188e-01 -1.17737921e-02
4.03055012e-01 -6.25513643e-02 -1.17826498e+00 1.90666437e+00
3.21749121e-01 6.26508534e-01 -1.21178076e-01 7.61504114e-01
1.23144770e+00 3.56017113e-01 5.88088855e-02 2.99634397e-01
1.23715913e+00 -8.02354395e-01 -4.02285010e-01 -1.09490827e-01
7.34455824e-01 -2.47422859e-01 1.14363122e+00 1.87560603e-01
-9.87959325e-01 -1.09748304e-01 -1.03020048e+00 -3.59129429e-01
-7.15133011e-01 -3.53290170e-01 1.25069821e+00 7.16817856e-01
-1.42665577e+00 3.65114391e-01 -6.99538410e-01 -8.80266905e-01
7.27683425e-01 5.34946859e-01 -6.46172702e-01 -3.95523489e-01
-1.26570380e+00 6.82554424e-01 3.43530327e-01 -7.47284666e-02
-8.12951148e-01 -9.11907911e-01 -9.63020027e-01 1.51133463e-01
4.94173735e-01 -1.17942977e+00 5.60586333e-01 -4.50339735e-01
-1.07105327e+00 1.12896216e+00 3.48341048e-01 -3.85339379e-01
2.01802790e-01 3.74383211e-01 -3.03931534e-01 1.70741767e-01
-3.55189443e-01 7.09176779e-01 1.78336591e-01 -1.22724879e+00
-1.60767347e-01 -4.59357113e-01 5.06631434e-01 1.30486742e-01
-5.59279084e-01 -3.43817085e-01 -8.00224066e-01 -2.29261771e-01
-1.61233917e-01 -1.28783405e+00 -1.21622533e-01 -2.23066434e-02
-7.97628760e-01 -4.78147149e-01 4.92976844e-01 -1.05489649e-01
1.18690681e+00 -1.72737658e+00 5.69044530e-01 3.68554354e-01
1.00692475e+00 3.34597081e-01 -5.03920257e-01 7.93408990e-01
-2.54477859e-01 3.52733940e-01 -3.67931455e-01 5.30893914e-02
2.52798617e-01 2.94868767e-01 7.81086236e-02 5.75945199e-01
3.40472430e-01 1.49311495e+00 -1.20416331e+00 -2.92465419e-01
3.78657371e-01 7.09686041e-01 -6.05395436e-01 -8.90514180e-02
-3.31458926e-01 1.33171864e-02 -2.86594629e-01 5.51741779e-01
7.36461341e-01 -1.03943503e+00 8.64257097e-01 -5.95807657e-02
4.40058887e-01 1.44067958e-01 -9.64941621e-01 1.72524178e+00
-7.23147169e-02 7.72809744e-01 -9.04757008e-02 -1.15635049e+00
4.93039191e-01 1.04390875e-01 5.08944571e-01 -3.72044981e-01
5.83241023e-02 -1.58599600e-01 3.11480433e-01 -6.24841273e-01
1.45746082e-01 2.25708276e-01 1.50391072e-01 5.03059447e-01
2.37867579e-01 9.06203017e-02 3.72884244e-01 7.44168580e-01
1.74842596e+00 1.03103071e-02 3.17135602e-01 -2.73693740e-01
1.29627332e-01 -1.09235771e-01 8.43237787e-02 6.22800112e-01
-1.66539624e-01 4.68884170e-01 1.04743552e+00 -4.24575746e-01
-1.17771661e+00 -1.20321655e+00 -1.12278968e-01 1.28408742e+00
1.08339362e-01 -3.85265499e-01 -5.71271420e-01 -9.19151366e-01
5.05163193e-01 1.18821822e-01 -9.37008560e-01 2.30051763e-02
-3.79718751e-01 -1.27390921e+00 5.69294810e-01 5.79365432e-01
-1.93634242e-01 -1.10768521e+00 2.63960600e-01 5.77166751e-02
3.35211456e-01 -1.15178156e+00 -3.46293837e-01 1.93871439e-01
-5.83809435e-01 -1.57961619e+00 -3.89435232e-01 -1.13289070e+00
6.53498232e-01 4.48715866e-01 1.74767065e+00 3.77703398e-01
-5.53439498e-01 4.59707528e-01 -1.79628536e-01 3.63827012e-02
-2.98246384e-01 3.29099625e-01 -3.20255339e-01 -3.41146767e-01
5.91809928e-01 -6.28258765e-01 -6.52570128e-01 6.71962649e-02
-1.04869258e+00 1.31367962e-03 5.59451401e-01 9.95744348e-01
3.45230520e-01 -3.03809047e-01 8.25407505e-01 -1.76858687e+00
6.71841204e-01 -7.62694299e-01 -6.51261151e-01 3.32837313e-01
-7.46088147e-01 5.77442050e-02 4.08018589e-01 3.65320174e-03
-4.85689670e-01 -2.74404198e-01 7.21632317e-03 -1.86184481e-01
1.67767316e-01 7.37829208e-01 -5.29029667e-02 -3.33353311e-01
6.28081679e-01 -1.50182709e-01 -5.84604070e-02 -1.10304192e-01
9.03239071e-01 2.12735996e-01 4.43746001e-01 -3.02917063e-01
8.47055912e-01 5.24102092e-01 5.13942897e-01 -6.56065583e-01
-5.85626781e-01 -4.51923132e-01 -7.08937943e-01 2.57326965e-03
8.77099633e-01 -6.61342263e-01 -1.13656056e+00 2.34968513e-01
-7.94247508e-01 -3.66917998e-01 1.70223832e-01 1.45316795e-01
-4.06785011e-01 5.89717150e-01 -7.83125758e-01 -3.45572412e-01
-2.09229499e-01 -1.12079477e+00 1.07937694e+00 -8.81673396e-02
1.87907353e-01 -1.61144125e+00 4.12243903e-01 2.65615106e-01
1.51285410e-01 6.06250167e-01 1.39762115e+00 -7.42971301e-01
-1.01252580e+00 3.17125246e-02 -6.22932673e-01 -1.53221926e-02
1.09690435e-01 1.93509996e-01 -8.22748363e-01 -5.91045141e-01
-1.07082295e+00 -7.84261346e-01 1.29451621e+00 3.32813740e-01
1.12416971e+00 7.35137463e-02 -7.22119391e-01 9.14746761e-01
1.64510465e+00 -3.95760119e-01 8.47144663e-01 -2.76792003e-03
1.19194484e+00 3.21134359e-01 -7.30621517e-02 5.18960282e-02
6.92677319e-01 3.60346228e-01 7.43402898e-01 -6.28281593e-01
-3.06757331e-01 -2.75748760e-01 -2.05245614e-02 8.80495965e-01
-3.16138417e-02 -9.56432819e-01 -8.27724814e-01 6.43587947e-01
-1.72928071e+00 -9.55578864e-01 -3.93490106e-01 1.95899224e+00
6.96184397e-01 4.97174391e-04 1.96931586e-01 -2.42872030e-01
7.98326612e-01 4.32138860e-01 -7.22860396e-01 -3.02208453e-01
-3.06036800e-01 3.99080843e-01 5.26525497e-01 3.41377974e-01
-1.24522340e+00 1.06669402e+00 6.72749901e+00 4.56876218e-01
-6.50996029e-01 -2.21930414e-01 4.79238749e-01 1.98446527e-01
-7.58864880e-01 -1.18139545e-02 -3.69724959e-01 1.42949566e-01
6.68266177e-01 6.71241283e-02 8.71986091e-01 3.87829751e-01
-3.94045472e-01 4.56569046e-01 -1.47660875e+00 8.33982646e-01
8.14098269e-02 -1.59458125e+00 -2.58900854e-03 3.55852455e-01
1.08531678e+00 4.63928789e-01 2.71449983e-01 3.78218710e-01
1.08070982e+00 -1.28644097e+00 -4.30647671e-01 3.06863666e-01
8.98451090e-01 -5.97671628e-01 4.36718822e-01 -2.23920181e-01
-1.27599502e+00 1.40724450e-01 -4.08746481e-01 3.42400432e-01
-1.84984967e-01 4.86741602e-01 -1.35786712e+00 7.65360534e-01
3.47390294e-01 8.84202063e-01 -6.02321982e-01 9.57993925e-01
-3.10689926e-01 6.79637909e-01 -1.07432559e-01 -5.25291152e-02
3.96787614e-01 -2.43664384e-01 4.56640959e-01 1.37503004e+00
-1.32252276e-01 -1.08895890e-01 3.97436678e-01 9.10294592e-01
-7.65432715e-01 9.22624320e-02 -1.10459292e+00 -6.49348319e-01
4.59971458e-01 1.56288040e+00 -9.49961364e-01 -9.83129665e-02
-7.03785479e-01 6.92051709e-01 8.50427806e-01 4.83762205e-01
-8.77474129e-01 -3.99725050e-01 6.78694606e-01 -8.19290429e-03
4.73685920e-01 -8.99936706e-02 8.56058523e-02 -1.02011752e+00
-2.86469102e-01 -9.60246921e-01 8.11068237e-01 -6.20642066e-01
-1.81586552e+00 3.61364096e-01 -2.71126330e-01 -5.89278042e-01
2.28908029e-03 -9.85181570e-01 -4.27162379e-01 5.85914373e-01
-1.52771544e+00 -1.31962049e+00 -3.19453776e-01 3.02872479e-01
-9.85025242e-02 -1.02385372e-01 8.00920784e-01 2.45970428e-01
-4.87235606e-01 6.95673168e-01 2.04751894e-01 2.50130504e-01
6.30382538e-01 -1.82772899e+00 1.00886369e+00 5.64725578e-01
4.38519031e-01 6.24865592e-01 3.72524917e-01 -6.24928534e-01
-2.05623698e+00 -1.17657804e+00 4.70049798e-01 -8.28520656e-01
9.47015166e-01 -7.77869761e-01 -9.02115643e-01 1.14660239e+00
4.74429876e-01 3.12058568e-01 7.66271174e-01 7.27706313e-01
-6.34517252e-01 9.21485275e-02 -9.71765935e-01 6.35058463e-01
1.56081927e+00 -6.20988786e-01 -1.59598976e-01 6.81505382e-01
9.02789772e-01 -2.13603675e-01 -1.14723933e+00 4.90906239e-01
4.62150633e-01 -7.66856670e-01 1.13650060e+00 -1.04282939e+00
3.50437611e-01 -1.69410244e-01 1.37996733e-01 -1.62568426e+00
-7.73954630e-01 -7.67070711e-01 -7.40439296e-02 9.32929754e-01
5.37489474e-01 -7.09587157e-01 9.57945764e-01 8.79259631e-02
-1.75021395e-01 -9.14170504e-01 -5.11536658e-01 -5.28214812e-01
1.62268192e-01 -4.26575579e-02 5.75473607e-01 1.20258164e+00
1.54568538e-01 7.48248696e-01 -2.68011063e-01 1.76925987e-01
5.75370133e-01 2.82470852e-01 1.07690299e+00 -1.41341519e+00
-3.83249134e-01 -5.39665699e-01 -7.57478535e-01 -8.59734476e-01
3.82404923e-01 -1.70533836e+00 -2.91553199e-01 -2.13430595e+00
8.12465668e-01 -2.52113551e-01 -5.36266565e-01 4.84916657e-01
-4.57860678e-01 5.67783892e-01 -6.18006214e-02 -3.49269181e-01
-8.72528076e-01 3.01039070e-01 1.64587665e+00 -3.17595512e-01
9.19538084e-03 -4.11306173e-01 -8.53496253e-01 2.87607819e-01
4.02452409e-01 -9.39531103e-02 -7.24510849e-01 -5.38824081e-01
5.69737911e-01 -1.71728104e-01 4.05601889e-01 -4.82907921e-01
4.00247201e-02 3.98953930e-02 2.58334666e-01 -3.07188392e-01
1.52801082e-01 -4.63679433e-01 -1.69290621e-02 2.33024418e-01
-4.61084992e-01 1.75703600e-01 6.03462867e-02 1.07198334e+00
3.83661449e-01 3.47004265e-01 6.01974368e-01 -3.72053823e-03
-5.97655416e-01 7.80198514e-01 8.72646645e-02 6.07937813e-01
9.71510708e-01 1.49629610e-02 -9.18065727e-01 -3.50382745e-01
-6.00779772e-01 5.64843953e-01 5.42098284e-01 3.62252116e-01
4.65744704e-01 -1.27396131e+00 -8.56565952e-01 6.65643513e-02
4.46123898e-01 -3.08277696e-01 1.37313470e-01 6.36788547e-01
-5.34450352e-01 4.74554569e-01 -2.47119050e-02 -7.02942312e-01
-1.21257138e+00 9.66834426e-01 9.11152512e-02 -3.50820690e-01
-7.59206116e-01 9.23579514e-01 5.53435147e-01 -7.53592491e-01
2.60818869e-01 -3.88937235e-01 -1.76820382e-01 -1.35074198e-01
3.47366706e-02 4.04772729e-01 3.09890416e-02 -4.24846321e-01
-3.73665780e-01 2.68793792e-01 -1.62861958e-01 4.93004173e-01
1.58132768e+00 2.26032630e-01 -2.15893060e-01 1.87978044e-01
1.38210678e+00 -1.73823610e-01 -1.08361220e+00 -3.31598133e-01
-5.61050326e-02 -2.55012006e-01 -1.48991868e-01 -8.90388906e-01
-1.27396822e+00 8.72144878e-01 3.18854541e-01 6.02992475e-01
8.06690335e-01 2.54716158e-01 6.73297584e-01 5.70859313e-01
5.74822687e-02 -6.38492942e-01 1.43602833e-01 4.61751938e-01
5.46010315e-01 -1.30522919e+00 1.93616405e-01 -6.46061122e-01
-4.82218325e-01 8.70468915e-01 5.96408904e-01 -2.85107821e-01
7.29018509e-01 2.57008642e-01 -3.18586260e-01 -8.34478259e-01
-9.54380572e-01 -5.71918368e-01 3.23052764e-01 8.30725491e-01
6.31136894e-01 2.57789224e-01 -2.60336474e-02 -1.24009937e-01
1.21126279e-01 -2.98119616e-02 4.55659956e-01 5.12347579e-01
-2.85471737e-01 -1.10841668e+00 3.90057474e-01 6.20368063e-01
-3.28427345e-01 -3.70475948e-01 -7.90269434e-01 1.12235785e+00
-4.23822179e-02 7.11966336e-01 2.97578722e-02 -3.68988544e-01
1.09490976e-01 -1.61198571e-01 8.38811159e-01 -8.83781672e-01
-4.95276809e-01 -2.37221509e-01 1.99214101e-01 -3.31031084e-01
-3.02334487e-01 -2.91565329e-01 -1.18452215e+00 -6.43422723e-01
-3.02506328e-01 -3.13952833e-01 2.64695823e-01 5.46914995e-01
5.20800650e-01 7.28334904e-01 4.90911752e-01 -5.71459115e-01
-3.67566884e-01 -7.74053276e-01 -7.52623320e-01 8.20271730e-01
8.87079015e-02 -7.33641565e-01 -1.71263605e-01 -3.07561457e-01] | [6.958550930023193, 6.238804817199707] |
c5a732b3-ffe9-4709-9327-ef6f81432509 | tart-a-plug-and-play-transformer-module-for | 2306.07536 | null | https://arxiv.org/abs/2306.07536v1 | https://arxiv.org/pdf/2306.07536v1.pdf | TART: A plug-and-play Transformer module for task-agnostic reasoning | Large language models (LLMs) exhibit in-context learning abilities which enable the same model to perform several tasks without any task-specific training. In contrast, traditional adaptation approaches, such as fine-tuning, modify the underlying models for each specific task. In-context learning, however, consistently underperforms task-specific tuning approaches even when presented with the same examples. While most existing approaches (e.g., prompt engineering) focus on the LLM's learned representations to patch this performance gap, our analysis actually reveal that LLM representations contain sufficient information to make good predictions. As such, we focus on the LLM's reasoning abilities and demonstrate that this performance gap exists due to their inability to perform simple probabilistic reasoning tasks. This raises an intriguing question: Are LLMs actually capable of learning how to reason in a task-agnostic manner? We answer this in the affirmative and propose TART which generically improves an LLM's reasoning abilities using a synthetically trained Transformer-based reasoning module. TART trains this reasoning module in a task-agnostic manner using only synthetic logistic regression tasks and composes it with an arbitrary real-world pre-trained model without any additional training. With a single inference module, TART improves performance across different model families (GPT-Neo, Pythia, BLOOM), model sizes (100M - 6B), tasks (14 NLP binary classification tasks), and even across different modalities (audio and vision). Additionally, on the RAFT Benchmark, TART improves GPT-Neo (125M)'s performance such that it outperforms BLOOM (176B), and is within 4% of GPT-3 (175B). Our code and models are available at https://github.com/HazyResearch/TART . | ['Christopher Ré', 'Christopher De Sa', 'Avanika Narayan', 'Kush Bhatia'] | 2023-06-13 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 2.64739454e-01 2.66913325e-01 -1.46668896e-01 -3.53482872e-01
-1.07895744e+00 -6.43982828e-01 8.02526474e-01 -1.63460538e-01
-3.22290570e-01 6.43098652e-01 9.61999968e-02 -8.33845794e-01
-1.66508317e-01 -7.39941001e-01 -9.14050519e-01 -2.38003030e-01
3.97015154e-01 8.61544847e-01 3.07193995e-01 -3.60117704e-01
1.95854738e-01 2.24899471e-01 -1.27642357e+00 7.54294395e-01
9.29256380e-01 8.02960873e-01 1.60897285e-01 8.36933315e-01
-1.42774582e-01 1.11237419e+00 -5.55208266e-01 -6.91774309e-01
9.00356621e-02 1.09678740e-02 -9.05790389e-01 -3.39478910e-01
5.98406732e-01 -2.20688302e-02 -2.09700525e-01 6.54997051e-01
3.82232189e-01 -1.33352950e-02 8.47842574e-01 -1.42741561e+00
-8.33795130e-01 9.68201518e-01 -3.30430448e-01 3.76569219e-02
1.52397141e-01 3.38329196e-01 1.06464982e+00 -9.83183920e-01
1.54929549e-01 1.41067851e+00 1.00841975e+00 8.73934627e-01
-1.49940145e+00 -9.27453637e-01 1.67317614e-01 8.56938064e-02
-1.26637053e+00 -3.91914338e-01 3.40418041e-01 -3.68970305e-01
1.34066129e+00 1.36500552e-01 2.11121053e-01 1.59544981e+00
2.31524929e-01 8.06328654e-01 1.33293891e+00 -5.24754882e-01
2.67854631e-01 1.60469368e-01 2.17590675e-01 5.30310988e-01
2.07657546e-01 1.23917863e-01 -5.26490390e-01 -1.88469440e-01
4.73565787e-01 -1.45891547e-01 1.25573769e-01 -3.75238769e-02
-1.31733239e+00 6.52811050e-01 2.84531206e-01 3.41293931e-01
-1.30393371e-01 5.23728967e-01 2.46927947e-01 4.87344772e-01
1.00179669e-02 8.72630954e-01 -1.03617251e+00 -1.80122927e-01
-8.24971557e-01 3.10455918e-01 9.35204446e-01 9.91994381e-01
6.49130404e-01 2.07560733e-01 -5.63171208e-01 8.90958190e-01
1.26738727e-01 4.34213847e-01 6.63097620e-01 -1.15716314e+00
5.17593503e-01 4.73175377e-01 -1.38858214e-01 -4.76229727e-01
-4.98727500e-01 -6.37307346e-01 -7.55200446e-01 2.46194869e-01
5.93478799e-01 -1.92789540e-01 -8.77602160e-01 2.07711601e+00
-1.79862112e-01 1.23253547e-01 3.85486066e-01 3.70709866e-01
8.85738611e-01 5.21262288e-01 3.60251963e-01 2.42484987e-01
1.56641269e+00 -1.15748823e+00 -7.59667754e-02 -7.15768278e-01
5.22434771e-01 -7.34456599e-01 1.61449075e+00 5.15170097e-01
-1.21113336e+00 -6.81748033e-01 -7.90916383e-01 -7.77160674e-02
-5.20363569e-01 -4.50554490e-02 7.51829088e-01 5.97044528e-01
-1.16046631e+00 2.96646327e-01 -4.80132967e-01 -4.53858882e-01
3.73923063e-01 2.74656624e-01 -2.00065508e-01 -2.28976429e-01
-1.34742820e+00 1.22969985e+00 5.15182912e-01 -3.85561585e-01
-1.06569791e+00 -9.81050789e-01 -6.55795574e-01 2.30242059e-01
4.47535276e-01 -1.14602542e+00 1.75283432e+00 -7.93209672e-01
-1.38138974e+00 7.97088683e-01 -1.70115709e-01 -7.28025079e-01
5.92713356e-01 -1.06941588e-01 -3.84442270e-01 -1.63415775e-01
5.45079075e-02 8.87432337e-01 8.97859871e-01 -1.13809431e+00
-6.42092407e-01 1.71691686e-01 5.00888407e-01 -5.28107695e-02
-1.11880474e-01 6.99228933e-03 -2.78690189e-01 -7.94178069e-01
-1.49866119e-01 -9.66233253e-01 -1.03501894e-01 -1.91420048e-01
-4.13137138e-01 -3.45974863e-01 5.30606508e-01 -3.60039413e-01
1.00778592e+00 -1.95519662e+00 -1.26894414e-01 -4.12901156e-02
2.39199057e-01 3.24635983e-01 -4.34191883e-01 2.34716564e-01
-1.83152780e-01 2.77693033e-01 -2.83510596e-01 -4.18285340e-01
2.93204606e-01 4.28027898e-01 -4.35193777e-01 -2.95999348e-01
3.40559781e-01 1.14915311e+00 -7.21943080e-01 -4.81311798e-01
7.13127553e-02 3.30413043e-01 -8.93426716e-01 1.94882154e-01
-6.95796967e-01 4.41828609e-01 -2.65211165e-01 4.87093419e-01
3.14984322e-01 -4.98709530e-01 1.83492720e-01 -4.88622040e-02
3.18505824e-01 4.20526892e-01 -8.04092526e-01 1.52975249e+00
-9.88448679e-01 6.39254689e-01 -3.33795696e-01 -1.07202363e+00
7.95474887e-01 2.96243072e-01 -5.90133704e-02 -7.59796381e-01
-1.88653156e-01 2.77736604e-01 2.47705206e-01 -3.94567072e-01
3.01701397e-01 -3.36452305e-01 -3.15154552e-01 5.56248069e-01
1.73027709e-01 -3.54835868e-01 1.15400583e-01 1.61311299e-01
1.44165981e+00 1.13532275e-01 3.28633428e-01 4.69142618e-03
4.40193951e-01 -4.61388230e-02 4.31425393e-01 1.21312582e+00
-3.16244811e-02 4.47082847e-01 3.81749779e-01 -2.99889445e-01
-9.45003450e-01 -1.29300535e+00 -1.26123667e-01 1.54846287e+00
-4.03534532e-01 -5.25398552e-01 -5.77418387e-01 -7.44063914e-01
1.19932815e-01 1.35823846e+00 -6.86693907e-01 -4.41763222e-01
-5.20327151e-01 -7.68100321e-01 9.15958881e-01 6.85480475e-01
7.26622105e-01 -1.34479606e+00 -3.32740813e-01 8.11342970e-02
-3.09861243e-01 -1.22345066e+00 -6.04947358e-02 3.93032908e-01
-8.22722197e-01 -9.18931007e-01 -3.46799821e-01 -5.62943101e-01
4.23572868e-01 3.98235582e-02 1.52518678e+00 3.27819996e-02
8.57647061e-02 5.05525291e-01 -1.24189951e-01 -4.14092600e-01
-8.17034543e-01 4.01792407e-01 -1.25320151e-01 -4.84203428e-01
4.38089430e-01 -8.42024446e-01 -2.25155771e-01 4.26442444e-01
-5.91503859e-01 3.03444892e-01 9.20452714e-01 9.43470299e-01
3.72792691e-01 1.17429413e-01 7.98753321e-01 -1.01334107e+00
7.11128056e-01 -5.58472157e-01 -2.98626900e-01 6.04660988e-01
-8.12303126e-01 3.49576920e-01 6.68557703e-01 -7.11246252e-01
-1.20793879e+00 -4.23076838e-01 -2.05228329e-01 -3.21773022e-01
-1.43062577e-01 4.94106531e-01 -6.06089421e-02 2.00990513e-01
1.05755150e+00 3.01644117e-01 -3.73170078e-01 -4.04973686e-01
3.29181582e-01 3.85831833e-01 6.25345170e-01 -1.16692042e+00
1.02812946e+00 -3.98786142e-02 -1.44293398e-01 -1.80029631e-01
-1.22701633e+00 1.10857300e-01 -1.57665461e-01 2.62268573e-01
6.56126618e-01 -1.03913879e+00 -9.70217586e-01 2.34421358e-01
-1.00992417e+00 -9.38025177e-01 -2.13185757e-01 2.88686693e-01
-7.81071603e-01 -6.99751973e-02 -6.30942345e-01 -7.43136048e-01
-3.66463631e-01 -1.09397984e+00 9.24154818e-01 1.02193989e-01
-6.56699359e-01 -1.04846537e+00 -2.08872721e-01 6.66018486e-01
7.02287853e-01 -1.31372705e-01 1.34560549e+00 -8.40492308e-01
-4.79422361e-01 -1.36462506e-02 -3.74096215e-01 4.53315467e-01
-1.05125450e-01 -8.09941813e-02 -1.24271393e+00 -1.06764724e-02
-1.64747179e-01 -5.24766624e-01 9.74915385e-01 2.18995601e-01
1.32078719e+00 -3.10014367e-01 -2.19449639e-01 5.55304408e-01
1.06175828e+00 2.90305931e-02 5.34666896e-01 5.24889767e-01
5.21137357e-01 3.56914490e-01 2.96834558e-01 3.46521400e-02
6.05926812e-01 6.26380146e-01 3.38981658e-01 2.48309553e-01
-2.94008940e-01 -4.73021239e-01 6.40749454e-01 4.50739533e-01
-3.48355025e-02 -9.39850137e-02 -1.18757010e+00 3.18329901e-01
-1.84088540e+00 -9.13244903e-01 2.60023683e-01 1.94513619e+00
1.23642743e+00 6.21772051e-01 -1.60604209e-01 1.07206367e-01
3.05617750e-01 -1.23084538e-01 -7.57875323e-01 -7.07576692e-01
-2.33097613e-01 5.20414531e-01 9.76389647e-02 5.27552128e-01
-7.58656621e-01 1.18622315e+00 6.68223906e+00 1.04106522e+00
-9.18003738e-01 2.84176320e-01 5.13452113e-01 -4.77255881e-03
-5.61836898e-01 9.89702567e-02 -8.88452530e-01 3.06239963e-01
1.35188031e+00 -1.11196004e-01 7.18157709e-01 8.35672736e-01
-2.07336023e-01 9.40216632e-05 -1.33933771e+00 8.57873261e-01
-8.59106481e-02 -1.34624147e+00 1.53531745e-01 -2.49150708e-01
6.64129257e-01 1.08852021e-01 3.40330482e-01 1.32665420e+00
7.28188753e-01 -1.41738999e+00 9.05005991e-01 5.98392963e-01
6.37632608e-01 -4.47483063e-01 4.70488042e-01 6.76898658e-01
-9.03560817e-01 -3.93034637e-01 -2.19747931e-01 -2.16009006e-01
-2.84105122e-01 4.50580060e-01 -1.06221914e+00 2.04593360e-01
8.27986002e-01 4.91160452e-01 -1.05668175e+00 6.31682098e-01
-6.17640615e-01 9.26620603e-01 -1.46452263e-01 1.70206040e-01
6.90613687e-02 4.64166790e-01 3.37054640e-01 1.31835437e+00
2.53055811e-01 -2.04183206e-01 -8.49050563e-03 1.07078481e+00
-2.74360359e-01 -3.72119963e-01 -3.58567744e-01 -7.16133714e-02
5.50600171e-01 1.13403952e+00 -1.41768157e-01 -5.06448328e-01
-4.10767883e-01 5.86827338e-01 5.88520706e-01 4.60085034e-01
-1.02194524e+00 8.46415237e-02 6.45698309e-01 5.40016964e-03
1.10044807e-01 5.57543188e-02 -4.92411941e-01 -1.22486711e+00
-1.37296930e-01 -1.29908824e+00 5.23810863e-01 -1.28633320e+00
-1.54437017e+00 5.55911541e-01 1.36830300e-01 -8.97436976e-01
-5.26948214e-01 -7.18068659e-01 -6.60102248e-01 9.03432190e-01
-1.52754056e+00 -1.39877164e+00 -2.87261307e-01 7.08840370e-01
5.38085520e-01 -4.95891511e-01 1.05927682e+00 9.61503163e-02
-4.74283308e-01 9.97503519e-01 -2.26936519e-01 2.03683197e-01
8.82583678e-01 -1.36370134e+00 6.09983325e-01 4.47937012e-01
1.07085392e-01 8.10089827e-01 7.80867755e-01 -3.09698880e-01
-1.00041616e+00 -1.18271339e+00 1.00560308e+00 -9.19140100e-01
8.54023516e-01 -4.31444257e-01 -8.98981988e-01 1.10649407e+00
7.99732506e-02 1.02459760e-02 6.79378450e-01 5.67210317e-01
-1.01501155e+00 -1.09797433e-01 -1.14442444e+00 8.96067023e-01
1.13685048e+00 -7.40512908e-01 -1.05339253e+00 4.05712068e-01
9.62190151e-01 -3.71805847e-01 -9.08271432e-01 4.52178478e-01
7.02389836e-01 -1.02518046e+00 1.19767094e+00 -8.73337090e-01
6.08301401e-01 -8.62298831e-02 -5.92402875e-01 -1.15345597e+00
-2.93515891e-01 -4.25409645e-01 -2.62380958e-01 1.21800661e+00
6.48757577e-01 -1.08891141e+00 5.57083964e-01 5.75130939e-01
-6.41418993e-02 -8.86702061e-01 -6.41897857e-01 -8.63895297e-01
5.87081671e-01 -1.02536404e+00 7.39416003e-01 7.87750959e-01
-2.17340082e-01 6.13085568e-01 -1.08443081e-01 7.26805851e-02
4.26945537e-01 1.42660484e-01 8.94885302e-01 -1.34182441e+00
-7.23205388e-01 -6.17459238e-01 1.31236359e-01 -7.81664371e-01
5.31116664e-01 -1.12398601e+00 -1.45554811e-01 -1.46926284e+00
2.91261882e-01 -7.37067282e-01 -4.29919183e-01 1.04588544e+00
-1.58407241e-01 1.21138714e-01 3.14462304e-01 2.24011078e-01
-4.81451631e-01 2.10488826e-01 1.09825587e+00 -3.42239469e-01
1.76328629e-01 2.34125420e-01 -1.14557076e+00 9.00768816e-01
9.20014501e-01 -5.12976825e-01 -4.81275797e-01 -4.41524029e-01
5.78114152e-01 -5.92066608e-02 7.55455256e-01 -1.28085077e+00
9.87278447e-02 -2.32237130e-01 4.81605470e-01 -1.75183967e-01
4.39036399e-01 -6.80271149e-01 6.51315674e-02 4.73834604e-01
-6.33255720e-01 -4.24603894e-02 5.46717346e-01 2.32160479e-01
-4.05974612e-02 -3.06535482e-01 7.52337694e-01 -2.27773696e-01
-6.90212011e-01 -8.27343110e-03 -3.44203144e-01 3.10871840e-01
5.82738936e-01 -1.53950959e-01 -6.71622038e-01 -3.94263148e-01
-8.62176239e-01 1.96679801e-01 2.28302225e-01 4.33156997e-01
2.86088735e-01 -1.08498955e+00 -8.53297889e-01 1.34269267e-01
3.30895275e-01 2.07905751e-02 2.40664333e-01 8.67778242e-01
5.12933685e-03 4.32929158e-01 -7.06771091e-02 -5.46324968e-01
-8.35315645e-01 5.37330508e-01 5.47123730e-01 -6.85311556e-01
-3.26865733e-01 8.27698946e-01 4.55662102e-01 -1.01052463e+00
3.76912542e-02 -6.16953850e-01 6.05270341e-02 -2.43546605e-01
4.08053130e-01 -3.16067450e-02 -7.99831599e-02 -9.41116139e-02
-2.88218409e-01 4.26214188e-01 -7.47513846e-02 3.70034296e-03
1.10070121e+00 2.22107142e-01 2.16721669e-01 5.42188764e-01
7.94356227e-01 4.90812920e-02 -1.23789144e+00 -4.75661129e-01
4.90316153e-02 1.27663827e-02 -3.18809122e-01 -1.44462478e+00
-7.71476686e-01 1.04070807e+00 1.29140228e-01 1.26903541e-02
1.15135264e+00 5.37439659e-02 5.08663297e-01 7.97828913e-01
5.13621747e-01 -7.71773517e-01 3.62840444e-01 9.01907444e-01
1.08135223e+00 -1.34189212e+00 -2.28988931e-01 1.86870750e-02
-7.93289363e-01 9.73037362e-01 8.64734769e-01 1.06155254e-01
5.60528994e-01 2.89779782e-01 -1.40612558e-01 2.36530006e-01
-1.38041818e+00 -7.23781735e-02 2.50465780e-01 7.10711896e-01
4.49544251e-01 1.09639071e-01 3.32396418e-01 8.77805054e-01
-6.12483501e-01 1.23080730e-01 2.02408925e-01 6.77882254e-01
-3.36666346e-01 -1.09182310e+00 -4.53384757e-01 4.86781031e-01
-2.47420415e-01 -4.84056562e-01 -7.41403028e-02 9.79074478e-01
1.91074356e-01 9.54475045e-01 -1.04090281e-01 -4.03392643e-01
3.26003432e-01 4.66138959e-01 4.65047181e-01 -8.20126176e-01
-7.60100484e-01 -3.76450270e-01 2.32131675e-01 -3.61030489e-01
-1.22708552e-01 -4.60616946e-01 -1.25344253e+00 -4.81629461e-01
2.62339383e-01 -8.30339268e-02 2.95206726e-01 1.11098731e+00
3.32619131e-01 7.34499931e-01 3.25877778e-02 -5.06976068e-01
-8.03105295e-01 -1.15564811e+00 -1.12188719e-01 1.46774575e-01
1.36257768e-01 -6.65484369e-01 -3.28003585e-01 -1.20339662e-01] | [10.20919418334961, 7.88517951965332] |
0f170e26-f7d4-46c2-b9c8-d24a291ca3c9 | person-re-identification-based-on-res2net | 1910.04061 | null | https://arxiv.org/abs/1910.04061v2 | https://arxiv.org/pdf/1910.04061v2.pdf | Improved Res2Net model for Person re-identification | Person re-identification has become a very popular research topic in the computer vision community owing to its numerous applications and growing importance in visual surveillance. Person re-identification remains challenging due to occlusion, illumination and significant intra-class variations across different cameras. In this paper, we propose a multi-task network base on an improved Res2Net model that simultaneously computes the identification loss and verification loss of two pedestrian images. Given a pair of pedestrian images, the system predicts the identities of the two input images and whether they belong to the same identity. In order to obtain deeper feature information of pedestrians, we propose to use the latest Res2Net model for feature extraction of each input image. Experiments on several large-scale person re-identification benchmark datasets demonstrate the accuracy of our approach. For example, rank-1 accuracies are 83.18% (+1.38) and 93.14% (+0.84) for the DukeMTMC and Market-1501 datasets, respectively. The proposed method shows encouraging improvements compared with state-of-the-art methods. | ['Hyo Jong Lee', 'Zongjing Cao'] | 2019-10-08 | null | null | null | null | ['large-scale-person-re-identification'] | ['computer-vision'] | [ 1.15994904e-02 -6.84227109e-01 1.51001751e-01 -4.60950613e-01
-4.61917907e-01 -3.69596153e-01 6.75434709e-01 7.87596256e-02
-9.75523770e-01 8.56440842e-01 1.43076986e-01 2.60130793e-01
2.51906663e-01 -4.37285602e-01 -4.69351172e-01 -5.53897500e-01
2.88744509e-01 3.17227900e-01 1.79458678e-01 1.14989579e-02
7.06753507e-02 3.89951676e-01 -1.76088893e+00 1.00520507e-01
5.92506766e-01 9.38852966e-01 -1.28081664e-01 6.15691304e-01
5.18827617e-01 3.56950462e-01 -5.20315766e-01 -9.16310728e-01
5.31392932e-01 -6.96295127e-02 -6.11500680e-01 1.38494540e-02
9.36594844e-01 -5.50458908e-01 -6.03069663e-01 1.30859625e+00
7.64657795e-01 2.81068265e-01 4.00939167e-01 -1.34458649e+00
-6.91984475e-01 5.16473055e-02 -9.51468647e-01 4.13204432e-01
3.34942877e-01 6.75102696e-02 6.62229955e-01 -8.39692414e-01
2.15243250e-01 1.25674808e+00 8.48381102e-01 7.22679138e-01
-1.22013438e+00 -1.11040986e+00 1.17978141e-01 7.49139845e-01
-1.85537994e+00 -5.89851379e-01 3.49004269e-01 -4.15746778e-01
6.27883136e-01 2.68501073e-01 4.17383075e-01 1.03646874e+00
-1.32846460e-01 7.17942536e-01 1.15193582e+00 -2.73065656e-01
-3.38894278e-01 4.18908626e-01 4.13112640e-01 4.82237101e-01
4.98095095e-01 1.59437299e-01 -3.06100368e-01 -1.08021889e-02
5.35924852e-01 2.72151858e-01 -1.81907520e-01 1.44999921e-02
-1.02593362e+00 4.51250762e-01 5.15517831e-01 9.12950113e-02
-1.94100291e-01 7.73651851e-03 5.09646118e-01 3.98178287e-02
3.70991826e-01 -5.29306494e-02 -9.99235436e-02 1.03611991e-01
-7.91640401e-01 4.03813541e-01 3.48075241e-01 8.05608571e-01
5.36044657e-01 -3.20708573e-01 -2.47782573e-01 1.17719567e+00
1.71293870e-01 5.49917877e-01 5.53994298e-01 -4.67034519e-01
4.62506682e-01 5.16864300e-01 3.73629987e-01 -1.18621290e+00
-2.89545804e-01 -5.92932045e-01 -1.35903609e+00 -3.08178756e-02
6.23748422e-01 1.82438884e-02 -6.93063140e-01 1.68685079e+00
2.07891732e-01 6.29681945e-01 -8.47318675e-03 9.84961510e-01
1.00518274e+00 3.71085465e-01 3.09722692e-01 2.15609297e-01
1.60153949e+00 -1.03054142e+00 -2.38326460e-01 -3.12204093e-01
5.86494803e-02 -6.57153785e-01 4.23341185e-01 8.41881409e-02
-7.66715109e-01 -1.16530430e+00 -9.23033595e-01 9.11825150e-02
-3.88429701e-01 6.81502879e-01 8.43609422e-02 9.26509678e-01
-1.15618026e+00 4.71216172e-01 -3.34352672e-01 -7.82008588e-01
3.80548149e-01 6.13764644e-01 -6.31483674e-01 -1.85001880e-01
-9.88642395e-01 8.29103947e-01 3.43496233e-01 2.01713502e-01
-5.82836032e-01 -6.21552765e-01 -6.71884716e-01 2.09740251e-02
4.67783585e-02 -6.98820055e-01 9.35874581e-01 -8.95837545e-01
-1.12118948e+00 1.24367952e+00 -3.01877230e-01 -5.76054275e-01
8.80557835e-01 -3.46075714e-01 -6.86397254e-01 -1.53505132e-01
3.90833169e-01 6.93927467e-01 6.50888681e-01 -1.15147340e+00
-1.13634515e+00 -5.57620287e-01 -1.11243486e-01 1.94162682e-01
-5.49766243e-01 4.38849211e-01 -5.39979577e-01 -5.53633988e-01
-2.77950466e-01 -1.08347952e+00 -5.58172613e-02 -1.57431856e-01
-6.66659415e-01 -3.24330598e-01 5.89482844e-01 -1.01794243e+00
7.53421426e-01 -2.00071812e+00 -9.99213606e-02 5.17497025e-02
3.24970812e-01 6.54753804e-01 -1.14944369e-01 -9.95451212e-02
-3.01321357e-01 -2.31612958e-02 3.23410362e-01 -8.09975028e-01
-2.14992136e-01 -4.77350563e-01 4.27853726e-02 6.76366270e-01
-4.96785119e-02 7.94905782e-01 -6.27124548e-01 -3.11885029e-01
5.15249908e-01 5.88023782e-01 -4.23993990e-02 2.48404860e-01
7.29393065e-01 5.64574301e-01 -1.98267519e-01 4.81291533e-01
9.42200124e-01 -2.75474221e-01 -7.96846300e-02 -6.15652084e-01
-1.40445158e-01 -4.60823208e-01 -1.37565458e+00 1.08438933e+00
-1.60087720e-01 6.86734974e-01 -2.59166270e-01 -7.91669250e-01
7.53891528e-01 1.80451274e-01 3.67301792e-01 -6.83196723e-01
3.11764508e-01 8.28873441e-02 -9.78313312e-02 -1.63728341e-01
7.53353655e-01 1.29774258e-01 5.30143548e-03 2.02659398e-01
-9.86591950e-02 8.73759985e-01 3.45501214e-01 -1.11546747e-01
5.72344363e-01 -1.82981431e-01 4.30028856e-01 -2.09470063e-01
1.15600252e+00 -5.64768612e-01 6.45913720e-01 8.21160853e-01
-6.76537871e-01 7.88239479e-01 -2.70336509e-01 -8.40816736e-01
-1.23913682e+00 -8.95451427e-01 -1.27565295e-01 9.33945119e-01
5.60332835e-01 -1.81272447e-01 -6.85799897e-01 -4.53313947e-01
9.01702493e-02 2.43851721e-01 -6.20209277e-01 -3.70117184e-03
-7.32463598e-01 -1.10674465e+00 7.50673652e-01 6.34784818e-01
1.24808133e+00 -5.37750661e-01 -1.76897377e-01 -9.36457887e-03
-4.82622266e-01 -1.49538875e+00 -7.80326128e-01 -8.48000705e-01
-2.51249462e-01 -1.26272452e+00 -1.29603446e+00 -1.02693892e+00
7.80255914e-01 6.38176024e-01 7.41173089e-01 3.67375702e-01
-3.96476120e-01 2.51886785e-01 -1.00837879e-01 -4.73790318e-02
-1.76450998e-01 5.42450603e-03 5.82380533e-01 6.59503639e-01
5.34107268e-01 -2.00378343e-01 -8.20294321e-01 7.44281173e-01
-3.05364341e-01 8.62628147e-02 2.52830684e-01 8.65356982e-01
4.40470517e-01 8.11353549e-02 5.36536038e-01 -3.92798632e-01
4.58731949e-01 -1.31104887e-01 -6.11151338e-01 6.41758978e-01
-3.73476565e-01 -4.74346042e-01 6.27393901e-01 -4.39215690e-01
-1.00241959e+00 1.69578820e-01 2.36976519e-02 -2.36689791e-01
-4.20929372e-01 -1.93587869e-01 -1.23417288e-01 -3.28111500e-01
2.64094651e-01 2.88458705e-01 -2.32035518e-01 -4.57345039e-01
4.90097776e-02 8.04239690e-01 8.75535250e-01 -2.61917204e-01
1.02584755e+00 4.28802848e-01 -1.67615414e-01 -8.81566346e-01
-6.15220606e-01 -7.45380163e-01 -6.49122298e-01 -3.34579945e-01
9.58374143e-01 -1.36489522e+00 -1.27077019e+00 1.09931993e+00
-1.23706865e+00 4.16214198e-01 1.51810855e-01 4.24673945e-01
5.83392270e-02 7.39668548e-01 -5.46002626e-01 -8.10191393e-01
-6.43282413e-01 -1.31769371e+00 8.74544680e-01 7.86230445e-01
3.03731449e-02 -6.81279302e-01 -2.25198686e-01 6.08855784e-01
4.57940191e-01 8.61973688e-02 1.86184555e-01 -7.68621266e-01
-5.21381915e-01 -6.44198239e-01 -9.06381309e-01 2.89870858e-01
3.14612478e-01 -2.88928479e-01 -1.13085639e+00 -7.09313393e-01
-5.18991649e-01 -9.45591182e-02 9.50874627e-01 2.99643487e-01
1.18602538e+00 -1.91460326e-01 -5.62671483e-01 6.93745375e-01
1.42246604e+00 3.19249555e-02 5.28522849e-01 6.19343877e-01
9.30840850e-01 4.83543068e-01 1.91318884e-01 5.29424727e-01
6.86602116e-01 9.40367997e-01 1.38383359e-01 -1.58751607e-01
-1.87994406e-01 -1.80340305e-01 3.60820070e-02 2.96105921e-01
-3.92385453e-01 -2.43103519e-01 -7.81651258e-01 5.03356934e-01
-1.98538268e+00 -1.20706201e+00 -1.69656143e-01 2.50526762e+00
3.18593651e-01 -1.81785241e-01 4.72393095e-01 -8.10669363e-02
1.53677559e+00 -5.62951863e-02 -5.73908508e-01 3.32702398e-01
-3.25111777e-01 -1.44467413e-01 9.14202213e-01 2.92428702e-01
-1.61932290e+00 8.81038308e-01 5.31011343e+00 7.14280128e-01
-6.74525082e-01 6.23181723e-02 1.03889859e+00 7.98097849e-02
6.99980915e-01 -5.15897512e-01 -1.38868284e+00 7.17041969e-01
6.72920883e-01 -5.16352534e-01 3.17348242e-01 4.62305903e-01
4.39044572e-02 -9.04887393e-02 -1.08078909e+00 1.61419630e+00
4.20620412e-01 -1.14195478e+00 1.70798209e-02 8.31879396e-03
8.12993586e-01 -2.36719549e-01 2.44979024e-01 7.34222680e-02
2.00646967e-01 -1.09617352e+00 5.91350853e-01 5.10761201e-01
9.08351064e-01 -8.92770886e-01 1.25920403e+00 5.19290902e-02
-1.62100554e+00 -7.94531256e-02 -4.52063918e-01 7.36466646e-02
2.78743595e-01 2.22776413e-01 -5.74899733e-01 6.45720184e-01
1.17306697e+00 9.99209106e-01 -9.59698856e-01 1.40042388e+00
1.13537274e-01 1.02422662e-01 -1.95990682e-01 1.41949490e-01
-3.24860394e-01 -9.56056044e-02 3.66706997e-01 1.13259578e+00
1.76290452e-01 -2.58288644e-02 4.02870744e-01 6.07504487e-01
-3.12592804e-01 1.23970071e-02 -2.58156508e-01 5.35785019e-01
2.60956764e-01 1.25250256e+00 -4.48890924e-01 -4.36237931e-01
-3.97768766e-01 1.31762946e+00 2.18831286e-01 2.92012125e-01
-9.12245393e-01 -2.44122386e-01 1.07496297e+00 -2.26621822e-01
1.69562504e-01 -3.25413309e-02 -7.81729966e-02 -1.28650272e+00
1.32377103e-01 -7.58055925e-01 5.09841084e-01 -3.61453891e-01
-1.72145092e+00 6.91744864e-01 -1.35256678e-01 -1.32989931e+00
-5.34685329e-03 -6.05867982e-01 -4.54237878e-01 1.16120994e+00
-1.57005930e+00 -1.46047068e+00 -7.79974818e-01 6.43791497e-01
5.51776886e-01 -6.73077345e-01 6.33567035e-01 7.16192842e-01
-9.89530802e-01 1.27024508e+00 2.79320776e-01 6.85461581e-01
9.23854470e-01 -8.41894448e-01 8.07487249e-01 1.13420844e+00
-3.50967348e-01 4.79665011e-01 4.74642485e-01 -5.70745111e-01
-9.81129348e-01 -1.34618413e+00 9.23944712e-01 -4.07809794e-01
3.55245441e-01 -1.16717182e-01 -7.18509138e-01 5.46328843e-01
9.49124340e-03 9.00465697e-02 7.24499345e-01 -6.92696944e-02
-4.10869330e-01 -4.13768113e-01 -1.31721413e+00 5.35369575e-01
1.04869771e+00 -4.04534847e-01 -2.13213399e-01 2.34959468e-01
2.79087365e-01 -2.21187785e-01 -6.57558680e-01 3.02349538e-01
8.69833171e-01 -9.58716035e-01 1.54992938e+00 -5.31760514e-01
-6.35611042e-02 -4.10609663e-01 -1.63796604e-01 -1.12515152e+00
-6.66124105e-01 -1.46766335e-01 3.55226457e-01 1.45464921e+00
-5.83545938e-02 -8.71143043e-01 8.25468302e-01 9.58244801e-01
5.25502205e-01 -1.08250670e-01 -1.05694497e+00 -7.72906184e-01
-2.63875276e-01 2.29787175e-02 6.78098083e-01 6.53199613e-01
-5.55907547e-01 2.33699411e-01 -9.54470456e-01 4.14681524e-01
1.15003264e+00 -1.03528142e-01 1.01303697e+00 -1.42312431e+00
8.57448019e-03 -4.29547191e-01 -8.23736072e-01 -9.55411434e-01
3.80278856e-01 -7.83281386e-01 -1.95685074e-01 -1.16459239e+00
8.60180616e-01 -3.49895835e-01 -5.80221236e-01 3.05587500e-01
-5.01807034e-01 7.66432047e-01 4.63614285e-01 3.56256425e-01
-6.60975873e-01 5.81613243e-01 6.22371137e-01 -4.74802643e-01
2.28513740e-02 2.59446651e-01 -6.97875679e-01 5.91280758e-01
9.49516058e-01 -9.12121311e-02 9.69160125e-02 -4.99369740e-01
-3.98534745e-01 -4.55077499e-01 1.00760758e+00 -1.35508943e+00
5.03975034e-01 2.10281551e-01 9.28166807e-01 -5.45290470e-01
5.03146470e-01 -7.00765848e-01 3.47174138e-01 4.83787686e-01
-2.85324663e-01 4.32195604e-01 3.16813439e-01 6.22218788e-01
-4.95863222e-02 2.48275921e-02 1.04443955e+00 -5.00096828e-02
-9.61949050e-01 6.06940687e-01 5.75787611e-02 -1.86284125e-01
1.12956810e+00 -3.25370759e-01 -5.02701640e-01 -2.84421146e-01
-3.85750204e-01 2.80757129e-01 5.08477330e-01 7.14257002e-01
7.45144486e-01 -1.41971636e+00 -1.08615017e+00 1.74307063e-01
1.93909228e-01 -5.83800018e-01 4.67423469e-01 4.54825222e-01
-1.78824499e-01 3.81162733e-01 -5.28249323e-01 -5.40743172e-01
-1.91239381e+00 2.73322433e-01 6.50980532e-01 -2.03513712e-01
-4.18184787e-01 8.05438757e-01 3.25129628e-01 -2.62103409e-01
2.22489521e-01 3.66077662e-01 -4.47430938e-01 -1.28695965e-01
1.02378559e+00 8.39198112e-01 -1.48558453e-01 -1.31430030e+00
-6.37404442e-01 8.28055203e-01 -5.84408820e-01 1.13654636e-01
1.05094528e+00 -2.00036034e-01 -5.37134297e-02 -8.19496214e-02
1.33266032e+00 -4.21127588e-01 -1.05214143e+00 -4.79237288e-01
-1.91087291e-01 -8.38955402e-01 -3.53755623e-01 -6.60355866e-01
-1.07909846e+00 6.25048041e-01 1.25043929e+00 -1.06902972e-01
9.04308438e-01 -3.33630562e-01 9.48152483e-01 2.09815145e-01
4.64104682e-01 -1.06305039e+00 -9.55365375e-02 2.41986200e-01
6.10470712e-01 -1.64208436e+00 1.27735674e-01 -3.13056886e-01
-3.32426965e-01 9.08312440e-01 8.62439334e-01 -2.75681820e-02
4.45181012e-01 -4.82007444e-01 -9.67238694e-02 4.86009508e-01
5.93014993e-02 -2.41987765e-01 4.66369569e-01 7.01005399e-01
1.96726620e-01 2.62070119e-01 6.02892786e-03 4.60151613e-01
-1.32528648e-01 7.97051471e-03 2.60205209e-01 3.77170593e-01
-1.67916328e-01 -1.08240473e+00 -5.75418293e-01 4.95824188e-01
-4.34230328e-01 6.46283254e-02 -1.84536785e-01 4.54283893e-01
1.59311026e-01 1.19916987e+00 4.40389551e-02 -7.46885240e-01
3.61061662e-01 -2.57995486e-01 3.23769599e-01 -1.83568373e-01
-6.89845800e-01 -5.22237778e-01 -2.53720153e-02 -2.76059598e-01
-6.41015828e-01 -8.19922984e-01 -5.66027403e-01 -7.32831776e-01
-1.78991139e-01 -1.40805662e-01 4.48601574e-01 8.14883113e-01
4.58653241e-01 2.36552775e-01 6.64614499e-01 -9.65510070e-01
-5.12160122e-01 -8.05930972e-01 -3.53328824e-01 7.35938787e-01
3.92112881e-01 -7.29606926e-01 -4.44028489e-02 1.31159708e-01] | [14.695645332336426, 0.9311878681182861] |
c441cd5d-3d8a-437b-ba2b-5d211f01a0a7 | the-spike-gating-flow-a-hierarchical | 2206.01910 | null | https://arxiv.org/abs/2206.01910v2 | https://arxiv.org/pdf/2206.01910v2.pdf | The Spike Gating Flow: A Hierarchical Structure Based Spiking Neural Network for Online Gesture Recognition | Action recognition is an exciting research avenue for artificial intelligence since it may be a game changer in the emerging industrial fields such as robotic visions and automobiles. However, current deep learning faces major challenges for such applications because of the huge computational cost and the inefficient learning. Hence, we develop a novel brain-inspired Spiking Neural Network (SNN) based system titled Spiking Gating Flow (SGF) for online action learning. The developed system consists of multiple SGF units which assembled in a hierarchical manner. A single SGF unit involves three layers: a feature extraction layer, an event-driven layer and a histogram-based training layer. To demonstrate the developed system capabilities, we employ a standard Dynamic Vision Sensor (DVS) gesture classification as a benchmark. The results indicate that we can achieve 87.5% accuracy which is comparable with Deep Learning (DL), but at smaller training/inference data number ratio 1.5:1. And only a single training epoch is required during the learning process. Meanwhile, to the best of our knowledge, this is the highest accuracy among the non-backpropagation algorithm based SNNs. At last, we conclude the few-shot learning paradigm of the developed network: 1) a hierarchical structure-based network design involves human prior knowledge; 2) SNNs for content based global dynamic feature detection. | ['Yuan Xie', 'Junwen Luo', 'C. -J. Richard Shi', 'Xiaoan Wang', 'Jiansong Zhang', 'Fangbo Tao', 'Tie XU', 'Qiaosha Zou', 'Yanhong Wang', 'Zihao Zhao'] | 2022-06-04 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 4.92036700e-01 -2.96251625e-01 -5.89190759e-02 -8.90008669e-05
1.41425747e-02 -1.06419787e-01 6.57434702e-01 -3.98106091e-02
-7.96230316e-01 7.43379951e-01 -2.00608820e-01 -6.05859570e-02
-1.91781282e-01 -9.39863801e-01 -7.12614298e-01 -1.15591979e+00
1.29104868e-01 1.07448407e-01 1.00413203e+00 -1.92878246e-01
5.58587909e-01 6.10773444e-01 -1.95265734e+00 2.94040948e-01
5.74846745e-01 1.40194118e+00 4.86431301e-01 5.25451481e-01
-2.21348435e-01 1.00064659e+00 -3.88776839e-01 1.96190163e-01
2.03241169e-01 -5.71908832e-01 -3.06810826e-01 -1.68070108e-01
-9.61390659e-02 -2.90974647e-01 -4.32810217e-01 9.87172782e-01
6.99320912e-01 3.17914516e-01 6.60300255e-01 -1.27604997e+00
-3.15845370e-01 2.68861264e-01 -3.69171590e-01 5.33895552e-01
-1.47326365e-02 5.65259337e-01 5.11110663e-01 -6.39997780e-01
6.07352912e-01 8.54993641e-01 4.19067949e-01 7.98940718e-01
-6.25565767e-01 -7.33213365e-01 7.39916563e-02 5.63131630e-01
-7.50467241e-01 -2.47050911e-01 8.02690446e-01 -4.17901665e-01
1.35342956e+00 -3.15070361e-01 1.07277608e+00 1.27567732e+00
5.62941074e-01 8.63864541e-01 1.23838425e+00 -1.27534285e-01
7.67189026e-01 -4.36025739e-01 2.75522143e-01 7.39430964e-01
3.43590319e-01 2.32530415e-01 -8.36847842e-01 4.18477595e-01
1.00727832e+00 2.98157245e-01 7.79934824e-02 -1.34950265e-01
-9.86017585e-01 7.00485647e-01 5.68813980e-01 4.53646421e-01
-6.14698350e-01 5.58017850e-01 5.72621286e-01 1.68784425e-01
-1.79574817e-01 8.19478855e-02 -3.39183599e-01 -3.89285475e-01
-6.94369614e-01 1.75346345e-01 6.83681846e-01 3.87847513e-01
4.60551381e-01 5.50589979e-01 -3.14977765e-02 5.38609087e-01
2.63722718e-01 3.54083389e-01 9.74235117e-01 -8.13804328e-01
1.96186360e-02 7.27872968e-01 -1.62907138e-01 -8.40856135e-01
-5.30008197e-01 -1.56158969e-01 -8.10227454e-01 7.72200048e-01
4.06689137e-01 -1.69698477e-01 -1.08752060e+00 1.44585359e+00
2.96959616e-02 4.08495486e-01 1.10686317e-01 8.94011676e-01
9.12892103e-01 6.38978183e-01 3.36475462e-01 -2.92956293e-01
1.17056131e+00 -8.46037149e-01 -6.70328319e-01 -2.17310652e-01
2.43550226e-01 -2.44478106e-01 8.40758264e-01 5.54598987e-01
-8.38598073e-01 -5.95149994e-01 -1.33950102e+00 4.14164603e-01
-7.54464865e-01 2.30985712e-02 1.01719022e+00 4.96526688e-01
-8.72105837e-01 7.05596805e-01 -1.04275787e+00 -7.13381946e-01
6.52853847e-01 6.16708517e-01 -1.34354115e-01 2.23899171e-01
-8.93027723e-01 7.32582629e-01 8.01375866e-01 -1.50163220e-02
-9.83002245e-01 -7.30556846e-02 -6.01186693e-01 6.83636665e-02
2.33585566e-01 -5.23139060e-01 1.03647780e+00 -8.70520294e-01
-1.96205175e+00 6.05661273e-01 1.06263254e-02 -7.67734528e-01
1.23678565e-01 5.15042767e-02 -2.38872379e-01 2.37088934e-01
-1.86270207e-01 7.11858869e-01 7.67968833e-01 -7.99282074e-01
-7.34610975e-01 -4.85486716e-01 -7.84449801e-02 1.85066000e-01
-3.99604172e-01 -1.71036318e-01 1.07666040e-02 -5.86232483e-01
1.72366768e-01 -6.74704194e-01 -2.98616383e-02 2.53639877e-01
1.85534850e-01 -4.93698299e-01 1.00697947e+00 -3.16486686e-01
9.52167988e-01 -2.00621653e+00 -1.20373957e-01 -7.64621720e-02
1.30775645e-02 5.99536180e-01 6.84448238e-03 2.41309538e-01
2.58044988e-01 -3.98207337e-01 -3.35274875e-01 3.16360384e-01
-1.67219639e-01 2.73935854e-01 6.20914288e-02 1.98705673e-01
1.29129693e-01 1.09262884e+00 -8.74761164e-01 -3.46706897e-01
5.14820814e-01 8.13088194e-02 -2.76717067e-01 6.81055337e-02
-2.73477346e-01 3.77218068e-01 -3.49965960e-01 9.44905460e-01
2.76304930e-01 -3.94658148e-02 -1.01252124e-01 -2.78761655e-01
-4.23854262e-01 -2.04545096e-01 -1.10660517e+00 1.80520368e+00
2.37225788e-03 6.72938287e-01 -1.86089069e-01 -1.46534979e+00
1.26438344e+00 9.59150791e-02 7.48120785e-01 -1.15872562e+00
5.26823938e-01 3.21866840e-01 2.64081359e-01 -7.70889044e-01
-8.51912424e-02 -1.70637786e-01 1.03439666e-01 2.55366117e-01
5.15805066e-01 3.39966625e-01 2.52736121e-01 -2.47758582e-01
1.51718020e+00 2.09822446e-01 2.96280295e-01 1.16568329e-02
4.20405477e-01 -2.05109511e-02 6.89603925e-01 8.15227807e-01
-7.53864169e-01 1.69042483e-01 1.68150842e-01 -5.42840421e-01
-7.03705728e-01 -1.19174123e+00 5.10478653e-02 1.13048398e+00
4.12711561e-01 1.69644594e-01 -6.14144921e-01 -3.35374981e-01
-1.60735965e-01 3.18258554e-01 -4.57583636e-01 -2.56972283e-01
-5.35836756e-01 -8.30182672e-01 7.08569407e-01 7.96832383e-01
1.15767813e+00 -1.77505648e+00 -1.15066242e+00 5.52597880e-01
4.47086483e-01 -9.10573542e-01 2.34474823e-01 5.29194653e-01
-1.04662836e+00 -1.20700002e+00 -4.98336405e-01 -1.13124311e+00
1.88268006e-01 -1.29393523e-03 4.59356070e-01 -2.50143826e-01
-4.99430120e-01 2.22130075e-01 -3.59938174e-01 -6.25026524e-01
1.61296308e-01 -2.09429994e-01 5.18627279e-02 -1.42873460e-02
7.81626523e-01 -9.28428113e-01 -8.79121721e-01 -4.56575258e-03
-7.42722511e-01 1.21074859e-02 9.44488883e-01 9.22435284e-01
5.94285250e-01 1.08970396e-01 7.77618706e-01 -5.78257918e-01
5.11506259e-01 -4.16727901e-01 -5.70296526e-01 2.73825638e-02
-4.84434366e-01 -4.95715514e-02 7.18968809e-01 -5.24685919e-01
-1.21411514e+00 4.45077330e-01 -1.75471097e-01 -1.08472906e-01
-4.57423896e-01 2.85034180e-01 -2.41177678e-02 -3.83252949e-01
6.30585611e-01 6.74542606e-01 9.96725336e-02 -2.80972332e-01
-5.60119003e-02 6.88457847e-01 6.85291588e-01 -1.60852224e-01
1.87198922e-01 4.66736197e-01 5.36057614e-02 -7.71854460e-01
-2.43309975e-01 -5.20780027e-01 -4.60743785e-01 -6.69650614e-01
1.11850321e+00 -6.52585626e-01 -1.11718154e+00 1.14542127e+00
-1.12817907e+00 -3.81114066e-01 -1.58729300e-01 6.23919725e-01
-7.27717578e-01 -4.82109152e-02 -7.90768325e-01 -9.12819862e-01
-5.08034408e-01 -8.71215880e-01 6.39466465e-01 7.66441762e-01
1.48367494e-01 -7.38480031e-01 1.89196974e-01 2.31247962e-01
4.84447658e-01 6.14359200e-01 7.15770543e-01 -6.95160031e-01
-5.15230238e-01 5.71990609e-02 -2.55109072e-01 1.74671322e-01
1.68971419e-01 -1.69306949e-01 -1.00049031e+00 -9.17963833e-02
1.02869511e-01 -4.83753353e-01 1.23610640e+00 5.60116291e-01
1.08650827e+00 4.98576313e-02 -2.62424976e-01 2.94983685e-01
1.74988711e+00 1.01444423e+00 7.49207914e-01 4.58152562e-01
4.91618276e-01 1.57245755e-01 3.51105362e-01 6.30049706e-01
-1.38695240e-01 2.42150590e-01 6.48115456e-01 3.10145557e-01
-1.78544551e-01 -6.36707023e-02 5.15762150e-01 9.13821995e-01
-3.61154288e-01 -1.64507970e-01 -8.53430212e-01 4.37200695e-01
-2.11851144e+00 -1.16832054e+00 -1.82430614e-02 1.96290064e+00
4.99571294e-01 5.00902534e-01 1.30224511e-01 4.09369379e-01
5.96974492e-01 1.30047187e-01 -1.15059674e+00 -4.10242289e-01
-8.77844170e-02 6.98261976e-01 3.54843915e-01 -7.36883879e-02
-1.13988709e+00 1.01557386e+00 5.43451786e+00 7.27237105e-01
-1.40037632e+00 3.03737018e-02 1.23395480e-01 2.22868097e-04
3.47128958e-01 -3.19248170e-01 -7.14806497e-01 8.01452994e-01
8.65611792e-01 -9.65646096e-03 4.81380135e-01 8.32130611e-01
2.18414500e-01 -3.87754947e-01 -5.46907306e-01 1.32315218e+00
4.97358525e-03 -1.58081758e+00 1.91003338e-01 -2.08290070e-01
4.91775781e-01 1.66033894e-01 -3.09508115e-01 3.88858408e-01
1.05919922e-02 -9.46547985e-01 2.79292881e-01 7.78865159e-01
3.93294126e-01 -7.52228975e-01 8.00498009e-01 4.13731635e-01
-1.36446762e+00 -5.87920249e-01 -3.69409919e-01 -5.25712252e-01
-4.11045318e-03 2.44268447e-01 -3.05633694e-01 9.18863937e-02
8.24440420e-01 8.67423415e-01 -3.55559617e-01 1.34587395e+00
9.43622589e-02 5.62441647e-01 -1.44168735e-01 -6.26023650e-01
3.57735127e-01 -4.32311483e-02 3.78839225e-01 1.04988062e+00
2.29432985e-01 2.54501104e-01 9.00000855e-02 5.88442206e-01
-3.79502438e-02 -1.39471158e-01 -8.09318602e-01 -1.88383639e-01
4.19917494e-01 1.11912286e+00 -1.15239584e+00 -2.95691311e-01
-3.92591506e-01 1.02295744e+00 1.85749963e-01 2.14951530e-01
-6.32713079e-01 -7.81453371e-01 4.28737819e-01 -1.44897550e-01
4.67861027e-01 -1.47982135e-01 -3.80817324e-01 -8.99759948e-01
-1.33469895e-01 -4.38517541e-01 2.28471518e-01 -6.98690355e-01
-1.00009012e+00 2.53308535e-01 -3.50554883e-01 -1.18550301e+00
-1.97776824e-01 -1.13117623e+00 -9.03989851e-01 2.56981730e-01
-1.33835495e+00 -8.56425405e-01 -6.15458846e-01 7.62696981e-01
7.70754755e-01 -5.22177994e-01 7.63638914e-01 1.44360229e-01
-5.84791183e-01 1.72420278e-01 1.02528408e-01 2.80398935e-01
2.65480548e-01 -1.07402802e+00 1.07092075e-02 8.97360384e-01
-2.62689292e-02 2.42587417e-01 2.48358443e-01 -5.91050684e-01
-1.47278678e+00 -1.02495706e+00 3.53813648e-01 1.43611476e-01
6.45060003e-01 -1.38188720e-01 -8.09894264e-01 2.02426255e-01
1.14239700e-01 1.09433174e-01 5.46178937e-01 -5.83278000e-01
1.36179432e-01 -3.68930131e-01 -1.17844975e+00 3.60578746e-01
1.20574450e+00 -2.18614489e-01 -6.34008288e-01 4.34195288e-02
4.31542218e-01 -1.73857167e-01 -5.83440423e-01 4.37862217e-01
8.32475841e-01 -1.13205254e+00 5.40635109e-01 -4.87250298e-01
2.03001529e-01 -4.64367986e-01 -4.62713912e-02 -8.39880586e-01
-4.46440756e-01 -3.26915920e-01 -4.82708037e-01 8.98955822e-01
-1.40665695e-01 -6.51476860e-01 1.11842501e+00 2.73436666e-01
-2.31121778e-01 -1.04331744e+00 -1.10436380e+00 -1.00539219e+00
-2.53245413e-01 -2.66267478e-01 -1.94409341e-02 4.75746125e-01
-5.72733134e-02 2.50583678e-01 1.10129956e-02 -2.18969733e-01
6.92835867e-01 -1.29975423e-01 2.42883176e-01 -1.44339299e+00
-2.98092335e-01 -5.19294977e-01 -9.88000214e-01 -8.20981205e-01
-1.59880430e-01 -5.10238111e-01 2.22422495e-01 -1.65555716e+00
1.19997658e-01 -5.28072678e-02 -8.14020097e-01 5.64185202e-01
2.78461963e-01 1.24897420e-01 3.32629532e-02 5.57668023e-02
-6.77481532e-01 5.63675463e-01 1.03171837e+00 -1.40176937e-01
-3.11785042e-01 -9.00117680e-03 -2.59193003e-01 8.01749110e-01
1.09201324e+00 -3.49783421e-01 -5.85374117e-01 -6.92265853e-02
-2.38944426e-01 -1.34846181e-01 5.11536658e-01 -1.57347298e+00
7.86442339e-01 -2.82699347e-01 5.09799004e-01 -4.47550178e-01
2.03645900e-01 -6.53845906e-01 -1.64747834e-01 1.03894520e+00
1.43017843e-02 -2.88882762e-01 1.56345293e-01 7.44917512e-01
-2.05882117e-01 -1.77626297e-01 9.00670111e-01 -3.84646505e-01
-1.55378902e+00 3.15321147e-01 -7.70192862e-01 -1.63751006e-01
1.26776397e+00 -6.82223260e-01 -3.21470588e-01 8.13263282e-02
-5.26253700e-01 -1.03101879e-01 5.67241339e-03 3.15996379e-01
8.90103459e-01 -1.29061854e+00 -1.08940341e-01 1.82437330e-01
1.32366214e-02 -1.53441086e-01 2.29312211e-01 6.65795386e-01
-5.98811328e-01 3.87041450e-01 -9.66577113e-01 -4.59694028e-01
-1.01662123e+00 2.94524759e-01 2.45166004e-01 -1.05993949e-01
-5.22439599e-01 6.96744084e-01 -1.14251763e-01 -9.71035659e-02
5.15551746e-01 -2.56110221e-01 -5.30668795e-01 8.95808712e-02
5.47002912e-01 6.97628498e-01 1.33793652e-02 -2.38657013e-01
-4.81513679e-01 4.61153597e-01 1.16727106e-01 -9.31058899e-02
1.54043615e+00 3.42494816e-01 7.70079568e-02 6.57028377e-01
8.53715360e-01 -9.28651452e-01 -1.61646533e+00 6.51724562e-02
1.32327735e-01 7.79395327e-02 -2.27835737e-02 -8.25131357e-01
-9.72899139e-01 9.23016965e-01 1.04945362e+00 -1.15245029e-01
1.24771845e+00 -3.73625308e-01 1.03694272e+00 8.31306219e-01
5.52119553e-01 -1.47120750e+00 5.77288032e-01 6.29300177e-01
5.88262975e-01 -1.20890546e+00 -3.30658197e-01 1.61244541e-01
-3.58355463e-01 1.41668117e+00 1.05324149e+00 -6.12342954e-01
8.05810928e-01 2.83756077e-01 -1.50001109e-01 -2.81292647e-01
-7.05398619e-01 -4.44369286e-01 -9.71658900e-02 7.57531583e-01
4.40367498e-02 -1.86296046e-01 -5.49483597e-01 7.45117962e-01
2.32255489e-01 6.29298985e-01 1.71539307e-01 1.25015712e+00
-9.85776246e-01 -7.61252403e-01 1.03684612e-01 7.92840898e-01
-9.39744562e-02 1.30593134e-02 -1.36296988e-01 5.52177787e-01
4.81753170e-01 7.60397553e-01 7.56790414e-02 -6.82722688e-01
2.46076480e-01 2.11488724e-01 6.04609549e-01 -3.73720467e-01
-5.11060953e-01 -3.37537616e-01 -3.35295945e-01 -7.36736715e-01
-7.86835670e-01 -4.09871489e-01 -1.72431052e+00 -1.53350711e-01
2.87171360e-02 -3.95483315e-01 6.89670384e-01 1.20520890e+00
2.62090445e-01 5.51795304e-01 3.45372796e-01 -8.92276168e-01
-1.16766237e-01 -7.88187325e-01 -6.73816919e-01 2.08377913e-01
-5.88534847e-02 -9.30147529e-01 -2.08262220e-01 1.75389349e-01] | [8.22718334197998, 2.395636796951294] |
63648821-3e49-407e-b557-4e986943673c | exploring-large-scale-unlabeled-faces-to | 2303.08617 | null | https://arxiv.org/abs/2303.08617v2 | https://arxiv.org/pdf/2303.08617v2.pdf | Exploring Large-scale Unlabeled Faces to Enhance Facial Expression Recognition | Facial Expression Recognition (FER) is an important task in computer vision and has wide applications in human-computer interaction, intelligent security, emotion analysis, and other fields. However, the limited size of FER datasets limits the generalization ability of expression recognition models, resulting in ineffective model performance. To address this problem, we propose a semi-supervised learning framework that utilizes unlabeled face data to train expression recognition models effectively. Our method uses a dynamic threshold module (\textbf{DTM}) that can adaptively adjust the confidence threshold to fully utilize the face recognition (FR) data to generate pseudo-labels, thus improving the model's ability to model facial expressions. In the ABAW5 EXPR task, our method achieved excellent results on the official validation set. | ['Wangyuan Zhu', 'Jichao Zhu', 'Guochen Xie', 'Gongpeng Zhao', 'Renda Li', 'Zhongpeng Cai', 'Jun Yu'] | 2023-03-15 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [ 2.97610015e-01 -1.75934732e-01 -3.07728320e-01 -1.07662308e+00
-3.67544204e-01 -2.92290717e-01 9.52118561e-02 -6.25747621e-01
-4.13400441e-01 7.76062548e-01 -4.77720588e-01 -6.49117082e-02
3.49876851e-01 -4.92654294e-01 -1.32369593e-01 -7.83361197e-01
7.32000619e-02 -4.07762080e-02 -3.29236031e-01 -2.44993284e-01
-1.20562702e-01 6.73064530e-01 -1.58760405e+00 5.58230162e-01
8.11927676e-01 1.54807901e+00 -3.41385335e-01 2.76056796e-01
-2.62236744e-01 1.04474497e+00 -6.25712931e-01 -5.23182631e-01
-7.55303875e-02 -4.66389954e-01 -5.06776571e-01 3.08445007e-01
5.52671850e-02 -2.87460923e-01 -4.79916371e-02 1.19378519e+00
3.15886199e-01 1.65704340e-01 6.54704213e-01 -1.63211358e+00
-5.68672061e-01 1.90334767e-02 -9.16794837e-01 1.21937478e-02
2.22937495e-01 -1.65315732e-01 4.45227891e-01 -1.14694810e+00
4.67910528e-01 1.21887791e+00 4.29960579e-01 1.11593175e+00
-9.74265158e-01 -1.27637613e+00 9.95999482e-03 2.64823854e-01
-1.70016992e+00 -6.60646141e-01 9.17209566e-01 -2.18164071e-01
5.88321924e-01 1.34316415e-01 4.13240373e-01 9.90304112e-01
-2.40429074e-01 8.40280533e-01 1.44042325e+00 -5.39936483e-01
3.45854580e-01 2.44307265e-01 1.63643509e-01 8.57562363e-01
-5.70268631e-01 -3.26284803e-02 -6.33267105e-01 -3.33300591e-01
6.58305407e-01 -1.42752245e-01 -6.66467100e-02 2.54342377e-01
-1.21072002e-01 8.82143676e-01 2.06301928e-01 2.03977898e-01
-4.30300266e-01 -2.38503322e-01 3.83893549e-01 3.99523616e-01
7.86222577e-01 -1.91499162e-02 -5.10859311e-01 -1.90288275e-01
-6.65856898e-01 -1.66453764e-01 5.28267801e-01 5.06974638e-01
8.85588229e-01 3.59235883e-01 -8.00153464e-02 1.58073211e+00
3.92655820e-01 4.28599626e-01 4.38578814e-01 -1.11361492e+00
-1.27414748e-01 8.06007266e-01 -8.70573521e-02 -8.20579112e-01
-9.05378088e-02 2.24911228e-01 -8.59672546e-01 3.98075342e-01
1.57490075e-01 -3.35342050e-01 -9.65909183e-01 1.99841726e+00
2.50662506e-01 2.55093664e-01 3.25103551e-02 7.94824600e-01
7.77108312e-01 6.34436011e-01 6.28744423e-01 -5.92203438e-01
1.16452754e+00 -7.36499846e-01 -9.68167603e-01 -1.75961241e-01
6.79536581e-01 -4.09163237e-01 8.80189776e-01 6.08675897e-01
-5.72685659e-01 -5.12758315e-01 -8.07742059e-01 3.32886726e-01
-2.32330367e-01 5.09607732e-01 1.06726706e+00 8.00496876e-01
-8.49371493e-01 -4.71315533e-03 -5.69394767e-01 -4.60591763e-02
8.59513521e-01 6.52239740e-01 -7.26542473e-01 -1.34399161e-01
-1.21104002e+00 6.28801405e-01 1.85569391e-01 3.74539852e-01
-4.04136449e-01 -1.18327670e-01 -8.91093493e-01 -1.27998525e-02
8.04352015e-02 4.50510718e-02 1.33843398e+00 -1.92194283e+00
-1.74523032e+00 1.13317573e+00 -3.85821790e-01 -1.46413803e-01
1.57944500e-01 6.60144314e-02 -7.05927551e-01 1.30395740e-01
-4.27477688e-01 8.06801915e-01 1.04394972e+00 -9.20284331e-01
-2.19117329e-01 -7.01799750e-01 -4.15390998e-01 -5.82520440e-02
-5.89460909e-01 5.40531874e-01 -4.69527841e-01 -4.84815031e-01
-7.93328136e-02 -9.52448070e-01 -1.67628765e-01 3.00465882e-01
1.53018340e-01 -5.70933461e-01 1.17790818e+00 -3.91851395e-01
1.18253553e+00 -2.53213239e+00 -3.03615123e-01 4.26128536e-01
-1.60881460e-01 6.94812059e-01 -1.17209487e-01 -4.29861784e-01
-3.70552748e-01 -4.40934710e-02 -2.56401718e-01 -2.33843982e-01
-3.97838652e-01 3.96997154e-01 -2.20101595e-01 1.39087304e-01
5.13846636e-01 6.85902655e-01 -4.95099068e-01 -6.67445898e-01
7.11831897e-02 5.53792894e-01 -3.52866471e-01 4.64723915e-01
-1.78429574e-01 3.27388614e-01 -7.22949982e-01 9.73416209e-01
8.43225837e-01 -1.40506625e-01 2.30214745e-01 -6.00158349e-02
3.83435279e-01 -7.08610177e-01 -9.38319147e-01 1.17200661e+00
-4.07238513e-01 5.33616781e-01 2.15851679e-01 -9.45986986e-01
1.38277769e+00 3.73449743e-01 6.40606344e-01 -6.37804031e-01
4.83156979e-01 8.57147723e-02 -8.62670243e-02 -6.19178116e-01
1.01110421e-01 -2.31792256e-01 1.11781381e-01 2.79223859e-01
2.30465993e-01 2.80252248e-01 -5.42055862e-03 -2.28920773e-01
7.29180694e-01 8.88530537e-02 1.86567739e-01 1.28550962e-01
8.60948145e-01 -2.92110801e-01 8.38085055e-01 6.04878701e-02
-5.55226862e-01 1.92738578e-01 4.49347049e-01 -4.66221511e-01
-4.46057498e-01 -7.22251296e-01 -3.39627415e-01 1.34334922e+00
-3.49046141e-01 -2.52282768e-01 -9.57988620e-01 -9.33902800e-01
-1.94175616e-01 6.06817782e-01 -8.74676406e-01 -5.39381087e-01
-1.89699158e-01 -9.24953103e-01 7.30051398e-01 6.18387222e-01
6.93269730e-01 -1.37251854e+00 -9.86193120e-02 -4.74118665e-02
-2.32458934e-01 -1.12041235e+00 -1.09095849e-01 1.43294679e-02
-5.71666718e-01 -1.07112241e+00 -4.87454712e-01 -7.20457017e-01
1.08769858e+00 -8.11600536e-02 6.19787514e-01 2.06453875e-01
-3.49332392e-01 1.91499934e-01 -4.07658219e-01 -6.21691585e-01
-5.47689259e-01 -4.74527776e-01 1.40198410e-01 4.46075678e-01
9.44624066e-01 -1.45315796e-01 -2.19618514e-01 5.77018261e-01
-9.23587382e-01 -2.89221287e-01 3.06559891e-01 1.04993773e+00
5.81431508e-01 -8.31315592e-02 9.05086398e-01 -8.27450514e-01
7.09800303e-01 -3.65657151e-01 -5.23312330e-01 4.48374659e-01
-5.49292684e-01 -2.76246548e-01 5.21151841e-01 -7.78022051e-01
-1.53594410e+00 3.24065566e-01 -5.00214994e-01 -6.81800127e-01
-9.37210172e-02 4.50752050e-01 -3.77252936e-01 -4.52276051e-01
5.64694166e-01 -7.63684884e-03 4.82941866e-01 -2.44294301e-01
-2.26612408e-02 1.17183697e+00 4.70843583e-01 -5.38174510e-01
7.95040950e-02 1.74286976e-01 4.97035570e-02 -8.74526560e-01
-9.77159202e-01 -1.49404868e-01 -3.81285042e-01 -4.67372686e-01
6.07979715e-01 -9.62216258e-01 -8.42395842e-01 7.41236627e-01
-1.01703751e+00 -1.90174684e-01 2.87905455e-01 3.07135850e-01
-3.70562464e-01 1.01948209e-01 -7.27779746e-01 -1.21056235e+00
-4.31487232e-01 -1.04519296e+00 9.60344017e-01 5.09554744e-01
-2.75380343e-01 -5.92500448e-01 -1.20632641e-01 5.02633870e-01
2.46083438e-01 2.63958573e-01 7.30591536e-01 -5.60577095e-01
1.27569124e-01 -3.51468325e-01 -3.07360858e-01 1.03660738e+00
2.70124227e-01 3.66088420e-01 -1.26194739e+00 1.17636494e-01
-1.30509228e-01 -1.05105150e+00 7.09867954e-01 -1.98908038e-02
1.79011273e+00 -4.46029156e-02 -7.63829350e-02 4.85654801e-01
8.35069120e-01 5.14790595e-01 8.91668916e-01 -1.97952598e-01
2.08141029e-01 6.62914634e-01 9.07298088e-01 6.60893917e-01
-1.06768221e-01 6.07591987e-01 1.90130398e-02 -2.79160410e-01
3.08859378e-01 -7.31990114e-02 4.44560051e-01 1.32647172e-01
-1.63948730e-01 -1.22983366e-01 -5.89051008e-01 -1.66010752e-01
-1.68131280e+00 -9.38778162e-01 2.19327405e-01 1.85610807e+00
8.87982249e-01 -2.68899798e-01 -1.85270637e-01 6.92385584e-02
6.00456417e-01 -9.93567333e-02 -6.57148480e-01 -7.34147012e-01
-1.54267892e-01 5.56947708e-01 -2.60724932e-01 1.38670042e-01
-1.06313980e+00 1.14270139e+00 6.62877035e+00 9.81318176e-01
-1.46763694e+00 -8.08047876e-02 1.18865108e+00 1.58114240e-01
3.28362048e-01 -5.66354513e-01 -6.69655740e-01 3.76255721e-01
7.80727148e-01 -5.70315458e-02 3.03091228e-01 1.32561994e+00
3.00014287e-01 -5.66754229e-02 -7.42084622e-01 1.17950845e+00
2.26142481e-01 -6.53928697e-01 7.39881992e-02 -1.59366786e-01
4.43262756e-01 -3.82717550e-01 2.21285090e-01 6.27534688e-01
6.72733560e-02 -1.28066146e+00 6.90161064e-02 3.08500230e-01
1.25387681e+00 -8.50776792e-01 7.16428578e-01 1.75233126e-01
-7.03095794e-01 -1.40281871e-01 -3.01517069e-01 8.38122517e-02
-2.30980478e-02 1.66988939e-01 -7.50848472e-01 1.09311147e-02
7.07821310e-01 3.71625930e-01 -4.36839402e-01 3.47259134e-01
-2.89536834e-01 9.20441449e-01 -3.19572419e-01 1.24708400e-04
-7.99475014e-02 -3.88718992e-01 -1.16814554e-01 9.83275354e-01
2.19195336e-01 5.24534106e-01 9.28790122e-02 5.67357838e-01
-3.81618291e-01 3.62531304e-01 -3.23006302e-01 -2.80675262e-01
2.39050254e-01 1.55382228e+00 -4.22232747e-01 -2.07580343e-01
-4.16670948e-01 9.15609598e-01 5.24799585e-01 4.66861039e-01
-9.00622308e-01 -7.83184990e-02 6.73462451e-01 -1.99873015e-01
-1.45193875e-01 8.53193924e-03 1.16180606e-01 -1.05909944e+00
-1.27302542e-01 -9.94037390e-01 6.56566024e-01 -9.31047320e-01
-1.29052365e+00 7.91317046e-01 -1.06206693e-01 -1.04359043e+00
-4.65675950e-01 -8.79211128e-01 -5.65614700e-01 6.25594854e-01
-1.25353837e+00 -1.17545772e+00 -4.45614964e-01 8.41960788e-01
3.36458713e-01 -4.74490225e-01 1.16492605e+00 3.42210054e-01
-8.82638991e-01 9.55153644e-01 -3.53383362e-01 4.54548538e-01
8.25053334e-01 -5.44603586e-01 -4.74690825e-01 5.15691578e-01
1.24037161e-01 5.21989882e-01 3.77818614e-01 -1.95275053e-01
-1.05287707e+00 -1.06726325e+00 4.77002770e-01 1.03889935e-01
3.50029349e-01 -2.17279464e-01 -1.07862413e+00 4.95285124e-01
-2.43489504e-01 4.98414367e-01 1.36634874e+00 7.69087523e-02
-5.87988496e-01 -3.17125261e-01 -1.50955939e+00 4.65503186e-01
6.42279565e-01 -5.05180955e-01 -6.33038506e-02 7.50632584e-02
1.36877447e-01 -1.24266975e-01 -7.31262922e-01 8.07423174e-01
7.50196338e-01 -7.71027863e-01 5.93861938e-01 -9.52429116e-01
2.85619974e-01 -1.25321701e-01 -2.78136462e-01 -1.00270879e+00
-5.98520301e-02 -3.73583496e-01 -2.17291698e-01 1.30899882e+00
3.29269022e-01 -5.29847980e-01 9.52642143e-01 1.19851768e+00
4.23686296e-01 -9.59756613e-01 -9.71099436e-01 -4.28685725e-01
-3.35703701e-01 -4.46777105e-01 5.76519489e-01 1.02417445e+00
-5.44636212e-02 2.66033769e-01 -3.24017912e-01 -2.32206017e-01
3.16415280e-01 -3.34261470e-02 5.93802810e-01 -1.11232436e+00
-7.72842467e-02 -2.34920233e-01 -4.60925907e-01 -6.66269004e-01
9.04889762e-01 -7.40060270e-01 -6.88774213e-02 -5.79250813e-01
1.99598148e-01 -5.12591183e-01 -7.02136159e-01 1.03426194e+00
-2.08007485e-01 6.53282881e-01 -9.94445011e-02 -1.03049003e-01
-6.42303348e-01 8.46284926e-01 9.80102539e-01 -3.43166254e-02
-1.19964473e-01 2.34117687e-01 -6.06889486e-01 8.78380597e-01
8.28972876e-01 -2.20460519e-01 -3.91326636e-01 -9.43762623e-03
-3.33754629e-01 9.45122242e-02 1.05182745e-01 -4.19585377e-01
-1.55331329e-01 -4.49348748e-01 9.03309584e-01 7.74538293e-02
5.44021606e-01 -7.44085252e-01 -7.56721348e-02 4.15641479e-02
-2.44929060e-01 -1.46275058e-01 4.09589082e-01 1.58517793e-01
-4.54751045e-01 -2.26433352e-01 1.02875543e+00 1.45460695e-01
-1.04155433e+00 7.03656793e-01 -5.07537425e-01 -3.85750979e-01
1.13326716e+00 -1.21508799e-01 4.34041489e-03 -4.55372810e-01
-7.68737078e-01 2.17459142e-01 3.81753683e-01 5.46414375e-01
8.56331289e-01 -1.40151978e+00 -5.82346261e-01 5.59935331e-01
4.77060109e-01 -3.85842949e-01 2.67697513e-01 4.64330167e-01
-4.20947038e-02 -2.35539079e-01 -3.65884602e-01 -4.57023114e-01
-1.78871632e+00 3.04624200e-01 4.37672645e-01 -4.37649488e-02
2.14054972e-01 8.40870917e-01 6.58772290e-02 -2.79944897e-01
2.23483816e-01 5.92529535e-01 -4.28445607e-01 -1.64962113e-01
9.08934653e-01 -4.11713962e-03 -1.93119228e-01 -9.27348077e-01
-2.80420244e-01 3.15032750e-01 -1.92649409e-01 5.32025434e-02
1.06237614e+00 1.55218720e-01 -3.97191137e-01 1.93603933e-01
1.36036229e+00 -3.24220479e-01 -1.06901753e+00 -1.48242921e-01
1.13608064e-02 -4.31859314e-01 1.23243019e-01 -7.95435905e-01
-1.13422680e+00 8.03793669e-01 8.07727456e-01 -3.37503165e-01
1.62864530e+00 -1.93252832e-01 4.02906030e-01 5.65959275e-01
4.06961113e-01 -1.22653639e+00 2.54297942e-01 2.54781842e-01
6.75262094e-01 -1.48026383e+00 -2.09934637e-01 -5.82046330e-01
-9.66882408e-01 1.22299433e+00 1.24190271e+00 3.41150314e-01
7.39776611e-01 4.63879883e-01 5.29380083e-01 4.70435880e-02
-7.44467974e-01 1.05186090e-01 1.51514024e-01 4.68646318e-01
5.54498672e-01 -1.52706373e-02 -3.05884063e-01 9.24963713e-01
2.26325959e-01 6.69065237e-01 3.61033380e-02 9.24091220e-01
-3.54390204e-01 -1.28109860e+00 -1.35057747e-01 6.02519453e-01
-7.43515909e-01 3.09153169e-01 -6.39880478e-01 3.44474584e-01
7.14327544e-02 1.03915107e+00 3.23509201e-02 -4.99416411e-01
7.86624998e-02 4.46629226e-01 4.93753552e-01 -2.24385694e-01
-1.84597015e-01 -2.34851055e-02 1.02839716e-01 -4.73357469e-01
-6.70617759e-01 -3.88053089e-01 -1.42233706e+00 1.26315672e-02
-3.81808132e-01 2.95127720e-01 4.94131029e-01 9.25299108e-01
3.88726324e-01 -1.18821613e-01 1.20387113e+00 -1.97031617e-01
-5.14551044e-01 -1.03842628e+00 -6.83369994e-01 6.09270692e-01
-9.89456922e-02 -9.23832297e-01 -1.92079797e-01 2.90850312e-01] | [13.587821960449219, 1.7268437147140503] |
9fcdb66c-d5ec-4208-a1f5-c81849bf8e10 | end-to-end-adversarial-text-to-speech | 2006.03575 | null | https://arxiv.org/abs/2006.03575v3 | https://arxiv.org/pdf/2006.03575v3.pdf | End-to-End Adversarial Text-to-Speech | Modern text-to-speech synthesis pipelines typically involve multiple processing stages, each of which is designed or learnt independently from the rest. In this work, we take on the challenging task of learning to synthesise speech from normalised text or phonemes in an end-to-end manner, resulting in models which operate directly on character or phoneme input sequences and produce raw speech audio outputs. Our proposed generator is feed-forward and thus efficient for both training and inference, using a differentiable alignment scheme based on token length prediction. It learns to produce high fidelity audio through a combination of adversarial feedback and prediction losses constraining the generated audio to roughly match the ground truth in terms of its total duration and mel-spectrogram. To allow the model to capture temporal variation in the generated audio, we employ soft dynamic time warping in the spectrogram-based prediction loss. The resulting model achieves a mean opinion score exceeding 4 on a 5 point scale, which is comparable to the state-of-the-art models relying on multi-stage training and additional supervision. | ['Mikołaj Bińkowski', 'Sander Dieleman', 'Jeff Donahue', 'Karen Simonyan', 'Erich Elsen'] | 2020-06-05 | null | https://openreview.net/forum?id=rsf1z-JSj87 | https://openreview.net/pdf?id=rsf1z-JSj87 | iclr-2021-1 | ['adversarial-text'] | ['adversarial'] | [ 8.12605977e-01 4.62440848e-01 1.33700877e-01 -2.85398096e-01
-1.27216554e+00 -6.78186893e-01 7.01240242e-01 -2.16534995e-02
-1.70311511e-01 5.62539160e-01 2.20094323e-01 -2.64769375e-01
3.99145842e-01 -5.59261262e-01 -9.31231320e-01 -5.78639090e-01
7.54616864e-04 2.48159051e-01 -7.24350661e-02 -8.50723386e-02
-5.27190194e-02 1.77604735e-01 -1.40193152e+00 5.14308333e-01
6.13919556e-01 1.20488286e+00 2.88497210e-02 1.21802068e+00
4.15720910e-01 8.15946817e-01 -9.19450879e-01 -6.40832961e-01
4.39013511e-01 -8.09282541e-01 -4.19666499e-01 -3.21642235e-02
3.80007535e-01 -2.42606342e-01 -4.65371847e-01 8.75108242e-01
7.18541861e-01 2.56904453e-01 6.60890520e-01 -1.03019547e+00
-3.80461156e-01 7.30177462e-01 1.51387110e-01 -1.73185572e-01
4.58462834e-01 3.39787871e-01 1.24221635e+00 -1.00689912e+00
3.43730986e-01 1.29095435e+00 5.97776294e-01 5.65832794e-01
-1.53404915e+00 -6.88962221e-01 -7.57544413e-02 -1.25570953e-01
-1.19906616e+00 -9.72761095e-01 7.47092903e-01 -4.41692293e-01
1.21957707e+00 3.12818259e-01 3.98279458e-01 1.34919643e+00
2.21321881e-01 4.77221280e-01 8.28691423e-01 -5.84765911e-01
2.38594025e-01 -2.43209470e-02 -1.15035427e+00 3.21925759e-01
-7.60316789e-01 3.63468975e-01 -6.93538666e-01 4.45926599e-02
6.39845967e-01 -4.28722322e-01 -2.09141761e-01 1.64067432e-01
-1.29898512e+00 6.44994080e-01 2.90367365e-01 -7.91077986e-02
-3.75694543e-01 3.41875911e-01 5.94882131e-01 6.67555749e-01
4.11558568e-01 4.83411402e-01 -3.96452725e-01 -2.33814061e-01
-1.37993252e+00 4.40391362e-01 7.04050004e-01 7.29748011e-01
3.65500331e-01 7.10144162e-01 -4.66681927e-01 7.99163103e-01
2.34385997e-01 6.44612610e-01 6.95831954e-01 -9.47931767e-01
6.74278796e-01 -1.65381491e-01 -4.52503711e-02 -5.64731300e-01
9.74342749e-02 -3.84324610e-01 -8.39195549e-01 4.16673213e-01
2.81822264e-01 -3.89896393e-01 -1.07097864e+00 1.83456254e+00
8.63026902e-02 4.71151739e-01 1.85427949e-01 7.23418653e-01
2.20443174e-01 1.03199327e+00 -1.37536302e-01 -2.99254090e-01
8.96752536e-01 -1.14552963e+00 -7.88157403e-01 -1.58994272e-01
1.34582639e-01 -1.15192473e+00 1.17377102e+00 5.80566943e-01
-1.50555980e+00 -8.47520709e-01 -1.20247400e+00 -9.09611583e-02
-1.18698493e-01 2.26472542e-01 -2.15332806e-01 5.69438457e-01
-9.87069726e-01 9.28137600e-01 -8.18523109e-01 4.13391829e-01
-3.25427949e-02 3.57230783e-01 -1.06621824e-01 3.21463645e-01
-1.41018474e+00 6.70902491e-01 3.82226378e-01 -8.87656882e-02
-1.19674981e+00 -9.37667966e-01 -9.61925387e-01 1.82834625e-01
3.25180292e-02 -7.40710676e-01 1.71485245e+00 -1.21337104e+00
-2.39303875e+00 4.47858036e-01 -6.43590167e-02 -9.14577186e-01
8.31363976e-01 -1.19712979e-01 -6.37079179e-01 1.57745466e-01
-6.85883537e-02 7.34773874e-01 1.59360421e+00 -8.13117206e-01
-6.10496223e-01 3.42521161e-01 -3.73982936e-01 6.25551119e-02
-1.17913201e-01 -7.90107921e-02 -1.67394385e-01 -1.33153760e+00
-3.96801621e-01 -9.26483870e-01 -1.46748930e-01 -1.15347400e-01
-3.79559517e-01 6.35861233e-02 6.64321005e-01 -9.92276132e-01
1.26842272e+00 -2.15955162e+00 3.36272836e-01 1.67828143e-01
-2.66731650e-01 2.70828456e-01 -2.08626255e-01 6.93823338e-01
-2.09527999e-01 -1.67177066e-01 -4.63699281e-01 -8.95121992e-01
2.95519203e-01 -8.40295181e-02 -9.40446854e-01 3.31787109e-01
6.31552517e-01 6.90816104e-01 -9.97899652e-01 -9.61701795e-02
3.16085964e-01 7.10764408e-01 -7.34775424e-01 6.72489226e-01
-5.51661253e-01 6.49647892e-01 2.74028867e-01 3.16133767e-01
9.79915187e-02 3.62318248e-01 6.11598529e-02 7.06853718e-02
3.00411391e-03 9.60042477e-01 -8.73903632e-01 1.78454876e+00
-1.05445945e+00 8.38268816e-01 1.33168325e-01 -8.10102403e-01
1.08107674e+00 9.19953644e-01 1.35723934e-01 -4.53178287e-01
2.15358310e-03 4.24905270e-01 2.54977457e-02 -1.29691631e-01
3.77557188e-01 -3.59986424e-01 -1.57248415e-02 4.27992404e-01
2.87202775e-01 -7.14793026e-01 -6.86219335e-02 -2.57669330e-01
1.02034235e+00 2.16039836e-01 -2.01323703e-02 3.35623413e-01
5.43163836e-01 -6.33858502e-01 2.76585072e-01 3.81282598e-01
2.48383775e-01 9.59713697e-01 4.65416342e-01 1.57797605e-01
-1.69655657e+00 -1.16802216e+00 1.03812285e-01 1.10146189e+00
-6.77063107e-01 -4.39735591e-01 -9.50889528e-01 -3.63595694e-01
-2.34834924e-01 9.22744393e-01 -2.60773033e-01 -3.05220783e-01
-7.57433832e-01 5.04056402e-02 9.85462546e-01 3.34494710e-01
-7.51476884e-02 -1.21900284e+00 -3.57337803e-01 6.54617608e-01
-3.78765054e-02 -1.17360020e+00 -8.26476038e-01 2.63100863e-01
-6.09906614e-01 -3.18858027e-01 -8.92807245e-01 -6.60050094e-01
2.63162136e-01 -5.30271709e-01 1.04295206e+00 -4.07786191e-01
-8.37179646e-03 -1.91858858e-01 -2.24401683e-01 -4.43703532e-01
-1.05279684e+00 2.18810782e-01 2.32400477e-01 3.57816070e-01
-3.91299427e-01 -8.77012670e-01 -4.95445043e-01 1.28275231e-01
-9.74938691e-01 1.55280471e-01 5.27966142e-01 1.04583418e+00
6.51781142e-01 1.93855800e-02 9.56060886e-01 -4.54264611e-01
6.21175826e-01 -3.53414088e-01 -5.97503901e-01 -7.70644844e-02
-4.72972333e-01 8.04511830e-02 1.37920272e+00 -6.75724149e-01
-7.59994626e-01 1.95035771e-01 -5.62132418e-01 -8.49525869e-01
3.67419198e-02 2.17468232e-01 -2.21894711e-01 3.97428513e-01
7.51550138e-01 3.39653164e-01 -3.70910801e-02 -3.28612089e-01
5.16067564e-01 9.41493154e-01 9.83469307e-01 -2.86921561e-01
8.88631463e-01 3.66859958e-02 -9.84567031e-02 -7.09616184e-01
-6.99799776e-01 1.33149568e-02 -4.28106725e-01 -8.29052329e-02
6.02106988e-01 -1.16751504e+00 -3.81643921e-01 4.86049294e-01
-1.19093549e+00 -7.19695687e-01 -5.85725069e-01 5.19948006e-01
-1.06677306e+00 -4.63876985e-02 -5.67363799e-01 -9.08682942e-01
-6.04459286e-01 -1.01586556e+00 1.25953925e+00 -3.24846238e-01
-3.72707397e-01 -8.88898194e-01 7.32779652e-02 7.79682100e-02
4.16096717e-01 4.06275421e-01 7.10208893e-01 -5.53681791e-01
-3.86160403e-01 -3.46431345e-01 3.16875041e-01 8.99488926e-01
1.07976899e-01 1.77830979e-01 -1.29814780e+00 -4.02238876e-01
1.49505381e-02 -3.89247835e-01 6.92122698e-01 1.41382515e-01
1.00836766e+00 -7.01266110e-01 3.29096735e-01 6.42397821e-01
9.44723070e-01 -6.12056023e-03 5.79214811e-01 -7.99158216e-02
5.06765068e-01 5.58141410e-01 5.49674571e-01 4.75579202e-01
-6.49239197e-02 7.74109662e-01 4.14373666e-01 5.81623465e-02
-4.28519368e-01 -7.37006068e-01 8.17262709e-01 1.06194890e+00
4.47408825e-01 -3.72272074e-01 -5.72733343e-01 6.68493509e-01
-1.54650795e+00 -1.03787720e+00 3.98351043e-01 2.33115983e+00
1.38787591e+00 5.43566167e-01 2.25699410e-01 6.51357710e-01
6.04932427e-01 2.19416514e-01 -6.41704500e-01 -7.81399310e-01
2.36970812e-01 7.38758326e-01 3.34668845e-01 7.05975711e-01
-8.51946235e-01 8.10959101e-01 6.35874414e+00 9.54334795e-01
-1.62289643e+00 -8.26806873e-02 5.45641780e-01 -3.22087646e-01
-3.24096143e-01 -2.37156972e-01 -3.89984548e-01 5.61277509e-01
1.74147582e+00 -2.20199972e-01 9.20343041e-01 5.70104539e-01
5.46906352e-01 5.66495955e-01 -1.37599409e+00 7.12365448e-01
-2.83564359e-01 -1.23246539e+00 4.61704470e-02 -2.98837095e-01
9.06176984e-01 -1.10753000e-01 4.26392734e-01 3.47388387e-01
1.25096127e-01 -1.52206051e+00 1.31338608e+00 4.11046267e-01
1.46885061e+00 -8.12334895e-01 2.90284306e-01 4.56986070e-01
-1.13785684e+00 1.00859646e-02 1.01975791e-01 -1.36951953e-01
3.69140565e-01 4.49644625e-01 -1.37668288e+00 2.90303260e-01
2.45070189e-01 4.75287110e-01 1.28569350e-01 6.69703066e-01
-4.22553629e-01 9.60681498e-01 -3.02488267e-01 2.96457082e-01
2.97084183e-01 8.93736556e-02 5.14880478e-01 1.58008432e+00
6.15245104e-01 -3.96616906e-01 -4.16932777e-02 7.56025434e-01
-3.46503645e-01 -1.26901790e-01 -3.45519036e-01 -1.96431026e-01
6.96702659e-01 9.02828693e-01 -7.00669140e-02 -2.81201720e-01
-2.51088262e-01 1.07957160e+00 8.26655105e-02 2.69356698e-01
-8.01828146e-01 -8.16162229e-01 7.11548924e-01 1.96659595e-01
6.86060727e-01 -1.38891473e-01 -2.59734422e-01 -8.40897024e-01
1.59822777e-01 -1.16437769e+00 -1.64445996e-01 -5.82694590e-01
-1.05698681e+00 8.81452799e-01 -4.14668858e-01 -1.35209548e+00
-1.16524482e+00 -3.24853897e-01 -7.65412629e-01 1.34153056e+00
-1.46076310e+00 -1.15414941e+00 2.33970955e-01 4.75387573e-01
8.53276074e-01 -2.87420988e-01 9.90288794e-01 2.90423274e-01
-3.15603018e-01 1.01796174e+00 8.75828192e-02 -8.27635871e-04
8.50022554e-01 -1.34522927e+00 1.11531341e+00 9.25288022e-01
2.52251089e-01 2.11836949e-01 1.01567483e+00 -2.11211175e-01
-9.62818146e-01 -1.48588586e+00 9.88930047e-01 -3.33455890e-01
7.65590549e-01 -5.61623156e-01 -7.83014357e-01 4.83154029e-01
3.26919764e-01 7.71990716e-02 5.77754676e-01 -6.20634139e-01
-4.57075119e-01 -9.90645811e-02 -9.98438179e-01 5.06042361e-01
5.10466456e-01 -1.03764009e+00 -5.17322421e-01 1.11067735e-01
9.62209821e-01 -8.16587090e-01 -9.88835931e-01 2.26965904e-01
5.27626097e-01 -8.12120736e-01 9.05660391e-01 -4.70233589e-01
7.21902072e-01 -3.22383940e-01 -2.26383671e-01 -1.60305810e+00
5.34425639e-02 -1.39377284e+00 -4.13082629e-01 1.24418616e+00
6.33215606e-01 -1.88976467e-01 4.85297382e-01 7.30307698e-02
-5.23486912e-01 -7.74535775e-01 -1.10857475e+00 -7.18929827e-01
1.61016449e-01 -5.50819635e-01 5.80133915e-01 4.86964017e-01
-4.27570790e-02 4.46609646e-01 -6.18604362e-01 2.41289034e-01
3.58431697e-01 -1.85550779e-01 6.58974767e-01 -6.38321698e-01
-6.98489308e-01 -4.76039618e-01 -9.52621400e-02 -1.12589324e+00
3.36927742e-01 -7.58486092e-01 3.97452176e-01 -9.29935157e-01
-7.99725235e-01 -2.21101969e-01 -9.98095572e-02 2.20912158e-01
-3.17896642e-02 3.50170821e-01 2.66710132e-01 -4.34103198e-02
3.35263722e-02 7.17765927e-01 1.08417976e+00 -8.79135951e-02
-1.87455744e-01 3.92857730e-01 -1.83683291e-01 3.84321272e-01
7.28561044e-01 -4.20727938e-01 -4.92808670e-01 -2.81138182e-01
9.27750766e-02 4.06864434e-01 3.40721905e-01 -1.07175744e+00
6.46021962e-02 -2.71959212e-02 1.92186326e-01 -2.48375222e-01
7.95731723e-01 -7.16734409e-01 1.77183628e-01 3.63535076e-01
-8.66755724e-01 -8.01206306e-02 1.99760407e-01 5.37584066e-01
-5.69350719e-01 7.34641915e-03 1.00278831e+00 1.08540662e-01
1.34575114e-01 2.02637076e-01 -3.71115714e-01 9.44899172e-02
6.84192240e-01 1.71888515e-01 2.10988045e-01 -6.62468791e-01
-6.35757685e-01 -1.83572277e-01 1.86551854e-01 4.49102521e-01
4.89202350e-01 -1.54302418e+00 -1.06649685e+00 5.02814949e-01
-2.08616957e-01 6.62377626e-02 -8.02546088e-03 3.51220816e-01
-3.45834523e-01 4.98763770e-01 -5.48162311e-02 -3.43356311e-01
-9.14771736e-01 4.13731337e-01 3.65169108e-01 -3.23247671e-01
-4.01916772e-01 7.92938948e-01 -1.74910575e-01 -3.70739043e-01
5.24715722e-01 -4.94234324e-01 3.39079052e-01 -1.13898702e-01
5.09170115e-01 2.11475343e-01 3.87506694e-01 -5.30408561e-01
1.41079060e-03 1.38306707e-01 2.94282138e-01 -7.93800116e-01
1.05091584e+00 5.96872065e-04 3.00295949e-01 5.76784492e-01
1.35794318e+00 2.27333173e-01 -1.81808269e+00 -2.50882775e-01
-3.50809783e-01 -2.95184016e-01 2.03339711e-01 -7.79964566e-01
-8.04954469e-01 1.03316414e+00 3.63894135e-01 3.60558122e-01
1.18095922e+00 -4.69405711e-01 1.26093745e+00 4.98295836e-02
-1.93227157e-01 -1.08282578e+00 2.74641484e-01 6.04802847e-01
1.09346211e+00 -8.24928164e-01 -4.33232546e-01 -3.76509130e-02
-5.29414356e-01 1.15272701e+00 1.22748584e-01 -3.45554858e-01
4.26842451e-01 5.25599122e-01 1.35598078e-01 6.46181762e-01
-1.26881421e+00 2.25483149e-01 5.65635443e-01 6.47965908e-01
5.66012263e-01 2.74679214e-02 2.56505519e-01 4.20194417e-01
-8.61138105e-01 -2.89918095e-01 3.15980881e-01 3.90742838e-01
-1.80252284e-01 -1.23254061e+00 -4.06669468e-01 1.82070836e-01
-7.65404046e-01 -3.02220792e-01 -1.82393089e-01 1.11924067e-01
-1.24348126e-01 1.06513929e+00 2.29519710e-01 -5.15169442e-01
4.90698636e-01 2.14897931e-01 2.81593502e-01 -6.45393312e-01
-8.39559555e-01 2.76677012e-01 1.19356997e-01 -3.53285611e-01
8.97151828e-02 -6.84552848e-01 -1.25126934e+00 -2.35907555e-01
-1.96857423e-01 -4.65175249e-02 8.51524472e-01 7.44700909e-01
1.27223119e-01 9.47427094e-01 1.16133249e+00 -1.26185548e+00
-1.08094895e+00 -1.20118177e+00 -2.91610003e-01 2.74697065e-01
9.93381083e-01 1.41694602e-02 -3.71540964e-01 5.16606927e-01] | [15.449807167053223, 6.098127365112305] |
a1d0b92d-e774-4288-8828-e42d86ce7007 | a-two-stream-amr-enhanced-model-for-document-1 | 2205.00241 | null | https://arxiv.org/abs/2205.00241v1 | https://arxiv.org/pdf/2205.00241v1.pdf | A Two-Stream AMR-enhanced Model for Document-level Event Argument Extraction | Most previous studies aim at extracting events from a single sentence, while document-level event extraction still remains under-explored. In this paper, we focus on extracting event arguments from an entire document, which mainly faces two critical problems: a) the long-distance dependency between trigger and arguments over sentences; b) the distracting context towards an event in the document. To address these issues, we propose a Two-Stream Abstract meaning Representation enhanced extraction model (TSAR). TSAR encodes the document from different perspectives by a two-stream encoding module, to utilize local and global information and lower the impact of distracting context. Besides, TSAR introduces an AMR-guided interaction module to capture both intra-sentential and inter-sentential features, based on the locally and globally constructed AMR semantic graphs. An auxiliary boundary loss is introduced to enhance the boundary information for text spans explicitly. Extensive experiments illustrate that TSAR outperforms previous state-of-the-art by a large margin, with 2.54 F1 and 5.13 F1 performance gain on the public RAMS and WikiEvents datasets respectively, showing the superiority in the cross-sentence arguments extraction. We release our code in https://github.com/ PKUnlp-icler/TSAR. | ['Zhifang Sui', 'Baobao Chang', 'Shuang Zeng', 'Tianyu Liu', 'Peiyi Wang', 'Runxin Xu'] | 2022-04-30 | null | https://aclanthology.org/2022.naacl-main.370 | https://aclanthology.org/2022.naacl-main.370.pdf | naacl-2022-7 | ['document-level-event-extraction'] | ['natural-language-processing'] | [ 3.11457187e-01 1.79120582e-02 -2.51397461e-01 -4.06048864e-01
-1.01911294e+00 -6.39741182e-01 7.95401692e-01 7.25199103e-01
-5.19398570e-01 6.86171174e-01 7.38340914e-01 -2.36417606e-01
-2.31649265e-01 -7.92186081e-01 -6.23376608e-01 -4.49950248e-01
-1.74264945e-02 3.24917920e-02 5.05063295e-01 -2.16572881e-01
3.71179253e-01 1.06103443e-01 -1.43208253e+00 7.12261677e-01
7.22027779e-01 8.57338846e-01 1.16413765e-01 4.47612733e-01
-4.62391049e-01 1.05389726e+00 -8.91986728e-01 -5.73293984e-01
-3.66866678e-01 -3.48851055e-01 -9.88297164e-01 -2.33929932e-01
-2.13598058e-01 7.23042488e-02 -3.70550543e-01 8.92757833e-01
5.38653910e-01 8.91821161e-02 3.16822052e-01 -1.09284365e+00
-2.59900182e-01 1.21319139e+00 -8.19871068e-01 6.50071025e-01
4.71077651e-01 -1.59688935e-01 1.27254391e+00 -8.49797785e-01
6.37346089e-01 1.36815214e+00 2.78993070e-01 1.82407856e-01
-7.47669160e-01 -5.82124114e-01 6.15165710e-01 2.60455400e-01
-1.03453016e+00 -2.63925701e-01 1.03287923e+00 -3.16341594e-02
1.18672812e+00 4.73021030e-01 2.45960101e-01 1.31337857e+00
6.02951981e-02 1.12387812e+00 8.89390707e-01 -2.51568884e-01
9.54882205e-02 -2.38424420e-01 6.16150260e-01 3.52720499e-01
1.84034631e-01 -2.78934032e-01 -8.49352896e-01 -3.02801967e-01
2.98394382e-01 6.22008229e-03 -2.37794653e-01 5.13080359e-01
-1.20778286e+00 6.96495891e-01 1.37048885e-01 5.13183594e-01
-4.19194132e-01 -2.18962431e-02 9.53622043e-01 8.03758353e-02
4.91049886e-01 4.24971171e-02 -8.62536013e-01 -4.40042228e-01
-6.95726037e-01 2.19584972e-01 7.15843558e-01 9.64904368e-01
2.45203465e-01 -3.26069504e-01 -5.18964827e-01 7.58573472e-01
3.16171438e-01 1.82864666e-01 4.82249260e-01 -2.28267327e-01
1.07734859e+00 8.13831329e-01 -1.29020661e-01 -1.16984546e+00
-4.22488153e-01 -3.92732382e-01 -6.07527018e-01 -6.03826702e-01
1.42659664e-01 -1.99313909e-01 -4.46543455e-01 1.82836986e+00
6.19388521e-01 2.39569142e-01 2.65410274e-01 7.48129368e-01
1.22238946e+00 9.49910581e-01 2.97177255e-01 -3.46574068e-01
1.95856059e+00 -8.94816339e-01 -1.05536604e+00 -5.43103933e-01
6.81836307e-01 -9.24395978e-01 1.14088964e+00 2.26946443e-01
-1.11456096e+00 -1.49214029e-01 -9.85105693e-01 -3.74931723e-01
-4.51413453e-01 2.46679917e-01 6.36390567e-01 8.99797603e-02
-1.40776664e-01 3.58217865e-01 -6.76813543e-01 2.49357689e-02
4.14972067e-01 8.98137838e-02 -1.76780328e-01 1.09265402e-01
-1.89093411e+00 5.74543417e-01 7.04576015e-01 9.06312466e-02
-4.07775134e-01 -7.01232374e-01 -9.35530484e-01 2.18001917e-01
9.61221457e-01 -2.99436808e-01 1.18082952e+00 -4.87168103e-01
-1.18092120e+00 5.92041254e-01 -3.46247047e-01 -2.88623750e-01
2.35318452e-01 -5.63426137e-01 -6.94831848e-01 2.07623407e-01
2.48061046e-01 9.32678767e-03 5.73717892e-01 -8.58899117e-01
-5.96248507e-01 -5.30174494e-01 7.42248073e-02 4.02808309e-01
-2.40800917e-01 5.51516235e-01 -6.14308119e-01 -1.08514047e+00
-7.53592029e-02 -4.12442982e-01 1.80436164e-01 -6.01684928e-01
-6.51351392e-01 -6.28951669e-01 9.31275070e-01 -7.15774477e-01
1.81072485e+00 -2.29083538e+00 6.01986423e-02 -1.49095729e-01
7.32432008e-02 2.14765131e-01 1.12006674e-02 7.50984192e-01
-2.39380181e-01 1.32435143e-01 -2.66925156e-01 -2.76584506e-01
-5.03894538e-02 1.77609444e-01 -6.63143158e-01 1.91318750e-01
4.03242022e-01 8.69530916e-01 -1.03191590e+00 -8.36362660e-01
-7.86099434e-02 3.41488540e-01 -2.15774760e-01 9.83267203e-02
-9.26201046e-02 9.32387188e-02 -8.70323777e-01 5.15645325e-01
6.40532613e-01 -3.36996704e-01 2.53362119e-01 -2.77850658e-01
-1.11547001e-01 1.04640949e+00 -1.37038994e+00 1.85949659e+00
-4.36365932e-01 1.72928318e-01 -6.62228167e-02 -1.03508019e+00
7.84220576e-01 4.55319762e-01 1.84534878e-01 -6.61920309e-01
2.14442119e-01 9.96405855e-02 -2.38530457e-01 -4.32678163e-01
4.49346721e-01 7.91247096e-03 -6.00481570e-01 3.66514087e-01
-5.83099984e-02 4.20033753e-01 3.36161405e-01 6.49034619e-01
1.24037349e+00 2.21583843e-02 5.04208863e-01 -2.01574251e-01
6.40739202e-01 -1.88205913e-01 7.38850713e-01 5.14049232e-01
3.20650116e-02 4.40806925e-01 8.98466587e-01 -7.84207731e-02
-3.16613942e-01 -7.93824732e-01 -1.44526765e-01 1.08117235e+00
4.43928689e-01 -9.90817189e-01 -5.89795530e-01 -1.23579812e+00
-1.19621940e-01 9.57466960e-01 -5.97507775e-01 6.63397908e-02
-8.69166970e-01 -9.26844060e-01 7.46109664e-01 7.69349635e-01
6.01224184e-01 -1.21738303e+00 -7.20159471e-01 4.00457680e-01
-6.45738125e-01 -1.20172668e+00 -5.73477328e-01 2.80300587e-01
-6.11222982e-01 -1.07046151e+00 -1.46407411e-01 -4.91567075e-01
3.13113004e-01 8.82192776e-02 1.14621150e+00 5.29386140e-02
-2.24277630e-01 -7.78129324e-02 -6.35570109e-01 -5.02414823e-01
5.48759811e-02 1.35666922e-01 -4.83200938e-01 -9.71786454e-02
5.90735316e-01 -4.62019205e-01 -6.09951615e-01 2.17817709e-01
-1.00459492e+00 1.53776750e-01 5.26750326e-01 8.65684509e-01
6.34754837e-01 2.70355672e-01 8.32443297e-01 -1.01485455e+00
7.70380795e-01 -8.32901061e-01 -2.18673885e-01 2.19604701e-01
-4.33202147e-01 1.32235408e-01 6.88313484e-01 -4.53184694e-01
-1.55914283e+00 -4.05258089e-01 -1.21153243e-01 1.53810278e-01
-9.89177525e-02 6.76543474e-01 -5.85573494e-01 1.04550695e+00
2.40547895e-01 1.61299706e-01 -4.97775048e-01 -4.03807431e-01
3.91293436e-01 7.63323784e-01 5.00005007e-01 -9.28324699e-01
4.31185126e-01 4.80147630e-01 -2.25300938e-01 -6.34517789e-01
-1.20959747e+00 -6.51104271e-01 -1.69410795e-01 1.58760130e-01
7.38181949e-01 -8.38323176e-01 -6.34763718e-01 2.52694637e-01
-1.48810947e+00 4.27188389e-02 -2.82537073e-01 3.04653138e-01
-3.47118676e-02 4.12312061e-01 -7.91280031e-01 -7.57057667e-01
-6.82920754e-01 -8.94091129e-01 1.31904018e+00 2.24552855e-01
-3.72904360e-01 -7.83361316e-01 -1.12368673e-01 2.24794671e-01
-9.67421904e-02 2.82443494e-01 8.01084936e-01 -9.98904824e-01
-2.35923856e-01 -4.70556803e-02 -4.01012480e-01 1.59184933e-02
2.13110402e-01 -2.40885317e-01 -8.63624215e-01 1.95725877e-02
2.59912044e-01 -1.67057276e-01 9.34069037e-01 3.27552931e-04
1.29764473e+00 -4.80548322e-01 -5.51501811e-01 1.70050487e-01
1.10163474e+00 1.98924258e-01 6.36160791e-01 3.59237373e-01
6.07759774e-01 6.60429657e-01 1.02046871e+00 5.96643984e-01
3.82622361e-01 4.60518479e-01 2.86153018e-01 1.03922427e-01
-1.76153734e-01 -4.54738915e-01 3.99494618e-01 8.22852671e-01
1.81353226e-01 -6.11176252e-01 -7.21367240e-01 6.22657716e-01
-2.18428063e+00 -9.55774486e-01 -3.40143919e-01 1.88121438e+00
1.08213377e+00 5.44903338e-01 -1.64375544e-01 4.88637716e-01
8.19935203e-01 5.25107145e-01 -2.90208817e-01 -2.41346672e-01
-2.25843668e-01 2.30239689e-01 1.72673926e-01 2.84601003e-01
-1.18918073e+00 8.18597078e-01 4.50778866e+00 1.24414515e+00
-8.38950276e-01 1.55831218e-01 5.37954569e-01 -1.76112279e-01
-4.21718270e-01 1.57378569e-01 -1.11231554e+00 7.73431242e-01
8.66415441e-01 -6.86669871e-02 -7.40103647e-02 5.49537659e-01
1.26209527e-01 -1.01615675e-01 -8.77529144e-01 7.83906102e-01
-1.09326197e-02 -1.28811932e+00 6.47736192e-02 -2.22261190e-01
9.33798701e-02 -2.59427249e-01 -2.71387309e-01 3.32385451e-01
-1.34426385e-01 -6.61771059e-01 7.77131975e-01 2.03851029e-01
5.31121135e-01 -9.65541065e-01 7.98115075e-01 2.82953650e-01
-1.69057035e+00 6.53002709e-02 6.22388572e-02 -9.38334092e-02
4.12426412e-01 9.93463814e-01 -4.23405141e-01 1.10407889e+00
6.64367378e-01 8.14358652e-01 -4.34746474e-01 3.29939902e-01
-7.03178644e-01 8.47073674e-01 -4.23413575e-01 -2.74577975e-01
3.03574353e-01 8.77936408e-02 8.05803061e-01 1.61065531e+00
-7.06470087e-02 4.73998278e-01 5.88020235e-02 7.75261700e-01
-1.75222486e-01 2.57385880e-01 -2.52561539e-01 -1.26957759e-01
6.83270037e-01 1.38690972e+00 -9.19661939e-01 -4.80307192e-01
-4.56493437e-01 9.60391581e-01 3.60981166e-01 1.60567582e-01
-1.21602929e+00 -7.66612470e-01 4.06120449e-01 -5.41602932e-02
3.87573898e-01 9.13644359e-02 -2.82556355e-01 -1.30477929e+00
4.51852232e-01 -8.00203145e-01 9.69955385e-01 -5.36905229e-01
-1.26415563e+00 5.69787025e-01 2.41428375e-01 -7.84725726e-01
6.51049614e-02 -4.94738072e-01 -7.91180909e-01 7.17597485e-01
-1.52011096e+00 -1.13032401e+00 -2.64087971e-02 4.15341705e-01
7.44709432e-01 2.67954618e-01 5.75049996e-01 4.66709882e-01
-8.86634707e-01 5.01847029e-01 -5.14045000e-01 2.95117795e-01
6.67151213e-01 -1.20273471e+00 3.42318773e-01 9.96542871e-01
1.31005347e-01 8.34404290e-01 4.74185318e-01 -8.44571233e-01
-1.40460718e+00 -9.56647098e-01 1.25701416e+00 -4.64881867e-01
7.53580272e-01 -3.95818561e-01 -1.10465801e+00 6.15930319e-01
3.07619005e-01 -3.94709371e-02 6.61429405e-01 3.39696348e-01
-5.16723037e-01 6.67829514e-02 -1.00314569e+00 6.55661047e-01
1.21050012e+00 -4.84993756e-01 -1.08944142e+00 1.20285682e-01
1.04824698e+00 -5.10936677e-01 -7.35396385e-01 5.13104737e-01
7.75363967e-02 -6.00251198e-01 9.61026669e-01 -5.29977024e-01
6.34515166e-01 -4.31553632e-01 -8.60580504e-02 -7.44925916e-01
3.39587219e-02 -8.33379328e-01 -4.40127492e-01 1.79587400e+00
4.85538125e-01 -5.49662948e-01 2.03917950e-01 2.26183623e-01
-8.86811912e-02 -1.02096224e+00 -8.24817836e-01 -5.93613207e-01
-2.51100063e-01 -7.06600070e-01 8.13748658e-01 8.91388655e-01
2.99490869e-01 9.69539404e-01 -5.47856465e-02 2.62046814e-01
3.30285579e-01 3.99885029e-01 2.17422605e-01 -9.60737348e-01
-3.12661648e-01 -3.23276013e-01 1.60170525e-01 -1.00273132e+00
1.94266498e-01 -9.24351990e-01 -5.07444702e-02 -1.52674520e+00
4.17880714e-01 -2.23963425e-01 -4.39395040e-01 5.21247387e-01
-5.35394430e-01 -4.47624922e-01 -2.59695984e-02 -2.35204305e-02
-7.86596537e-01 6.08716726e-01 9.84129369e-01 1.43310338e-01
-1.51558295e-01 -2.84121335e-01 -8.38105083e-01 8.76368344e-01
8.54136765e-01 -6.96284473e-01 -5.42750418e-01 -4.39402103e-01
3.33329856e-01 8.67509544e-02 3.44663739e-01 -4.55434769e-01
1.87659845e-01 -9.71771851e-02 1.57888830e-01 -9.75515246e-01
1.50193468e-01 -4.64820564e-01 -2.91312933e-01 1.22449160e-01
-3.60316217e-01 2.69004971e-01 4.56423074e-01 7.61273861e-01
-4.62426484e-01 -3.59081775e-01 2.83703178e-01 2.25662924e-02
-5.53231299e-01 7.36844689e-02 -5.05828075e-02 6.15477860e-01
8.51319969e-01 3.02308708e-01 -6.50641739e-01 5.28191179e-02
-3.64629060e-01 2.46168837e-01 -2.05742344e-01 5.63785672e-01
6.03952527e-01 -1.15721202e+00 -7.05892324e-01 -4.48535830e-02
1.52246073e-01 2.65349299e-01 3.92868608e-01 7.55881786e-01
5.40816672e-02 2.16975659e-01 5.76296389e-01 -2.48080209e-01
-1.47853708e+00 6.08616352e-01 -2.24208072e-01 -7.51617312e-01
-8.71090293e-01 7.88598239e-01 2.14661032e-01 -9.23310146e-02
1.65532589e-01 -4.17053789e-01 -2.66290694e-01 2.86483735e-01
6.91952348e-01 3.11365068e-01 1.19693771e-01 -4.03587520e-01
-6.84734643e-01 2.13525742e-01 -3.22071820e-01 -1.98347196e-01
1.16709399e+00 -9.98338982e-02 -2.29262874e-01 3.04935396e-01
1.11926866e+00 3.41355115e-01 -8.29773128e-01 -2.92263508e-01
2.53706664e-01 -2.10849077e-01 7.60429502e-02 -9.41791415e-01
-6.90568388e-01 7.05311120e-01 -1.57581881e-01 3.57514083e-01
1.21194828e+00 3.79487187e-01 1.06450772e+00 9.81927961e-02
-8.52257684e-02 -1.02552259e+00 1.79153129e-01 6.40219212e-01
1.05529654e+00 -8.86499345e-01 5.58848605e-02 -8.63784254e-01
-7.06305861e-01 9.63600457e-01 4.13978130e-01 7.18803406e-02
6.06800795e-01 5.69994628e-01 -3.80615115e-01 -4.65671241e-01
-8.51673126e-01 -6.02240115e-02 2.73469239e-01 -1.06990613e-01
6.45842791e-01 -3.78521457e-02 -7.59979010e-01 1.40193307e+00
-1.69367658e-03 -2.61179507e-01 1.27667829e-01 1.21141720e+00
-1.23663761e-01 -1.21619272e+00 -1.13017276e-01 2.08177850e-01
-8.77896845e-01 -4.23695982e-01 -3.99176031e-01 8.37677836e-01
1.19171508e-01 1.04305542e+00 -5.30857630e-02 -3.45472880e-02
5.34817398e-01 1.66110620e-01 2.70560563e-01 -6.05706573e-01
-8.32775056e-01 2.50900030e-01 5.32757461e-01 -5.43080330e-01
-3.76883000e-01 -8.27733696e-01 -1.85335696e+00 7.69871399e-02
-4.11887944e-01 3.16167414e-01 4.33576077e-01 9.69340324e-01
6.87702596e-01 9.98752236e-01 4.72377390e-01 -1.46827325e-01
-5.28959692e-01 -9.28471744e-01 -3.55782330e-01 5.85972846e-01
4.26298194e-02 -6.05377555e-01 -4.52568382e-01 -1.38048708e-01] | [9.090896606445312, 9.168861389160156] |
859aab68-b994-420c-946a-8a6ce1f70593 | promptpose-language-prompt-helps-animal-pose | 2206.11752 | null | https://arxiv.org/abs/2206.11752v3 | https://arxiv.org/pdf/2206.11752v3.pdf | CLAMP: Prompt-based Contrastive Learning for Connecting Language and Animal Pose | Animal pose estimation is challenging for existing image-based methods because of limited training data and large intra- and inter-species variances. Motivated by the progress of visual-language research, we propose that pre-trained language models (e.g., CLIP) can facilitate animal pose estimation by providing rich prior knowledge for describing animal keypoints in text. However, we found that building effective connections between pre-trained language models and visual animal keypoints is non-trivial since the gap between text-based descriptions and keypoint-based visual features about animal pose can be significant. To address this issue, we introduce a novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively. The CLAMP attempts to bridge the gap by adapting the text prompts to the animal keypoints during network training. The adaptation is decomposed into spatial-aware and feature-aware processes, and two novel contrastive losses are devised correspondingly. In practice, the CLAMP enables the first cross-modal animal pose estimation paradigm. Experimental results show that our method achieves state-of-the-art performance under the supervised, few-shot, and zero-shot settings, outperforming image-based methods by a large margin. | ['DaCheng Tao', 'Jing Zhang', 'Yufei Xu', 'Zhe Chen', 'Wen Wang', 'Xu Zhang'] | 2022-06-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_CLAMP_Prompt-Based_Contrastive_Learning_for_Connecting_Language_and_Animal_Pose_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_CLAMP_Prompt-Based_Contrastive_Learning_for_Connecting_Language_and_Animal_Pose_CVPR_2023_paper.pdf | cvpr-2023-1 | ['animal-pose-estimation'] | ['computer-vision'] | [ 1.03666365e-01 -3.16955775e-01 -2.08266139e-01 -4.63571578e-01
-5.35772085e-01 -4.67772752e-01 6.10054314e-01 2.35494286e-01
-8.17220390e-01 4.45857793e-01 7.27787763e-02 3.59637141e-01
-2.36497447e-02 -4.90275681e-01 -1.15946949e+00 -4.57836509e-01
-1.45999059e-01 2.08682910e-01 5.26951015e-01 -2.33135670e-01
1.49282441e-01 5.94360232e-01 -1.67563915e+00 4.58413223e-03
5.16993463e-01 1.07588398e+00 6.84564352e-01 5.34154952e-01
-1.32016554e-01 8.32076311e-01 -2.77338535e-01 -2.79594988e-01
1.98870689e-01 -5.12123942e-01 -5.08978128e-01 4.58314717e-02
6.64064348e-01 -5.63982725e-01 -4.33028132e-01 8.56745958e-01
3.57996255e-01 3.16124558e-01 7.03536034e-01 -1.48418188e+00
-6.28301561e-01 3.52185339e-01 -8.54702830e-01 3.41791391e-01
9.62790549e-02 3.71463805e-01 1.00443268e+00 -1.01312280e+00
6.77149653e-01 1.36711812e+00 8.24663460e-01 5.21572053e-01
-1.48983169e+00 -7.14267015e-01 3.87665361e-01 3.20129931e-01
-1.42396259e+00 -4.56889093e-01 9.40246344e-01 -6.31095231e-01
7.72160888e-01 -8.76627788e-02 8.85604739e-01 1.13886094e+00
4.21682224e-02 8.07507813e-01 8.11026514e-01 -3.17912310e-01
7.66712725e-02 1.89226940e-02 -2.36841843e-01 8.55967402e-01
9.79699343e-02 2.80868590e-01 -9.35582578e-01 -1.74196362e-01
9.57052648e-01 1.84499383e-01 -2.54284441e-01 -1.11691248e+00
-1.33335245e+00 8.65621686e-01 8.37465882e-01 -5.63277490e-02
-3.33617985e-01 4.71443623e-01 4.10827309e-01 -2.63334420e-02
4.38348562e-01 5.91072738e-01 -3.11583608e-01 1.84437424e-01
-8.61168981e-01 2.53167987e-01 4.86964226e-01 1.08824968e+00
8.96816313e-01 -5.55186272e-02 -2.67592520e-01 9.17586625e-01
5.06533206e-01 5.68589628e-01 2.12400287e-01 -8.15530419e-01
3.01066279e-01 2.40723655e-01 9.04669389e-02 -1.01301527e+00
-3.12683135e-01 -3.79242748e-01 -5.90799868e-01 1.61561146e-02
4.06322449e-01 1.22348107e-01 -9.80386078e-01 2.00811315e+00
4.48216259e-01 1.49031103e-01 -3.46009910e-01 1.05885589e+00
9.16976869e-01 6.55482948e-01 4.40932065e-01 -2.12623384e-02
1.39347577e+00 -1.19299912e+00 -6.12241983e-01 -8.16261888e-01
4.06053752e-01 -4.71172005e-01 1.10063446e+00 -2.21572265e-01
-9.41853642e-01 -4.88683611e-01 -1.02519047e+00 -2.29237050e-01
-4.45456266e-01 1.63836822e-01 5.87150931e-01 -2.94481069e-02
-7.68875360e-01 4.38869506e-01 -8.60583305e-01 -6.17423296e-01
4.92233336e-01 1.40345842e-01 -6.20434642e-01 1.10474594e-01
-9.66742754e-01 9.45415020e-01 3.04513454e-01 2.08275348e-01
-1.10975087e+00 -7.84957230e-01 -1.27352834e+00 -9.67446715e-02
6.64384782e-01 -7.29303658e-01 1.30865848e+00 -9.90219235e-01
-1.33415055e+00 1.15815616e+00 2.69984603e-02 -5.44822991e-01
5.86074710e-01 -3.85006964e-01 1.14379257e-01 5.78294575e-01
1.95570603e-01 1.27759051e+00 1.07760918e+00 -1.55571616e+00
-6.81596458e-01 -4.83131170e-01 6.30821362e-02 2.38057509e-01
-2.32716814e-01 7.84581807e-03 -6.99819565e-01 -8.37569594e-01
4.66957502e-02 -9.68337119e-01 -1.27957329e-01 1.04737270e+00
-2.70562340e-03 -3.34355347e-02 5.90565205e-01 -5.96206546e-01
8.68556917e-01 -2.15321207e+00 2.76736170e-01 -1.91094592e-01
2.29594469e-01 1.19115718e-01 -4.37946647e-01 5.10034800e-01
5.04205786e-02 -3.64111662e-01 -1.82969943e-01 -4.14632052e-01
-1.17925182e-01 2.96303481e-01 -3.65673780e-01 7.60585368e-01
4.00545686e-01 1.00874269e+00 -9.25391197e-01 -7.85987437e-01
3.90174359e-01 5.22476077e-01 -7.76286662e-01 4.30167586e-01
-2.61502892e-01 3.84723425e-01 -3.21493626e-01 7.09327996e-01
5.20076573e-01 -2.59191751e-01 -7.43694305e-02 -4.74691838e-01
-6.01596683e-02 -1.37893379e-01 -7.53530562e-01 1.97714925e+00
-4.18396473e-01 5.79728305e-01 8.10330287e-02 -1.09842968e+00
5.22811294e-01 -1.70430735e-01 3.38780910e-01 -5.43361902e-01
1.28945649e-01 -1.01518750e-01 -1.42604023e-01 -5.44417083e-01
4.09164935e-01 -1.08583882e-01 -3.31841521e-02 -1.40486136e-01
4.41076070e-01 -3.16522688e-01 1.83091849e-01 1.63756415e-01
6.95552588e-01 6.02289975e-01 5.48643529e-01 -1.95603043e-01
2.12206274e-01 3.20630544e-03 4.46147829e-01 6.67747617e-01
-4.23851818e-01 5.03457785e-01 3.21052819e-01 -3.12459052e-01
-1.04205418e+00 -1.01108038e+00 2.59990525e-03 1.52073860e+00
5.29254138e-01 -4.11359400e-01 -6.09649241e-01 -6.48388147e-01
2.36060768e-01 4.71217573e-01 -9.16757703e-01 -3.39570075e-01
-4.96425301e-01 -2.07437143e-01 3.25665414e-01 8.10804605e-01
5.01911104e-01 -1.00907016e+00 -9.24457252e-01 9.53544676e-02
-2.31042698e-01 -1.36157203e+00 -5.68521917e-01 4.62959915e-01
-5.85842729e-01 -7.62098968e-01 -9.07195091e-01 -9.91333604e-01
7.84430504e-01 5.61773896e-01 8.64066422e-01 7.54560456e-02
-5.49739540e-01 4.29794639e-01 -4.00832206e-01 -2.48204112e-01
5.85324578e-02 -1.93899855e-01 1.05505129e-02 -3.00399780e-01
1.99006408e-01 -4.82595295e-01 -7.39243686e-01 2.69507349e-01
-7.19284594e-01 2.09356025e-01 7.48333454e-01 9.77183282e-01
7.21059680e-01 -5.89481592e-01 1.47347689e-01 -2.93745309e-01
1.65070593e-02 -3.31512600e-01 -5.76532781e-01 2.70607740e-01
-1.75494626e-02 1.82489559e-01 5.46385348e-01 -8.42800617e-01
-6.76231921e-01 3.84855002e-01 -4.20678370e-02 -5.45930564e-01
-1.84393227e-01 3.71388853e-01 4.52801362e-02 -4.59705353e-01
5.70170522e-01 2.54903078e-01 5.02632186e-02 -4.53158379e-01
3.60176027e-01 2.51603216e-01 6.32632256e-01 -5.24954140e-01
9.70371664e-01 6.09984636e-01 -2.90437676e-02 -1.00496101e+00
-1.00786364e+00 -6.87367022e-01 -6.81379616e-01 -4.04280305e-01
1.06745303e+00 -1.15365720e+00 -6.16095185e-01 4.62948561e-01
-1.12457192e+00 -4.57285494e-01 -2.44285554e-01 7.04662144e-01
-8.48095894e-01 3.81191045e-01 -5.41100740e-01 -5.95012188e-01
-8.91611055e-02 -1.06123388e+00 1.39160109e+00 7.14743603e-03
-1.04185939e-01 -6.25060141e-01 -1.86164267e-02 1.99586824e-01
2.44634613e-01 1.12214625e-01 7.65118241e-01 -5.19417167e-01
-6.10223353e-01 -1.55653223e-01 -4.92496282e-01 9.79420394e-02
-2.60409772e-01 -2.36483455e-01 -8.19460988e-01 -4.20237660e-01
-2.40996912e-01 -7.77758121e-01 8.75170350e-01 4.19800043e-01
1.19939005e+00 -1.91005729e-02 -4.85953659e-01 8.08177769e-01
1.37544918e+00 -1.58609524e-01 2.20069945e-01 3.34173769e-01
8.22180748e-01 7.13998497e-01 6.75583899e-01 4.17934895e-01
4.71590728e-01 1.01007497e+00 5.67999125e-01 -1.10310748e-01
-2.10429251e-01 -8.03544819e-01 1.06984377e-01 4.36645687e-01
9.73562673e-02 -1.07742965e-01 -8.90221834e-01 6.57505333e-01
-1.89728582e+00 -9.68214512e-01 3.67079288e-01 2.02552390e+00
8.30106735e-01 -1.14570208e-01 2.20201135e-01 -3.68640929e-01
5.32486320e-01 2.84116417e-01 -7.33895540e-01 2.56130695e-01
6.97902143e-02 -3.21040630e-01 6.84656918e-01 2.21499324e-01
-1.30832601e+00 1.12178826e+00 6.46526432e+00 8.56350660e-01
-1.05909801e+00 -8.03998485e-02 1.74901351e-01 -4.60440069e-02
4.50198576e-02 -9.47192609e-02 -7.47672677e-01 3.14890802e-01
2.72863865e-01 -5.24307659e-04 3.78392786e-01 1.05668044e+00
-3.10003553e-02 -2.26591542e-01 -1.33424592e+00 1.19108796e+00
4.70332265e-01 -1.06875467e+00 1.16152115e-01 -9.86302793e-02
3.51094812e-01 1.01492330e-01 2.37218458e-02 2.65133679e-01
7.84978643e-02 -7.38992453e-01 1.21051085e+00 3.95124972e-01
6.48106754e-01 -5.42275488e-01 3.62065583e-01 5.55141687e-01
-1.45366848e+00 -1.25036702e-01 -3.93749088e-01 4.09002341e-02
3.01707555e-02 -1.04232363e-01 -3.31713974e-01 1.96898833e-01
9.71454084e-01 7.36310780e-01 -6.98465526e-01 1.20869148e+00
-1.24049239e-01 1.70747057e-01 -4.09387052e-01 -1.31182879e-01
3.75032753e-01 3.35684046e-02 5.66225648e-01 1.07332873e+00
1.41484976e-01 7.09833503e-02 3.91542405e-01 9.06372011e-01
-7.05679432e-02 1.60115346e-01 -5.65999210e-01 -1.08109668e-01
3.61172706e-01 1.22405612e+00 -8.83830249e-01 -2.15565488e-01
-4.34631497e-01 1.03026950e+00 6.68322086e-01 3.59144092e-01
-1.04536450e+00 -3.17568719e-01 5.86562395e-01 2.73640156e-01
6.90697908e-01 -3.20787251e-01 2.07646966e-01 -1.05535877e+00
-5.25752679e-02 -5.69134533e-01 5.27495593e-02 -9.72960532e-01
-1.29221106e+00 3.42941344e-01 4.72118467e-01 -1.38232911e+00
-2.02033117e-01 -6.29382908e-01 -1.27727270e-01 5.17912030e-01
-1.63236892e+00 -1.68729830e+00 -3.53197604e-01 3.82114947e-01
6.50771499e-01 8.85253847e-02 5.89416504e-01 1.35309324e-01
-9.71702412e-02 6.32404625e-01 -1.00887157e-01 1.02804229e-01
8.16169977e-01 -8.68258119e-01 1.99117675e-01 7.85373330e-01
2.01915011e-01 5.10656655e-01 9.33497012e-01 -5.10823309e-01
-1.44645107e+00 -1.00519311e+00 6.05943441e-01 -2.88083524e-01
8.10556114e-01 -6.63916171e-01 -8.28092337e-01 5.35537958e-01
-1.13015041e-01 4.51744825e-01 4.89883423e-01 -1.18876390e-01
-5.60610414e-01 -1.17017264e-02 -9.10033643e-01 6.80198848e-01
1.21438360e+00 -7.28173852e-01 -6.98765814e-01 2.99857467e-01
7.08118618e-01 -4.01037455e-01 -4.80170637e-01 4.84260619e-01
6.95082307e-01 -5.24117172e-01 1.23053801e+00 -5.10466456e-01
6.77000165e-01 -3.78732145e-01 -2.15455130e-01 -1.25012112e+00
-4.45406705e-01 -6.21736348e-02 -5.02560250e-02 9.83850002e-01
7.78087825e-02 -1.16921656e-01 4.30756062e-01 2.27703810e-01
-1.01033282e-02 -5.84157050e-01 -6.63036644e-01 -8.64431858e-01
1.15652301e-03 -1.75302878e-01 1.62167430e-01 7.06753850e-01
-1.86758727e-01 4.56444472e-01 -7.67065287e-01 1.56772792e-01
7.84988761e-01 1.82807848e-01 9.12056923e-01 -1.10620022e+00
-3.01782638e-01 -3.19132060e-01 -6.40522778e-01 -1.46920812e+00
3.31249505e-01 -6.75124764e-01 5.80888212e-01 -1.38431060e+00
5.21634936e-01 3.72138098e-02 5.30259917e-03 5.99292874e-01
-1.85025841e-01 4.81043547e-01 3.23505402e-01 2.06203908e-01
-8.48582268e-01 7.73221016e-01 1.25164795e+00 -1.32436365e-01
3.66746224e-02 -3.29673409e-01 -3.89413178e-01 9.98398840e-01
3.81602407e-01 -4.64341819e-01 -3.40617597e-01 -4.84172851e-01
1.27005085e-01 1.38555951e-02 9.21615601e-01 -9.93680239e-01
3.52873862e-01 -2.11300179e-01 3.77054602e-01 -7.05122352e-01
5.26095092e-01 -9.13754463e-01 -2.14703575e-01 4.01594073e-01
-6.47383928e-01 -3.39109093e-01 2.59070188e-01 1.00393450e+00
-1.80897012e-01 2.34622993e-02 1.02791250e+00 -1.41723171e-01
-1.08760369e+00 5.28140306e-01 -1.68308794e-01 2.51770139e-01
1.01549697e+00 -2.13831887e-01 -1.45534039e-01 -2.78241783e-01
-3.86706233e-01 3.91053617e-01 5.58518589e-01 5.84489763e-01
5.93683243e-01 -1.33619249e+00 -4.06554312e-01 2.36152276e-01
7.99466193e-01 -7.31501356e-02 4.50207233e-01 8.98199856e-01
-6.21808708e-01 1.79422960e-01 -4.03976381e-01 -7.81738281e-01
-1.38285661e+00 7.47397840e-01 3.15159172e-01 1.14803813e-01
-7.64152050e-01 1.18816423e+00 8.10894847e-01 -3.43224823e-01
6.43273592e-01 -3.76052678e-01 -1.06527224e-01 7.97514245e-02
5.47544301e-01 -7.86932111e-02 -3.83930326e-01 -9.78946567e-01
-4.37167257e-01 9.65152800e-01 -1.16858490e-01 -1.07479785e-02
1.44257307e+00 -2.56326556e-01 1.75907940e-01 6.37463987e-01
1.23943651e+00 -3.56063783e-01 -1.62361777e+00 -5.92007577e-01
-3.19084466e-01 -5.55272281e-01 1.08499654e-01 -5.75869799e-01
-7.87522554e-01 1.10380769e+00 6.21820211e-01 -2.48520195e-01
8.72504592e-01 2.80370951e-01 6.64519668e-01 5.57984114e-01
4.18814510e-01 -1.21330881e+00 4.43368375e-01 3.52577895e-01
1.09780073e+00 -1.59588969e+00 5.87092228e-02 -4.38551992e-01
-7.34217584e-01 7.28918493e-01 9.72495377e-01 -1.06475465e-01
5.68986475e-01 1.68320388e-01 -1.78749308e-01 -1.72965989e-01
-6.61965847e-01 -5.06144404e-01 5.18330693e-01 6.25660479e-01
1.95504755e-01 -2.08256885e-01 2.96114106e-02 4.63425964e-01
3.28996666e-02 -1.13858804e-01 -1.22587167e-01 1.03390372e+00
-6.43407583e-01 -4.28355634e-01 -2.32677430e-01 1.95610791e-01
-2.20983610e-01 -1.34969398e-01 -2.80139208e-01 8.61968994e-01
1.70545310e-01 3.87709588e-01 1.30850986e-01 -1.62839994e-01
3.64641219e-01 -2.15357244e-01 6.47854507e-01 -6.17133141e-01
-2.34847501e-01 1.28498495e-01 -1.67204902e-01 -6.93363965e-01
-7.19210744e-01 -3.15258265e-01 -9.77918029e-01 -1.95406619e-02
-5.17547011e-01 -1.79097503e-01 5.40038347e-01 9.55443203e-01
3.98079343e-02 4.44048941e-01 3.14601868e-01 -1.32913339e+00
-5.39584339e-01 -8.25573385e-01 -5.11167824e-01 4.72368866e-01
5.45611858e-01 -9.83891547e-01 -5.72248697e-01 2.58445948e-01] | [7.6684980392456055, -0.9397859573364258] |
adffc92c-e6f0-4750-8b09-ba458b7d83bc | high-dimensional-and-permutation-invariant | 2306.03933 | null | https://arxiv.org/abs/2306.03933v1 | https://arxiv.org/pdf/2306.03933v1.pdf | High-dimensional and Permutation Invariant Anomaly Detection | Methods for anomaly detection of new physics processes are often limited to low-dimensional spaces due to the difficulty of learning high-dimensional probability densities. Particularly at the constituent level, incorporating desirable properties such as permutation invariance and variable-length inputs becomes difficult within popular density estimation methods. In this work, we introduce a permutation-invariant density estimator for particle physics data based on diffusion models, specifically designed to handle variable-length inputs. We demonstrate the efficacy of our methodology by utilizing the learned density as a permutation-invariant anomaly detection score, effectively identifying jets with low likelihood under the background-only hypothesis. To validate our density estimation method, we investigate the ratio of learned densities and compare to those obtained by a supervised classification algorithm. | ['Benjamin Nachman', 'Vinicius Mikuni'] | 2023-06-06 | null | null | null | null | ['density-estimation'] | ['methodology'] | [ 7.63075352e-02 -2.90185124e-01 -2.12679263e-02 -2.25048900e-01
-6.08666658e-01 -6.21021926e-01 9.90886390e-01 2.49434114e-01
-2.96606660e-01 9.26006138e-01 -1.24026306e-01 -4.63651925e-01
-6.51174009e-01 -8.88658941e-01 -5.48449993e-01 -8.95040572e-01
-3.34058374e-01 8.36751878e-01 4.71708953e-01 4.73705888e-01
3.01042110e-01 1.07579827e+00 -1.30831909e+00 -3.37181568e-01
8.60209584e-01 1.02636468e+00 -1.23791479e-01 9.12379086e-01
-4.12892163e-01 3.53414357e-01 -4.77018327e-01 -4.83579427e-01
3.27774823e-01 -6.01762235e-01 -4.53629494e-01 -2.11269140e-01
5.24816394e-01 -4.68708836e-02 -4.37171996e-01 1.47924936e+00
9.21812356e-02 4.70786572e-01 1.38855922e+00 -1.15753078e+00
-5.29294789e-01 2.69228294e-02 -3.51996928e-01 8.15086782e-01
-4.48605753e-02 -1.43296178e-02 1.08524227e+00 -8.31261992e-01
6.36143208e-01 1.19573307e+00 4.09657657e-01 2.14891374e-01
-1.56495953e+00 -4.01981950e-01 -2.05363452e-01 1.68577030e-01
-1.07632995e+00 -1.72691137e-01 8.63499165e-01 -7.02246845e-01
7.34824538e-01 1.53643951e-01 4.07423139e-01 1.30870295e+00
6.36718631e-01 3.41644645e-01 1.19962788e+00 -4.41932619e-01
5.83425581e-01 1.47475302e-01 2.76822895e-01 6.45844698e-01
7.30640769e-01 3.27061146e-01 -4.05926764e-01 -4.96720880e-01
8.75603855e-01 -2.97249675e-01 7.55812749e-02 -9.73222911e-01
-9.00275290e-01 9.69017506e-01 -1.23094469e-01 2.83850729e-01
-2.87298650e-01 -1.23917863e-01 3.22336227e-01 4.89644669e-02
7.85233855e-01 5.47663867e-01 -4.45330918e-01 -3.87928486e-01
-8.88830662e-01 3.98693860e-01 1.09184039e+00 7.26280987e-01
6.20733142e-01 2.05412418e-01 -4.38376933e-01 5.88919461e-01
2.33993635e-01 7.78163195e-01 1.60784945e-01 -7.52728999e-01
9.11398605e-02 2.29712605e-01 1.82272792e-01 -7.64756203e-01
-3.67663205e-01 -5.07085204e-01 -9.30588484e-01 3.55888218e-01
8.20290387e-01 -1.05282977e-01 -8.59090209e-01 1.54290414e+00
3.54433060e-01 2.15529189e-01 2.16760859e-02 5.75173080e-01
1.48970503e-02 5.06595075e-01 1.75998166e-01 -3.07880580e-01
1.01554108e+00 -4.29971725e-01 -4.47549969e-01 9.97104049e-02
2.54025102e-01 -5.51004946e-01 8.32622945e-01 3.17686588e-01
-7.91160166e-01 -2.61387646e-01 -6.72579825e-01 3.34196895e-01
-2.16183200e-01 2.39133332e-02 6.90789282e-01 8.75648975e-01
-7.27716565e-01 9.11970258e-01 -1.08713984e+00 -4.80558187e-01
2.77480870e-01 -4.25425619e-02 -2.21384302e-01 2.92059362e-01
-7.65299857e-01 7.39102304e-01 4.35783774e-01 -3.87934268e-01
-7.74674654e-01 -9.26743090e-01 -7.06066608e-01 3.59959900e-01
1.63446918e-01 -5.82774043e-01 1.14456952e+00 -1.30588248e-01
-1.59068060e+00 3.25158149e-01 -2.57243723e-01 -7.27249026e-01
5.17386198e-01 -6.95229694e-02 -6.32401288e-01 3.34988087e-01
4.86099236e-02 2.20102770e-03 1.16364062e+00 -8.67777884e-01
-4.07385767e-01 -1.26782417e-01 -2.93915510e-01 -3.48461360e-01
-1.70702025e-01 -1.16713107e-01 -2.41154730e-02 -6.27362907e-01
2.80237973e-01 -7.66226649e-01 -9.42034200e-02 -7.46839568e-02
-3.42487961e-01 -8.63175765e-02 8.20311546e-01 -5.95684111e-01
8.02697599e-01 -2.23728442e+00 2.04923123e-01 7.14779675e-01
2.10295334e-01 9.67460126e-02 9.93108079e-02 1.22922316e-01
6.57657757e-02 -3.56016234e-02 -5.05559444e-01 -2.04841182e-01
2.62874573e-01 1.60506248e-01 -3.74012113e-01 5.61710775e-01
7.03340948e-01 5.45586050e-01 -9.45763230e-01 -1.37630627e-01
3.19455087e-01 1.72031060e-01 -5.67670822e-01 1.83141068e-01
-2.21095964e-01 7.65464544e-01 -4.35527444e-01 5.24046063e-01
9.06991482e-01 -6.33640513e-02 -1.01603135e-01 -3.07636652e-02
-1.27481371e-01 1.56817272e-01 -1.06688821e+00 1.25054681e+00
-1.51885480e-01 4.68733996e-01 4.43001464e-02 -1.05550897e+00
9.77780461e-01 -7.23518655e-02 4.18616980e-01 -4.93468404e-01
-3.22408453e-02 1.64385095e-01 2.21363097e-01 -1.89822152e-01
4.68089074e-01 -4.92702872e-01 -4.30168808e-02 3.28403443e-01
5.70406079e-01 -1.31871149e-01 4.06582117e-01 3.40599716e-01
1.59598446e+00 -3.15430090e-02 2.63313085e-01 -4.20689791e-01
5.17550826e-01 -3.94353151e-01 2.90068120e-01 1.34930170e+00
-2.90084749e-01 2.76889920e-01 9.74143803e-01 7.56136924e-02
-1.42372477e+00 -1.81299591e+00 -7.07599223e-01 6.55971527e-01
-2.83339113e-01 -4.16209042e-01 -2.64897197e-01 -8.77019465e-01
3.22901994e-01 8.88823628e-01 -3.05414349e-01 -3.95928293e-01
-2.55616456e-01 -1.02203107e+00 3.89584184e-01 3.39888632e-01
2.23972529e-01 -7.17933416e-01 1.35620058e-01 -1.81965269e-02
6.36425138e-01 -1.20723855e+00 1.55617744e-02 2.70261079e-01
-7.41577387e-01 -1.00609410e+00 -4.46087211e-01 -1.91225689e-02
5.55567324e-01 -2.44672477e-01 8.97852600e-01 -4.92881626e-01
-5.32279968e-01 6.98536634e-01 -2.90463418e-01 -2.61298001e-01
-7.06776559e-01 -2.35736221e-02 5.19232273e-01 -4.09843363e-02
4.97566283e-01 -7.03293204e-01 -2.07468942e-01 1.10994697e-01
-7.10836351e-01 -6.81047440e-01 8.88599277e-01 7.02856898e-01
3.54279548e-01 3.56762260e-01 4.88911957e-01 -7.78093457e-01
6.86891019e-01 -6.20521724e-01 -9.43573713e-01 -1.01247273e-01
-3.75925690e-01 3.54981631e-01 5.69504797e-01 -4.36771244e-01
-1.09556293e+00 -2.26140037e-01 -5.24323843e-02 -5.89537442e-01
-5.15126526e-01 1.34911343e-01 -1.92220472e-02 -3.12836796e-01
4.04997677e-01 1.99826360e-01 -9.52478349e-02 -5.82583308e-01
3.94579083e-01 1.42244846e-01 5.73759437e-01 -9.02107716e-01
1.26108909e+00 3.31490874e-01 5.91590106e-01 -1.04982686e+00
-9.54191327e-01 -5.24110675e-01 -8.62640023e-01 -1.17007211e-01
9.16227162e-01 -4.46309805e-01 -4.32615727e-01 3.28461796e-01
-8.96624148e-01 1.49138048e-01 -7.65647233e-01 1.11285198e+00
-5.59395134e-01 6.80611014e-01 -5.06551802e-01 -1.20230389e+00
1.59561440e-01 -7.16307878e-01 8.86114895e-01 1.48769587e-01
-9.88871977e-02 -1.28112352e+00 5.09169817e-01 -3.18079352e-01
4.49737817e-01 -8.36740956e-02 1.33798468e+00 -1.10259533e+00
-6.24773145e-01 -1.84646189e-01 -4.46948290e-01 5.88185787e-01
2.96449456e-02 1.89641416e-02 -7.86001146e-01 -2.64103323e-01
1.50533631e-01 1.01397730e-01 1.00522566e+00 4.23272818e-01
1.43073130e+00 1.76842704e-01 -2.46984079e-01 4.42543238e-01
1.02152312e+00 -1.19251937e-01 2.78606385e-01 7.06687802e-03
4.65288371e-01 2.96788454e-01 2.26758555e-01 4.73975658e-01
-4.28155363e-01 3.85900766e-01 -2.27524503e-03 3.42483908e-01
2.38315184e-02 -2.17090368e-01 3.37304413e-01 8.31236124e-01
1.10475935e-01 -1.43739298e-01 -7.91826963e-01 1.16668411e-01
-1.59046507e+00 -1.08612287e+00 -4.29723710e-02 2.37302494e+00
4.78323668e-01 5.74598253e-01 1.05231978e-01 -1.40415862e-01
6.78254008e-01 1.47537729e-02 -4.81633991e-01 -3.07401270e-01
1.95178054e-02 6.76518381e-01 4.26722586e-01 4.29494649e-01
-1.23225224e+00 5.81168115e-01 7.06555939e+00 9.59827900e-01
-5.70895910e-01 1.51836082e-01 7.75625929e-02 -8.63233134e-02
-1.68942079e-01 3.95665988e-02 -9.62039948e-01 6.51311278e-01
1.03828251e+00 -1.88831270e-01 1.39703766e-01 8.20702255e-01
-4.47035022e-02 -2.75030881e-01 -1.13472104e+00 6.49582148e-01
2.08231751e-02 -1.08675933e+00 1.53084844e-01 3.16730618e-01
5.66909075e-01 -1.20696612e-01 4.12745886e-02 6.54132068e-01
2.79987514e-01 -6.99236393e-01 2.61825681e-01 9.86359298e-01
3.44101369e-01 -7.38975644e-01 5.93335092e-01 1.94524169e-01
-8.02489400e-01 1.26703665e-01 -6.83237731e-01 -3.77943702e-02
3.72264415e-01 1.15050197e+00 -8.57991636e-01 5.76637924e-01
3.47724229e-01 5.76638281e-01 -6.04242802e-01 1.42803955e+00
5.22938743e-03 9.39760804e-01 -5.67055941e-01 -6.00894801e-02
9.06761065e-02 -8.43700230e-01 1.29691446e+00 1.24761581e+00
8.41336012e-01 -4.09550041e-01 9.26775336e-02 1.23225760e+00
-8.27491954e-02 -1.35707095e-01 -7.77654409e-01 -3.69822443e-01
2.00517878e-01 1.42610872e+00 -9.90293622e-01 -8.93697143e-02
-4.71682519e-01 8.00838232e-01 4.29762602e-01 3.35851610e-01
-6.14899874e-01 -2.29029849e-01 9.01876509e-01 1.31614894e-01
6.37252152e-01 -7.99457312e-01 1.62292406e-01 -1.23907614e+00
-5.67160510e-02 -2.36402869e-01 1.82582542e-01 -3.23254287e-01
-2.18045831e+00 2.28318051e-02 2.53512591e-01 -1.02297044e+00
-2.20501110e-01 -1.24329603e+00 -8.56478870e-01 9.62042451e-01
-1.30822778e+00 -6.89388275e-01 1.41410902e-01 3.28052282e-01
5.59246838e-02 -6.42295659e-01 7.55289972e-01 1.83555096e-01
-4.09635782e-01 2.36308217e-01 5.53997934e-01 -9.14915651e-02
6.45298541e-01 -1.74346185e+00 3.15946877e-01 1.00103128e+00
3.22461843e-01 5.70100844e-01 8.75283003e-01 -9.94535267e-01
-1.20911014e+00 -9.27749872e-01 2.10267782e-01 -7.04995036e-01
1.35899222e+00 -6.35856271e-01 -1.13581610e+00 3.89304429e-01
-3.77405196e-01 2.05987468e-01 6.62133217e-01 4.08919632e-01
-1.57003179e-01 2.91241705e-01 -1.22861290e+00 2.68798649e-01
9.67291236e-01 -6.23680890e-01 -6.29274905e-01 3.13615561e-01
4.71574724e-01 -1.25161424e-01 -7.39909649e-01 4.67391282e-01
8.41301158e-02 -6.40974522e-01 9.90712345e-01 -9.74564970e-01
4.50695632e-03 -1.98412910e-01 -1.14508040e-01 -1.30911219e+00
-5.75165033e-01 -3.36532056e-01 -5.73310614e-01 1.25119328e+00
2.03236461e-01 -6.95914984e-01 8.07373583e-01 3.28178287e-01
2.09228545e-01 -1.59518734e-01 -1.44735050e+00 -1.05977094e+00
2.12573618e-01 -5.14985502e-01 -1.86893400e-02 6.41500890e-01
-3.76091957e-01 2.41480470e-01 -2.64234602e-01 6.26023412e-01
8.21686745e-01 -1.17773190e-01 5.88388264e-01 -1.65341449e+00
-6.15637898e-01 -5.10459602e-01 -9.16394234e-01 -7.37812936e-01
6.09168231e-01 -1.08474422e+00 -1.72015905e-01 -8.43539417e-01
1.95821241e-01 -2.97099084e-01 -4.36211556e-01 -2.38431737e-01
9.30843726e-02 -1.04749054e-02 -1.04748078e-01 -5.79324923e-02
-5.13221085e-01 7.91156530e-01 8.63841712e-01 5.10481559e-02
2.13388190e-01 2.60923356e-01 -6.39372766e-02 9.10364807e-01
8.04927886e-01 -4.64094281e-01 -2.08353862e-01 3.58488202e-01
1.88019410e-01 -3.42732996e-01 6.32601082e-01 -1.35746741e+00
1.96714606e-02 -6.01703003e-02 8.51046622e-01 -4.83511567e-01
4.24324930e-01 -4.81145710e-01 -2.86657661e-01 -6.80501089e-02
-1.12470783e-01 -1.38458148e-01 3.07041496e-01 1.10385430e+00
2.20221560e-03 -7.20881581e-01 7.12521076e-01 7.30906576e-02
-5.22110403e-01 2.67619580e-01 -5.94949901e-01 2.43884876e-01
1.02516723e+00 4.01423246e-01 -2.55135149e-01 -2.26716608e-01
-1.01830232e+00 -2.42102310e-01 2.46689796e-01 6.43044561e-02
3.82470965e-01 -1.23350322e+00 -4.56653476e-01 4.21771914e-01
2.56765541e-02 -4.91755515e-01 6.15977049e-02 8.55214715e-01
-3.30565959e-01 3.14210862e-01 -1.69049263e-01 -7.52757311e-01
-8.05969656e-01 5.09283245e-01 2.38900065e-01 -3.83582503e-01
-7.11771607e-01 5.40285408e-01 1.08842351e-01 -5.92348456e-01
-9.15690064e-02 -3.55700105e-01 1.03863299e-01 -1.89853951e-01
3.61470729e-01 4.49281365e-01 7.71369636e-02 -1.92911431e-01
-1.45298585e-01 1.86405912e-01 -1.68948054e-01 -2.22162291e-01
8.73933256e-01 1.66042402e-01 -2.08465144e-01 7.84019530e-01
9.15374875e-01 4.29731339e-01 -1.21596265e+00 -2.39551410e-01
2.40652993e-01 -7.22222030e-01 -2.93724332e-02 -5.42377532e-01
-4.22821969e-01 8.13256204e-01 6.06768191e-01 5.95566988e-01
3.51502508e-01 3.25498462e-01 3.90702069e-01 6.01117730e-01
2.08641842e-01 -8.23043168e-01 1.16880484e-01 7.98750758e-01
4.41587210e-01 -1.02956796e+00 -2.31498964e-02 -3.41166884e-01
-1.72029957e-01 1.10292065e+00 5.00195980e-01 -2.49120459e-01
9.13712084e-01 2.84914196e-01 -4.94033068e-01 -1.36917904e-01
-4.25154030e-01 -2.53586620e-01 5.51394939e-01 6.95791066e-01
1.51866361e-01 2.32060533e-03 -1.59865603e-01 2.82057881e-01
-2.63232470e-01 -5.44546723e-01 4.60858524e-01 7.24837124e-01
-6.03847742e-01 -1.27875245e+00 -1.69367000e-01 9.74363744e-01
-3.58731180e-01 1.94786908e-03 -1.20166920e-01 6.88632846e-01
-2.80093759e-01 4.34661537e-01 5.06319642e-01 2.30637044e-01
8.76931027e-02 6.08664751e-01 6.56397760e-01 -5.71582258e-01
2.11373150e-01 -1.92001790e-01 -3.16548161e-02 -3.07525277e-01
-2.77430922e-01 -8.82596135e-01 -7.44421363e-01 -2.54283458e-01
-4.94475067e-02 3.61501992e-01 6.54661298e-01 1.06597590e+00
2.16034070e-01 6.48608625e-01 2.41367728e-01 -9.26487982e-01
-9.48860943e-01 -9.87125576e-01 -1.00350869e+00 5.23739815e-01
2.60378391e-01 -1.14799297e+00 -8.65742028e-01 -5.74420989e-01] | [7.333451747894287, 3.951925277709961] |
0afdccc9-f1e7-4488-96ba-8a521417a9ce | 3d-saliency-guided-deep-quality-predictor-for | null | null | https://www.sciencedirect.com/science/article/pii/S0925231222000029 | https://www.researchgate.net/publication/357645676_3D_Saliency_guided_Deep_Quality_predictor_for_No-Reference_Stereoscopic_Images | 3D Saliency guided Deep Quality predictor for No-Reference Stereoscopic Images | The use of 3D technologies is growing rapidly, and stereoscopic imaging is usually used to display the 3D contents. However, compression, transmission and other necessary treatments may reduce the quality of these images. Stereo Image Quality Assessment (SIQA) has attracted more attention to ensure good viewing experience for the users and thus several methods have been proposed in the literature with a clear improvement for deep learning-based methods. This paper introduces a new deep learning-based no-reference SIQA using cyclopean view hypothesis and human visual attention. First, the cyclopean image is constructed considering the presence of binocular rivalry that covers the asymmetric distortion case. Second, the saliency map is computed considering the depth information. The latter aims to extract patches on the most perceptual relevant regions. Finally, a modified version of the pre-trained Convolutional Neural Network (CNN) is fine-tuned and used to predict the quality score through the selected patches. Five distinct pre-trained models were analyzed and compared in term of results. The performance of the proposed metric has been evaluated on four commonly used datasets (3D LIVE phase I and phase II databases as well as Waterloo IVC 3D Phase 1 and Phase 2). Compared with the state-of-the-art metrics, the proposed method gives better outcomes. The implementation code will be made accessible to the public at: https://github.com/o-messai/3D-NR-SIQA | ['Zianou Ahmed seghir', 'Fella Hachouf', 'Aladine Chetouani', 'Oussama Messai'] | 2022-01-06 | null | null | null | journal-2022-1 | ['image-quality-estimation', 'blind-image-quality-assessment', 'stereoscopic-image-quality-assessment', 'no-reference-image-quality-assessment'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 1.33807242e-01 -2.89241165e-01 -2.33609732e-02 -2.29189113e-01
-8.61102104e-01 -2.04537749e-01 3.68959159e-01 -4.44725715e-02
-2.95995146e-01 5.71666241e-01 4.08636272e-01 2.72803791e-02
-1.28632009e-01 -6.09272301e-01 -5.51172853e-01 -7.52488911e-01
-1.04052350e-01 -8.03329498e-02 4.93552446e-01 -2.73680210e-01
6.62394643e-01 5.19122124e-01 -1.81906617e+00 3.07595372e-01
9.28526938e-01 1.28961575e+00 5.87098837e-01 4.87688094e-01
2.68505603e-01 4.12851155e-01 -3.87407094e-01 -2.31548175e-01
5.34616649e-01 -4.76970643e-01 -6.91742778e-01 1.11398019e-01
3.89177144e-01 -4.45531040e-01 -4.62050766e-01 1.16267645e+00
9.75245357e-01 3.03892866e-02 3.07537109e-01 -9.86209631e-01
-4.07871366e-01 -1.31104365e-01 -6.40971839e-01 7.47152090e-01
5.76503456e-01 4.15299594e-01 7.75052071e-01 -9.78569210e-01
4.90235716e-01 1.06004155e+00 6.35991171e-02 2.53376096e-01
-7.95544028e-01 -6.36418462e-01 -4.90430683e-01 9.18649971e-01
-1.40393472e+00 -3.54276747e-01 9.64939713e-01 -2.40320250e-01
9.08704102e-01 1.25922307e-01 7.30569184e-01 6.74832940e-01
5.87679803e-01 7.21187353e-01 1.40952051e+00 -3.92950505e-01
1.54619962e-01 1.07911669e-01 -2.64141053e-01 4.91133153e-01
4.74087568e-03 3.67492050e-01 -6.62532926e-01 2.39116803e-01
6.60726130e-01 -1.38182238e-01 -5.70615411e-01 -6.02858901e-01
-1.04762936e+00 5.51184177e-01 7.69147038e-01 3.40484083e-01
-5.00941157e-01 -2.89439827e-01 2.91740596e-01 1.61320373e-01
5.14766932e-01 3.78699154e-01 -2.46994630e-01 -2.33211949e-01
-9.78545308e-01 1.71840116e-01 1.69989675e-01 7.15641141e-01
6.60930574e-01 -2.61675590e-03 -1.64525226e-01 8.92432690e-01
2.79241830e-01 5.82318425e-01 4.45268065e-01 -9.72954273e-01
4.88712847e-01 4.98486936e-01 1.94744349e-01 -1.09593487e+00
-4.08839971e-01 -6.16393447e-01 -8.52794111e-01 5.18564939e-01
7.05701634e-02 1.99309260e-01 -7.48261392e-01 1.15081620e+00
1.65888965e-01 -1.17335528e-01 -7.16145430e-03 1.42668247e+00
8.24655235e-01 6.72375977e-01 -4.13021296e-01 -1.93325698e-01
1.12248373e+00 -7.84230173e-01 -7.04930782e-01 -9.53809768e-02
2.82495972e-02 -9.59701955e-01 9.76732075e-01 6.98100507e-01
-1.44580400e+00 -9.34404492e-01 -1.33129776e+00 -2.18398228e-01
-1.89990357e-01 -3.39826792e-02 2.22692609e-01 6.63620114e-01
-1.27154362e+00 4.56258684e-01 -6.10518157e-01 -2.24132031e-01
5.96492648e-01 3.47441018e-01 -2.94545382e-01 -4.52228159e-01
-1.37056243e+00 1.08804250e+00 1.77743256e-01 7.93444887e-02
-9.90582883e-01 -4.64405447e-01 -7.27408111e-01 9.99471620e-02
1.07736841e-01 -5.47925949e-01 9.64504778e-01 -9.58331823e-01
-1.48959088e+00 9.81025219e-01 -2.68487155e-01 -3.37405920e-01
3.27940673e-01 -2.27407262e-01 -3.14198583e-01 5.27380884e-01
1.16469443e-01 7.75391281e-01 7.21344113e-01 -1.22134829e+00
-8.99395406e-01 -5.10674179e-01 2.89914846e-01 7.30958998e-01
2.42678504e-02 4.20951657e-02 -7.45982587e-01 -4.19712007e-01
2.24198520e-01 -6.41812503e-01 1.15148842e-01 6.71603531e-02
-2.48162270e-01 -3.01455939e-03 5.22727191e-01 -7.96643138e-01
1.18159580e+00 -2.10781670e+00 2.84019202e-01 -7.81679079e-02
1.41246572e-01 3.57340127e-01 5.31092994e-02 1.12322740e-01
-1.03566125e-01 -2.81667322e-01 -1.84361503e-01 -1.54471725e-01
-3.36335152e-01 -3.37064236e-01 1.97757870e-01 6.42095804e-01
1.92930892e-01 5.53351820e-01 -7.52563655e-01 -4.19143856e-01
7.30615854e-01 4.89941806e-01 -5.84719241e-01 2.98674047e-01
2.65220016e-01 6.35052800e-01 -1.20562844e-01 7.53638864e-01
1.03416705e+00 -1.44238994e-01 -2.04580784e-01 -4.67618704e-01
-3.75862777e-01 1.88476339e-01 -1.07964432e+00 1.89521492e+00
-5.16996861e-01 7.89019704e-01 -1.03891283e-01 -8.61015797e-01
8.12472880e-01 2.42776245e-01 5.07239521e-01 -1.43699801e+00
2.25603804e-01 4.47703719e-01 1.54029742e-01 -7.35592961e-01
3.49170595e-01 6.38342127e-02 2.18291372e-01 3.93218882e-02
1.31633595e-01 -2.96498209e-01 1.79506969e-02 -1.95584670e-02
7.20411360e-01 3.61566208e-02 3.63884121e-01 -1.51485011e-01
8.92125249e-01 -1.20106906e-01 3.77464265e-01 3.31575245e-01
-4.47286218e-01 1.02617216e+00 2.63233304e-01 -3.12996656e-01
-1.21203935e+00 -1.07759881e+00 -1.98447466e-01 3.59308034e-01
7.18544900e-01 -5.97434379e-02 -6.01892591e-01 -3.13189536e-01
-4.28085387e-01 6.65711462e-01 -4.16593641e-01 -1.22738183e-01
-3.45024735e-01 -3.73887658e-01 -1.25030264e-01 7.31078014e-02
1.03530633e+00 -1.20097494e+00 -9.45833564e-01 -4.75139543e-02
-5.30760288e-01 -9.97778177e-01 -2.69707739e-01 -1.32143155e-01
-7.73324728e-01 -1.10541403e+00 -1.10214269e+00 -6.68078601e-01
3.40450168e-01 7.01156676e-01 1.01103878e+00 -2.40271300e-01
-2.00555176e-01 2.14779437e-01 -3.87834698e-01 -3.00703257e-01
-7.26633593e-02 -9.85586792e-02 -1.80708513e-01 2.25259662e-02
3.50594193e-01 -4.70030010e-01 -1.15217948e+00 2.99212307e-01
-8.46615791e-01 2.46312141e-01 7.86644876e-01 7.49416888e-01
6.88203931e-01 1.75774828e-01 2.21564844e-01 -2.52012879e-01
3.61416727e-01 -2.47471645e-01 -7.17698991e-01 -1.54964194e-01
-5.25033653e-01 -1.18410498e-01 3.08710843e-01 1.21763624e-01
-1.11954916e+00 -1.87237844e-01 -3.07844162e-01 -5.95797062e-01
-2.49068245e-01 1.82813913e-01 -4.60538775e-01 -1.57349631e-01
5.59582889e-01 3.46821308e-01 -4.23841141e-02 -2.99787134e-01
-1.08528748e-01 8.05486262e-01 3.60854954e-01 1.55353755e-01
6.93911433e-01 6.44410849e-01 1.11824991e-02 -7.12966084e-01
-7.36788988e-01 -5.76076150e-01 -5.72919369e-01 -5.62773287e-01
1.01331031e+00 -9.24200296e-01 -5.40070295e-01 6.89488590e-01
-1.09065557e+00 -6.41201437e-02 -6.92755505e-02 7.36322284e-01
-7.64095962e-01 3.32539797e-01 -2.29573622e-01 -4.84468490e-01
-4.81202185e-01 -1.64326108e+00 9.98382390e-01 4.80953127e-01
1.86079547e-01 -5.50025344e-01 -2.26429082e-03 5.23430884e-01
5.00305831e-01 2.05398966e-02 8.56808364e-01 -6.27195761e-02
-8.02207470e-01 2.42143702e-02 -4.29392576e-01 5.71208537e-01
1.17842235e-01 -5.66681385e-01 -1.05114090e+00 -2.65085071e-01
2.05869466e-01 -1.39489606e-01 5.24069369e-01 8.20897758e-01
1.16772962e+00 7.87266195e-02 -5.15166596e-02 6.76395476e-01
1.66384089e+00 5.15482128e-01 1.02845144e+00 5.40127397e-01
4.32983905e-01 5.35394788e-01 8.75709593e-01 3.97460282e-01
2.71668315e-01 9.91931081e-01 6.91183388e-01 -3.03040355e-01
-3.91591996e-01 6.93439096e-02 1.75895691e-01 6.24788165e-01
-1.28602669e-01 -2.45334595e-01 -8.42000246e-01 6.55010939e-01
-1.25155652e+00 -9.94721055e-01 -1.15534768e-03 2.31538057e+00
5.56145489e-01 3.12958211e-01 -1.65898874e-01 4.35355991e-01
6.01710796e-01 2.32239485e-01 -5.77322423e-01 -3.74423891e-01
-2.46964604e-01 4.91828978e-01 4.16134268e-01 3.64082783e-01
-1.05401063e+00 4.81101125e-01 5.23765039e+00 9.27421093e-01
-1.36183584e+00 2.80882210e-01 7.55420566e-01 -2.78064698e-01
-7.73050860e-02 -1.68754473e-01 -4.77371722e-01 5.94556391e-01
7.93614149e-01 -8.33392814e-02 3.16782534e-01 4.52913582e-01
6.35080516e-01 -6.28588378e-01 -7.36674368e-01 1.43881023e+00
2.91911632e-01 -1.20983064e+00 -8.33337530e-02 1.11180328e-01
8.61928940e-01 1.75055400e-01 3.67346853e-01 -6.52718768e-02
-5.12211740e-01 -8.36907327e-01 6.45921588e-01 6.31652832e-01
8.33300889e-01 -1.00977993e+00 1.00428355e+00 1.31875485e-01
-8.21094990e-01 -1.76887974e-01 -3.45364749e-01 1.37234136e-01
1.61529317e-01 5.59212387e-01 -4.30790544e-01 7.98152566e-01
1.16467249e+00 8.33859563e-01 -7.60547400e-01 1.52096450e+00
-1.08413123e-01 3.41163218e-01 6.37556016e-02 2.21007407e-01
2.07360610e-01 -4.58890165e-04 7.91599452e-01 7.44003296e-01
5.95622480e-01 1.63003236e-01 -3.92516851e-01 7.08481371e-01
1.39066458e-01 1.14116363e-01 -5.07913530e-01 5.98804891e-01
1.38620334e-02 1.03543961e+00 -4.58838701e-01 -2.03601360e-01
-7.17288673e-01 1.14505231e+00 -1.95253238e-01 2.38968715e-01
-6.98918700e-01 -4.12070155e-01 4.44853455e-01 4.45278287e-01
4.01909292e-01 6.77650347e-02 -2.29098991e-01 -1.04228222e+00
1.16019078e-01 -9.93372977e-01 1.64229304e-01 -1.28170705e+00
-9.67925906e-01 6.89907074e-01 7.82483146e-02 -1.81914175e+00
-1.39019340e-01 -4.22573090e-01 -3.95088702e-01 1.13472235e+00
-1.84093666e+00 -5.43657839e-01 -6.45376265e-01 6.32833064e-01
9.15364325e-01 -1.34735882e-01 4.62611049e-01 6.35661125e-01
-1.90932512e-01 3.75870764e-01 1.74510524e-01 -2.67149448e-01
8.04925025e-01 -9.93801236e-01 2.78983060e-02 9.64922011e-01
-3.31761450e-01 1.51048424e-02 6.90997303e-01 -3.45484018e-01
-1.21090102e+00 -8.03726017e-01 6.87485158e-01 -6.45091385e-02
4.37424071e-02 -4.04638387e-02 -7.80297935e-01 -6.27878681e-02
6.05538428e-01 7.54788071e-02 2.42300302e-01 -5.05043149e-01
1.30055338e-01 -3.58476430e-01 -1.23710394e+00 3.24196637e-01
7.76975274e-01 -3.81446779e-01 -3.39200139e-01 8.00332893e-03
4.38837945e-01 -5.65573871e-01 -6.86031759e-01 4.57054794e-01
4.71796215e-01 -1.61327446e+00 9.17731881e-01 3.18927646e-01
7.56741703e-01 -4.65646118e-01 -2.70838499e-01 -1.36182916e+00
-1.74890235e-01 -1.72212973e-01 9.83894691e-02 7.19393909e-01
1.69309095e-01 -4.09666598e-01 6.31362617e-01 6.52683303e-02
-3.24308574e-01 -7.50264406e-01 -1.06637430e+00 -3.78786057e-01
-2.92265475e-01 -2.25770071e-01 3.20503145e-01 5.13984740e-01
-2.06648558e-01 2.82648265e-01 -3.58060122e-01 2.32186541e-01
6.82345510e-01 1.21519893e-01 5.36684513e-01 -9.02898610e-01
-1.71277732e-01 -4.41498309e-01 -8.12011123e-01 -1.03103685e+00
-4.73581314e-01 -7.30392039e-01 -1.50600657e-01 -1.63586450e+00
1.22645721e-01 -1.71708852e-01 -4.08301681e-01 -2.40460053e-01
-6.14617765e-02 4.44138795e-01 1.92493975e-01 2.73739040e-01
-5.30347884e-01 7.32470095e-01 1.60632741e+00 -1.99760884e-01
-1.87627807e-01 2.65290231e-01 -4.00708705e-01 4.58120555e-01
9.94973838e-01 -1.72570527e-01 -4.85805780e-01 -4.13202494e-01
9.41988379e-02 3.21041018e-01 4.52973127e-01 -1.50804758e+00
1.91228300e-01 2.39353970e-01 6.85718060e-01 -1.04003179e+00
5.07758260e-01 -8.74347687e-01 -2.09655482e-02 4.81865287e-01
-2.10418835e-01 2.05192398e-02 1.82340503e-01 2.84090996e-01
-6.09342337e-01 -1.01238862e-01 1.15471876e+00 -1.93635285e-01
-9.93855417e-01 2.99193710e-01 -7.84615353e-02 -1.05115965e-01
9.21902835e-01 -4.62826014e-01 -5.43307029e-02 -5.07201254e-01
-4.62866604e-01 -6.14964403e-03 4.13000166e-01 3.56447607e-01
1.03167391e+00 -1.28431571e+00 -7.04490125e-01 3.01619351e-01
2.29860038e-01 -2.06925929e-01 7.29131579e-01 9.51670408e-01
-6.66368663e-01 6.83548033e-01 -7.69987404e-01 -8.23546112e-01
-1.14870107e+00 6.40239716e-01 5.18772840e-01 -1.49944052e-01
-4.10785496e-01 5.71253717e-01 2.99645185e-01 5.78411743e-02
1.70245379e-01 -1.90569833e-01 -5.77306032e-01 -1.66196123e-01
5.60878038e-01 3.94971400e-01 2.85926849e-01 -8.11265588e-01
-3.12298447e-01 6.93373024e-01 1.19980574e-01 -2.33598590e-01
1.29906201e+00 -5.78445673e-01 1.58193022e-01 2.31462032e-01
1.35594797e+00 -1.46070644e-01 -1.29028678e+00 -2.61821181e-01
-3.79595488e-01 -8.38516951e-01 3.82170022e-01 -7.96664655e-01
-1.20488536e+00 1.32504594e+00 1.45678866e+00 -7.71848261e-02
1.71940482e+00 -1.14451468e-01 7.10613191e-01 -2.74078459e-01
5.13659239e-01 -8.58772278e-01 1.68894634e-01 2.08599851e-01
1.00934589e+00 -1.54536986e+00 4.82781790e-02 -1.85485810e-01
-6.62995577e-01 9.20692086e-01 6.73240125e-01 -9.38556492e-02
6.54918611e-01 -1.82792410e-01 -4.94562685e-02 -3.28806460e-01
-4.80978847e-01 -3.46791476e-01 5.86139560e-01 7.03610122e-01
2.82316178e-01 -1.89027414e-01 -3.24774772e-01 1.66146100e-01
-9.84549522e-02 -1.07728713e-03 4.89661396e-01 5.98021388e-01
-3.52756470e-01 -6.27092421e-01 -4.28108752e-01 3.72406453e-01
-5.42618096e-01 -1.55020162e-01 1.43698692e-01 6.35362506e-01
3.52081895e-01 1.04776263e+00 6.35863468e-02 -3.06944162e-01
5.97300053e-01 -4.09598500e-01 4.80736583e-01 -3.79806161e-01
-4.15556312e-01 1.62021413e-01 -2.30762511e-01 -8.93622458e-01
-7.89931417e-01 -4.91574526e-01 -8.94627810e-01 -2.01217905e-01
-1.48708493e-01 -9.34137851e-02 7.18704045e-01 5.85225761e-01
3.38763207e-01 4.70645159e-01 9.10084844e-01 -1.11006868e+00
9.54816397e-03 -9.19922054e-01 -4.73366559e-01 3.99079859e-01
4.37120736e-01 -8.56632590e-01 -3.92403960e-01 -9.58766863e-02] | [11.789329528808594, -1.9554492235183716] |
819876b3-6017-4dbf-bf6e-aa2eb763c417 | are-negative-samples-necessary-in-entity | 2108.05278 | null | https://arxiv.org/abs/2108.05278v2 | https://arxiv.org/pdf/2108.05278v2.pdf | Are Negative Samples Necessary in Entity Alignment? An Approach with High Performance, Scalability and Robustness | Entity alignment (EA) aims to find the equivalent entities in different KGs, which is a crucial step in integrating multiple KGs. However, most existing EA methods have poor scalability and are unable to cope with large-scale datasets. We summarize three issues leading to such high time-space complexity in existing EA methods: (1) Inefficient graph encoders, (2) Dilemma of negative sampling, and (3) "Catastrophic forgetting" in semi-supervised learning. To address these challenges, we propose a novel EA method with three new components to enable high Performance, high Scalability, and high Robustness (PSR): (1) Simplified graph encoder with relational graph sampling, (2) Symmetric negative-free alignment loss, and (3) Incremental semi-supervised learning. Furthermore, we conduct detailed experiments on several public datasets to examine the effectiveness and efficiency of our proposed method. The experimental results show that PSR not only surpasses the previous SOTA in performance but also has impressive scalability and robustness. | ['Man Lan', 'Yuanbin Wu', 'Wenting Wang', 'Xin Mao'] | 2021-08-11 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 1.22891821e-01 1.05468892e-01 -3.98195654e-01 -1.63190737e-01
-8.15533698e-01 -2.42394656e-01 1.02596171e-01 3.73988837e-01
-3.67095083e-01 8.94367576e-01 1.60568982e-01 -2.54299134e-01
-2.18289807e-01 -8.56464863e-01 -7.47336924e-01 -4.12586182e-01
-2.39500985e-01 4.81906533e-01 5.67536414e-01 -2.41724521e-01
-1.29557371e-01 2.15369210e-01 -1.06811726e+00 -1.53776467e-01
1.29778683e+00 8.44713688e-01 1.17310718e-01 1.48008034e-01
-1.06645703e-01 1.06333148e+00 -3.70471388e-01 -1.02800190e+00
1.87929660e-01 -2.62068152e-01 -9.14030969e-01 -1.40153423e-01
1.92855865e-01 -2.97161579e-01 -5.18945932e-01 1.33883977e+00
6.02709591e-01 -1.24786839e-01 2.94301420e-01 -1.56427610e+00
-7.33553469e-01 1.06350851e+00 -8.86482775e-01 -2.76979804e-02
2.52002597e-01 -1.54581249e-01 1.28179061e+00 -6.06328964e-01
6.67739749e-01 1.03532124e+00 8.47827077e-01 2.34195456e-01
-9.83789980e-01 -8.19371045e-01 2.19102174e-01 2.98292696e-01
-1.81216812e+00 -2.40574211e-01 7.45227098e-01 -1.35447025e-01
9.39258933e-01 7.33610988e-02 5.45993149e-01 6.92149460e-01
-8.56171697e-02 8.96792829e-01 7.23255336e-01 -2.85558373e-01
-6.81313351e-02 1.58327803e-01 1.70311611e-02 9.09246922e-01
4.92539793e-01 -2.80415088e-01 -3.65363419e-01 -3.12046975e-01
5.75716615e-01 -8.28447193e-02 -3.21564198e-01 -5.90986431e-01
-1.10725319e+00 7.08243132e-01 6.98915660e-01 7.03591630e-02
-2.10690066e-01 4.28240746e-02 5.86813748e-01 3.59311104e-01
2.10849330e-01 1.96651831e-01 -5.23983181e-01 1.90519169e-01
-5.80342114e-01 1.14443116e-01 7.05535352e-01 1.39654458e+00
7.88921714e-01 4.21843603e-02 1.74266919e-01 7.50859618e-01
2.12660268e-01 5.00704706e-01 3.29095930e-01 -3.91145736e-01
1.05265749e+00 9.04150486e-01 -1.45019919e-01 -1.27821577e+00
-4.53412473e-01 -5.75243235e-01 -1.23004341e+00 -4.97273862e-01
-1.30310670e-01 -8.22439194e-02 -5.15418768e-01 2.03430772e+00
3.89338672e-01 9.30851623e-02 2.64899552e-01 5.05556881e-01
1.01542044e+00 3.99689347e-01 1.79082364e-01 -4.29463923e-01
1.15036702e+00 -1.21601152e+00 -8.68565381e-01 -2.37285897e-01
9.80114460e-01 -4.87623245e-01 9.40708876e-01 -1.10394873e-01
-9.15960371e-01 -4.12335604e-01 -1.27679300e+00 -1.20681912e-01
-4.15287793e-01 3.38964880e-01 9.20319617e-01 4.71176147e-01
-8.52225721e-01 4.92751628e-01 -5.10374844e-01 -2.85986811e-01
4.47942346e-01 6.17071331e-01 -6.53336287e-01 -1.01343868e-02
-1.50515270e+00 6.78626060e-01 7.67756939e-01 -5.93435131e-02
-2.24790931e-01 -5.72106361e-01 -1.02506220e+00 2.24171758e-01
7.44329512e-01 -7.62478828e-01 7.78185129e-01 -7.55563915e-01
-1.05356324e+00 5.13405859e-01 -7.51694441e-02 -3.85289907e-01
5.96524715e-01 -3.65856886e-01 -7.61546075e-01 -6.70546591e-02
2.07399458e-01 3.99742007e-01 2.00371042e-01 -1.02331626e+00
-6.37508214e-01 -4.46317285e-01 1.52495760e-03 4.17823732e-01
-5.07165432e-01 -1.69424221e-01 -7.28072882e-01 -7.10931182e-01
2.68047273e-01 -9.41257358e-01 -2.03533486e-01 -4.55615520e-01
-7.06206441e-01 -2.10776776e-01 4.25008386e-01 -7.26611614e-01
1.78033912e+00 -2.16204739e+00 2.20678568e-01 1.70132533e-01
3.18848521e-01 4.88673002e-01 -3.22104990e-02 6.12627268e-01
-2.43959263e-01 2.22375840e-01 -1.30447447e-01 -4.98884059e-02
-1.53972864e-01 1.91104449e-02 -1.57613039e-01 1.10727355e-01
8.44673961e-02 1.08688915e+00 -1.10782266e+00 -7.83748150e-01
-1.02927729e-01 2.68477678e-01 -4.30337369e-01 1.39580414e-01
6.34978414e-02 1.22748233e-01 -3.63351494e-01 7.89087534e-01
7.54179716e-01 -6.37179255e-01 5.75555682e-01 -5.97233355e-01
2.94235110e-01 2.86965013e-01 -1.63663745e+00 1.34474528e+00
1.79800969e-02 3.63825588e-03 -2.98347652e-01 -7.26359367e-01
8.51429760e-01 1.85656339e-01 3.56504619e-01 -7.03747034e-01
-7.29333535e-02 4.22916055e-01 -2.11005449e-01 -3.99292141e-01
7.57753372e-01 2.60944992e-01 -1.77558977e-02 2.04288840e-01
1.39595509e-01 5.58775306e-01 3.42306644e-01 5.03677249e-01
1.09082425e+00 1.88738462e-02 5.90774894e-01 1.45277726e-02
6.78198338e-01 -2.05411226e-01 1.04693401e+00 4.16566223e-01
-2.14202166e-01 3.97542790e-02 6.63091719e-01 -2.96343327e-01
-8.96032214e-01 -7.53233314e-01 3.45610052e-01 6.99780345e-01
4.46733236e-01 -8.37323308e-01 -4.17919099e-01 -1.01499033e+00
1.78064834e-02 4.34548378e-01 -2.69752920e-01 -4.76453453e-01
-5.90897560e-01 -9.61173356e-01 7.01747715e-01 7.14959621e-01
8.05449426e-01 -6.61807299e-01 2.14755878e-01 1.47413045e-01
-5.61198652e-01 -1.36173415e+00 -4.94617939e-01 -2.26362105e-02
-9.67804909e-01 -1.21815276e+00 -3.67435932e-01 -9.92556810e-01
9.96751189e-01 4.35419887e-01 9.66169178e-01 2.32410878e-01
1.74878493e-01 -2.05032945e-01 -3.24160159e-01 -1.81404009e-01
-2.66999424e-01 5.80046594e-01 1.13955587e-01 -2.24246085e-01
3.94522369e-01 -4.88347828e-01 -4.97690141e-01 3.42409372e-01
-8.51881623e-01 1.13184124e-01 1.01645851e+00 8.75527263e-01
7.00509131e-01 4.19192255e-01 8.55406523e-01 -1.41817808e+00
5.24107873e-01 -4.42978114e-01 -6.20462716e-01 7.99928665e-01
-1.15909767e+00 2.23145988e-02 7.62452781e-01 -1.78999379e-01
-8.68786395e-01 1.20342761e-01 -2.64640629e-01 -2.43295982e-01
4.40362573e-01 6.31956577e-01 -3.62637579e-01 -9.05552804e-02
4.43433046e-01 2.90199041e-01 -1.40827477e-01 -3.69147182e-01
2.95185655e-01 6.45765424e-01 5.30384421e-01 -4.11285400e-01
1.00081348e+00 1.70307845e-01 2.39006174e-03 -2.04276115e-01
-7.79439390e-01 -3.44836980e-01 -6.46386683e-01 2.40463883e-01
4.08313900e-01 -1.42482400e+00 -6.39773011e-01 3.84891570e-01
-7.69976318e-01 1.16668552e-01 3.51440571e-02 4.85745341e-01
-3.00564915e-01 8.10837746e-01 -8.01968098e-01 -6.32149398e-01
-8.22950661e-01 -9.67959821e-01 7.01479733e-01 3.98695529e-01
1.48810282e-01 -6.84568226e-01 -7.37969354e-02 5.07597089e-01
2.70686924e-01 1.56706795e-01 1.17085671e+00 -7.69027054e-01
-6.50117159e-01 -2.97821552e-01 -3.92126352e-01 1.41645640e-01
1.73654377e-01 1.56458002e-02 -4.83327776e-01 -5.68190813e-01
-7.17926741e-01 -4.58358973e-01 5.49495459e-01 -1.92293480e-01
8.59103084e-01 -4.62274730e-01 -5.06558120e-01 5.98125100e-01
1.68276691e+00 4.27335426e-02 7.04216003e-01 2.72624403e-01
9.79372442e-01 3.08386177e-01 9.49534714e-01 2.53976554e-01
7.94723094e-01 5.19204080e-01 3.02544892e-01 -3.03128570e-01
1.60971861e-02 -8.05768847e-01 3.01765084e-01 1.49303389e+00
-1.28383547e-01 -2.73686588e-01 -6.47026002e-01 6.69071555e-01
-2.20079088e+00 -6.90240979e-01 -1.00293398e-01 2.25522208e+00
7.48814285e-01 3.48523349e-01 1.61476418e-01 1.69645295e-01
8.77160132e-01 8.86538029e-02 -5.77839851e-01 7.10207820e-02
-2.50564426e-01 -1.80293128e-01 6.91590667e-01 1.09902352e-01
-1.13396120e+00 1.11323082e+00 6.22329807e+00 7.20638990e-01
-7.59511650e-01 1.63554382e-02 2.90496856e-01 2.65100539e-01
-3.37740958e-01 1.54720634e-01 -9.00594294e-01 5.07147551e-01
5.07371128e-01 -2.69269198e-01 3.40800762e-01 9.37580287e-01
-6.46054029e-01 3.72210681e-01 -8.13621879e-01 9.11596358e-01
1.70782749e-02 -1.17664385e+00 2.44130641e-01 -1.71907738e-01
6.44446969e-01 1.77186325e-01 -4.32220817e-01 4.92668092e-01
5.04396617e-01 -6.70225322e-01 3.70989889e-01 1.13590129e-01
8.46981883e-01 -1.01913130e+00 1.00762141e+00 2.32221693e-01
-1.60641253e+00 -6.91812858e-02 -6.20072663e-01 4.54295903e-01
8.81489292e-02 6.05346024e-01 -6.67691886e-01 1.34445965e+00
6.24513626e-01 8.19259942e-01 -6.71933949e-01 9.11377549e-01
-4.29257154e-01 2.46382758e-01 -3.07174206e-01 -1.73501689e-02
-1.20207109e-01 -2.48602509e-01 3.02570760e-01 1.05796301e+00
2.51623988e-01 -1.85100451e-01 1.41951814e-01 4.70497638e-01
-4.26720262e-01 4.12231445e-01 -6.03552461e-01 -2.05663696e-01
9.95217621e-01 1.20499218e+00 -4.97074246e-01 -2.44786292e-01
-7.34977961e-01 8.54226053e-01 8.23954105e-01 1.73417494e-01
-1.00306618e+00 -6.91130400e-01 2.17926025e-01 -9.77687016e-02
3.17468762e-01 -5.99696115e-02 2.24813551e-01 -1.25446284e+00
2.91321695e-01 -9.61222291e-01 8.86052251e-01 -5.99250555e-01
-1.18971050e+00 6.85645521e-01 -3.42342794e-01 -1.23597634e+00
6.29607961e-02 -9.08848569e-02 -2.09051445e-01 4.91441905e-01
-1.66438115e+00 -1.42911947e+00 -2.62653291e-01 6.16914988e-01
-8.67910311e-02 -2.04475716e-01 7.14160502e-01 8.89899969e-01
-9.01681244e-01 1.10541022e+00 8.19897186e-03 2.84367442e-01
8.24404955e-01 -1.22121787e+00 4.87185061e-01 9.19637859e-01
-1.65920071e-02 5.93675554e-01 2.38516703e-01 -8.63387167e-01
-1.36099803e+00 -1.23145974e+00 1.25164652e+00 -1.02345057e-01
5.84744573e-01 -2.15661466e-01 -9.49946642e-01 1.07200074e+00
-1.71695709e-01 -1.16123125e-01 7.77994752e-01 3.29375267e-01
-5.68832815e-01 -4.63548571e-01 -1.06531131e+00 7.10665405e-01
1.18597770e+00 -4.23162729e-01 -3.77333105e-01 2.84928381e-01
9.95370686e-01 -4.96067911e-01 -1.17247653e+00 9.28165674e-01
4.31115538e-01 -7.43454397e-01 8.84319544e-01 -6.82584167e-01
2.73872763e-01 -4.74823475e-01 -4.91350591e-02 -1.01769066e+00
-6.04336202e-01 -6.51492715e-01 -4.45044279e-01 1.66922164e+00
4.78473037e-01 -8.49490106e-01 6.15108669e-01 3.43126118e-01
1.73565373e-01 -9.35349822e-01 -6.12020075e-01 -7.84588873e-01
-3.52393806e-01 2.24027544e-01 1.08418512e+00 1.30844748e+00
1.59776896e-01 7.11071134e-01 -6.49207711e-01 4.94769871e-01
5.12096584e-01 2.69891798e-01 8.71194363e-01 -1.09592485e+00
-2.22692207e-01 -9.83267128e-02 -5.84398210e-01 -9.76494849e-01
-1.91814050e-01 -8.78615975e-01 -3.11133534e-01 -1.58641660e+00
6.66264951e-01 -6.81008220e-01 -5.02007306e-01 6.43347144e-01
-5.33869743e-01 -1.37868598e-01 -4.35429364e-02 4.37651783e-01
-1.02763307e+00 4.93245959e-01 8.82968485e-01 6.26832470e-02
-1.09848797e-01 -2.65127122e-01 -9.35886204e-01 4.46989149e-01
6.42219961e-01 -5.05533099e-01 -6.68718934e-01 -3.60203952e-01
6.76091492e-01 1.69007853e-01 -7.12390020e-02 -9.08696353e-01
4.52844739e-01 -1.17112361e-02 1.55048683e-01 -7.15783715e-01
-1.79916963e-01 -7.64585853e-01 3.96669507e-01 6.21986806e-01
-1.30245224e-01 4.31026638e-01 -1.52392790e-01 7.90022492e-01
-3.69852304e-01 -2.90782414e-02 5.83559752e-01 1.08195886e-01
-6.89023077e-01 5.28500021e-01 4.21264708e-01 1.82677388e-01
1.02647245e+00 1.11551709e-01 -5.58295667e-01 -2.80129433e-01
-2.82589257e-01 6.13585293e-01 5.22188365e-01 5.23318946e-01
3.16582143e-01 -1.51134157e+00 -6.32145464e-01 1.39668524e-01
3.12009007e-01 2.79967397e-01 2.55927414e-01 9.15441573e-01
-4.43494767e-01 3.32695216e-01 -3.20614688e-02 -1.89089194e-01
-1.43686783e+00 6.83831692e-01 1.52585208e-01 -7.10371792e-01
-5.81978798e-01 6.67208850e-01 -7.06268549e-02 -6.98751628e-01
3.90512943e-01 2.29944810e-01 -2.04989493e-01 -1.37975529e-01
2.05897823e-01 3.35016996e-01 4.55009975e-02 -6.32664204e-01
-4.00279433e-01 4.12674546e-01 -5.35607398e-01 6.34235263e-01
1.21365583e+00 -2.23507002e-01 -2.48360604e-01 1.00965209e-01
9.99639213e-01 1.91629618e-01 -6.55937254e-01 -6.56252682e-01
1.41511664e-01 -3.32422942e-01 -2.65325367e-01 -5.86358607e-01
-1.23279703e+00 4.38616514e-01 2.59718597e-01 1.22084834e-01
1.20021057e+00 -2.02630714e-01 1.25045979e+00 4.20315146e-01
5.08959830e-01 -1.27410972e+00 -1.51233763e-01 2.42805079e-01
3.97884578e-01 -1.17363858e+00 5.25712371e-01 -9.27263379e-01
-8.73488843e-01 7.58450389e-01 9.04721916e-01 3.48114192e-01
4.32007730e-01 1.32960090e-02 -2.80465245e-01 -2.20443889e-01
-7.83190370e-01 -2.95043260e-01 1.04414135e-01 3.31097662e-01
2.90695161e-01 -5.42985573e-02 -4.56907809e-01 8.59901965e-01
-1.36008814e-01 1.75971240e-01 2.43146390e-01 9.67490613e-01
-9.63719562e-02 -1.15653849e+00 3.84428531e-01 4.80973989e-01
-4.94610995e-01 -4.15803194e-02 -4.61807996e-01 9.10997093e-01
-3.37853208e-02 6.89444065e-01 -3.14721346e-01 -7.79873013e-01
4.65839893e-01 4.42854827e-03 3.07901204e-01 -2.24336535e-01
-3.56629819e-01 -2.38044485e-01 2.50720024e-01 -3.82923245e-01
-3.24089199e-01 -3.30663562e-01 -1.28456926e+00 -4.02114153e-01
-8.70318294e-01 2.04095155e-01 2.71919072e-01 6.55713558e-01
7.78629661e-01 4.03502941e-01 7.10167050e-01 2.79477477e-01
-7.10778534e-01 -8.27298939e-01 -6.41809881e-01 6.02876067e-01
-7.82961175e-02 -4.65844214e-01 -1.62356332e-01 -3.79688382e-01] | [8.74822998046875, 7.973048210144043] |
fbb7fa08-4afb-463b-9e38-e91ab00678b4 | one-class-kernel-spectral-regression | 1807.01085 | null | http://arxiv.org/abs/1807.01085v6 | http://arxiv.org/pdf/1807.01085v6.pdf | One-Class Kernel Spectral Regression | The paper introduces a new efficient nonlinear one-class classifier
formulated as the Rayleigh quotient criterion optimisation. The method,
operating in a reproducing kernel Hilbert space, minimises the scatter of
target distribution along an optimal projection direction while at the same
time keeping projections of positive observations distant from the mean of the
negative class. We provide a graph embedding view of the problem which can then
be solved efficiently using the spectral regression approach. In this sense,
unlike previous similar methods which often require costly eigen-computations
of dense matrices, the proposed approach casts the problem under consideration
into a regression framework which is computationally more efficient. In
particular, it is shown that the dominant complexity of the proposed method is
the complexity of computing the kernel matrix. Additional appealing
characteristics of the proposed one-class classifier are: 1-the ability to be
trained in an incremental fashion (allowing for application in streaming data
scenarios while also reducing the computational complexity in a non-streaming
operation mode); 2-being unsupervised, but providing the option for refining
the solution using negative training examples, when available; And last but not
the least, 3-the use of the kernel trick which facilitates a nonlinear mapping
of the data into a high-dimensional feature space to seek better solutions. | ['Shervin Rahimzadeh Arashloo', 'Josef Kittler'] | 2018-07-03 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 4.05199081e-01 3.18779141e-01 1.94948286e-01 -1.04785904e-01
-4.28154469e-01 -3.82752448e-01 5.71636140e-01 3.74087125e-01
-6.59843862e-01 5.40566325e-01 -2.36493155e-01 -3.23029995e-01
-5.77265799e-01 -7.42611349e-01 -3.20570409e-01 -1.16589773e+00
-3.52769911e-01 5.26968420e-01 1.41574740e-01 -1.95293069e-01
2.57249087e-01 7.73452997e-01 -1.72883141e+00 -3.01529169e-01
7.70727575e-01 1.18889201e+00 8.68605822e-02 7.33927786e-01
1.63060009e-01 5.85070491e-01 -2.77853966e-01 -1.82099655e-01
4.81709599e-01 -3.18873793e-01 -4.83160168e-01 3.38749647e-01
-5.21484837e-02 2.51316696e-01 1.89797416e-01 9.25303817e-01
4.03533012e-01 2.75296420e-01 8.47672582e-01 -1.10361087e+00
-7.14130476e-02 -6.23344304e-03 -3.90232891e-01 7.76439235e-02
3.48599255e-01 -3.76936555e-01 9.89899755e-01 -1.01618052e+00
3.67783397e-01 6.73607469e-01 4.94155198e-01 1.12775899e-01
-1.53965461e+00 -2.04419225e-01 -2.89230138e-01 2.22154602e-01
-1.63187945e+00 -2.49281555e-01 8.73613477e-01 -6.22917295e-01
7.31067121e-01 6.07731223e-01 8.45994055e-01 4.61031377e-01
9.87895057e-02 4.95702863e-01 1.06207252e+00 -7.01026678e-01
5.59568822e-01 6.26559794e-01 -4.77326885e-02 4.89262819e-01
2.35490233e-01 -6.54625744e-02 -1.72954082e-01 -3.82717311e-01
2.36830667e-01 -4.07776237e-02 -4.67240006e-01 -1.02212203e+00
-9.44123030e-01 1.10816002e+00 9.97895449e-02 3.76982689e-01
-4.94761109e-01 -4.36570793e-01 3.43422830e-01 3.28287989e-01
5.99250674e-01 2.80439615e-01 -1.58653438e-01 -7.32019693e-02
-1.06873858e+00 8.19689110e-02 1.18872988e+00 4.57382590e-01
9.93397057e-01 1.00805305e-01 5.75829864e-01 5.48347175e-01
3.99121404e-01 3.01624715e-01 6.23908758e-01 -3.11966479e-01
3.36191624e-01 5.97331047e-01 1.80328246e-02 -1.07898176e+00
-4.52694625e-01 -6.02135181e-01 -8.37117076e-01 3.96355718e-01
3.76844108e-01 -2.34031789e-02 -1.54217035e-01 1.39403534e+00
6.37244284e-01 2.53587067e-01 2.90407509e-01 6.53941512e-01
-2.20446922e-02 6.20360017e-01 -3.46525371e-01 -6.40248179e-01
8.64332318e-01 -5.92508793e-01 -4.61703360e-01 2.44693369e-01
7.48982072e-01 -7.60882914e-01 6.92377031e-01 5.99697590e-01
-6.89905584e-01 -2.58871138e-01 -1.22832561e+00 4.34045136e-01
-5.01001000e-01 1.89519197e-01 3.39433908e-01 8.80094349e-01
-1.14501607e+00 6.50577188e-01 -6.12401187e-01 -3.55045378e-01
-1.58257544e-01 6.43301904e-01 -4.98620838e-01 1.75819680e-01
-9.12942231e-01 9.31720376e-01 4.69767839e-01 2.82229275e-01
-9.56347063e-02 -5.49248695e-01 -1.04558706e+00 1.07093520e-01
2.73382157e-01 -6.78510070e-02 5.32699645e-01 -1.04131746e+00
-1.66099787e+00 5.89662015e-01 -1.44456372e-01 -4.10032839e-01
6.89471841e-01 -3.80241908e-02 -3.43474865e-01 3.25904757e-01
-1.08951658e-01 -8.84297043e-02 1.13263810e+00 -8.11634839e-01
-3.49893659e-01 -4.73797500e-01 -3.39293092e-01 3.83648127e-01
-6.76907599e-01 -4.16205049e-01 -8.37848634e-02 -4.44292188e-01
2.79822290e-01 -1.01699460e+00 -3.33917171e-01 -2.05732331e-01
-8.50895494e-02 -1.75979018e-01 1.00647509e+00 -4.55759108e-01
1.16853309e+00 -2.32410169e+00 6.02640986e-01 9.31925416e-01
1.11930460e-01 1.64764777e-01 2.36536726e-01 8.85434210e-01
-4.42638606e-01 -4.00652349e-01 -4.03686851e-01 -2.89441198e-02
-2.76197374e-01 6.43897057e-02 -1.83024123e-01 1.00573885e+00
4.18968290e-01 3.57574195e-01 -7.64375746e-01 -4.70234245e-01
4.80281442e-01 5.36599994e-01 -3.93811256e-01 2.32397601e-01
4.42645371e-01 2.74358362e-01 -1.64974079e-01 5.79655915e-02
4.50993091e-01 -1.30869880e-01 1.90020815e-01 -5.02747074e-02
-2.32474670e-01 -1.78895801e-01 -1.85249412e+00 1.19530857e+00
-5.32272518e-01 6.03264451e-01 1.62667960e-01 -1.61542928e+00
1.20503986e+00 3.74493688e-01 8.00523877e-01 -2.37780124e-01
2.72137336e-02 3.39806825e-01 -1.18066810e-01 -2.58832753e-01
4.88602281e-01 -3.44552338e-01 2.28652164e-01 4.91256088e-01
8.48034862e-03 -5.90581447e-02 1.64442584e-01 8.10163394e-02
7.91280866e-01 -2.11704858e-02 7.05688596e-01 -5.38153648e-01
9.93367910e-01 -2.99292624e-01 8.23246166e-02 2.28043094e-01
1.68169618e-01 2.88798422e-01 4.08124655e-01 -3.21152419e-01
-7.56803930e-01 -9.28760111e-01 -3.73759419e-01 6.79080725e-01
-8.01865384e-03 -2.43993744e-01 -5.88733852e-01 -5.17207623e-01
-7.92424679e-02 5.64155221e-01 -7.15007484e-01 -2.46246438e-02
-3.67384851e-01 -5.88925898e-01 6.99695386e-03 1.29252058e-02
-4.46010567e-02 -7.32742250e-01 -8.35739374e-01 1.89115733e-01
1.27434567e-01 -6.51136220e-01 -1.37178466e-01 6.20170951e-01
-1.11627042e+00 -1.04952860e+00 -8.17680597e-01 -6.69793725e-01
6.78600788e-01 1.17999390e-01 5.88547409e-01 -1.67956188e-01
-4.55727130e-01 6.61210537e-01 -3.33550036e-01 -2.51368493e-01
-2.51051337e-01 -6.26171231e-02 1.96969345e-01 5.90851426e-01
3.87513012e-01 -6.28747582e-01 -2.46823639e-01 2.95827031e-01
-9.82618988e-01 -2.75875121e-01 5.49502969e-01 9.87527192e-01
5.15433967e-01 3.40145290e-01 5.69947779e-01 -9.12172854e-01
5.51771939e-01 -5.85174680e-01 -7.63245106e-01 1.45461023e-01
-8.77783895e-01 1.75985768e-01 8.72673750e-01 -4.27282840e-01
-7.59150684e-01 4.45656002e-01 5.69249243e-02 -2.03709215e-01
-1.22980950e-02 3.79522711e-01 8.19180831e-02 -3.80899906e-01
7.61375844e-01 5.50969303e-01 4.16056961e-01 -3.10287148e-01
3.01145434e-01 7.66002893e-01 1.97843716e-01 -2.14065954e-01
1.06086266e+00 5.23262739e-01 4.03187722e-01 -1.46332526e+00
-2.63949692e-01 -1.09132171e+00 -9.87623334e-01 -4.10296738e-01
4.59279358e-01 -5.84369838e-01 -6.00115657e-01 2.47519258e-02
-5.27762949e-01 1.10055715e-01 -6.62639022e-01 7.15721846e-01
-8.40325058e-01 7.88483560e-01 -8.43582526e-02 -1.27718163e+00
-1.56013459e-01 -8.06320667e-01 7.41683602e-01 2.31825057e-02
-2.82128215e-01 -1.16185057e+00 2.59741902e-01 5.67072779e-02
1.46634117e-01 9.65595096e-02 7.41347849e-01 -9.07131374e-01
-4.59352545e-02 -7.68214643e-01 3.46913561e-02 5.56253672e-01
5.28186448e-02 -1.39541924e-01 -9.08150613e-01 -5.40159762e-01
3.78832579e-01 -1.00576118e-01 4.22031045e-01 1.08213589e-01
6.14234805e-01 -1.10241115e-01 -6.36623567e-03 4.04248804e-01
1.50353861e+00 -4.36021760e-02 3.83611262e-01 2.87039548e-01
3.45956773e-01 8.46587062e-01 7.51588464e-01 6.80714607e-01
-5.92795610e-02 7.44769335e-01 3.77949625e-01 -2.38786623e-01
6.21384561e-01 2.26345778e-01 2.59753615e-01 9.43180740e-01
-1.46647003e-02 1.62779555e-01 -6.55944943e-01 5.15441120e-01
-1.86930871e+00 -9.77939129e-01 -1.88707560e-01 2.72581577e+00
3.48135144e-01 -5.37616434e-03 3.16633284e-01 8.14116776e-01
5.43511868e-01 1.85625441e-02 -1.17667399e-01 -6.00397706e-01
9.38590243e-02 4.29976672e-01 4.87393409e-01 6.71765566e-01
-1.06553054e+00 1.29792407e-01 5.29322290e+00 5.89760840e-01
-1.19237053e+00 -2.21516088e-01 7.99591318e-02 2.01736987e-02
-8.42625573e-02 7.98536688e-02 -4.80650872e-01 3.17382634e-01
9.71166313e-01 -1.39956698e-01 3.97186637e-01 1.00009394e+00
1.55789644e-01 -3.71253788e-01 -1.10109794e+00 1.03913057e+00
2.67878532e-01 -8.45310271e-01 -2.78468579e-01 4.94197220e-01
2.61975855e-01 -2.46112645e-01 -1.08410142e-01 5.00933677e-02
-4.70672041e-01 -7.46531129e-01 6.09453082e-01 3.59472662e-01
5.27739048e-01 -9.94863033e-01 7.87813604e-01 6.47283137e-01
-1.11514890e+00 -1.60814628e-01 -2.98349202e-01 -2.39587501e-01
2.15750393e-02 6.90343142e-01 -9.61762846e-01 6.74420834e-01
4.41362798e-01 5.68944395e-01 -4.81303483e-01 9.88298774e-01
2.20251247e-01 3.82033259e-01 -6.50809109e-01 -2.66590089e-01
2.73622662e-01 -6.71615303e-01 6.56542242e-01 1.32535422e+00
5.73664069e-01 -7.28484690e-02 1.26981750e-01 3.89367700e-01
5.38434863e-01 7.44774938e-01 -9.52932298e-01 6.98388845e-04
-4.78096819e-03 1.37171912e+00 -8.97826254e-01 -1.19646285e-02
-5.13918459e-01 9.79362428e-01 3.24050546e-01 3.36871922e-01
-2.15950534e-01 -6.52753055e-01 1.42451301e-01 2.76784152e-01
5.44950962e-01 -1.60868332e-01 -8.95856172e-02 -9.95154262e-01
2.59221137e-01 -4.76883620e-01 3.95791203e-01 -1.80301681e-01
-9.27350938e-01 5.48454702e-01 -5.97909242e-02 -1.51656544e+00
-7.20736980e-01 -6.92417145e-01 -3.39914948e-01 1.06546628e+00
-1.32906282e+00 -6.26023114e-01 4.51895855e-02 7.15012312e-01
1.33387342e-01 -2.21575618e-01 1.18971598e+00 1.65120736e-01
-2.35705867e-01 3.37731034e-01 4.43302423e-01 -2.00817734e-01
3.37572843e-01 -1.43380916e+00 -3.81497592e-01 7.58989692e-01
2.57526726e-01 4.73017037e-01 8.36025298e-01 -2.61366278e-01
-1.20745850e+00 -6.36036992e-01 9.72142279e-01 5.23508899e-03
8.46147060e-01 -3.54369193e-01 -9.49555874e-01 1.66665792e-01
-1.32082731e-01 7.98684284e-02 1.05544043e+00 1.54708892e-01
-4.91412468e-02 -1.47332907e-01 -1.04296327e+00 3.09811980e-01
6.95276111e-02 -5.31859756e-01 -3.86868864e-01 3.65126312e-01
-1.27206799e-02 -3.49913649e-02 -9.07557011e-01 4.76728939e-03
5.68193555e-01 -1.07728589e+00 7.60887921e-01 -3.98352683e-01
-5.76421469e-02 -4.15958881e-01 -9.16226655e-02 -1.17449307e+00
-1.65146157e-01 -7.09855556e-01 -1.73888847e-01 8.00967693e-01
3.91562790e-01 -8.85021865e-01 8.96451414e-01 3.13229352e-01
3.71111721e-01 -1.05795252e+00 -1.26886821e+00 -7.18922436e-01
-4.13189590e-01 -4.76416290e-01 -1.59025550e-01 7.55946517e-01
2.76597857e-01 2.32382730e-01 -4.16157573e-01 2.18975410e-01
7.07769454e-01 8.43475983e-02 6.71522021e-01 -1.46502995e+00
-6.80187523e-01 -7.14196861e-02 -1.06245852e+00 -6.92280114e-01
3.56860198e-02 -7.78298914e-01 2.66864952e-02 -8.31753075e-01
-2.23730668e-01 -3.72866362e-01 -2.84926146e-01 -8.94290805e-02
1.69903159e-01 2.26532549e-01 7.99673144e-03 2.65821934e-01
-1.66595131e-01 4.39059347e-01 5.69355607e-01 2.52661079e-01
-3.09242904e-01 4.64942247e-01 -3.61335963e-01 5.99134207e-01
5.45680344e-01 -4.25823450e-01 -6.59003258e-01 3.45790803e-01
3.70108306e-01 4.83765788e-02 1.30675524e-01 -9.17454541e-01
2.24975124e-01 1.44214347e-01 2.63988554e-01 -3.88676584e-01
4.52041179e-01 -1.13649058e+00 1.11314312e-01 5.35529852e-01
-1.37383848e-01 3.80116999e-02 -1.36655299e-02 9.85731781e-01
-3.20426315e-01 -8.44658375e-01 8.96571398e-01 2.12637633e-01
-4.90613699e-01 -1.75716281e-01 -5.94188750e-01 -2.37831861e-01
1.32744443e+00 -5.99345803e-01 5.16174495e-01 -5.52342296e-01
-8.75672400e-01 -1.83310226e-01 5.01834631e-01 -6.75534680e-02
4.96012568e-01 -1.06763327e+00 -4.63328987e-01 5.65482199e-01
1.89218864e-01 -1.64129913e-01 1.04610890e-01 1.11591148e+00
-6.44975305e-01 3.54857415e-01 5.97959943e-02 -7.92688668e-01
-1.33706057e+00 5.41732788e-01 8.21368098e-02 -3.40238482e-01
-6.62369072e-01 5.43174922e-01 4.23420779e-03 -3.79615873e-01
1.23943590e-01 1.47769824e-01 -5.28987765e-01 3.44495982e-01
5.60031474e-01 5.81428170e-01 4.55254853e-01 -8.78389776e-01
-2.93576509e-01 4.98894870e-01 1.46602377e-01 -2.62108058e-01
1.45936561e+00 -9.25666839e-02 -2.26290300e-01 6.13915920e-01
1.45329726e+00 2.84390271e-01 -9.72944140e-01 -2.49450609e-01
3.00808668e-01 -3.66875410e-01 9.93613228e-02 -9.41718742e-02
-6.34216726e-01 8.11918914e-01 6.69610500e-01 5.63918114e-01
1.22340274e+00 -2.86843061e-01 1.18781976e-01 4.84963447e-01
1.07852764e-01 -1.11111832e+00 -2.23653615e-01 1.90143451e-01
7.31185794e-01 -9.81251061e-01 2.12671623e-01 -5.60767412e-01
-4.62368608e-01 1.62182415e+00 -4.09919582e-02 -3.20588529e-01
8.43875468e-01 1.32729232e-01 1.41633824e-01 -9.53584313e-02
-4.04492587e-01 -1.26487017e-01 3.26936990e-01 6.38581753e-01
3.94569129e-01 5.02023883e-02 -4.76560444e-01 -4.70075645e-02
-6.76719993e-02 -2.62010753e-01 5.53054631e-01 8.36095273e-01
-2.79377699e-01 -1.03940153e+00 -3.52836311e-01 5.38542449e-01
-2.51786470e-01 1.60307586e-01 -2.04727456e-01 9.09926295e-01
-2.33224511e-01 8.60695601e-01 -1.43091112e-01 -1.15837388e-01
3.29784572e-01 2.84515202e-01 1.92691356e-01 -6.91756904e-01
-2.40743935e-01 2.45155796e-01 -1.85279265e-01 -4.55324262e-01
-3.73522103e-01 -8.42029214e-01 -8.84581625e-01 2.16221899e-01
-6.49505079e-01 5.93677521e-01 9.20982599e-01 9.91156757e-01
-7.75418952e-02 6.18222281e-02 9.65919733e-01 -1.07756519e+00
-9.84600484e-01 -7.29561269e-01 -1.03269184e+00 2.28669256e-01
2.05117106e-01 -6.25638545e-01 -6.78114772e-01 -1.64540499e-01] | [7.823635578155518, 4.101921558380127] |
337caf5e-c329-47c6-b459-25168005dfff | multi-modal-egocentric-activity-recognition | 1807.00612 | null | https://arxiv.org/abs/1807.00612v3 | https://arxiv.org/pdf/1807.00612v3.pdf | Multi-modal Egocentric Activity Recognition using Audio-Visual Features | Egocentric activity recognition in first-person videos has an increasing importance with a variety of applications such as lifelogging, summarization, assisted-living and activity tracking. Existing methods for this task are based on interpretation of various sensor information using pre-determined weights for each feature. In this work, we propose a new framework for egocentric activity recognition problem based on combining audio-visual features with multi-kernel learning (MKL) and multi-kernel boosting (MKBoost). For that purpose, firstly grid optical-flow, virtual-inertia feature, log-covariance, cuboid are extracted from the video. The audio signal is characterized using a "supervector", obtained based on Gaussian mixture modelling of frame-level features, followed by a maximum a-posteriori adaptation. Then, the extracted multi-modal features are adaptively fused by MKL classifiers in which both the feature and kernel selection/weighing and recognition tasks are performed together. The proposed framework was evaluated on a number of egocentric datasets. The results showed that using multi-modal features with MKL outperforms the existing methods. | ['Alptekin Temizel', 'Peter Jančovič', 'Fatih Özkan', 'Mehmet Ali Arabaci', 'Elif Surer'] | 2018-07-02 | null | null | null | null | ['egocentric-activity-recognition'] | ['computer-vision'] | [ 3.87889892e-02 -5.50571978e-01 -8.87808483e-03 -2.14062199e-01
-8.57091188e-01 -1.46009997e-01 7.39253819e-01 1.82420149e-01
-6.62068665e-01 6.51135147e-01 7.96074867e-01 6.37750983e-01
-4.74766254e-01 -2.27682605e-01 -3.16582412e-01 -1.11970615e+00
-2.24266067e-01 -1.34183735e-01 2.35127985e-01 1.25928074e-01
4.52744961e-01 3.21829408e-01 -2.01712322e+00 3.36713493e-01
5.92204630e-01 1.02598214e+00 1.83707923e-01 1.09366620e+00
2.72476166e-01 9.57759082e-01 -3.08086604e-01 -1.08647421e-02
-9.62118879e-02 -3.25471103e-01 -4.85854715e-01 2.62228340e-01
6.41168505e-02 -1.91982627e-01 4.39635254e-02 5.73994815e-01
7.50939608e-01 6.98100507e-01 9.32517827e-01 -1.34483683e+00
2.54431367e-01 -8.12616795e-02 -5.35415590e-01 4.96457905e-01
8.65598321e-01 -1.32772595e-01 7.28290439e-01 -9.72458363e-01
3.18731576e-01 1.21043110e+00 6.05336905e-01 1.29539341e-01
-9.24051702e-01 -4.23538446e-01 4.14969027e-02 1.04261124e+00
-1.41404760e+00 -6.09054685e-01 1.04917014e+00 -7.40663230e-01
7.80732870e-01 1.19203083e-01 1.09917307e+00 9.76457596e-01
2.81849116e-01 8.77685666e-01 8.65630388e-01 -5.17669261e-01
4.01755542e-01 2.63116181e-01 -5.47887832e-02 5.49489379e-01
-5.50010987e-02 -4.23649281e-01 -8.25743198e-01 -4.51376587e-01
3.14056158e-01 5.79349175e-02 -2.27319479e-01 -8.01099539e-01
-1.36348689e+00 6.64616942e-01 -1.96573377e-01 3.77380103e-01
-7.82671809e-01 1.60207316e-01 6.89637065e-01 -5.17580472e-02
6.35845542e-01 -3.40446293e-01 -2.59923875e-01 -5.70157409e-01
-8.38639200e-01 1.86791390e-01 6.08133376e-01 5.15086353e-01
7.35767007e-01 -6.10966049e-02 -2.53345698e-01 8.54945064e-01
5.83221912e-01 3.12062651e-01 9.57910478e-01 -8.78752828e-01
1.93220362e-01 5.03451347e-01 1.04827888e-01 -1.02673423e+00
-6.20763838e-01 -2.51448691e-01 -6.53546929e-01 1.71387747e-01
3.23952675e-01 -2.05959573e-01 -4.61127013e-02 1.55420899e+00
9.00441349e-01 6.82471752e-01 1.98775753e-01 8.60624433e-01
4.32113290e-01 6.19872093e-01 2.17941031e-01 -5.41965365e-01
1.50035334e+00 -7.94995666e-01 -8.76051664e-01 2.82536417e-01
7.12907374e-01 -8.12962651e-01 5.21515608e-01 7.05770016e-01
-7.09124029e-01 -7.98295796e-01 -1.03315842e+00 4.24322516e-01
-3.05409551e-01 3.58263582e-01 3.57988834e-01 8.38134587e-01
-7.75665998e-01 3.83577645e-01 -7.40607858e-01 -6.58472419e-01
8.68619531e-02 1.68633714e-01 -6.01048768e-01 3.65518987e-01
-9.32295799e-01 7.25339055e-01 6.88707292e-01 -9.79536772e-02
-8.59691501e-01 -1.88137233e-01 -9.31351125e-01 -9.41084027e-02
9.19834524e-02 -6.49526596e-01 8.66393209e-01 -1.11688447e+00
-1.78355443e+00 5.25849164e-01 -2.13568166e-01 -3.93191755e-01
3.91652137e-01 -5.53620994e-01 -5.16069412e-01 4.83917773e-01
-1.30675301e-01 5.38955480e-02 1.28781331e+00 -7.04730153e-01
-9.31163490e-01 -6.58437371e-01 -4.04010480e-03 8.18017542e-01
-6.21790707e-01 6.44366741e-02 -1.20686173e-01 -5.64743996e-01
-9.06060040e-02 -6.77894771e-01 2.34493181e-01 -3.18883866e-01
6.65375367e-02 -4.93553907e-01 8.63107681e-01 -8.11087251e-01
1.30987203e+00 -2.16682220e+00 3.96556139e-01 6.91149309e-02
-7.79508501e-02 2.06218585e-02 4.68796670e-01 3.60442281e-01
-4.36994294e-03 -7.85264254e-01 2.90085495e-01 -3.32924366e-01
-1.78026468e-01 -2.75060594e-01 3.38610351e-01 1.00947976e+00
-1.27629176e-01 2.22060934e-01 -9.92353439e-01 -1.03301108e+00
8.42439651e-01 7.27438450e-01 -6.60992622e-01 3.24582636e-01
4.26297843e-01 5.71383357e-01 -4.29191232e-01 4.50503170e-01
4.10332322e-01 2.59546787e-01 -1.02702543e-01 -3.50915492e-01
-1.77981168e-01 -4.49915320e-01 -1.74759102e+00 1.95815873e+00
-4.33463663e-01 4.56267029e-01 4.86335764e-03 -1.30205882e+00
6.43515766e-01 6.35625482e-01 1.09363055e+00 2.33755000e-02
2.01859146e-01 -9.47883353e-02 -2.37013862e-01 -9.73480523e-01
4.02995557e-01 3.49874645e-01 6.03245832e-02 3.53421539e-01
5.71389496e-01 3.82121325e-01 2.97140956e-01 8.95248540e-03
8.76927435e-01 5.78546226e-01 7.59500027e-01 -1.12381622e-01
1.43407130e+00 -4.20002609e-01 4.55974132e-01 3.82650167e-01
-4.94315892e-01 1.90341651e-01 3.77829783e-02 -1.83426902e-01
-6.89225614e-01 -8.83191228e-01 6.32508621e-02 1.34803760e+00
1.68528706e-02 -5.31910002e-01 -9.81060803e-01 -4.14541602e-01
-1.76507145e-01 4.90660876e-01 -4.59810108e-01 -3.44775707e-01
-3.72637123e-01 -6.29905105e-01 3.61232758e-01 2.86378771e-01
5.29108524e-01 -1.05063641e+00 -9.50046360e-01 4.40985024e-01
-2.72007167e-01 -9.85797763e-01 -2.45011643e-01 -1.28850222e-01
-8.94025624e-01 -1.10828888e+00 -8.96456897e-01 -3.88351411e-01
2.74421632e-01 3.75452250e-01 4.46048349e-01 -6.06541812e-01
-4.00870860e-01 1.18656218e+00 -4.31403905e-01 -2.36791119e-01
5.77571020e-02 -2.02752367e-01 5.02790153e-01 9.19217885e-01
4.76117879e-01 -8.18123877e-01 -7.69590616e-01 1.80748329e-01
-5.14717460e-01 -1.90190703e-01 5.14908433e-01 7.20146894e-01
2.52542496e-01 -1.33593753e-01 6.30215228e-01 -1.55774593e-01
5.82880020e-01 -5.80001712e-01 -1.37240142e-01 2.17680082e-01
-1.64406985e-01 -1.50704980e-01 3.66077602e-01 -6.95752800e-01
-1.33723605e+00 2.10432991e-01 1.43704966e-01 -4.45424736e-01
-3.67704868e-01 2.26526663e-01 -3.59861523e-01 1.05917811e-01
6.33095801e-01 4.89534885e-01 -1.15559518e-01 -2.93863326e-01
4.98346657e-01 8.75043988e-01 3.62103492e-01 -4.82927799e-01
4.00816083e-01 6.19852781e-01 -3.16653624e-02 -1.41933656e+00
-4.24236804e-01 -1.07700205e+00 -7.93991208e-01 -9.53643382e-01
1.13699675e+00 -1.17794442e+00 -1.02542806e+00 8.00189734e-01
-8.63676906e-01 3.61075610e-01 -1.46387815e-01 1.32910478e+00
-1.01128721e+00 7.39220023e-01 -9.56419110e-02 -1.44843102e+00
-3.83924872e-01 -8.61475468e-01 1.07198763e+00 3.59695017e-01
-2.36868396e-01 -9.91572976e-01 4.02951837e-01 7.54980683e-01
1.19475715e-01 2.20585361e-01 1.76903471e-01 -5.93969882e-01
-9.92360860e-02 -3.65097493e-01 2.06727788e-01 4.60745782e-01
2.09295645e-01 -2.88053125e-01 -1.20122027e+00 -1.85905188e-01
1.16220430e-01 -3.22230101e-01 4.86999869e-01 6.88606858e-01
8.99401188e-01 -9.65371877e-02 -2.08913386e-01 4.59528953e-01
1.05087698e+00 1.55788913e-01 3.88085216e-01 3.31929773e-01
6.21222198e-01 4.43783700e-01 9.06864285e-01 1.17952478e+00
4.16226745e-01 7.61887193e-01 3.65703613e-01 4.51708049e-01
3.38472784e-01 2.21787974e-01 6.72414958e-01 7.58424819e-01
-5.15583038e-01 1.38430037e-02 -3.06342661e-01 4.99168158e-01
-2.21149421e+00 -1.39436972e+00 2.07348898e-01 2.32795119e+00
2.29380101e-01 2.31660474e-02 5.24540305e-01 5.50242841e-01
9.05659497e-01 2.50827730e-01 -3.59862447e-01 -1.27914295e-01
9.44565162e-02 -2.33654663e-01 2.51943916e-01 2.96251953e-01
-1.51121318e+00 4.09428507e-01 4.64300776e+00 9.85734999e-01
-6.71204507e-01 3.98769200e-01 -6.02969639e-02 -1.56860381e-01
4.06888813e-01 -4.68074642e-02 -7.14050174e-01 5.89803517e-01
8.52807701e-01 -1.83815181e-01 2.40518779e-01 1.02426660e+00
5.74535310e-01 -6.30715430e-01 -7.06148207e-01 1.55801666e+00
5.84315658e-01 -7.37653673e-01 -2.86462158e-01 -3.25056240e-02
3.67254287e-01 -3.25075328e-01 -4.23162997e-01 2.12183908e-01
-3.52486402e-01 -1.73394397e-01 5.92021823e-01 1.02821684e+00
1.41797796e-01 -9.03476059e-01 5.60867190e-01 4.87319946e-01
-1.52254808e+00 -3.60802233e-01 -1.18193902e-01 -7.87234604e-02
2.52897233e-01 4.77886796e-01 -5.23116887e-01 6.55567527e-01
7.08582282e-01 9.33462799e-01 -2.58480072e-01 1.32065260e+00
1.36828676e-01 4.12769467e-01 -3.44573766e-01 -1.89846754e-01
-1.18248582e-01 -3.41601610e-01 7.71690667e-01 1.45472515e+00
4.69519675e-01 -1.23278148e-01 2.07995400e-01 1.05835594e-01
5.13756752e-01 6.44401908e-01 -5.42288661e-01 3.17360908e-01
3.50650661e-02 1.55256140e+00 -5.76964796e-01 -5.75291336e-01
-5.22544503e-01 1.10365164e+00 1.23098731e-01 1.79115042e-01
-8.71459126e-01 -3.29088688e-01 6.56630397e-01 -1.55417293e-01
1.92680985e-01 -2.56863415e-01 5.33696473e-01 -1.44617295e+00
-3.96568887e-02 -3.39020938e-01 7.11798608e-01 -9.13907588e-01
-6.44476175e-01 1.17719308e-01 3.62987489e-01 -1.67620516e+00
-6.68829560e-01 -3.41045886e-01 -5.15550137e-01 4.95661825e-01
-1.14205456e+00 -1.25394320e+00 -5.60550570e-01 1.07408059e+00
8.17220926e-01 -3.63063574e-01 6.93531096e-01 5.51677048e-01
-5.15425742e-01 1.27267510e-01 2.18579531e-01 -1.60026222e-01
8.97737384e-01 -1.09139359e+00 -6.08757913e-01 6.85412765e-01
3.76491733e-02 1.51706114e-01 6.68836594e-01 -4.26985115e-01
-1.29191303e+00 -7.19139397e-01 5.98429859e-01 -9.42173526e-02
5.27563691e-01 -1.27169639e-01 -3.50765616e-01 4.43224579e-01
2.04153471e-02 7.34680817e-02 9.07380939e-01 -1.29134566e-01
1.61668584e-01 -4.38495815e-01 -1.11902595e+00 2.39276320e-01
8.20052624e-01 -4.67731088e-01 -5.76594293e-01 3.27591419e-01
-1.04749814e-01 -3.70251872e-02 -1.01750982e+00 1.64887786e-01
1.11970532e+00 -1.18170822e+00 1.07417512e+00 -3.96077126e-01
-3.10408652e-01 -4.98312473e-01 -2.66937405e-01 -1.28119230e+00
-4.08384800e-01 -4.16844487e-01 -4.89422530e-01 1.17043912e+00
-4.79752779e-01 -4.25437480e-01 6.89857304e-01 1.27244517e-01
-3.31670605e-02 -2.44323418e-01 -1.18295181e+00 -5.11602640e-01
-9.77805197e-01 -5.84556758e-01 3.72810811e-02 6.39405191e-01
2.82204658e-01 2.88089663e-01 -8.01798940e-01 -1.20949168e-02
9.33511257e-01 -3.59890163e-01 1.08853042e+00 -1.34126818e+00
-3.39217544e-01 -1.39462531e-01 -1.15996897e+00 -6.46060288e-01
1.70235127e-01 -5.19548595e-01 -3.09161156e-01 -1.16976452e+00
2.22663894e-01 2.82802165e-01 -6.72197342e-01 -1.37857556e-01
-1.17690697e-01 1.46806940e-01 -9.79500636e-03 5.01983240e-02
-9.54742372e-01 7.17653692e-01 7.34066069e-01 -7.42315948e-02
-4.13785517e-01 3.19617957e-01 -1.10674910e-01 9.64251280e-01
6.08775258e-01 -2.67085910e-01 -5.29099643e-01 3.08743954e-01
-1.42403349e-01 1.95122182e-01 4.42278773e-01 -1.64505005e+00
2.94519037e-01 -2.73841992e-02 6.27882421e-01 -7.54694760e-01
9.00853455e-01 -8.64105225e-01 1.39352158e-01 4.58901823e-01
-1.26172498e-01 -2.29094699e-01 -2.84875453e-01 8.04616034e-01
-2.29309782e-01 -2.69116879e-01 6.48955405e-01 -1.06756195e-01
-1.04444361e+00 -9.02655870e-02 -8.23042214e-01 -2.54897386e-01
1.38238025e+00 -6.42075837e-01 2.78858542e-01 -6.11181676e-01
-9.70363200e-01 -6.03327863e-02 -5.98608740e-02 3.71918470e-01
6.04378760e-01 -1.48423862e+00 -6.18163645e-01 1.13061123e-01
3.90213072e-01 -4.85207409e-01 7.53375232e-01 1.27911651e+00
-2.29971603e-01 2.22392142e-01 -4.26808208e-01 -7.80879259e-01
-1.57805777e+00 4.13393438e-01 1.78352401e-01 -1.92309365e-01
-3.03874373e-01 4.34009582e-01 -5.65973334e-02 1.36671625e-02
2.67884940e-01 -8.75588804e-02 -9.35766399e-01 6.44615531e-01
6.10899091e-01 1.04996097e+00 -6.79508522e-02 -1.22998667e+00
-4.90407795e-01 9.28179622e-01 3.08447987e-01 -3.25267583e-01
1.09610891e+00 -3.56659442e-01 1.98878825e-01 7.17357993e-01
1.07359767e+00 -1.07089795e-01 -1.06799054e+00 -2.59850383e-01
-5.40907495e-02 -4.27548259e-01 5.97143508e-02 2.87586404e-03
-5.27049899e-01 7.54288256e-01 1.05912089e+00 -1.81504842e-02
1.15792382e+00 -3.11057448e-01 3.92827630e-01 4.26438779e-01
3.32580417e-01 -1.51369393e+00 3.76762450e-01 1.01183817e-01
6.37564719e-01 -1.00776756e+00 3.10239941e-01 8.66713002e-02
-6.37682915e-01 1.11524546e+00 4.39904630e-01 -2.41131723e-01
1.00747025e+00 -2.70121008e-01 -2.25761145e-01 1.64600998e-01
-5.44919789e-01 -4.23497170e-01 3.00978899e-01 7.13329852e-01
2.76712656e-01 -1.15722761e-01 -4.95133191e-01 4.64871526e-01
1.74073055e-01 2.27417886e-01 2.94352651e-01 1.02931392e+00
-6.82405889e-01 -5.82051754e-01 -8.30655992e-01 3.58681113e-01
-4.47426558e-01 5.03037751e-01 2.51265168e-01 3.69238973e-01
3.19649577e-01 1.13494837e+00 -9.54444185e-02 -4.32594061e-01
1.80995911e-01 3.90797257e-01 6.15907967e-01 -1.56224310e-01
-4.11385864e-01 4.10457313e-01 8.07441026e-02 -6.34547770e-01
-1.10299742e+00 -1.06625748e+00 -7.08469748e-01 1.89548671e-01
-4.59919721e-01 2.78819293e-01 8.95959437e-01 1.06966460e+00
1.91824630e-01 3.54451299e-01 7.24455833e-01 -1.35397780e+00
-3.90758485e-01 -1.18105090e+00 -7.86931515e-01 6.66305661e-01
1.78086936e-01 -1.03508198e+00 -5.89738309e-01 3.30158234e-01] | [7.976809978485107, 0.375964492559433] |
7acb243b-99de-4cbd-92e0-bc7c1f281ff7 | neural-best-buddies-sparse-cross-domain | 1805.04140 | null | http://arxiv.org/abs/1805.04140v2 | http://arxiv.org/pdf/1805.04140v2.pdf | Neural Best-Buddies: Sparse Cross-Domain Correspondence | Correspondence between images is a fundamental problem in computer vision,
with a variety of graphics applications. This paper presents a novel method for
sparse cross-domain correspondence. Our method is designed for pairs of images
where the main objects of interest may belong to different semantic categories
and differ drastically in shape and appearance, yet still contain semantically
related or geometrically similar parts. Our approach operates on hierarchies of
deep features, extracted from the input images by a pre-trained CNN.
Specifically, starting from the coarsest layer in both hierarchies, we search
for Neural Best Buddies (NBB): pairs of neurons that are mutual nearest
neighbors. The key idea is then to percolate NBBs through the hierarchy, while
narrowing down the search regions at each level and retaining only NBBs with
significant activations. Furthermore, in order to overcome differences in
appearance, each pair of search regions is transformed into a common
appearance. We evaluate our method via a user study, in addition to comparisons
with alternative correspondence approaches. The usefulness of our method is
demonstrated using a variety of graphics applications, including cross-domain
image alignment, creation of hybrid images, automatic image morphing, and more. | ['Daniel Cohen-Or', 'Mingyi Shi', 'Jing Liao', 'Kfir Aberman', 'Dani Lischinski', 'Baoquan Chen'] | 2018-05-10 | null | null | null | null | ['image-morphing'] | ['computer-vision'] | [ 3.00448537e-01 -3.63499187e-02 2.12145343e-01 -3.67232144e-01
-3.85797888e-01 -4.56508577e-01 5.90796828e-01 5.36223590e-01
-2.94906348e-01 3.99326771e-01 1.50049962e-02 2.34445766e-01
-1.10741839e-01 -8.54600549e-01 -7.80402720e-01 -5.39391160e-01
1.55343086e-01 4.59057182e-01 7.22049952e-01 -4.41232294e-01
4.13645834e-01 9.23626125e-01 -1.69193828e+00 5.00405848e-01
6.45094812e-01 9.90958095e-01 3.11945826e-01 1.98387489e-01
-1.99559152e-01 1.42497629e-01 -3.34642380e-01 -4.69840109e-01
5.68762541e-01 -5.53412199e-01 -1.05774844e+00 2.18276381e-01
1.06278443e+00 2.28796806e-02 1.32421963e-02 1.32570624e+00
2.41392359e-01 3.12615216e-01 4.88621682e-01 -1.13987553e+00
-5.96666574e-01 2.07443103e-01 -8.17575932e-01 8.87708664e-02
2.81121433e-01 -1.63121551e-01 9.48582232e-01 -8.71993601e-01
6.37815595e-01 1.35105610e+00 8.38549972e-01 5.66539586e-01
-1.67500281e+00 -5.20349622e-01 8.85267258e-02 3.30676474e-02
-1.50619590e+00 -3.50468457e-01 9.58320081e-01 -4.18762594e-01
7.53263831e-01 2.75339961e-01 8.02598119e-01 4.83604729e-01
1.04491465e-01 4.52865809e-01 1.09276867e+00 -5.61442256e-01
1.88533962e-01 2.56149411e-01 -3.16448063e-02 8.57099175e-01
-2.81756129e-02 -8.96284655e-02 -5.27239859e-01 -2.38377929e-01
1.08658493e+00 -1.30399734e-01 -3.46424103e-01 -9.83411551e-01
-1.24394000e+00 6.60268366e-01 7.78331399e-01 6.72932327e-01
-2.38127917e-01 3.70594151e-02 1.55340463e-01 1.26054913e-01
2.85148084e-01 7.64523447e-01 -2.62718856e-01 4.66386706e-01
-9.77519870e-01 2.30357647e-01 6.07763469e-01 7.89602697e-01
1.22434413e+00 -4.03720826e-01 1.66126713e-02 1.10875344e+00
5.30765317e-02 -2.94900715e-01 5.43653429e-01 -9.48837399e-01
2.50065297e-01 7.95369387e-01 -6.52825087e-02 -1.44906354e+00
-2.31903031e-01 -2.54226148e-01 -9.55691516e-01 4.94543076e-01
3.63638431e-01 5.14080822e-01 -8.82918537e-01 1.69795334e+00
4.07989711e-01 3.78504358e-02 -8.72548148e-02 6.77831471e-01
7.14760721e-01 5.25431871e-01 -5.23136519e-02 1.93224609e-01
1.35195017e+00 -1.07301116e+00 -4.36029322e-02 -2.30880052e-01
3.31198931e-01 -8.93991828e-01 1.18574131e+00 1.59695297e-01
-1.45335639e+00 -8.61024201e-01 -8.64483297e-01 -1.65944368e-01
-5.61548948e-01 -1.04197010e-01 2.75451988e-01 3.72364640e-01
-1.48243594e+00 9.14206326e-01 -3.97832930e-01 -7.07406580e-01
4.18842822e-01 4.17224944e-01 -6.14173532e-01 1.35553882e-01
-9.11632299e-01 8.22363913e-01 3.93907934e-01 -7.69275054e-02
-3.28266025e-01 -5.71079373e-01 -9.59329128e-01 1.46502465e-01
-2.00016156e-01 -8.76337647e-01 9.77553725e-01 -1.46445537e+00
-1.15532708e+00 1.39040399e+00 -8.31771493e-02 -2.29216442e-01
4.10685956e-01 6.31154105e-02 -1.89585954e-01 4.26779054e-02
3.01879436e-01 1.19645929e+00 8.74782562e-01 -1.55407429e+00
-8.21798801e-01 -5.74777544e-01 3.76647152e-02 4.36109781e-01
-3.74249279e-01 1.62203237e-02 -8.81031573e-01 -7.72587955e-01
5.32259107e-01 -8.56218934e-01 -3.50598842e-01 2.97517836e-01
-1.50465533e-01 -1.30998388e-01 7.81857550e-01 -4.88898635e-01
8.43651354e-01 -2.29635954e+00 3.03715467e-01 4.97325063e-01
1.94293708e-01 -2.98570730e-02 -2.51942784e-01 1.17140703e-01
-2.83558279e-01 7.54034072e-02 -3.98980498e-01 -2.70723641e-01
-3.97876620e-01 8.74034017e-02 1.62995644e-02 2.50679761e-01
2.49416724e-01 7.80152678e-01 -8.98232818e-01 -6.14557445e-01
1.70041844e-01 1.79298177e-01 -5.68107843e-01 1.84921056e-01
8.66069347e-02 5.91264367e-01 -1.74408257e-01 3.46306175e-01
7.23301172e-01 -1.56825170e-01 -1.00639015e-01 -6.43222392e-01
-2.22182497e-01 1.50634602e-01 -1.19729114e+00 1.79108715e+00
-4.68415767e-01 6.78308666e-01 4.57504950e-02 -1.16177392e+00
1.06744075e+00 -1.69623688e-01 4.22172606e-01 -8.66609156e-01
5.70323803e-02 1.64935276e-01 -8.50016400e-02 -2.30951265e-01
6.34804189e-01 -3.51592004e-02 1.57742668e-03 3.64761114e-01
-1.09389694e-02 -5.57132185e-01 2.08041146e-01 -1.88349281e-02
5.67911088e-01 7.23683182e-03 5.47907829e-01 -3.76776606e-01
5.70835352e-01 8.12000409e-02 2.98756063e-01 6.25999868e-01
-6.76083416e-02 9.81782556e-01 2.73974448e-01 -7.35420704e-01
-1.19308972e+00 -1.07125580e+00 -1.42605469e-01 8.86094213e-01
6.03975475e-01 -1.21237375e-01 -9.22886014e-01 -5.24971962e-01
4.40901779e-02 3.23209167e-01 -8.53937209e-01 -1.59355596e-01
-8.52840066e-01 -2.73050606e-01 2.34010741e-01 4.83686984e-01
6.47979021e-01 -1.24240375e+00 -9.06102300e-01 1.75757691e-01
2.22617760e-02 -8.36917639e-01 -6.58097029e-01 1.06889695e-01
-1.08913088e+00 -8.85681391e-01 -8.61766934e-01 -1.32397461e+00
9.45779562e-01 4.83325988e-01 1.34121156e+00 2.40403637e-01
-5.27697802e-01 2.35338077e-01 -1.45554319e-01 1.79520145e-01
-4.99998868e-01 -4.70095314e-02 -2.41513312e-01 1.87154353e-01
-7.18701514e-04 -6.18151724e-01 -6.83981538e-01 5.28174698e-01
-8.94483745e-01 2.43817359e-01 5.05068600e-01 7.74217129e-01
8.15093875e-01 2.34169886e-02 -2.81657968e-02 -7.30202973e-01
7.20573425e-01 -1.47953987e-01 -6.75137937e-01 3.30776274e-01
-2.00557768e-01 1.79768428e-01 4.99106497e-01 -3.25686008e-01
-7.35688806e-01 2.26491407e-01 8.79746750e-02 -5.26432216e-01
-3.80248040e-01 1.08594358e-01 -1.87319189e-01 -3.52471888e-01
7.18036652e-01 1.86878592e-01 7.82710090e-02 -4.12991762e-01
4.23579544e-01 1.94293797e-01 7.78298616e-01 -4.96026874e-01
6.07687831e-01 5.43991208e-01 -1.74296871e-01 -6.08052611e-01
-4.61708158e-01 -4.08921361e-01 -1.09780157e+00 -4.17039730e-02
8.08060467e-01 -5.07168055e-01 -2.69310325e-01 4.71289396e-01
-1.43938839e+00 -1.72573313e-01 -3.69610667e-01 1.67063653e-01
-6.38458788e-01 2.90923774e-01 -3.06442589e-01 -1.96922630e-01
-9.59479362e-02 -1.21783447e+00 1.00622797e+00 3.95945728e-01
-4.08702224e-01 -9.67426300e-01 1.20863207e-01 -2.36671492e-02
2.20426053e-01 1.35041118e-01 1.36962533e+00 -5.22190213e-01
-3.29582155e-01 -3.49580832e-02 -4.44742948e-01 2.10683495e-01
4.40373033e-01 -7.28920028e-02 -7.75759220e-01 -4.06070113e-01
-2.09733680e-01 -1.39597684e-01 7.57329226e-01 3.35545421e-01
1.30244100e+00 -1.51905388e-01 -6.22276306e-01 7.01819539e-01
1.42580497e+00 4.00919229e-01 6.44121647e-01 4.97127831e-01
6.94767535e-01 1.09132028e+00 2.60969669e-01 5.98173812e-02
3.30323651e-02 1.00416839e+00 4.90652680e-01 -6.22492969e-01
-1.75107270e-01 -1.22026257e-01 -1.81534290e-01 3.88981432e-01
9.52613875e-02 7.55819380e-02 -8.56823325e-01 6.78097486e-01
-1.70526683e+00 -7.99100816e-01 -2.37053651e-02 2.50104809e+00
7.40202487e-01 5.28845377e-02 1.00142874e-01 -1.08912721e-01
1.10702634e+00 1.91966705e-02 -4.93151248e-01 -5.45412540e-01
-7.12396875e-02 4.48687732e-01 1.64554924e-01 3.07789892e-01
-1.04793620e+00 1.01866937e+00 6.27279377e+00 7.03832984e-01
-1.14186120e+00 -4.55894619e-02 9.11249518e-01 2.59190530e-01
-1.99930817e-01 2.09472366e-02 -6.20736897e-01 1.64961249e-01
5.52843027e-02 1.24947898e-01 2.71549195e-01 5.88279545e-01
-1.39122963e-01 -2.18213707e-01 -1.34785438e+00 1.06180108e+00
1.69948205e-01 -1.53263402e+00 2.58129656e-01 -1.34361222e-01
9.66558814e-01 -2.54658103e-01 1.35034293e-01 -2.17761040e-01
2.27466419e-01 -1.02305007e+00 8.80237341e-01 2.93616027e-01
6.72912955e-01 -7.58732021e-01 3.80360872e-01 6.79841563e-02
-1.38062716e+00 1.61066681e-01 -3.72417271e-01 3.71307373e-01
-6.65390417e-02 2.64950097e-01 -4.95424539e-01 4.12318647e-01
1.07930481e+00 5.04271388e-01 -6.05411768e-01 1.22436643e+00
2.20005825e-01 -2.63503045e-01 -1.37411445e-01 7.85036385e-02
3.27768266e-01 -3.28969866e-01 2.55918443e-01 1.14177728e+00
3.61769974e-01 -1.83692891e-02 8.08522776e-02 1.22268653e+00
-4.54404205e-02 2.11050525e-01 -7.42505789e-01 5.10227799e-01
4.81196791e-01 1.24698031e+00 -1.12329948e+00 -3.55771959e-01
-4.51042354e-01 1.43619406e+00 5.79129755e-01 7.24642798e-02
-6.95306957e-01 -4.77608234e-01 5.50103605e-01 2.49568582e-01
3.09266716e-01 -9.00053233e-02 -2.30353981e-01 -7.41359651e-01
-1.14041297e-02 -9.54769611e-01 3.26858461e-01 -8.36811483e-01
-1.29334688e+00 1.05154812e+00 -2.10997202e-02 -1.48189270e+00
1.68987233e-02 -3.67223859e-01 -6.77311182e-01 1.12757421e+00
-1.15584385e+00 -1.04201460e+00 -6.16953313e-01 7.01509953e-01
6.54821754e-01 -9.16669145e-02 7.15305388e-01 1.96267694e-01
-2.17017651e-01 4.98930633e-01 9.10089612e-02 4.88582358e-04
5.76489568e-01 -1.02367496e+00 7.30578184e-01 5.27180135e-01
3.82261038e-01 6.82708502e-01 4.73517776e-01 -4.75344837e-01
-5.19958973e-01 -1.07305777e+00 7.90589273e-01 -1.37992464e-02
3.35475892e-01 -3.05353433e-01 -1.17165613e+00 1.82920605e-01
3.29351962e-01 6.71826955e-03 3.10277492e-01 -1.03099503e-01
-4.85258132e-01 -1.14341527e-01 -1.07529521e+00 8.88605177e-01
1.02476776e+00 -4.53596503e-01 -6.35038197e-01 1.02997452e-01
2.98169315e-01 -3.83452654e-01 -6.09623492e-01 2.90386438e-01
7.06173539e-01 -1.37604320e+00 1.09227514e+00 -6.31679416e-01
4.53089833e-01 -4.15540963e-01 -5.42647094e-02 -1.46050763e+00
-5.09233892e-01 -4.47463453e-01 6.58119857e-01 1.08835971e+00
3.39201301e-01 -3.73510569e-01 9.32535231e-01 5.05488217e-01
6.04007998e-03 -6.82438672e-01 -8.26910853e-01 -7.77706027e-01
-1.91166892e-03 1.37393251e-01 5.98628342e-01 1.03950882e+00
-2.40709171e-01 2.54248589e-01 1.09104943e-02 5.94502613e-02
4.13590610e-01 5.06177485e-01 8.15099359e-01 -1.40324950e+00
-1.63474023e-01 -8.96561503e-01 -5.33267736e-01 -1.08633173e+00
9.81132090e-02 -8.84296536e-01 1.44520953e-01 -1.37157261e+00
2.90262043e-01 -7.60101080e-01 -1.58166885e-01 4.79338825e-01
3.14980149e-02 6.52252078e-01 1.66836441e-01 3.67842317e-01
-1.56648904e-01 4.06858653e-01 1.16006672e+00 -2.17256740e-01
-4.44495142e-01 1.14972433e-02 -5.01324117e-01 9.13743377e-01
7.64510632e-01 -4.33213145e-01 -2.53052950e-01 -5.45770884e-01
-7.27662370e-02 -1.79758459e-01 5.93428493e-01 -1.06319141e+00
2.31797412e-01 -1.00389242e-01 4.55855906e-01 -5.24499416e-01
3.69234681e-01 -8.71470034e-01 3.58933151e-01 5.96880853e-01
-4.90884066e-01 5.06281793e-01 3.63322347e-01 2.96886832e-01
-4.31364894e-01 -4.59983230e-01 1.30838859e+00 -4.02256072e-01
-8.54892731e-01 1.78988650e-01 2.80869342e-02 -1.50236592e-01
1.17117345e+00 -8.78860533e-01 1.41657770e-01 -1.34309307e-01
-7.71551013e-01 -1.49090022e-01 8.35430384e-01 4.57112134e-01
8.35368276e-01 -1.45603943e+00 -4.67351973e-01 4.61674273e-01
3.78558248e-01 6.99083731e-02 7.52546564e-02 5.19156694e-01
-5.91613173e-01 1.69416815e-01 -7.42330372e-01 -7.21184850e-01
-1.45690000e+00 4.16403085e-01 5.98533750e-01 -7.49343485e-02
-5.37087262e-01 1.02877140e+00 7.61564016e-01 -2.44954467e-01
1.21521421e-01 -3.14724803e-01 -3.60201716e-01 4.65684682e-02
2.03600198e-01 9.27595496e-02 3.20820451e-01 -9.35255527e-01
-2.31337219e-01 1.05661285e+00 -2.20067561e-01 -1.56845719e-01
1.22415686e+00 -9.05712098e-02 -4.61065888e-01 2.55838186e-01
1.42395604e+00 -4.43577282e-02 -1.15206075e+00 -3.53145719e-01
4.81346324e-02 -8.06473970e-01 -2.43877292e-01 -2.90269643e-01
-1.33078218e+00 7.68156469e-01 7.66659617e-01 2.36178577e-01
1.26943159e+00 3.09139848e-01 5.96906066e-01 -2.05149669e-02
3.46320927e-01 -8.47361147e-01 3.01255673e-01 3.38718861e-01
1.13265967e+00 -1.06038558e+00 -1.67692199e-01 -4.15981352e-01
-3.91977549e-01 1.19281757e+00 9.60806251e-01 -2.16798916e-01
3.99486184e-01 6.71118312e-03 -8.82051215e-02 -3.83207560e-01
-2.80007869e-01 -1.46071106e-01 4.53233659e-01 6.59985662e-01
4.23215479e-01 -2.55290031e-01 -3.18904608e-01 -2.55422760e-02
-1.27884343e-01 -6.02056444e-01 2.28490978e-01 7.91208386e-01
-4.11851764e-01 -1.18962049e+00 -4.13775742e-01 2.30188325e-01
4.71642762e-02 -2.61438906e-01 -6.66204810e-01 9.22315657e-01
4.00060982e-01 4.06459540e-01 6.00369215e-01 -1.96947917e-01
5.72090566e-01 -2.96856970e-01 6.06462419e-01 -7.16987610e-01
-7.59663343e-01 3.81129817e-03 -3.22623879e-01 -5.66709757e-01
-5.06486177e-01 -5.93432605e-01 -1.03434932e+00 -3.25412825e-02
-2.62506902e-01 -2.94951745e-03 5.31429052e-01 5.97244382e-01
2.83633471e-01 2.38048732e-01 7.18415499e-01 -1.07635200e+00
8.71128775e-03 -4.84992921e-01 -4.31306303e-01 8.09253514e-01
1.55929089e-01 -6.35727048e-01 3.67812701e-02 3.20138395e-01] | [8.430933952331543, -1.9154328107833862] |
550562f2-25a7-45d5-bc65-474ff9c6af9c | robust-semi-supervised-learning-for | 2303.09930 | null | https://arxiv.org/abs/2303.09930v1 | https://arxiv.org/pdf/2303.09930v1.pdf | Robust Semi-Supervised Learning for Histopathology Images through Self-Supervision Guided Out-of-Distribution Scoring | Semi-supervised learning (semi-SL) is a promising alternative to supervised learning for medical image analysis when obtaining good quality supervision for medical imaging is difficult. However, semi-SL assumes that the underlying distribution of unaudited data matches that of the few labeled samples, which is often violated in practical settings, particularly in medical images. The presence of out-of-distribution (OOD) samples in the unlabeled training pool of semi-SL is inevitable and can reduce the efficiency of the algorithm. Common preprocessing methods to filter out outlier samples may not be suitable for medical images that involve a wide range of anatomical structures and rare morphologies. In this paper, we propose a novel pipeline for addressing open-set supervised learning challenges in digital histology images. Our pipeline efficiently estimates an OOD score for each unlabelled data point based on self-supervised learning to calibrate the knowledge needed for a subsequent semi-SL framework. The outlier score derived from the OOD detector is used to modulate sample selection for the subsequent semi-SL stage, ensuring that samples conforming to the distribution of the few labeled samples are more frequently exposed to the subsequent semi-SL framework. Our framework is compatible with any semi-SL framework, and we base our experiments on the popular Mixmatch semi-SL framework. We conduct extensive studies on two digital pathology datasets, Kather colorectal histology dataset and a dataset derived from TCGA-BRCA whole slide images, and establish the effectiveness of our method by comparing with popular methods and frameworks in semi-SL algorithms through various experiments. | ['Amit Sethi', 'Shashikant Khade', 'Abhijit PATIL', 'Varsha S', 'Nikhil Cherian Kurian'] | 2023-03-17 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 4.62894380e-01 2.13882148e-01 -4.89264816e-01 -5.28289199e-01
-1.31429207e+00 -4.78962868e-01 2.68379360e-01 5.39968967e-01
-5.29084086e-01 6.33459568e-01 -1.24429323e-01 -3.79344881e-01
-1.17164738e-01 -3.92736971e-01 -7.00915635e-01 -1.07395554e+00
1.71152145e-01 7.86297381e-01 3.21007818e-01 4.47168499e-01
-1.68732496e-03 4.89345223e-01 -1.12247968e+00 7.36219525e-01
6.60378695e-01 6.68805242e-01 1.07827365e-01 6.40205503e-01
-1.26133353e-01 6.95514143e-01 -5.68204820e-01 -1.22549169e-01
1.97152495e-01 -5.26935816e-01 -7.73208201e-01 1.65261403e-01
4.98017877e-01 -1.90589949e-01 1.40682563e-01 1.18180549e+00
7.44917274e-01 -5.12001395e-01 8.21968019e-01 -1.28951848e+00
1.17930360e-01 6.93280697e-01 -6.08786225e-01 5.14001667e-01
-6.31649345e-02 3.66259128e-01 7.87355483e-01 -8.99769425e-01
1.16026235e+00 7.47721195e-01 8.32157433e-01 4.49536324e-01
-1.35595572e+00 -5.94405591e-01 -1.80991605e-01 -1.63763031e-01
-1.25872719e+00 -3.77414584e-01 3.21268737e-01 -4.85736281e-01
4.28900689e-01 1.91653773e-01 3.83320868e-01 9.71091390e-01
4.40868437e-01 8.58978748e-01 1.05441344e+00 -3.53471279e-01
5.55062711e-01 2.83991605e-01 1.34221911e-01 6.89578176e-01
3.37720722e-01 -1.26060052e-02 -4.71731752e-01 -6.34097517e-01
3.14417243e-01 1.88242808e-01 -2.66870353e-02 -5.91029823e-01
-1.21147954e+00 5.50961792e-01 1.98548451e-01 7.72269964e-02
-1.49477363e-01 -3.67417008e-01 6.07465684e-01 2.61354297e-01
3.37560326e-01 1.65915310e-01 -3.79401773e-01 2.44581282e-01
-1.36556065e+00 -1.67342156e-01 6.13800108e-01 7.03020334e-01
5.38478255e-01 -6.51144028e-01 -2.34616145e-01 6.73585474e-01
4.80221570e-01 1.16194353e-01 7.78265178e-01 -5.85242510e-01
-1.24632791e-01 8.67621064e-01 -2.93577939e-01 -2.26915866e-01
-5.70124626e-01 -3.88938934e-01 -8.93879473e-01 3.30593675e-01
8.53367150e-01 1.09093353e-01 -1.12452388e+00 1.29075742e+00
7.69327641e-01 1.85549989e-01 -3.65388989e-02 7.19531953e-01
9.45342422e-01 3.85624766e-02 7.10615516e-02 -4.09177780e-01
1.24143732e+00 -5.29270768e-01 -5.94429493e-01 5.55652231e-02
1.29939413e+00 -6.14205658e-01 1.19222689e+00 4.33337301e-01
-5.63050032e-01 -1.02465726e-01 -9.76369381e-01 1.25373587e-01
-2.00107992e-01 2.67192066e-01 4.72664475e-01 5.74848592e-01
-8.03288698e-01 4.54391986e-01 -1.06411600e+00 -4.99495327e-01
1.09373856e+00 4.67686504e-01 -5.56381464e-01 -3.42890173e-01
-6.37037933e-01 4.85318989e-01 3.45829636e-01 1.24919839e-01
-1.04127085e+00 -1.03586185e+00 -7.31064856e-01 -4.19712394e-01
3.65197361e-01 -3.61580223e-01 1.02742422e+00 -7.95979679e-01
-1.00512755e+00 1.45863712e+00 -1.09040454e-01 -3.47332776e-01
8.03028703e-01 3.53264630e-01 -2.10622579e-01 2.84078181e-01
2.86477566e-01 5.89684725e-01 6.87625229e-01 -1.06543672e+00
-5.57559013e-01 -4.58838671e-01 -6.25565469e-01 -7.86732435e-02
1.04237244e-01 -1.37697473e-01 -2.86158025e-01 -5.24974465e-01
2.75455505e-01 -1.03409064e+00 -4.74500000e-01 4.61460829e-01
-5.98221242e-01 -2.97316276e-02 7.15749621e-01 -1.59712568e-01
9.83824313e-01 -2.33503842e+00 -4.10078943e-01 4.83209819e-01
2.88535476e-01 2.67261028e-01 1.03375979e-01 -1.01277985e-01
-2.99089283e-01 1.96752563e-01 -3.25808138e-01 -2.46206298e-01
-2.75697112e-01 2.40254804e-01 3.02123632e-02 9.79742229e-01
2.34521285e-01 6.06938362e-01 -1.07468331e+00 -9.70166028e-01
-8.86958241e-02 -5.46472743e-02 -5.55442512e-01 3.13914806e-01
-1.29537150e-01 6.27750456e-01 -7.61208907e-02 1.06132364e+00
6.63463414e-01 -4.06332135e-01 3.07986856e-01 -1.77106678e-01
3.38581711e-01 1.54651366e-02 -1.05739260e+00 1.55036938e+00
2.79606730e-02 2.34458238e-01 2.54663620e-02 -6.30467594e-01
5.82659245e-01 1.47154093e-01 5.83854735e-01 -2.28468478e-01
7.64175355e-02 3.95235360e-01 2.61211246e-01 -6.43850386e-01
-1.36899486e-01 -3.68667990e-01 1.46889493e-01 6.08601987e-01
3.02591681e-01 -1.23519637e-01 2.23184749e-01 3.98420572e-01
1.38003302e+00 -2.14977693e-02 5.36056697e-01 -3.59290421e-01
5.06024837e-01 1.51401401e-01 7.92675436e-01 8.49904120e-01
-7.70713389e-01 8.62850845e-01 9.60060358e-01 -3.15075547e-01
-8.48939538e-01 -1.23063719e+00 -6.98111951e-01 6.40531242e-01
-2.11234629e-01 -2.50490069e-01 -5.29828072e-01 -1.35619843e+00
-4.95019592e-02 2.36442447e-01 -7.60986030e-01 -5.45084998e-02
-1.47576064e-01 -1.09021378e+00 6.47396505e-01 9.10893977e-02
-1.84690118e-01 -1.02145302e+00 -4.26843941e-01 -6.80409446e-02
1.55734301e-01 -9.34267282e-01 -3.89310390e-01 6.34807587e-01
-9.73253191e-01 -1.55384207e+00 -3.79400522e-01 -7.64525950e-01
1.42493820e+00 -1.05315857e-01 9.63562429e-01 2.32646689e-01
-7.29058027e-01 -6.43369853e-02 -2.90719599e-01 -5.87696910e-01
-9.60345864e-01 1.46019891e-01 -8.45512971e-02 5.72735071e-02
5.59686303e-01 -1.00283109e-01 -5.15660524e-01 4.68993366e-01
-1.28226912e+00 -2.86996841e-01 7.61386991e-01 1.25365496e+00
1.27201891e+00 1.52947009e-01 5.22998810e-01 -1.66003454e+00
-7.45025203e-02 -5.38549185e-01 -4.91086453e-01 2.11948425e-01
-4.29703951e-01 -9.82730687e-02 5.89671671e-01 -4.78849947e-01
-7.54537642e-01 4.15737510e-01 -2.26874858e-01 -2.95525104e-01
-3.47626507e-01 5.35361826e-01 -1.34239167e-01 1.52026517e-02
8.68840277e-01 -1.15579173e-01 4.47663486e-01 -4.86883111e-02
-1.33745417e-01 9.78267550e-01 3.98450404e-01 -1.53825164e-01
5.12978911e-01 9.33072269e-01 1.89539809e-02 -6.83024228e-01
-1.12711692e+00 -7.80919433e-01 -7.82320797e-01 -7.73796216e-02
3.39585692e-01 -7.49401748e-01 -3.86493564e-01 5.51839352e-01
-2.77255356e-01 -5.84263265e-01 -5.74312389e-01 5.53211212e-01
-3.62292796e-01 2.94428825e-01 -6.57967925e-01 -5.66480577e-01
-2.43848145e-01 -1.54837072e+00 1.21867311e+00 1.28633887e-01
-4.84298110e-01 -1.00833273e+00 1.80181414e-01 3.26790065e-01
-5.76225780e-02 3.23166072e-01 8.73786986e-01 -1.20218432e+00
-1.93964154e-01 -4.69822735e-01 1.60009757e-01 1.16632089e-01
2.29535282e-01 2.68877983e-01 -9.90149915e-01 -4.76586282e-01
-1.71089321e-02 -6.80313230e-01 8.50777388e-01 5.30426145e-01
1.28855669e+00 1.83642477e-01 -4.85055894e-01 7.60110557e-01
1.29167938e+00 -3.48561943e-01 5.57129204e-01 2.61139452e-01
4.04849499e-01 5.33706367e-01 1.01191688e+00 3.66492391e-01
9.28196311e-02 1.99165642e-01 3.87017131e-01 -4.29097176e-01
-1.07032038e-01 -2.09771305e-01 2.52062589e-01 2.73268163e-01
8.59280467e-01 6.34700507e-02 -1.10884857e+00 6.82275653e-01
-1.54771399e+00 -6.14396930e-01 -2.21346229e-01 2.26991725e+00
1.28788447e+00 3.99428695e-01 9.79996622e-02 1.81620091e-01
6.60193980e-01 -2.43963912e-01 -7.50233769e-01 1.02815591e-01
-9.03805047e-02 1.50189415e-01 3.73518705e-01 1.16926268e-01
-1.10677397e+00 6.25574410e-01 6.75930643e+00 8.51038277e-01
-1.17404532e+00 1.54476911e-01 1.06843245e+00 -2.30281726e-01
-8.65385458e-02 -3.89496572e-02 -9.43245173e-01 4.47383553e-01
8.17507923e-01 1.94727913e-01 -2.53427267e-01 6.95354164e-01
2.96655059e-01 -6.08695626e-01 -1.49476314e+00 8.81454766e-01
4.62251678e-02 -1.36149955e+00 -1.07901521e-01 1.97240815e-01
6.85109973e-01 1.76122576e-01 6.73103705e-02 -2.89524086e-02
3.49295735e-01 -1.05510044e+00 2.44880930e-01 3.33856940e-01
1.05709755e+00 -4.15233672e-01 1.28908098e+00 4.78025973e-01
-3.62727165e-01 2.19178572e-01 -3.16317230e-01 5.60683072e-01
-1.38090730e-01 1.22233498e+00 -1.39211786e+00 4.00436521e-01
6.50817811e-01 6.83368981e-01 -8.56945992e-01 1.17753029e+00
-1.21775009e-01 8.16592872e-01 -3.67630988e-01 3.73887748e-01
-1.09132864e-01 2.20190406e-01 3.69520485e-01 1.22830796e+00
1.11446436e-02 -2.32993260e-01 1.38324484e-01 6.30508661e-01
-5.60074970e-02 2.04512388e-01 -2.43560299e-01 -6.59301803e-02
4.69164163e-01 1.57581067e+00 -1.24707913e+00 -3.47542971e-01
-3.18513006e-01 4.53890651e-01 2.84886032e-01 -4.62865196e-02
-6.48355365e-01 7.52636045e-02 8.15856159e-02 1.85636431e-01
2.01897398e-02 5.02919137e-01 -2.93185115e-01 -1.11605656e+00
-2.67641515e-01 -1.18869483e+00 1.11298001e+00 -4.03836787e-01
-1.73391902e+00 3.07598621e-01 -3.77964944e-01 -1.53295779e+00
4.06510867e-02 -6.17066145e-01 -4.63927925e-01 4.97416526e-01
-1.43684506e+00 -1.11081684e+00 -1.82669580e-01 4.41488355e-01
2.33595371e-01 -8.61726180e-02 8.36677969e-01 1.97790608e-01
-7.53713310e-01 8.31152856e-01 2.75075547e-02 1.44110277e-01
1.31847835e+00 -1.38749731e+00 -2.74390846e-01 7.24192381e-01
-1.40326535e-02 5.25985718e-01 4.54203367e-01 -7.64876902e-01
-1.07346869e+00 -1.26020157e+00 4.91058558e-01 -5.41254580e-01
6.06694162e-01 -2.14832336e-01 -9.83739078e-01 7.45669961e-01
-3.52169991e-01 7.31629968e-01 1.51510894e+00 -2.64695555e-01
-9.48626250e-02 -7.29679912e-02 -1.61195302e+00 3.65370065e-01
6.41732454e-01 -3.09722751e-01 -3.12993109e-01 4.67032641e-01
5.27105778e-02 -7.26513743e-01 -1.00811982e+00 3.10101748e-01
4.01611865e-01 -8.73853505e-01 6.27635002e-01 -5.16926229e-01
2.09725603e-01 -5.55924416e-01 7.02537149e-02 -1.07694113e+00
5.88679910e-02 -5.03289640e-01 3.58294882e-02 1.19522321e+00
4.88337755e-01 -5.38951933e-01 1.08872092e+00 3.18663150e-01
-8.33335817e-02 -7.71265268e-01 -1.04907978e+00 -4.39651579e-01
1.49935931e-01 -1.93432599e-01 2.47199669e-01 1.02872300e+00
2.01696202e-01 -2.58495986e-01 2.79681772e-01 3.34264219e-01
9.04478312e-01 -1.04176365e-01 8.10182393e-01 -1.08381510e+00
-2.56629854e-01 -1.54883236e-01 -5.04852116e-01 -3.06391835e-01
8.12773183e-02 -1.35341418e+00 2.06774652e-01 -1.09855282e+00
7.16680586e-01 -6.99291825e-01 -3.34343940e-01 5.99627972e-01
-4.11247700e-01 5.76770782e-01 -4.67301160e-01 4.26079571e-01
-7.36313760e-01 3.25032463e-03 1.02188969e+00 -3.61532532e-02
-2.01721027e-01 2.52291765e-02 -6.61975861e-01 9.07680213e-01
5.06788611e-01 -9.36700940e-01 -2.01263577e-01 2.37872407e-01
8.05592164e-02 -2.79583603e-01 2.78722465e-01 -7.68572628e-01
3.39576066e-01 -6.08516186e-02 6.31282926e-01 -8.94419789e-01
-4.69886929e-01 -7.35157847e-01 1.44172519e-01 6.41409814e-01
-5.57355404e-01 -5.35654843e-01 -3.42448950e-02 5.44447005e-01
-1.79977134e-01 -3.46383810e-01 1.17556965e+00 -1.51488900e-01
-2.87958115e-01 3.07924747e-01 -3.00370544e-01 2.05712289e-01
1.17552269e+00 -3.14544052e-01 -2.91412413e-01 1.32691473e-01
-9.65369165e-01 4.91340488e-01 6.85629785e-01 -1.46090791e-01
4.44826692e-01 -1.01441836e+00 -6.91113353e-01 4.59510058e-01
5.83171904e-01 4.93349224e-01 4.60196942e-01 1.39718759e+00
-5.89765131e-01 2.58849971e-02 7.76793063e-02 -1.22547913e+00
-1.32820439e+00 4.24232930e-01 4.76954341e-01 -5.93635917e-01
-4.96438801e-01 8.17507863e-01 1.76218465e-01 -9.00253177e-01
1.96098149e-01 -3.17501456e-01 6.17941096e-02 1.00869089e-01
6.13530099e-01 2.64311694e-02 4.77126390e-01 -4.89815891e-01
-6.18745506e-01 -1.43254116e-01 -5.84410906e-01 1.63509265e-01
1.37467432e+00 7.74469003e-02 -3.05988103e-01 7.40405023e-01
1.10622227e+00 1.03865422e-01 -1.25547087e+00 -3.52931917e-01
1.31114826e-01 -3.97563517e-01 1.22076906e-01 -8.34839582e-01
-9.05311465e-01 5.21348238e-01 6.64075136e-01 -3.01777244e-01
9.96383488e-01 1.69723123e-01 3.02106500e-01 -1.35316312e-01
3.00940841e-01 -9.71084177e-01 9.02936161e-02 2.42567882e-02
2.70723403e-01 -1.54475605e+00 3.37231100e-01 -5.92517018e-01
-5.84704757e-01 1.00547647e+00 5.77565372e-01 1.01179458e-01
5.99776745e-01 6.99234724e-01 5.63067436e-01 -3.54889959e-01
-8.25831056e-01 6.90581948e-02 -3.49694267e-02 5.88916421e-01
4.51752007e-01 -2.18912780e-01 -1.19919442e-01 5.99154592e-01
1.71323754e-02 2.25663677e-01 7.29239523e-01 1.05938303e+00
-9.21918452e-02 -1.10370255e+00 -4.68379319e-01 8.69261503e-01
-7.67377079e-01 1.49075374e-01 -3.48685145e-01 6.82753980e-01
2.27403358e-01 5.51864505e-01 6.41550822e-03 1.58972606e-01
7.03557283e-02 5.49501851e-02 3.68864775e-01 -1.04909432e+00
-6.13226593e-01 3.83650512e-01 -2.19599724e-01 -4.71664876e-01
-4.80848163e-01 -1.04207623e+00 -1.48888683e+00 2.22399354e-01
-5.61197639e-01 -1.23243615e-01 2.37549841e-01 9.62184608e-01
2.10943252e-01 5.22293806e-01 6.40487313e-01 -3.64553779e-01
-5.69179118e-01 -7.72744000e-01 -7.60807216e-01 5.72674394e-01
4.19841200e-01 -4.49249178e-01 -6.27604961e-01 2.78216243e-01] | [15.067413330078125, -2.787741184234619] |
b0e5ef59-639b-459b-b11a-144199e456dd | defending-against-adversarial-attack-in-ecg | 2203.09487 | null | https://arxiv.org/abs/2203.09487v1 | https://arxiv.org/pdf/2203.09487v1.pdf | Defending Against Adversarial Attack in ECG Classification with Adversarial Distillation Training | In clinics, doctors rely on electrocardiograms (ECGs) to assess severe cardiac disorders. Owing to the development of technology and the increase in health awareness, ECG signals are currently obtained by using medical and commercial devices. Deep neural networks (DNNs) can be used to analyze these signals because of their high accuracy rate. However, researchers have found that adversarial attacks can significantly reduce the accuracy of DNNs. Studies have been conducted to defend ECG-based DNNs against traditional adversarial attacks, such as projected gradient descent (PGD), and smooth adversarial perturbation (SAP) which targets ECG classification; however, to the best of our knowledge, no study has completely explored the defense against adversarial attacks targeting ECG classification. Thus, we did different experiments to explore the effects of defense methods against white-box adversarial attack and black-box adversarial attack targeting ECG classification, and we found that some common defense methods performed well against these attacks. Besides, we proposed a new defense method called Adversarial Distillation Training (ADT) which comes from defensive distillation and can effectively improve the generalization performance of DNNs. The results show that our method performed more effectively against adversarial attacks targeting on ECG classification than the other baseline methods, namely, adversarial training, defensive distillation, Jacob regularization, and noise-to-signal ratio regularization. Furthermore, we found that our method performed better against PGD attacks with low noise levels, which means that our method has stronger robustness. | ['Shenda Hong', 'Tong Liu', 'Weilun Xu', 'Zhaoji Fu', 'Shijia Geng', 'Jiahao Shao'] | 2022-03-14 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 1.28026292e-01 -8.77713785e-02 2.32037783e-01 -2.63818473e-01
-5.14410138e-01 -6.47702992e-01 8.28825310e-03 -8.90953243e-02
-3.50027263e-01 7.12958515e-01 1.04627192e-01 -5.09165466e-01
4.78765368e-02 -8.53585541e-01 -3.94055128e-01 -7.51221418e-01
-3.46281588e-01 -1.62421122e-01 -8.00468177e-02 -4.43090469e-01
-1.32886559e-01 7.34377146e-01 -5.20051479e-01 2.98288733e-01
1.00188911e+00 7.44698524e-01 -8.70869040e-01 7.88563371e-01
3.94295305e-01 7.14734256e-01 -1.06409323e+00 -5.25730848e-01
6.32817864e-01 -7.22597718e-01 -3.69515687e-01 -7.34783649e-01
9.77260768e-02 -5.60514927e-01 -8.50562990e-01 1.23781931e+00
1.35884905e+00 -1.29286841e-01 2.95997202e-01 -9.43310440e-01
-5.60043812e-01 7.19731331e-01 -3.98531079e-01 4.57098842e-01
1.44354686e-01 4.10752922e-01 3.60415012e-01 -4.48844075e-01
1.38477117e-01 1.12083507e+00 1.00581646e+00 1.09408939e+00
-1.13732219e+00 -1.11436570e+00 -8.02764297e-02 -1.52826697e-01
-1.23831308e+00 1.81359500e-02 1.10323048e+00 -1.53620556e-01
3.40872347e-01 5.72266817e-01 4.90912706e-01 1.65693212e+00
5.84555089e-01 5.84233224e-01 1.16139352e+00 -8.07965621e-02
1.60291716e-01 -3.71801332e-02 -4.90256213e-02 3.75226408e-01
2.24074617e-01 5.27019143e-01 -3.78728546e-02 -8.02149534e-01
8.90114129e-01 9.51631144e-02 -6.10513806e-01 3.21544230e-01
-9.53196108e-01 8.65218282e-01 4.88115877e-01 2.61034787e-01
-3.07893515e-01 3.17341276e-02 1.00889003e+00 4.18042660e-01
2.60349631e-01 6.72030330e-01 -5.02776265e-01 4.21260074e-02
-4.13106620e-01 1.28455713e-01 8.17172885e-01 2.71428883e-01
-1.92959040e-01 7.51419485e-01 -3.33709151e-01 6.76454902e-01
-3.54946963e-03 7.69920230e-01 6.35030031e-01 -8.25311065e-01
4.41265434e-01 2.52217054e-01 -4.03735131e-01 -1.36186862e+00
-4.64086652e-01 -7.04473376e-01 -1.46362972e+00 1.86260089e-01
3.09784681e-01 -8.53976130e-01 -8.08959126e-01 1.84088981e+00
8.50478858e-02 5.00144064e-01 3.05983365e-01 9.17489231e-01
8.75841498e-01 2.39491686e-01 1.38468176e-01 -1.43797591e-01
9.19712007e-01 -2.63415337e-01 -8.60668838e-01 -1.32664330e-02
4.78647977e-01 -5.75572908e-01 1.03088295e+00 5.28393388e-01
-7.53910542e-01 -6.25380337e-01 -1.22974229e+00 5.85875630e-01
-3.11278924e-02 -4.98230398e-01 6.36098981e-01 1.49558628e+00
-5.29609084e-01 9.36073124e-01 -9.66875136e-01 1.52975410e-01
6.99910820e-01 3.74845922e-01 -1.60067901e-01 1.78472534e-01
-1.82074678e+00 7.16670454e-01 2.70147268e-02 2.48279959e-01
-1.02011704e+00 -7.48259306e-01 -6.50553882e-01 -1.00532383e-01
-1.32787516e-02 -6.15100503e-01 6.09615266e-01 -9.36518610e-01
-1.41527665e+00 6.33230150e-01 6.75575256e-01 -8.87696028e-01
5.61246395e-01 -4.43871111e-01 -7.36617446e-01 2.42435113e-01
-4.14192498e-01 6.98455945e-02 8.59283626e-01 -8.24999511e-01
2.02629045e-01 -4.59191740e-01 1.94511816e-01 -2.18760073e-01
-6.28843188e-01 5.51940352e-02 2.85834372e-01 -1.18576169e+00
-2.11024832e-04 -1.06524229e+00 -5.62973142e-01 -1.18635237e-01
-5.63829362e-01 4.02618051e-01 8.30623448e-01 -6.79356515e-01
1.40278125e+00 -2.57509375e+00 -2.97747105e-01 3.60647559e-01
3.03342551e-01 8.75558972e-01 -7.92474672e-02 2.30713218e-01
-4.05933112e-01 4.63189363e-01 -3.73690814e-01 3.39934975e-01
-2.91535646e-01 3.99925470e-01 -6.78769052e-01 5.98333895e-01
-9.75880250e-02 9.92762208e-01 -8.14944148e-01 -2.47927725e-01
5.91239631e-02 5.45073807e-01 -6.97815418e-01 1.85568824e-01
3.71162683e-01 7.96171188e-01 -5.74626684e-01 6.19888663e-01
7.34563351e-01 2.25704536e-01 2.75282443e-01 -2.41853416e-01
6.56254232e-01 -2.18740731e-01 -1.01546383e+00 1.12483144e+00
-5.81978895e-02 4.24895227e-01 -7.38393813e-02 -1.22170639e+00
1.04084587e+00 5.89622498e-01 5.49261808e-01 -3.27303886e-01
4.24268365e-01 1.33919999e-01 4.80808377e-01 -4.46284473e-01
-4.21272665e-01 -3.16686153e-01 -2.26410821e-01 9.43458453e-02
-4.71267939e-01 3.30970995e-02 -6.18254066e-01 2.66621560e-01
1.40781236e+00 -2.34076872e-01 1.57269016e-01 -1.85264662e-01
5.46814382e-01 -4.77472991e-01 1.00562990e+00 9.50064182e-01
-5.75721860e-01 6.59870744e-01 5.27521312e-01 -8.68259549e-01
-4.87021565e-01 -1.17365754e+00 -2.46254519e-01 4.78836805e-01
1.76255722e-02 -2.56783515e-01 -7.88470626e-01 -1.05692124e+00
8.21207017e-02 3.57476026e-01 -6.00331187e-01 -8.32406044e-01
-9.25690591e-01 -1.09952962e+00 1.66756451e+00 8.42753172e-01
8.50129485e-01 -9.57464218e-01 -4.71245646e-01 2.06849977e-01
-5.87310903e-02 -9.50853586e-01 -5.56102931e-01 5.93549907e-02
-1.16521716e+00 -1.10484052e+00 -7.42941558e-01 -4.20295924e-01
5.46351850e-01 -4.62816596e-01 8.74325812e-01 3.04673612e-03
-5.41293621e-01 2.93673486e-01 -2.99339205e-01 -8.28481913e-01
-6.63352847e-01 -3.98141026e-01 5.27061522e-01 7.42750196e-03
1.02061398e-01 -8.58863771e-01 -8.07031333e-01 3.24959636e-01
-8.95192146e-01 -7.14904904e-01 4.52343524e-01 9.81435061e-01
3.52790356e-01 1.74038947e-01 7.30138302e-01 -1.21987152e+00
1.11324620e+00 -2.12700188e-01 -1.86799809e-01 -1.02705836e-01
-5.30810416e-01 -1.61537424e-01 1.08713281e+00 -9.00800645e-01
-5.54880261e-01 -3.13686967e-01 -6.59455776e-01 -7.64685273e-01
-1.50941059e-01 3.70914817e-01 -1.41141817e-01 -4.48470920e-01
1.22497177e+00 -2.40961630e-02 2.50481181e-02 -3.27826649e-01
-4.91545945e-02 5.99034488e-01 5.28424323e-01 -5.74632108e-01
9.54091191e-01 5.39177537e-01 1.79556981e-01 -5.20861149e-01
-7.10425854e-01 2.75407583e-01 -5.97242154e-02 5.58556914e-02
8.33054841e-01 -6.70984924e-01 -9.01504338e-01 6.40305996e-01
-8.95433068e-01 -1.66693643e-01 -3.22076827e-01 6.05175555e-01
-1.22618517e-02 8.13281178e-01 -1.00092173e+00 -6.27044380e-01
-8.94536793e-01 -8.51428747e-01 2.09562227e-01 3.17064822e-02
-2.49121264e-01 -9.59100425e-01 2.01387238e-02 8.52484182e-02
6.83889925e-01 1.12678468e+00 8.80458832e-01 -1.02854776e+00
3.65902297e-02 -4.60904300e-01 3.00471693e-01 1.04784667e+00
1.70912415e-01 -2.12374538e-01 -1.01671815e+00 -5.40060699e-01
6.33199394e-01 -1.45905823e-01 3.20080131e-01 4.78923827e-01
1.51548517e+00 -4.22830462e-01 -1.60059631e-01 1.05228972e+00
1.01427734e+00 4.37671065e-01 1.03316355e+00 2.06385195e-01
7.43995130e-01 8.71397182e-02 2.86745578e-01 3.82849574e-01
-4.09891427e-01 2.95826703e-01 6.17249429e-01 -5.29740214e-01
2.50935733e-01 1.29185349e-01 2.79241741e-01 5.09402812e-01
-2.87593871e-01 -4.53724200e-03 -8.50881875e-01 -1.90836098e-02
-1.47202456e+00 -8.72628093e-01 -1.23960920e-01 2.15897107e+00
7.97637224e-01 4.79880154e-01 -5.86638004e-02 4.48456705e-01
6.17027462e-01 4.13874276e-02 -9.09487307e-01 -6.38728201e-01
-2.16375768e-01 5.84393561e-01 5.22548437e-01 -1.32027995e-02
-1.20275438e+00 4.99067754e-01 6.77902126e+00 6.19847953e-01
-1.35456073e+00 -3.01951356e-02 7.38676369e-01 -5.64774126e-03
9.21010301e-02 -4.43089187e-01 -2.64295280e-01 5.52045703e-01
9.02661443e-01 -1.05678446e-04 2.24387541e-01 9.54310179e-01
3.90683897e-02 7.85029650e-01 -8.51991951e-01 1.10365701e+00
-7.23477751e-02 -1.05371428e+00 7.36410543e-02 5.32174669e-03
5.70087016e-01 -2.27409318e-01 1.61947846e-01 5.04362822e-01
1.65114254e-01 -1.01974452e+00 -1.71161175e-01 2.10928053e-01
7.50111222e-01 -1.00198865e+00 1.07325709e+00 2.91244835e-01
-7.07797766e-01 -2.24337563e-01 -2.63604790e-01 -2.49326434e-02
-3.62941362e-02 5.62318981e-01 -3.58819366e-01 5.50973475e-01
7.32104957e-01 3.94296229e-01 -2.48355269e-01 5.83436787e-01
-4.00411516e-01 1.06966388e+00 -9.79557410e-02 3.17212880e-01
-4.51343367e-03 9.31163505e-02 8.49534333e-01 9.53313112e-01
-1.50626838e-01 3.71777564e-01 3.20357680e-01 5.70227921e-01
-5.87597601e-02 -4.49901819e-03 -7.65414298e-01 4.55293432e-02
3.93552274e-01 8.76485288e-01 -4.75468159e-01 -1.43826857e-01
-1.53233543e-01 8.35176766e-01 -3.61356020e-01 3.29706401e-01
-1.13432670e+00 -8.02555859e-01 8.15560937e-01 8.04879814e-02
-3.20875555e-01 4.96220179e-02 -5.62261760e-01 -1.14680326e+00
-1.39779979e-02 -1.64289725e+00 7.27934182e-01 -2.80030191e-01
-1.60770726e+00 8.07821333e-01 -2.03358307e-01 -1.38864791e+00
9.07380357e-02 -4.14080024e-01 -8.03102493e-01 9.86611843e-01
-8.79792571e-01 -6.56620622e-01 -1.04827248e-01 1.04616952e+00
1.53127566e-01 -3.61976355e-01 1.18258786e+00 5.11027515e-01
-5.24928927e-01 1.12287796e+00 -1.05372690e-01 9.13248479e-01
7.42312312e-01 -9.33969438e-01 4.14881408e-01 1.07936490e+00
-4.42180000e-02 1.01235247e+00 5.34901202e-01 -6.71351671e-01
-1.12077308e+00 -1.03492630e+00 -2.05062181e-02 -2.98478782e-01
3.26113254e-01 -2.16455445e-01 -1.10914767e+00 4.81542408e-01
-1.56393722e-02 4.08219129e-01 9.75252330e-01 -1.28270909e-01
-3.22492599e-01 -2.75800616e-01 -1.42466891e+00 7.77458727e-01
7.74423301e-01 -5.51425636e-01 -5.76966763e-01 3.27805132e-01
6.67271078e-01 -7.82629430e-01 -1.04851174e+00 9.96901631e-01
5.35658956e-01 -7.97847331e-01 1.38515759e+00 -8.70166957e-01
2.32832521e-01 -2.12355927e-02 2.31095590e-03 -1.27584159e+00
-2.60335475e-01 -8.72761250e-01 -1.50279909e-01 8.98436189e-01
5.73117025e-02 -1.25272667e+00 6.62185490e-01 6.05315328e-01
-1.23147003e-01 -9.08483505e-01 -7.72443473e-01 -9.21773076e-01
2.02615187e-01 -2.52299488e-01 5.03571689e-01 1.32631731e+00
-1.76386282e-01 6.75339699e-02 -6.89375401e-01 5.17471135e-01
5.45675755e-01 -3.07199210e-01 6.66114151e-01 -9.11696315e-01
-5.85187316e-01 -1.56573236e-01 -7.47464240e-01 -4.00556803e-01
-3.35664637e-02 -7.70133674e-01 -4.78771061e-01 -9.80198801e-01
-3.41842443e-01 -3.29880714e-01 -6.54867113e-01 4.96977955e-01
-6.57232285e-01 5.13362646e-01 1.10021450e-01 1.16314925e-01
1.12856947e-01 1.25614613e-01 1.25968134e+00 -1.73187762e-01
-3.60762566e-01 3.10827047e-01 -9.50797319e-01 1.00908017e+00
1.02341223e+00 -7.27669418e-01 -3.32699805e-01 -1.43031031e-01
-5.30951358e-02 1.29371390e-01 2.73116052e-01 -1.02517200e+00
-1.04826599e-01 2.02303380e-01 5.88809311e-01 -1.11119956e-01
3.82255577e-02 -6.97419703e-01 1.27650350e-01 1.17093551e+00
-3.00409436e-01 -2.09953263e-02 4.47811306e-01 4.14956629e-01
-1.38738245e-01 2.09832415e-01 1.00144911e+00 -7.45557696e-02
-9.07986835e-02 3.51643085e-01 -3.80342603e-01 3.90939504e-01
1.04919791e+00 -1.54982507e-01 -1.47758365e-01 -4.01194066e-01
-1.02923632e+00 -5.83398482e-03 -1.32637709e-01 1.51844144e-01
7.96356380e-01 -1.24055624e+00 -8.78592432e-01 4.37489420e-01
-4.69471902e-01 -2.06025884e-01 3.90282035e-01 8.24741483e-01
-7.11967230e-01 -1.16307966e-01 -2.98778206e-01 -4.76956367e-01
-1.58312380e+00 7.60347843e-01 7.33077765e-01 -2.56497532e-01
-9.68700528e-01 8.15749049e-01 3.14477175e-01 -4.52779144e-01
4.49017227e-01 -6.83126599e-02 -7.70798773e-02 -4.81202722e-01
4.54235613e-01 5.78192353e-01 3.03669348e-02 2.00153477e-02
-5.90547979e-01 4.86027390e-01 -3.01496144e-02 4.00318205e-01
1.02880096e+00 5.00009894e-01 -2.84061935e-02 -2.91063674e-02
9.84572291e-01 2.50455886e-01 -7.55447268e-01 6.66976497e-02
-5.22589684e-01 -4.76268500e-01 -1.00312755e-01 -8.40224743e-01
-1.35058630e+00 1.03927648e+00 1.00166285e+00 3.68263602e-01
1.42953026e+00 -7.31874049e-01 1.00354528e+00 3.89161766e-01
2.55806357e-01 -6.20708525e-01 2.33279184e-01 1.69783637e-01
6.66131914e-01 -9.15922582e-01 1.57263000e-02 -2.98765212e-01
-8.00619006e-01 9.22585487e-01 6.20141447e-01 -4.53082889e-01
8.35234821e-01 5.12437403e-01 6.79107785e-01 2.72616278e-02
-5.32553382e-02 5.30854404e-01 1.02754384e-01 7.30848432e-01
1.98261440e-01 1.00425497e-01 -4.35891330e-01 7.66110122e-01
-5.93794659e-02 -2.03127503e-01 2.56191552e-01 8.76224101e-01
2.49782026e-01 -1.27323401e+00 -6.12602890e-01 4.84458596e-01
-1.30903006e+00 -2.21912801e-01 -3.92371088e-01 7.20107675e-01
1.06769785e-01 8.78316224e-01 -6.47993505e-01 -7.07969010e-01
6.49536729e-01 -6.54699504e-02 2.13561431e-01 -4.63941306e-01
-1.11758006e+00 -6.63136244e-02 -1.91560820e-01 -5.94239414e-01
-1.89284496e-02 -1.86257631e-01 -1.05853879e+00 -4.01185602e-01
-3.46622854e-01 2.11189330e-01 3.22524726e-01 6.32251918e-01
3.59273463e-01 1.03096378e+00 8.83498192e-01 -3.92137825e-01
-1.13999999e+00 -8.13128948e-01 -6.84040308e-01 6.57474458e-01
2.00768486e-01 -1.11866668e-01 -5.81742465e-01 -2.18530491e-01] | [14.302435874938965, 3.1592235565185547] |
835de342-8e08-4ce7-8f24-8685d77f5742 | learning-structural-information-for-syntax | null | null | https://aclanthology.org/2022.findings-naacl.160 | https://aclanthology.org/2022.findings-naacl.160.pdf | Learning Structural Information for Syntax-Controlled Paraphrase Generation | Syntax-controlled paraphrase generation aims to produce paraphrase conform to given syntactic patterns. To address this task, recent works have started to use parse trees (or syntactic templates) to guide generation.A constituency parse tree contains abundant structural information, such as parent-child relation, sibling relation, and the alignment relation between words and nodes.Previous works have only utilized parent-child and alignment relations, which may affect the generation quality.To address this limitation, we propose a Structural Information-augmented Syntax-Controlled Paraphrasing (SI-SCP) model. Particularly, we design a syntax encoder based on tree-transformer to capture parent-child and sibling relations. To model the alignment relation between words and nodes, we propose an attention regularization objective, which makes the decoder accurately select corresponding syntax nodes to guide the generation of words.Experiments show that SI-SCP achieves state-of-the-art performances in terms of semantic and syntactic quality on two popular benchmark datasets.Additionally, we propose a Syntactic Template Retriever (STR) to retrieve compatible syntactic structures. We validate that STR is capable of retrieving compatible syntactic structures. We further demonstrate the effectiveness of SI-SCP to generate diverse paraphrases with retrieved syntactic structures. | ['Yufeng Chen', 'Jinan Xu', 'Yao Meng', 'Yujie Zhang', 'Deyi Xiong', 'Chenglin Bai', 'Erguang Yang'] | null | null | null | null | findings-naacl-2022-7 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 3.90817761e-01 1.10468566e-02 -2.96845406e-01 -6.47786617e-01
-6.29741669e-01 -4.43155318e-01 3.62730056e-01 1.14050500e-01
-2.04992041e-01 4.64597970e-01 7.46547997e-01 -3.03270221e-01
1.40560493e-01 -1.00449228e+00 -7.74668932e-01 -2.68876523e-01
6.22931540e-01 3.16268384e-01 1.02017701e-01 -5.01493871e-01
3.43250960e-01 2.88666543e-02 -1.38801825e+00 7.21501946e-01
1.32261360e+00 5.55634677e-01 6.87653124e-01 -1.12188691e-02
-7.47866154e-01 6.02716804e-01 -6.58662140e-01 -7.56039917e-01
-1.50546553e-02 -8.20442259e-01 -8.96692872e-01 -2.23016158e-01
-1.43978205e-02 -3.14506404e-02 -1.51233301e-01 1.12284672e+00
3.51633370e-01 -1.37448817e-01 4.58682686e-01 -8.86262953e-01
-1.15165627e+00 1.02652109e+00 -2.87781268e-01 2.73689747e-01
7.10471570e-01 6.50045974e-03 1.46393692e+00 -1.01055110e+00
6.62982702e-01 1.45051944e+00 2.57537335e-01 5.91931283e-01
-1.03030014e+00 -6.39270782e-01 3.69886070e-01 3.70210469e-01
-1.18249428e+00 -3.36052239e-01 9.74657238e-01 2.76611391e-02
1.16646910e+00 7.48474970e-02 5.73894918e-01 1.19949675e+00
3.59722793e-01 1.01772547e+00 7.46134222e-01 -6.44825995e-01
1.24804862e-01 -2.05837917e-02 3.79013032e-01 5.48166454e-01
1.42868951e-01 -8.45232531e-02 -5.50641060e-01 5.77139221e-02
4.42215621e-01 -2.52959162e-01 -3.64688963e-01 -1.20272994e-01
-1.00381243e+00 9.90872383e-01 5.42278171e-01 3.70109826e-01
-1.40763447e-01 -1.93890125e-01 6.21587813e-01 4.54553187e-01
2.03743830e-01 5.29208243e-01 -2.95840830e-01 2.84519456e-02
-6.59266710e-01 2.55106330e-01 6.39736712e-01 1.38529909e+00
5.80562532e-01 -6.19028173e-02 -5.23058534e-01 1.26272798e+00
3.10573876e-01 5.37991822e-01 9.30308461e-01 -7.51755059e-01
9.91756618e-01 8.74341011e-01 -2.78944463e-01 -9.65550601e-01
-1.16303116e-01 -4.76595938e-01 -8.58598709e-01 -6.14565790e-01
-2.49826133e-01 2.02469543e-01 -7.93757975e-01 1.95475185e+00
9.38779712e-02 -1.89142406e-01 2.90225625e-01 8.97044063e-01
1.05913234e+00 9.20897901e-01 9.39655229e-02 -3.14056396e-01
1.42985392e+00 -1.08586037e+00 -6.34977639e-01 -5.03471911e-01
8.20493579e-01 -6.82310283e-01 1.60291183e+00 -1.00682944e-01
-1.24454927e+00 -5.95223606e-01 -9.22559142e-01 -2.62666821e-01
-3.56370807e-02 1.48806483e-01 2.83197761e-01 2.99668491e-01
-5.97813189e-01 5.10579765e-01 -5.17508030e-01 -1.38489828e-01
2.12542832e-01 1.83520634e-02 -1.79808646e-01 -2.21799448e-01
-1.51690090e+00 7.66417027e-01 6.35568321e-01 2.35060006e-02
-3.59320253e-01 -5.61062753e-01 -1.14908457e+00 3.83252501e-01
1.20851301e-01 -1.03473067e+00 1.24565387e+00 -9.70324457e-01
-1.35521936e+00 7.53156424e-01 -6.01904273e-01 -2.90052593e-01
-1.27478495e-01 -1.09705262e-01 -2.71993220e-01 -6.37667701e-02
3.61688703e-01 5.27964056e-01 5.33200443e-01 -9.52393353e-01
-5.56745708e-01 -3.43495876e-01 -4.09296900e-03 4.76548135e-01
-4.22767609e-01 2.53753006e-01 -5.61051011e-01 -1.06685078e+00
3.73104930e-01 -6.96795702e-01 -8.97661820e-02 -6.06203020e-01
-5.53841412e-01 -3.97494853e-01 3.57048035e-01 -7.29698300e-01
1.69698632e+00 -2.00743628e+00 5.59683084e-01 1.39669348e-02
-2.32668146e-01 4.16597784e-01 -4.53346461e-01 5.31600118e-01
-2.20508389e-02 2.47893617e-01 -3.80287379e-01 -3.51051956e-01
-5.28659672e-03 4.79807079e-01 -5.73695898e-01 -3.87285888e-01
2.50203192e-01 1.25582683e+00 -8.85867774e-01 -4.53529149e-01
-6.62057400e-02 8.09446722e-02 -6.83405876e-01 3.78636569e-01
-2.94455677e-01 1.51298400e-02 -6.17016912e-01 4.62018609e-01
5.37061810e-01 -2.33004332e-01 2.10034132e-01 -2.95896292e-01
3.39334548e-01 9.81335878e-01 -5.91309190e-01 1.81687605e+00
-9.73476708e-01 2.41153017e-02 -2.85817921e-01 -8.13467145e-01
1.21086252e+00 1.44237131e-02 -1.27083272e-01 -1.20726585e+00
1.91792808e-02 3.72951061e-01 -5.17786928e-02 -3.81061077e-01
5.04068196e-01 -2.26094469e-01 -2.87127733e-01 3.12516719e-01
-1.06335543e-02 -1.71230927e-01 3.27722043e-01 3.63494456e-01
1.20587826e+00 1.17610008e-01 2.74916679e-01 -1.10734738e-01
9.19713736e-01 -1.11925602e-01 8.24038327e-01 4.73644644e-01
2.49864832e-01 6.83153510e-01 5.36121130e-01 -2.78208226e-01
-1.11401880e+00 -1.08110344e+00 1.17174357e-01 9.30321276e-01
3.63938898e-01 -8.45682800e-01 -7.51381516e-01 -7.60867059e-01
-2.40107164e-01 1.11411047e+00 -2.84630448e-01 -5.67614198e-01
-1.14122832e+00 -5.20466685e-01 4.09914762e-01 6.92213178e-01
5.08726656e-01 -1.51592863e+00 -2.46729732e-01 5.61870456e-01
-5.64657390e-01 -1.14664841e+00 -7.72322834e-01 -1.04653254e-01
-7.89322793e-01 -9.03619409e-01 -5.71595848e-01 -1.18717325e+00
6.66050673e-01 2.44777173e-01 1.32409942e+00 5.52202873e-02
1.60361215e-01 -3.43841404e-01 -8.34794223e-01 1.92904826e-02
-7.12797046e-01 4.72243845e-01 -3.12050134e-01 -2.96646863e-01
3.90629858e-01 -8.55826855e-01 -4.72739488e-01 3.36949080e-01
-8.81605327e-01 4.17134792e-01 7.37324238e-01 9.60990071e-01
6.50328994e-01 -4.34791535e-01 6.73450887e-01 -9.51965690e-01
9.88428473e-01 -4.30134267e-01 -3.96779031e-01 5.92078686e-01
-5.61663508e-01 5.02698123e-01 1.24299550e+00 -2.18183640e-02
-1.19848216e+00 -1.02758162e-01 -4.04375225e-01 -4.36389476e-01
6.00736849e-02 6.33887947e-01 -4.96096641e-01 5.49936235e-01
4.32893038e-01 7.12170839e-01 -3.84380996e-01 -6.17980957e-01
4.19188291e-01 7.83266306e-01 5.22661150e-01 -7.85485983e-01
5.63981414e-01 -9.12009329e-02 -1.20949373e-01 -2.36027405e-01
-1.11179817e+00 -1.16072252e-01 -4.64512020e-01 4.08643931e-01
6.62182808e-01 -8.41732204e-01 1.73498224e-02 2.10612565e-01
-1.60722804e+00 4.85957377e-02 -1.74321786e-01 3.29637602e-02
-4.57301497e-01 6.34560049e-01 -7.90204883e-01 -2.11995438e-01
-9.67951477e-01 -1.19424248e+00 1.22121954e+00 9.93338004e-02
-2.84297764e-01 -6.84989035e-01 5.45102358e-02 4.10633415e-01
4.55725938e-01 -1.74298719e-01 1.58143723e+00 -6.54387236e-01
-5.19579768e-01 1.97281405e-01 -3.63505632e-01 3.52605253e-01
2.26669714e-01 -3.31988096e-01 -3.03631902e-01 -1.35345131e-01
-6.41886964e-02 -1.55440673e-01 8.46669495e-01 -1.77162997e-02
9.71358120e-01 -5.65236151e-01 -3.27440053e-01 7.79312670e-01
1.03997612e+00 2.32148215e-01 6.62774920e-01 2.99993157e-01
7.68826306e-01 7.09881306e-01 6.69080257e-01 2.64029890e-01
5.75928390e-01 8.29898179e-01 3.42718028e-02 3.68222952e-01
-2.98540980e-01 -8.03042591e-01 5.17884731e-01 1.23479748e+00
5.72666407e-01 -4.11157846e-01 -7.03060329e-01 5.06958783e-01
-1.79842114e+00 -6.04164183e-01 -1.63426042e-01 1.87926340e+00
1.05581117e+00 2.48547077e-01 -3.29990834e-01 -4.89504747e-02
7.77383387e-01 2.66667306e-01 -3.54046285e-01 -5.64446330e-01
-2.13648558e-01 4.37013179e-01 -4.24764752e-02 3.72738659e-01
-5.15559018e-01 1.35771489e+00 5.22870255e+00 1.08276761e+00
-7.75506437e-01 -4.22577895e-02 3.25819939e-01 -1.03941998e-02
-9.41073120e-01 1.68231905e-01 -9.64846671e-01 8.61436665e-01
6.41938090e-01 -5.22558570e-01 3.51942807e-01 8.70722175e-01
2.19927460e-01 3.09423417e-01 -1.07961917e+00 8.77150238e-01
2.06590265e-01 -1.32855487e+00 6.69082940e-01 -3.22423965e-01
5.00051737e-01 -2.10440010e-01 -3.12127382e-01 4.49765503e-01
2.06739038e-01 -8.65090013e-01 5.69115520e-01 1.28550157e-01
7.04068005e-01 -8.62809718e-01 7.52465367e-01 3.13788801e-01
-1.48894095e+00 -8.90416130e-02 -6.82373285e-01 2.28533402e-01
4.91140068e-01 6.46456420e-01 -4.41982657e-01 7.59517491e-01
3.78812551e-01 7.85683215e-01 -6.00480258e-01 7.03422606e-01
-9.04818177e-01 5.91848075e-01 -1.25498623e-01 -1.61796242e-01
2.32471481e-01 -3.23849171e-01 6.05443478e-01 1.20493996e+00
4.51165318e-01 3.82270440e-02 8.97270963e-02 1.12156463e+00
-1.64075911e-01 4.81979668e-01 -4.81437892e-01 2.75150478e-01
9.56082284e-01 8.59596312e-01 -2.72100389e-01 -2.88873136e-01
-2.86539137e-01 1.20331395e+00 7.21630871e-01 9.43876207e-02
-7.04403937e-01 -4.69022781e-01 5.12008607e-01 6.68850914e-02
3.13563794e-01 2.38327193e-03 -3.31921428e-01 -1.42728126e+00
6.03032410e-01 -1.04771221e+00 4.38658506e-01 -9.86321330e-01
-1.16184044e+00 7.89141953e-01 8.25552940e-02 -9.70994830e-01
-1.79149896e-01 -2.26248711e-01 -8.33501697e-01 9.09391105e-01
-1.49936140e+00 -1.22572577e+00 -1.70583323e-01 2.81628221e-01
1.08245611e+00 -2.87230343e-01 7.24831522e-01 1.11194447e-01
-8.34935784e-01 8.32732797e-01 -1.98686704e-01 1.17993988e-01
3.33941519e-01 -9.50864315e-01 9.61601079e-01 9.18008029e-01
2.64465600e-01 8.50265205e-01 3.54149669e-01 -7.09087431e-01
-1.26350427e+00 -1.29596722e+00 1.37835467e+00 -1.60925910e-01
3.07462662e-01 -3.57850462e-01 -1.15520799e+00 5.03831863e-01
1.86371908e-01 -4.72776413e-01 4.79914635e-01 -6.82236552e-02
-5.83714902e-01 -1.26803532e-01 -7.26764798e-01 8.39591086e-01
1.52420545e+00 -4.40497667e-01 -1.16587639e+00 2.89776921e-01
1.20771575e+00 -3.11345369e-01 -4.44108307e-01 6.42880321e-01
2.47352034e-01 -9.34429586e-01 7.01464236e-01 -5.94326079e-01
9.80520546e-01 -6.69415966e-02 -1.78504035e-01 -1.50672770e+00
-4.53902036e-01 -3.94477397e-01 -1.81219772e-01 1.53444552e+00
5.83811343e-01 -3.64970177e-01 7.82358110e-01 3.57230633e-01
-5.09691358e-01 -9.68477905e-01 -9.10881937e-01 -8.96559179e-01
2.60249197e-01 -1.99837357e-01 8.70445848e-01 6.09766424e-01
2.42972732e-01 8.94558668e-01 -1.65591553e-01 -1.61875442e-01
2.57949084e-01 5.68962216e-01 5.10506213e-01 -7.87177920e-01
-4.79715705e-01 -4.90124106e-01 4.24810499e-02 -1.45056164e+00
7.29035497e-01 -1.28826809e+00 -9.68384445e-02 -1.68210697e+00
2.46502519e-01 -4.72383261e-01 -7.06092566e-02 4.95477021e-01
-5.22341311e-01 -3.68726820e-01 2.92269230e-01 4.13493842e-01
-4.62311983e-01 1.00732934e+00 1.22213793e+00 -4.13611829e-02
-2.88221300e-01 -8.14682543e-02 -7.95980155e-01 4.76257265e-01
9.14505064e-01 -5.16088367e-01 -5.59916079e-01 -7.83202410e-01
3.86475950e-01 2.54527181e-01 -6.50178418e-02 -7.59508193e-01
1.14410266e-01 -2.42500156e-01 -9.34049562e-02 -5.07700741e-01
9.13293660e-02 -4.73060399e-01 -2.80710049e-02 4.06772256e-01
-5.88188946e-01 5.25195718e-01 -8.16083997e-02 3.36739093e-01
-4.24598992e-01 -6.01825356e-01 5.64069390e-01 -2.46116266e-01
-5.64226270e-01 2.53100712e-02 -9.00749192e-02 4.53074574e-01
6.92235351e-01 -1.92028861e-02 -3.20144087e-01 -1.45304039e-01
-2.37825334e-01 5.01247942e-01 5.09049714e-01 7.14645088e-01
8.23443651e-01 -1.52805555e+00 -8.00491512e-01 4.40935552e-01
3.92104685e-01 6.50157109e-02 1.12983142e-03 1.24728628e-01
-3.81162167e-01 4.64808851e-01 -3.98281775e-02 -1.83723733e-01
-1.25998616e+00 6.37344718e-01 2.21041054e-01 -5.17967939e-01
-6.77892506e-01 8.35446239e-01 3.88559461e-01 -5.19758463e-01
3.57652940e-02 -5.27760684e-01 -1.58597454e-01 -3.34836334e-01
4.55713153e-01 -2.31768172e-02 8.12901258e-02 -5.29190600e-01
-2.71834791e-01 5.94754755e-01 -4.81019109e-01 5.79054020e-02
9.73505616e-01 -1.71031207e-01 -2.06532523e-01 -1.84527084e-01
1.23204517e+00 -6.49327412e-02 -5.87751925e-01 -4.44648236e-01
1.68426052e-01 -4.10235524e-01 -2.62579203e-01 -6.58671856e-01
-9.30194438e-01 8.19864810e-01 -1.11582480e-01 -2.33491540e-01
1.15135431e+00 8.08612779e-02 1.41101921e+00 4.95400429e-01
4.76035625e-01 -8.86332095e-01 1.65571928e-01 8.89336288e-01
1.07519507e+00 -8.09584200e-01 -3.59618574e-01 -8.86229396e-01
-7.54743040e-01 8.71200025e-01 8.70401740e-01 -4.63285670e-02
1.94353744e-01 3.81664112e-02 -2.41137564e-01 -3.40534076e-02
-8.98521304e-01 -1.26927316e-01 2.64748067e-01 3.49063337e-01
5.71973145e-01 -1.79503709e-01 -8.13702464e-01 8.28917980e-01
-5.90822279e-01 -1.58318788e-01 2.47201309e-01 8.69433939e-01
-5.76812625e-01 -1.65049458e+00 1.22719526e-01 3.58728290e-01
-2.06599876e-01 -5.78821182e-01 -5.67974091e-01 2.50505000e-01
-8.08554143e-02 8.89451683e-01 -5.94360456e-02 -3.91920298e-01
6.77595377e-01 1.62486210e-01 3.94610226e-01 -1.04923320e+00
-7.44234204e-01 -1.59632280e-01 1.98451370e-01 -5.13849258e-01
5.58404345e-03 -3.19871038e-01 -1.32298613e+00 -1.16854556e-01
-3.49097878e-01 3.81110728e-01 2.36482829e-01 8.95500422e-01
6.69831634e-01 4.60671991e-01 9.34554279e-01 -3.71223241e-01
-7.41372168e-01 -8.58252406e-01 -2.95364290e-01 7.67653704e-01
-2.67232150e-01 -3.09928149e-01 -2.40212858e-01 -1.74628735e-01] | [11.630940437316895, 9.315764427185059] |
c7420447-0014-42a9-af59-65489e1e4122 | pseudo-value-based-deep-neural-networks-for | 2207.05291 | null | https://arxiv.org/abs/2207.05291v1 | https://arxiv.org/pdf/2207.05291v1.pdf | Pseudo value-based Deep Neural Networks for Multi-state Survival Analysis | Multi-state survival analysis (MSA) uses multi-state models for the analysis of time-to-event data. In medical applications, MSA can provide insights about the complex disease progression in patients. A key challenge in MSA is the accurate subject-specific prediction of multi-state model quantities such as transition probability and state occupation probability in the presence of censoring. Traditional multi-state methods such as Aalen-Johansen (AJ) estimators and Cox-based methods are respectively limited by Markov and proportional hazards assumptions and are infeasible for making subject-specific predictions. Neural ordinary differential equations for MSA relax these assumptions but are computationally expensive and do not directly model the transition probabilities. To address these limitations, we propose a new class of pseudo-value-based deep learning models for multi-state survival analysis, where we show that pseudo values - designed to handle censoring - can be a natural replacement for estimating the multi-state model quantities when derived from a consistent estimator. In particular, we provide an algorithm to derive pseudo values from consistent estimators to directly predict the multi-state survival quantities from the subject's covariates. Empirical results on synthetic and real-world datasets show that our proposed models achieve state-of-the-art results under various censoring settings. | ['Sanjay Purushotham', 'Md Mahmudur Rahman'] | 2022-07-12 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [-7.91528299e-02 -3.14957201e-01 -5.50538957e-01 -7.03274846e-01
-1.20800626e+00 1.02111794e-01 2.46083990e-01 2.90937930e-01
-1.37183875e-01 1.20439231e+00 1.81451559e-01 -7.27736652e-01
-2.60890454e-01 -6.78129613e-01 -5.09106100e-01 -8.14529479e-01
-3.10013205e-01 6.72902703e-01 -2.98851520e-01 1.05904276e-02
-1.84979364e-01 2.54846185e-01 -7.68791676e-01 -5.83139397e-02
8.83286417e-01 7.24044263e-01 -4.86140221e-01 8.07446182e-01
6.52289614e-02 8.29747677e-01 -2.68361688e-01 -2.38888279e-01
-2.85859585e-01 -5.76226234e-01 -4.17672068e-01 -3.69078040e-01
-8.40700939e-02 -6.23855293e-01 -4.96106505e-01 6.02955282e-01
4.79844332e-01 -1.59283265e-01 1.23301518e+00 -1.45785451e+00
-3.86499912e-01 3.66152048e-01 -3.00241709e-01 1.35129705e-01
-6.78618923e-02 1.06329195e-01 6.68512940e-01 -4.29913998e-01
3.71874161e-02 1.18689644e+00 1.36072183e+00 6.83060944e-01
-1.43456900e+00 -5.13663232e-01 -1.01495646e-01 -5.00395484e-02
-1.29155397e+00 -4.30235386e-01 5.20808995e-01 -5.31569064e-01
6.10065997e-01 2.03368366e-01 5.15322447e-01 1.61843836e+00
1.13015759e+00 7.18664527e-01 1.11111772e+00 -2.96122152e-02
4.46909159e-01 -3.51524383e-01 6.56831861e-01 5.92420816e-01
3.00206095e-01 5.18043041e-01 -2.37192959e-01 -8.22305143e-01
1.15897167e+00 7.14573264e-01 1.06711112e-01 -5.40960670e-01
-1.24730480e+00 1.04187560e+00 -5.08620869e-03 -2.09263220e-01
-6.82138741e-01 6.02656484e-01 4.09615785e-01 1.90018922e-01
5.13057947e-01 -3.67727369e-01 -7.57670283e-01 -8.00255761e-02
-1.15491247e+00 1.06683947e-01 9.20219660e-01 7.52162337e-01
1.55905589e-01 1.11979678e-01 -7.59201348e-01 4.10845250e-01
3.87335211e-01 6.70371056e-01 3.51971090e-01 -7.37261236e-01
-2.06540842e-02 1.75821066e-01 4.24796283e-01 -1.47140533e-01
-9.84403610e-01 -9.09515738e-01 -1.58177471e+00 -8.44871104e-02
7.56609142e-01 -4.32725072e-01 -1.16266310e+00 2.13886833e+00
2.82461017e-01 6.09413445e-01 1.61158830e-01 2.81970918e-01
4.68016505e-01 4.74824160e-01 3.50015670e-01 -5.81358075e-01
1.27766669e+00 -4.76171225e-01 -8.92903328e-01 -6.31323308e-02
7.69607604e-01 9.16513205e-02 7.23850727e-01 7.63043063e-03
-9.89093125e-01 -2.24275794e-02 -5.23496151e-01 1.00816451e-01
2.74539795e-02 -4.60954197e-02 7.85279691e-01 6.25538886e-01
-8.35502982e-01 8.49504113e-01 -1.49103653e+00 -3.12888473e-01
5.90573192e-01 4.15659100e-01 -1.17990689e-03 2.29364172e-01
-1.40741277e+00 8.03587914e-01 -1.88546598e-01 1.78099722e-01
-1.44396424e+00 -8.30385447e-01 -7.03996181e-01 3.74911785e-01
-6.29448742e-02 -1.29308486e+00 1.35477674e+00 -4.72848684e-01
-1.32381308e+00 4.56747144e-01 -3.05108815e-01 -6.56585097e-01
6.18423462e-01 1.09099358e-01 -4.85312313e-01 -2.83007383e-01
1.49733767e-01 7.08520338e-02 7.36390710e-01 -6.15574181e-01
-3.41754675e-01 -4.41104442e-01 -6.33818984e-01 -2.06698090e-01
-1.11433841e-01 -8.76801088e-02 4.53459434e-02 -6.58832490e-01
5.15215546e-02 -9.82013404e-01 -6.01763189e-01 1.09902181e-01
-5.83435416e-01 -5.05761951e-02 3.74127664e-02 -8.88913274e-01
1.25028718e+00 -1.86297905e+00 -1.89719275e-01 -2.00929776e-01
2.44161114e-01 -2.67494947e-01 3.60740647e-02 5.58814764e-01
-3.15218084e-02 -1.54902384e-01 -6.40088260e-01 -6.59297407e-01
-2.78455496e-01 2.37047300e-01 -7.76209980e-02 8.21508229e-01
-6.63087796e-03 1.08708906e+00 -7.96800911e-01 -7.28938222e-01
-3.99759412e-02 6.05171621e-01 -3.66073132e-01 1.76786557e-01
2.43409440e-01 6.60116434e-01 -4.50834781e-01 8.18306327e-01
4.72138286e-01 -7.50068486e-01 -1.92581816e-03 3.66900116e-01
3.35228086e-01 4.76406403e-02 -7.08076119e-01 1.41684842e+00
-4.73945886e-01 2.21596271e-01 -3.00986141e-01 -1.15538871e+00
6.37236714e-01 5.80832422e-01 6.40895963e-01 4.60196361e-02
3.79490167e-01 1.31273568e-01 -1.74714774e-01 -1.88906312e-01
-1.50285691e-01 -8.52013528e-01 -6.19723082e-01 3.82398844e-01
-2.28872932e-02 2.93741792e-01 -4.04457897e-01 -2.87668072e-02
1.39994395e+00 -2.50928402e-01 7.23555028e-01 -1.78786024e-01
1.78926855e-01 -2.04504535e-01 1.09200513e+00 1.16268361e+00
-6.60026491e-01 5.35923779e-01 7.68292427e-01 -5.97537756e-01
-8.32405627e-01 -1.73991728e+00 -7.47165799e-01 4.80135620e-01
-3.60438913e-01 3.34871054e-01 -3.27055871e-01 -6.88475609e-01
2.68898070e-01 7.88382709e-01 -9.40173090e-01 -6.96237504e-01
-1.49290949e-01 -1.43818676e+00 6.53118432e-01 7.89303422e-01
-7.64245465e-02 -7.75588334e-01 -3.80199879e-01 5.55723429e-01
-2.94432670e-01 -4.19254482e-01 -4.54060882e-01 4.88710701e-01
-1.42655694e+00 -8.46431375e-01 -1.21109903e+00 -6.36636496e-01
4.88314390e-01 -4.75350469e-01 1.00570726e+00 -2.28712589e-01
7.33753890e-02 6.51177913e-02 2.71108121e-01 -3.49699855e-01
-6.86423302e-01 3.60988081e-02 2.59136140e-01 4.60208058e-02
2.39676073e-01 -6.09129965e-01 -7.82150209e-01 1.24205478e-01
-7.48263061e-01 5.52666038e-02 6.66376472e-01 1.36469579e+00
7.56562352e-01 -4.69300717e-01 1.34466338e+00 -9.76275146e-01
4.11069751e-01 -8.43266487e-01 -3.68872374e-01 4.03149903e-01
-9.16137993e-01 2.13565677e-01 5.79330027e-01 -7.11430371e-01
-9.02817488e-01 2.24050224e-01 -4.91572209e-02 -4.06175286e-01
-1.12928547e-01 7.64227450e-01 2.52255291e-01 4.93281484e-01
2.71374106e-01 4.41438615e-01 1.41190201e-01 -4.43322748e-01
-4.19949144e-02 6.99957609e-01 7.00082898e-01 -2.54204303e-01
1.22520708e-01 6.80180669e-01 6.99634492e-01 -1.86939567e-01
-1.12531626e+00 -3.06312948e-01 -5.78835368e-01 1.43646449e-01
7.36275375e-01 -1.12719119e+00 -8.68135393e-01 9.29444432e-01
-8.60938311e-01 -5.71796536e-01 -2.96836406e-01 7.23734081e-01
-9.83037293e-01 2.38564223e-01 -1.16818249e+00 -1.00586593e+00
-5.43632209e-01 -8.66637349e-01 1.03152466e+00 1.54691175e-01
-9.35017467e-02 -1.52995169e+00 5.31459749e-01 -6.73995018e-02
3.35868627e-01 5.65386593e-01 1.24191248e+00 -5.76946318e-01
-3.88276465e-02 -5.57326376e-01 3.85378394e-03 1.04995400e-01
3.70894045e-01 -3.41558069e-01 -5.60972095e-01 -5.59750080e-01
7.29715079e-02 6.95615485e-02 8.22235525e-01 1.37339079e+00
1.05483389e+00 -1.28728136e-01 -7.12810397e-01 7.90381432e-01
1.28006756e+00 1.48153231e-01 3.77339423e-01 -4.70279381e-02
3.81976247e-01 1.26493722e-01 3.40795934e-01 8.34069729e-01
7.06321359e-01 3.47917527e-01 4.01686609e-01 -6.20790422e-02
2.11821213e-01 -3.63037497e-01 3.86346459e-01 6.05996668e-01
5.86659968e-01 -2.31909961e-01 -1.06129658e+00 8.33932579e-01
-2.05165863e+00 -7.52693236e-01 -6.78963065e-01 2.54354405e+00
1.10556519e+00 8.32976401e-02 3.18364650e-01 -2.93822825e-01
7.93810666e-01 -2.26013854e-01 -1.12947524e+00 -1.72834456e-01
-1.15667984e-01 2.06507556e-02 5.80169559e-01 3.38374466e-01
-1.03462934e+00 5.61385393e-01 7.52382421e+00 3.58408451e-01
-6.92629933e-01 3.17191929e-01 9.63076115e-01 -1.99518964e-01
4.54892777e-02 2.43727073e-01 -8.45593512e-01 6.01510882e-01
1.67112601e+00 -2.08097354e-01 -2.28039678e-02 5.22955477e-01
2.97871351e-01 1.05159611e-01 -1.35196805e+00 7.48727798e-01
-3.79995883e-01 -1.11763382e+00 -1.91789418e-01 1.63349882e-01
7.34208405e-01 -1.89616047e-02 1.91813871e-01 5.90812922e-01
5.83564341e-01 -7.73904562e-01 4.36346561e-01 1.17612553e+00
1.08304608e+00 -7.87411571e-01 8.62658799e-01 3.77148598e-01
-6.73332751e-01 -1.93558037e-01 -2.15355247e-01 -4.94470857e-02
7.41825044e-01 9.47512865e-01 -6.36409342e-01 2.35428646e-01
2.95844167e-01 6.90290093e-01 -3.68297756e-01 1.14445460e+00
1.06289148e-01 1.08935523e+00 -2.17008233e-01 1.96683705e-01
-9.86395031e-02 1.24376148e-01 4.25437778e-01 7.71521389e-01
6.16816282e-01 -1.84903398e-01 -2.15628549e-01 8.95214856e-01
2.33259439e-01 -3.36299986e-01 -2.73714215e-01 1.05699398e-01
2.52731740e-01 8.27567279e-01 -6.20011747e-01 -4.64190483e-01
-4.50926542e-01 9.37908232e-01 1.88458696e-01 3.54981124e-01
-1.02397621e+00 -1.66448712e-01 7.54169106e-01 9.39622521e-02
4.86098379e-02 -1.69519871e-01 -5.91462016e-01 -1.44184613e+00
-5.07474542e-01 -5.12759328e-01 8.87585223e-01 -5.82270384e-01
-1.64550579e+00 2.19360128e-01 -1.00764513e-01 -1.09962952e+00
-5.16643107e-01 -2.45429724e-01 -7.00518191e-01 9.61457670e-01
-1.36904538e+00 -1.13359499e+00 1.61044121e-01 7.37372339e-01
2.75370508e-01 1.39060944e-01 9.05588388e-01 9.62401628e-02
-9.97169256e-01 6.99178874e-01 7.38740921e-01 9.06982347e-02
7.37094223e-01 -1.27270746e+00 6.55965626e-01 6.64672375e-01
-2.14258969e-01 3.94090623e-01 6.94286048e-01 -1.05241919e+00
-1.10125756e+00 -1.09150982e+00 9.92553949e-01 -7.98663318e-01
5.84329903e-01 -2.02411190e-01 -1.06878364e+00 8.61463845e-01
-5.77382565e-01 1.90604836e-01 8.22365940e-01 4.09693390e-01
1.90705329e-01 -1.69408675e-02 -1.10242116e+00 4.17709649e-01
6.96722925e-01 -4.81344908e-01 -4.47164178e-01 3.41797709e-01
5.58108211e-01 -1.22643679e-01 -8.94396424e-01 3.51161242e-01
6.96219265e-01 -6.58258557e-01 1.09675682e+00 -1.28748035e+00
3.81893963e-01 2.46191442e-01 1.08431794e-01 -1.39467955e+00
-5.94345927e-01 -5.40036321e-01 -5.45036912e-01 1.01559484e+00
3.09013635e-01 -8.07335377e-01 7.55033791e-01 8.94715965e-01
-3.05189714e-02 -8.31688464e-01 -1.45856416e+00 -1.11353242e+00
5.88129163e-01 -3.62781048e-01 8.59778285e-01 7.62119949e-01
-9.99936685e-02 4.72731926e-02 -7.73366868e-01 3.86430562e-01
1.02237415e+00 1.80966273e-01 3.51409227e-01 -1.65992248e+00
-2.44337723e-01 -2.16443706e-02 -1.12435482e-01 -5.71309745e-01
3.76396865e-01 -7.08700240e-01 2.16116756e-02 -1.67377830e+00
7.56570995e-01 -4.75826740e-01 -9.80692923e-01 5.30497491e-01
-3.96830082e-01 -2.60995299e-01 -4.63266999e-01 3.38913500e-01
-3.01235378e-01 7.33535111e-01 9.51556087e-01 1.31033093e-01
-2.91157782e-01 7.24829495e-01 -5.73274612e-01 4.69214618e-01
7.28957355e-01 -1.02981079e+00 -8.70196894e-02 7.02002272e-02
1.10924028e-01 1.32422805e+00 7.40050077e-01 -8.80370617e-01
3.40453498e-02 -5.49382031e-01 3.88848990e-01 -8.11170816e-01
3.70679855e-01 -5.33888400e-01 4.74677175e-01 9.45978045e-01
-3.77518594e-01 6.63463846e-02 -3.04686725e-02 1.13319743e+00
1.60787731e-01 1.60830021e-01 7.80153394e-01 2.18551502e-01
5.24867922e-02 6.37337327e-01 -7.66534567e-01 -2.63397139e-03
7.27748990e-01 1.97030112e-01 -1.20652743e-01 -8.61414909e-01
-9.85251844e-01 3.34446400e-01 2.15165943e-01 7.95358717e-02
4.71529007e-01 -1.49601555e+00 -9.38366711e-01 1.47968635e-01
9.81056876e-03 -2.45839745e-01 6.23182476e-01 1.21647620e+00
-1.39333144e-01 5.20048738e-01 -1.24996945e-01 -6.01223528e-01
-7.64218390e-01 7.30796397e-01 5.21324515e-01 -7.41995156e-01
-5.82525015e-01 4.41419750e-01 4.84499335e-01 -2.97377557e-01
-1.45993486e-01 -5.99272668e-01 1.87053710e-01 -7.91626275e-02
2.48320907e-01 4.49376613e-01 -2.81795323e-01 -4.58308667e-01
-4.44773883e-01 -3.67934088e-04 -2.22117752e-02 -2.32458845e-01
1.34730923e+00 -3.85593146e-01 3.55482250e-01 1.02907264e+00
1.10354960e+00 -8.59795690e-01 -1.58286536e+00 -2.13826105e-01
-1.11900076e-01 -3.18798944e-02 3.78487378e-01 -9.25183654e-01
-7.81619668e-01 1.03871310e+00 7.68774748e-01 -3.37810703e-02
9.50370729e-01 -1.67885795e-02 1.04048920e+00 1.13149710e-01
3.09531003e-01 -5.87733626e-01 -4.03417081e-01 2.11893842e-01
4.59777892e-01 -1.33042324e+00 -1.66354835e-01 1.47113219e-01
-4.26786929e-01 7.24384367e-01 1.77898794e-01 -1.64303984e-02
1.04467797e+00 1.46455348e-01 6.76040575e-02 -7.28177577e-02
-1.06005478e+00 1.38793766e-01 9.26245600e-02 5.35722911e-01
3.17611754e-01 3.09427351e-01 -2.55877912e-01 1.21670401e+00
1.57380700e-01 5.51161766e-01 7.56507516e-01 7.32402027e-01
1.48781300e-01 -9.74125683e-01 -3.08239460e-01 1.07354259e+00
-7.04363644e-01 -1.31979793e-01 2.96559125e-01 4.47971910e-01
-7.13240564e-01 7.44809151e-01 7.87137672e-02 1.58763304e-01
1.18581345e-02 3.86783659e-01 1.20766856e-01 -3.91857207e-01
-2.75291532e-01 -7.34053329e-02 -2.06103101e-01 -2.17046365e-01
8.11050832e-02 -1.13860321e+00 -1.10117280e+00 -3.44308227e-01
-4.39604104e-01 -1.35354027e-01 4.48185295e-01 9.53746259e-01
4.25845474e-01 7.38312006e-01 7.23496079e-01 -3.71281028e-01
-1.24216843e+00 -8.59613478e-01 -9.71758425e-01 2.20887735e-01
8.59802127e-01 -7.24178195e-01 -6.14336789e-01 -7.12399781e-02] | [7.801756381988525, 5.567692756652832] |
b560115a-cedd-42de-9e79-a6a1c4ec80ac | 100-things-you-always-wanted-to-know-about-1 | null | null | https://aclanthology.org/P18-5001 | https://aclanthology.org/P18-5001.pdf | 100 Things You Always Wanted to Know about Semantics \& Pragmatics But Were Afraid to Ask | Meaning is a fundamental concept in Natural Language Processing (NLP), given its aim to build systems that mean what they say to you, and understand what you say to them. In order for NLP to scale beyond partial, task-specific solutions, it must be informed by what is known about how humans use language to express and understand communicative intents. The purpose of this tutorial is to present a selection of useful information about semantics and pragmatics, as understood in linguistics, in a way that{'}s accessible to and useful for NLP practitioners with minimal (or even no) prior training in linguistics. The tutorial content is based on a manuscript in progress I am co-authoring with Prof. Alex Lascarides of the University of Edinburgh. | ['Emily M. Bender'] | 2018-07-01 | null | null | null | acl-2018-7 | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 3.47551554e-01 5.87242067e-01 -1.42803714e-01 -8.07152331e-01
-4.45845723e-01 -7.81012774e-01 5.86635113e-01 4.28332627e-01
-4.58757460e-01 7.00369000e-01 6.42148972e-01 -8.20822060e-01
-1.71597242e-01 -3.92511994e-01 -2.06728503e-02 -1.10215031e-01
2.14036137e-01 3.26229423e-01 9.78391021e-02 -6.00529134e-01
4.87451762e-01 4.31125373e-01 -1.21048093e+00 2.92361438e-01
5.82217395e-01 3.31897497e-01 4.73529041e-01 5.32466650e-01
-7.26000786e-01 9.31619346e-01 -4.48076397e-01 -3.07749540e-01
-3.15323561e-01 -3.68109912e-01 -1.44885302e+00 1.29227098e-02
-2.37242982e-01 1.69737011e-01 3.10509175e-01 1.09790194e+00
-8.40918124e-02 1.60589084e-01 5.98749816e-01 -5.96941650e-01
-5.05362689e-01 8.01740170e-01 1.24380767e-01 2.97993690e-01
8.03649962e-01 1.06087640e-01 1.20326436e+00 -3.98838103e-01
6.60809219e-01 1.44825363e+00 3.10914665e-01 8.04728329e-01
-1.02090883e+00 3.46476696e-02 3.45702052e-01 -3.51106167e-01
-1.02706337e+00 -8.10703695e-01 5.11213362e-01 -4.46221620e-01
1.30339003e+00 2.57681966e-01 4.00616765e-01 6.77756011e-01
1.43118605e-01 9.35710371e-01 8.60521674e-01 -9.60317910e-01
1.80212200e-01 5.54120958e-01 6.77138269e-01 2.42214397e-01
1.92192703e-01 -1.83063582e-01 -5.61763048e-01 1.30031835e-02
3.47634137e-01 -5.70208013e-01 -2.79846370e-01 7.13247135e-02
-1.29578412e+00 8.62687945e-01 -9.22276452e-02 9.04163778e-01
-3.58907491e-01 1.56934604e-01 4.39834386e-01 5.12928724e-01
1.89700499e-01 6.75741911e-01 -8.92267466e-01 -7.69149125e-01
-3.43471646e-01 1.69662103e-01 1.55870533e+00 9.17888999e-01
4.72375453e-01 -6.19028583e-02 6.27389431e-01 8.27677548e-01
7.90232778e-01 5.82389474e-01 5.53363800e-01 -1.11845195e+00
1.55652404e-01 7.68977404e-01 3.75993848e-01 -8.71631622e-01
-3.62881809e-01 1.85860366e-01 1.25176683e-01 -1.33090287e-01
5.00284433e-01 -3.63249093e-01 -2.43604124e-01 1.61290741e+00
1.64372593e-01 -8.28905463e-01 5.74511349e-01 6.39357805e-01
1.06011856e+00 7.57806778e-01 5.67369401e-01 -5.81256390e-01
1.78547812e+00 -2.07703084e-01 -8.92520070e-01 -6.91940188e-01
1.11019838e+00 -1.23402166e+00 1.55705726e+00 3.38874280e-01
-1.17938542e+00 -5.44241741e-02 -8.12671363e-01 -4.91079718e-01
-5.35637438e-01 -5.50330691e-02 1.03233349e+00 9.85138535e-01
-8.90075088e-01 3.02984327e-01 -6.89484060e-01 -9.79993045e-01
-1.82432532e-01 1.45828307e-01 -1.12074971e-01 2.70161033e-01
-1.19271970e+00 1.12002420e+00 6.64112985e-01 2.50528306e-01
3.29085328e-02 -1.30248085e-01 -1.09728181e+00 -2.42341787e-01
4.48278666e-01 -4.64041173e-01 1.82651758e+00 -1.16764975e+00
-1.53359437e+00 1.22780812e+00 -7.06416249e-01 -2.25244418e-01
-9.69706401e-02 -3.46532106e-01 -5.00023007e-01 2.26508841e-01
2.67065436e-01 5.39950669e-01 2.25476861e-01 -1.10628510e+00
-9.51018870e-01 -4.89901632e-01 3.82152945e-01 5.31076491e-01
1.76604707e-02 8.46839488e-01 -5.76678570e-03 -2.03509480e-01
3.51383477e-01 -6.38988256e-01 9.70820636e-02 -2.41174668e-01
-1.17443912e-01 -6.93335712e-01 7.36034691e-01 -6.52584851e-01
1.32157874e+00 -2.10309792e+00 -2.31654853e-01 2.51177028e-02
5.25640808e-02 2.89092124e-01 2.67710000e-01 1.04369688e+00
1.61079779e-01 4.02020693e-01 -3.34768444e-01 1.52409434e-01
4.97365117e-01 7.14155555e-01 -5.59179842e-01 2.45600224e-01
5.86679019e-02 7.57984459e-01 -1.01401746e+00 -4.53734368e-01
3.63227665e-01 3.51860762e-01 -9.01035815e-02 -3.03830020e-02
-3.58107299e-01 2.21420288e-01 -9.30872440e-01 3.92740965e-01
3.66995707e-02 -1.48263767e-01 3.94489408e-01 4.46503460e-01
-4.41894293e-01 1.23800993e+00 -1.00564492e+00 1.37652600e+00
-5.45498669e-01 7.21466124e-01 2.07350358e-01 -9.80895698e-01
7.74204850e-01 7.14986622e-01 -8.21434557e-02 -8.12849626e-02
2.44310319e-01 5.27877092e-01 3.81569535e-01 -7.10086524e-01
4.51094061e-01 -7.11074054e-01 -2.48847321e-01 8.72172296e-01
-3.23063314e-01 -5.86937368e-01 3.84187251e-01 2.77660728e-01
6.20388031e-01 1.44752651e-01 8.61384511e-01 -7.70316720e-01
9.26605701e-01 1.99875712e-01 2.91141033e-01 4.65267062e-01
-1.85342774e-01 -1.52721837e-01 6.99719608e-01 -5.88087916e-01
-7.41340935e-01 -5.72205544e-01 -3.89444321e-01 1.10261619e+00
-1.51424110e-01 -3.81237626e-01 -8.80013525e-01 -2.12703094e-01
-6.79433227e-01 1.53418219e+00 -1.81947187e-01 3.86467367e-01
-6.54665291e-01 -1.27586871e-01 2.43828088e-01 1.00893810e-01
3.20042282e-01 -1.67644691e+00 -9.65508759e-01 2.82166570e-01
-6.81745484e-02 -1.21830356e+00 -1.69651002e-01 1.80360883e-01
-8.01196754e-01 -9.29343462e-01 -2.01006636e-01 -1.01772833e+00
4.61537421e-01 -2.07490916e-03 1.02855754e+00 2.58643717e-01
3.64476413e-01 8.49974215e-01 -6.67923927e-01 -7.93397188e-01
-8.39276731e-01 5.94168156e-02 -2.45027944e-01 -4.37021106e-01
1.14105558e+00 -2.74053156e-01 -3.96861583e-02 -5.37867136e-02
-9.59430337e-01 -5.91170900e-02 1.97231352e-01 3.56636554e-01
7.97787011e-02 -9.44681466e-02 3.76947135e-01 -1.27145016e+00
9.63235557e-01 -1.43869713e-01 9.79998149e-03 3.41698974e-01
-2.54929483e-01 -8.24840739e-02 2.48034030e-01 1.91775367e-01
-1.33106399e+00 -1.03252128e-01 -4.72686678e-01 8.41413558e-01
-7.52510190e-01 6.39661014e-01 -5.19623578e-01 6.51259497e-02
6.29891574e-01 1.14627369e-01 1.31719217e-01 -4.01008427e-01
4.79975939e-01 9.31838274e-01 1.31153986e-01 -9.62675631e-01
3.27189296e-01 1.76832676e-01 -2.48882473e-01 -1.53894043e+00
-9.53468263e-01 -7.27483928e-01 -6.02020741e-01 8.92821625e-02
8.37795734e-01 -4.27438736e-01 -7.37496316e-01 -5.80550916e-02
-1.29994452e+00 -3.49124163e-01 -5.42837739e-01 6.20642006e-01
-7.10075498e-01 5.50188959e-01 -4.55989629e-01 -1.07156849e+00
-1.61446080e-01 -8.25776339e-01 8.96224797e-01 1.93156093e-01
-1.02315807e+00 -1.54029131e+00 -4.00814384e-01 2.96417534e-01
1.53467283e-01 -2.53231674e-01 1.14783466e+00 -1.13179672e+00
5.98567538e-04 -1.85883701e-01 2.73968242e-02 3.57115716e-01
3.06010246e-01 7.60885924e-02 -7.42870808e-01 2.02953130e-01
4.70290095e-01 -3.40821236e-01 -3.12488433e-02 4.25297141e-01
4.33767855e-01 -5.26362896e-01 -2.01861665e-01 -2.00758368e-01
1.31765270e+00 4.11662430e-01 5.12402177e-01 3.36161107e-01
1.02192059e-01 1.32645202e+00 5.70912361e-01 1.22227401e-01
6.30050719e-01 1.26523718e-01 -3.19815993e-01 4.70592469e-01
1.12488776e-01 -1.74850851e-01 3.06985617e-01 7.71211267e-01
1.25964403e-01 -2.29222357e-01 -1.25673282e+00 6.70489728e-01
-1.93539155e+00 -8.67239475e-01 -2.32284769e-01 1.91088080e+00
1.00040817e+00 1.54104397e-01 -1.37621894e-01 7.83606246e-02
5.63370645e-01 2.27827743e-01 1.28665462e-01 -1.20234895e+00
5.47845103e-02 -5.97865181e-03 -4.93102372e-02 9.89453793e-01
-8.81738961e-01 1.47967803e+00 6.56186581e+00 4.24605638e-01
-9.17764783e-01 -1.86742410e-01 3.30081314e-01 6.12469196e-01
-7.01256454e-01 5.19289196e-01 -8.13041031e-01 2.55441461e-02
1.11341012e+00 -5.10749340e-01 4.85919893e-01 5.62101722e-01
7.35499680e-01 -3.54527235e-01 -1.18707526e+00 6.76828027e-01
-1.14603519e-01 -9.55025434e-01 -1.45696491e-01 -1.37340441e-01
5.29183559e-02 -1.27388418e-01 -3.77367765e-01 5.50865903e-02
4.84294176e-01 -9.21274304e-01 4.31311190e-01 2.61299461e-01
1.69138625e-01 -6.05228245e-01 7.59552419e-01 8.67093801e-01
-9.39912438e-01 3.61736864e-02 -3.49018872e-01 -7.17222631e-01
5.68068504e-01 2.35839531e-01 -7.46575415e-01 1.33989841e-01
2.47004747e-01 3.94790411e-01 7.38505572e-02 5.74125946e-01
-7.10868716e-01 6.94662690e-01 -6.52951717e-01 -8.69505286e-01
4.61527854e-01 -7.52553791e-02 7.17653215e-01 1.34838963e+00
-1.18467294e-01 8.53946388e-01 1.53144851e-01 6.11653507e-01
5.91801822e-01 6.14210665e-01 -8.46293807e-01 -5.32667935e-01
4.88502562e-01 1.05725789e+00 -8.48609805e-01 -3.12488586e-01
-6.39270008e-01 4.10832316e-01 -2.19728187e-01 2.97873318e-01
9.37359631e-02 -5.13676882e-01 5.50244808e-01 2.03268185e-01
-7.72492886e-02 -4.21600640e-01 -4.57222283e-01 -8.88230443e-01
-2.12562103e-02 -9.05360043e-01 3.09788913e-01 -7.57356882e-01
-1.08138335e+00 4.27403808e-01 3.43921095e-01 -5.47967553e-01
-6.97551906e-01 -9.21390474e-01 -5.69415748e-01 1.09409654e+00
-1.25741160e+00 -1.22023237e+00 4.13894355e-01 -1.84675341e-03
6.13402903e-01 3.83447260e-02 1.09183073e+00 -4.82215673e-01
8.62974487e-03 -5.20305447e-02 -5.51994979e-01 4.01916094e-02
1.38256818e-01 -1.27525949e+00 3.64819646e-01 7.28312135e-01
3.21026117e-01 1.27049243e+00 1.33350277e+00 -4.50926363e-01
-1.43756890e+00 -1.78711891e-01 1.90400767e+00 -6.39800549e-01
9.91410792e-01 1.50239274e-01 -9.23239112e-01 1.00303173e+00
5.63483655e-01 -4.99069840e-01 9.84569311e-01 2.94888884e-01
1.91012695e-01 1.56358778e-01 -1.28985023e+00 6.95379734e-01
6.45434856e-01 -5.56016743e-01 -1.64992917e+00 5.39614737e-01
9.21483517e-01 -2.86974400e-01 -6.66081905e-01 -3.16100568e-02
4.37201619e-01 -4.77499396e-01 4.80280519e-01 -7.05401719e-01
-6.53613210e-02 -3.98511112e-01 -2.18801364e-01 -7.88487673e-01
2.73289651e-01 -1.02183700e+00 7.49322593e-01 1.18264794e+00
7.45573103e-01 -1.16077030e+00 5.31975508e-01 1.32928038e+00
-2.75747299e-01 -2.50733048e-01 -7.96634197e-01 -2.93315262e-01
3.77742290e-01 -9.30449426e-01 3.17396104e-01 8.01370561e-01
9.43538368e-01 8.03817153e-01 3.53698581e-01 -6.98839501e-02
3.54938626e-01 -1.31836250e-01 6.42795205e-01 -1.30451989e+00
-7.58548602e-02 -4.31954414e-01 -8.71313065e-02 -1.47422552e+00
2.78246641e-01 -6.24572456e-01 2.38547936e-01 -1.73085070e+00
-3.13711107e-01 -2.73259878e-01 2.80346692e-01 6.06797040e-01
3.18566293e-01 -3.90279651e-01 1.34905308e-01 3.11625034e-01
-2.69873649e-01 2.81057626e-01 1.23780179e+00 4.16061759e-01
-4.84273195e-01 2.14179546e-01 -1.20446682e+00 1.33481991e+00
8.23398292e-01 -3.76367390e-01 -6.83490276e-01 -3.97721261e-01
5.71343720e-01 -7.65454862e-03 -1.90653019e-02 -4.69649285e-01
3.07902575e-01 -6.08502626e-01 -2.15568990e-01 -4.24104750e-01
1.90583870e-01 -8.17276120e-01 -2.29685828e-01 3.18534881e-01
-5.91475964e-01 2.68401057e-02 2.99701661e-01 3.06216888e-02
-2.97174633e-01 -9.80044544e-01 3.55535209e-01 -6.37770653e-01
-1.06954086e+00 -3.09096426e-01 -7.68603861e-01 1.70600578e-01
9.61360037e-01 -3.89237165e-01 -4.99416627e-02 -6.12144113e-01
-8.16257775e-01 3.04276347e-01 4.45514470e-01 1.44336879e-01
5.86649835e-01 -6.36952579e-01 -3.64635170e-01 -4.96940687e-02
1.09713182e-01 -1.59055844e-01 -2.34974585e-02 5.68161130e-01
-8.67904186e-01 1.02461684e+00 2.59582192e-01 -1.76667720e-01
-1.22149217e+00 2.72079051e-01 1.99822858e-01 2.85782516e-01
-8.74311507e-01 8.50660861e-01 2.77372330e-01 -5.86455941e-01
1.52705591e-02 -4.09655601e-01 -6.87103152e-01 -1.60945147e-01
8.74538779e-01 -3.24001491e-01 -4.39696193e-01 -8.92725229e-01
-5.00484228e-01 4.59483087e-01 -1.38551280e-01 -5.47395110e-01
1.15968406e+00 -6.77976966e-01 -4.63654637e-01 9.25921023e-01
8.29515636e-01 1.05468452e-01 -5.83728135e-01 -3.02471906e-01
4.88434374e-01 -2.76276618e-01 -1.43616453e-01 -7.42675662e-01
-1.26504302e-01 9.93248105e-01 -6.62322342e-02 6.38390601e-01
7.05370843e-01 3.48758370e-01 7.83860028e-01 8.83571029e-01
3.80365223e-01 -1.40136671e+00 -3.96792829e-01 7.12559164e-01
1.07763815e+00 -9.93043780e-01 1.76300369e-02 -6.34956241e-01
-9.92868543e-01 1.48408091e+00 7.67701268e-02 2.47360051e-01
1.03456700e+00 8.79488513e-02 4.98862475e-01 -5.11551440e-01
-9.32762861e-01 -3.09745580e-01 -1.75858617e-01 7.04437733e-01
9.95594859e-01 9.19154361e-02 -1.14558673e+00 2.69133776e-01
-5.19740939e-01 4.44226637e-02 6.02998376e-01 1.23971975e+00
-1.10465014e+00 -1.46927702e+00 -1.68866619e-01 9.44692641e-02
-6.42394483e-01 -2.85137236e-01 -7.94761598e-01 1.17182815e+00
-2.43473127e-01 1.24129379e+00 -9.76778418e-02 3.86831999e-01
2.01941833e-01 3.01409364e-01 1.98206842e-01 -1.19194198e+00
-7.19660163e-01 1.77467883e-01 8.22090328e-01 -9.74753276e-02
-6.50636017e-01 -9.16937530e-01 -1.66087317e+00 -3.79294366e-01
3.27452854e-03 6.00690961e-01 7.68431544e-01 1.40924215e+00
-2.89402783e-01 -6.42447844e-02 -4.05752622e-02 -2.78379440e-01
-4.48893487e-01 -1.03883374e+00 -7.44880259e-01 -7.14441296e-04
6.93446994e-02 -8.33734497e-02 -3.86328459e-01 4.93184999e-02] | [10.40298843383789, 8.691010475158691] |
99dcee22-7701-46a1-b400-02b3fae7dd42 | robust-controlled-table-to-text-generation | null | null | https://openreview.net/forum?id=VBZCrsaUpsM | https://openreview.net/pdf?id=VBZCrsaUpsM | Robust (Controlled) Table-to-Text Generation with Structure-Aware Equivariance Learning | Controlled table-to-text generation seeks to generate natural language descriptions for highlighted subparts of a table. Previous SOTA systems still employ a sequence-to-sequence generation method, which merely captures the table as a linear structure and is brittle when table layouts change. We seek to go beyond this paradigm by (1) effectively expressing the relations of content pieces in the table, and (2) making our model robust to content-invariant structural transformations. Accordingly, we propose an equivariance learning framework, encoding tables with a structure-aware self-attention mechanism. This prunes the full self-attention structure into an order-invariant graph attention that captures the connected graph structure of cells belonging to the same row or column, and it differentiates between relevant cells and irrelevant cells from the structural perspective. Our framework also modifies the positional encoding mechanism to preserve the relative position of tokens in the same cell but enforce position invariance among different cells. Our technology is free to be plugged into existing table-to-text generation models, and has improved T5-based models to offer better performance on ToTTo and HiTab. Moreover, on a harder version of ToTTo, we preserve promising performance, while previous SOTA systems, even with transformation-based data augmentation, have seen significant performance drops. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 5.46176016e-01 6.03007138e-01 -1.88986391e-01 -9.38735902e-02
-5.78892827e-01 -7.71224797e-01 6.52489364e-01 3.05432111e-01
2.21713781e-01 8.24481428e-01 5.88096380e-01 -5.08212090e-01
2.56796330e-01 -1.34597921e+00 -1.01015842e+00 -3.50173354e-01
1.58861473e-01 7.25214243e-01 2.27898583e-01 -6.83813393e-01
1.85363457e-01 4.37154084e-01 -1.27986419e+00 7.11188436e-01
8.24208677e-01 6.70365930e-01 -8.67727026e-02 6.40969992e-01
-5.34293890e-01 1.00656474e+00 -6.00073576e-01 -6.29075825e-01
3.38097900e-01 -7.39706099e-01 -9.94286597e-01 1.48156807e-01
8.31308186e-01 -1.23192891e-01 -5.79487979e-01 8.76564860e-01
2.50792027e-01 -2.02989385e-01 4.91961509e-01 -1.28095543e+00
-1.23707569e+00 1.16201985e+00 -4.77120191e-01 -8.18362460e-02
4.09594119e-01 2.26316258e-01 1.34014499e+00 -5.98477006e-01
8.40573311e-01 1.36815226e+00 4.60180849e-01 4.02592391e-01
-1.48193657e+00 -3.68325621e-01 4.43878204e-01 6.58674538e-02
-1.39249611e+00 -5.15335917e-01 6.84577525e-01 -2.22284138e-01
1.31087327e+00 5.31128526e-01 7.08792627e-01 9.34407771e-01
5.68939745e-01 6.61992490e-01 5.93301892e-01 -4.62198764e-01
8.19255114e-02 -3.10232323e-02 -2.35307291e-02 8.21030974e-01
5.86125851e-01 -3.83324355e-01 -4.24811393e-01 1.82633519e-01
8.62349331e-01 -2.74189979e-01 -1.32088512e-01 -7.48683035e-01
-1.35915387e+00 6.27305984e-01 6.26861155e-01 3.14713001e-01
-4.33046110e-02 2.27343902e-01 4.18816924e-01 2.90200442e-01
1.31185204e-01 8.39337409e-01 -4.52965796e-01 2.24167388e-02
-5.85118413e-01 5.52497387e-01 8.81392956e-01 1.58609521e+00
8.23086381e-01 1.61376953e-01 -7.86080778e-01 3.17320347e-01
-2.02383354e-01 2.74947852e-01 4.06931549e-01 -5.10429263e-01
7.43346810e-01 1.04460466e+00 -4.40937243e-02 -1.07421851e+00
-1.81176886e-01 -5.80480874e-01 -8.75891507e-01 -1.67414844e-01
1.81910843e-01 3.32096592e-03 -1.05520308e+00 1.77727282e+00
-4.24371697e-02 -2.71954834e-01 -1.02625675e-01 4.38743293e-01
6.84760153e-01 5.28471768e-01 -1.96500063e-01 4.61653322e-02
1.35754883e+00 -1.05389631e+00 -8.49875927e-01 -4.40369517e-01
9.10253227e-01 -3.87164444e-01 1.37541783e+00 7.97578618e-02
-1.34712148e+00 -6.55737698e-01 -1.30207002e+00 -5.35460293e-01
-8.22772563e-01 -2.59421647e-01 6.41090870e-01 5.95330536e-01
-1.33373737e+00 4.22855407e-01 -4.57344323e-01 -2.56444603e-01
1.99606255e-01 3.18152815e-01 -3.66412669e-01 -5.38454838e-02
-1.29222524e+00 8.15134346e-01 4.84846860e-01 -1.09232858e-01
-2.70363837e-01 -6.83764398e-01 -1.32365882e+00 5.02198756e-01
7.70601571e-01 -1.05635583e+00 9.52054679e-01 -6.16357684e-01
-1.15595281e+00 6.91180766e-01 -3.23806554e-01 -4.62456524e-01
2.99326301e-01 7.63597935e-02 -4.02518123e-01 -2.59921283e-01
2.03577995e-01 8.26523781e-01 5.24472892e-01 -1.38206756e+00
-4.62996334e-01 -2.83263654e-01 1.36013523e-01 2.82145470e-01
-1.80311441e-01 -4.81649339e-01 -8.01287591e-01 -9.39406633e-01
1.36134356e-01 -9.07325387e-01 -1.24020576e-01 -2.53544480e-01
-9.63923395e-01 1.47167683e-01 6.46488070e-01 -4.67418700e-01
1.70302737e+00 -1.86113095e+00 2.29262695e-01 7.05986619e-02
3.55276674e-01 8.25825520e-03 -2.51370221e-01 6.62635624e-01
-2.58728653e-01 5.80392182e-01 -3.17537457e-01 -8.84305835e-02
1.74821138e-01 1.57556966e-01 -3.90350938e-01 -1.23691007e-01
5.16411185e-01 1.41987538e+00 -9.35330808e-01 -4.51478660e-01
-1.12448409e-01 1.12534173e-01 -1.07777596e+00 -1.12497538e-01
-4.57757413e-01 -1.62456855e-01 -2.55962163e-01 5.76820970e-01
5.61589897e-01 -2.51127571e-01 4.43666756e-01 -3.09939772e-01
-2.25503687e-02 6.04042411e-01 -1.12855899e+00 1.56649590e+00
-3.70927483e-01 3.51976097e-01 -1.93569928e-01 -4.85411495e-01
8.58270228e-01 -1.80304021e-01 2.05969617e-01 -8.73998225e-01
-1.68841034e-01 -1.59147963e-01 2.29425341e-01 -5.89222163e-02
1.10888338e+00 1.87224910e-01 -5.66805899e-01 3.16995531e-01
-2.14839652e-01 -3.91707808e-01 5.35024822e-01 6.64740145e-01
1.17827165e+00 2.76437551e-01 3.52399826e-01 -3.56268406e-01
3.66342694e-01 7.50468895e-02 3.88409585e-01 8.72808337e-01
2.32504323e-01 6.77735090e-01 8.79529536e-01 -4.71960545e-01
-1.14698195e+00 -1.00426853e+00 2.92344391e-01 1.01041508e+00
6.79211989e-02 -8.04014921e-01 -8.76849771e-01 -7.43115842e-01
2.15338990e-01 1.01013124e+00 -9.70212042e-01 -4.99412894e-01
-8.24274778e-01 -5.24239957e-01 4.91688251e-01 7.31026888e-01
2.84724712e-01 -1.11564064e+00 -8.52312148e-02 3.78359854e-01
-1.44994557e-01 -9.26391959e-01 -1.02500379e+00 3.70521843e-01
-5.42052507e-01 -8.01812410e-01 -1.32454008e-01 -7.62879610e-01
8.40895295e-01 4.17649150e-02 1.39301836e+00 8.63639861e-02
-2.55354610e-03 4.43002321e-02 -2.13674799e-01 -3.88200760e-01
-5.83684742e-01 6.23825431e-01 -2.72028387e-01 -2.57328991e-02
6.80621490e-02 -3.78682971e-01 -1.93104625e-01 2.05453441e-01
-1.03738546e+00 3.36957574e-01 5.33626437e-01 7.80169904e-01
5.89519203e-01 -5.16451448e-02 4.07037079e-01 -1.32064366e+00
6.78682029e-01 -2.38752499e-01 -3.46798122e-01 4.28531498e-01
-7.67962933e-01 5.88927090e-01 8.43013287e-01 -1.06285907e-01
-8.11169028e-01 -3.07917949e-02 1.52877256e-01 -1.93442360e-01
8.49749520e-02 4.35560971e-01 -8.09589326e-01 3.64890993e-01
5.16811728e-01 4.44023937e-01 -1.56659007e-01 -1.72864109e-01
7.44858086e-01 1.14731744e-01 7.22334564e-01 -6.22025132e-01
1.05604470e+00 2.90268242e-01 2.65437830e-02 -3.06050688e-01
-5.23406804e-01 2.05500230e-01 -8.64928246e-01 3.39425057e-01
6.57510936e-01 -6.64901495e-01 -5.14729559e-01 2.53736734e-01
-9.90165710e-01 -5.34514844e-01 -5.92611253e-01 -4.38860804e-01
-3.95932853e-01 3.02983433e-01 -7.82873452e-01 -3.97181988e-01
-3.16173971e-01 -1.01789415e+00 1.13751173e+00 -4.84020859e-02
-5.25974810e-01 -9.56513047e-01 -1.26971275e-01 2.74283066e-03
4.10713673e-01 2.07318768e-01 1.53359103e+00 -5.77844918e-01
-9.76900160e-01 -9.11524519e-02 -3.07920158e-01 -2.27006555e-01
4.50730920e-01 2.01321170e-01 -4.85655069e-01 -2.56548464e-01
-5.11819124e-01 9.03716087e-02 8.54900539e-01 2.18011178e-02
1.18321228e+00 -7.22477436e-01 -4.70791429e-01 7.39555240e-01
1.18808103e+00 3.64540517e-01 9.18505251e-01 3.22022885e-01
1.21523666e+00 5.07701397e-01 1.18933260e-01 2.21564502e-01
7.34131336e-01 7.20358849e-01 3.95369619e-01 -2.39474252e-01
-4.51359242e-01 -9.09429133e-01 2.43983299e-01 7.10879087e-01
5.05753696e-01 -8.01443458e-01 -7.37634301e-01 6.80613637e-01
-1.59118521e+00 -8.52363527e-01 -1.85550466e-01 2.05039191e+00
8.57978404e-01 5.28800428e-01 -1.72154188e-01 1.86283946e-01
6.20342255e-01 2.68677503e-01 -5.13310015e-01 -6.70110047e-01
-3.15838367e-01 2.03287557e-01 6.00408971e-01 5.24712384e-01
-7.15833008e-01 1.30951607e+00 6.72467852e+00 6.61649883e-01
-8.98580432e-01 -4.53941852e-01 7.93644309e-01 -5.59218042e-02
-7.87340820e-01 1.40910794e-03 -8.31803977e-01 2.74102360e-01
6.17782593e-01 -4.35610294e-01 5.96393287e-01 5.90319633e-01
-3.23525071e-02 2.24420264e-01 -1.46503651e+00 5.51298201e-01
2.80476898e-01 -1.56725097e+00 8.22600305e-01 1.76031828e-01
6.24554455e-01 -6.42016768e-01 5.18512279e-02 5.72936893e-01
4.89140481e-01 -1.18562639e+00 9.61638153e-01 3.69841486e-01
1.03875411e+00 -8.38477969e-01 4.90777493e-01 -9.91429538e-02
-1.42077780e+00 7.20338821e-02 -1.16019636e-01 -1.86620533e-01
-1.21371401e-02 1.39696971e-01 -9.60013747e-01 7.39808381e-01
3.93924922e-01 5.12524486e-01 -9.50606048e-01 3.82536054e-01
-1.34894893e-01 1.70886219e-01 1.74942896e-01 -5.14876284e-02
2.00490877e-01 -1.82220474e-01 4.12528962e-01 1.16327083e+00
3.00032437e-01 -1.53380737e-01 8.48065615e-02 1.14839745e+00
-3.68974626e-01 1.05949650e-02 -8.38309288e-01 -1.10538818e-01
5.46934903e-01 9.75829363e-01 -8.49230826e-01 -6.26132607e-01
-2.79793710e-01 1.08512437e+00 3.47804904e-01 2.66143113e-01
-7.36264110e-01 -5.01222432e-01 6.19719505e-01 6.02987826e-01
5.47860563e-01 -1.06780529e-01 -8.69701803e-01 -1.02889574e+00
-1.56540722e-02 -1.14075649e+00 3.63066286e-01 -9.88369465e-01
-8.71724010e-01 8.19457531e-01 -4.55459394e-02 -8.35235655e-01
-2.80335754e-01 -3.42037290e-01 -4.55213904e-01 9.68490839e-01
-9.91785467e-01 -1.26342511e+00 -8.43578801e-02 3.77205789e-01
4.94360954e-01 -7.92841334e-03 7.92999506e-01 2.72684153e-02
-7.30768025e-01 1.10756040e+00 -2.64451504e-01 3.52401108e-01
6.74362242e-01 -1.56260586e+00 1.30084062e+00 1.07694268e+00
2.04598457e-01 1.02606261e+00 5.15724063e-01 -8.92110407e-01
-1.64432800e+00 -1.25710511e+00 1.13715720e+00 -7.91216552e-01
6.13579512e-01 -8.44385982e-01 -1.07980835e+00 1.01251042e+00
3.68304908e-01 -2.34961078e-01 2.41148114e-01 2.24193484e-02
-5.32917142e-01 -1.51610374e-01 -7.42445827e-01 1.16899478e+00
1.44973457e+00 -4.80287254e-01 -4.25886571e-01 6.40312284e-02
1.06514120e+00 -6.40252650e-01 -5.89886606e-01 2.79085040e-01
3.63938361e-01 -7.87709117e-01 7.94068158e-01 -7.17846990e-01
6.74212813e-01 -4.14766610e-01 3.94699387e-02 -1.41868067e+00
-8.42008591e-01 -8.50580990e-01 -2.57692635e-01 1.26135182e+00
7.77084291e-01 -4.94243264e-01 1.03280175e+00 5.75883210e-01
-4.75423992e-01 -5.69658101e-01 -5.26440203e-01 -6.05321467e-01
1.91446155e-01 6.34261966e-03 1.12353265e+00 1.06658649e+00
2.35163286e-01 7.58790851e-01 -3.23114812e-01 -8.98028165e-02
9.03137550e-02 1.46657228e-01 8.86454880e-01 -9.57677841e-01
-2.80451238e-01 -7.36883342e-01 5.18780947e-02 -1.16847372e+00
-1.18432678e-01 -1.09358466e+00 -1.23467082e-02 -1.82555580e+00
1.23116635e-01 -2.13953882e-01 -9.29595679e-02 7.61177778e-01
-4.05321360e-01 6.72549009e-02 3.07571083e-01 -1.74684823e-01
-3.28138053e-01 4.54026937e-01 1.56943345e+00 -4.08857197e-01
-2.48149872e-01 -5.15540481e-01 -1.38257480e+00 1.33395389e-01
6.40852034e-01 -3.44845742e-01 -6.94032550e-01 -5.33875465e-01
3.88657153e-01 3.69909294e-02 -1.51589140e-01 -7.88410306e-01
1.02705285e-01 -1.68231562e-01 4.32031333e-01 -5.79901159e-01
-4.76575755e-02 -5.76682448e-01 1.78869709e-01 4.88164157e-01
-6.08142853e-01 6.11347377e-01 4.12669867e-01 3.56171042e-01
-2.55911499e-02 1.79578364e-01 4.95691478e-01 -2.89032459e-01
-5.04973114e-01 3.38134557e-01 -4.16075915e-01 1.59756437e-01
6.82160914e-01 -4.56781745e-01 -5.54017067e-01 -4.64528859e-01
-3.33323210e-01 2.37449437e-01 8.29430997e-01 5.92382312e-01
3.28526765e-01 -1.40542877e+00 -6.33525968e-01 5.79860091e-01
3.43670368e-01 2.29057133e-01 9.26375389e-02 1.62211969e-01
-5.98443329e-01 6.37949407e-01 -3.29638660e-01 -1.31560251e-01
-8.72377872e-01 9.19742465e-01 2.15609729e-01 -7.02491760e-01
-4.93497223e-01 5.94804943e-01 4.76104945e-01 -3.49777043e-01
-2.06449796e-02 -8.19707155e-01 5.75057231e-02 -4.30212170e-02
3.75765830e-01 2.25816425e-02 3.30042094e-01 -4.61248219e-01
-2.17294484e-01 1.58473283e-01 -4.28868890e-01 1.33506060e-02
8.85917366e-01 -4.20924462e-02 -1.08385742e-01 1.57568574e-01
8.02307367e-01 4.70921248e-01 -1.14367747e+00 -8.19569677e-02
1.84987392e-02 -1.82806030e-01 -1.24596134e-01 -1.06466091e+00
-9.18126404e-01 6.09911799e-01 -2.18797863e-01 3.71998310e-01
8.72531414e-01 -3.17664117e-01 8.21035743e-01 3.66691500e-01
2.15126678e-01 -7.71775901e-01 1.63147584e-01 7.46302068e-01
1.13859510e+00 -7.34162629e-01 -2.92122792e-02 -6.86383545e-01
-7.14871168e-01 7.66463578e-01 9.60219860e-01 1.67809233e-01
8.29758029e-03 7.19784915e-01 -2.20106214e-01 -9.13444981e-02
-1.03410661e+00 -1.94963306e-01 2.85993010e-01 7.25093365e-01
5.49435973e-01 -6.15856331e-03 -1.46680987e-02 5.53196490e-01
-5.03959894e-01 -3.22431862e-01 7.44635463e-01 8.49562228e-01
-1.40403062e-01 -1.33054185e+00 -2.48280987e-01 4.81912315e-01
-1.33156031e-01 -6.35122895e-01 -8.63832295e-01 1.15679538e+00
5.87171651e-02 6.77267849e-01 3.72126698e-01 -4.96435523e-01
6.80573165e-01 3.03738266e-01 4.08316225e-01 -9.06784475e-01
-6.88363612e-01 9.83627588e-02 2.58099884e-01 -3.78355950e-01
4.42272872e-01 -5.45264781e-01 -1.24295294e+00 -5.44540584e-01
1.22305207e-01 6.72946647e-02 5.04839420e-02 4.58155304e-01
6.88090801e-01 1.12325978e+00 3.58530432e-01 -4.21924680e-01
-1.84186816e-01 -8.80518377e-01 -4.92242992e-01 4.98661667e-01
1.86628163e-01 -2.93559968e-01 1.15356512e-01 9.78288651e-02] | [10.855363845825195, 8.48784351348877] |
59c27a33-ff68-4d82-944c-753f69ea2517 | a-distributional-view-on-multi-objective | 2005.07513 | null | https://arxiv.org/abs/2005.07513v1 | https://arxiv.org/pdf/2005.07513v1.pdf | A Distributional View on Multi-Objective Policy Optimization | Many real-world problems require trading off multiple competing objectives. However, these objectives are often in different units and/or scales, which can make it challenging for practitioners to express numerical preferences over objectives in their native units. In this paper we propose a novel algorithm for multi-objective reinforcement learning that enables setting desired preferences for objectives in a scale-invariant way. We propose to learn an action distribution for each objective, and we use supervised learning to fit a parametric policy to a combination of these distributions. We demonstrate the effectiveness of our approach on challenging high-dimensional real and simulated robotics tasks, and show that setting different preferences in our framework allows us to trace out the space of nondominated solutions. | ['Nicolas Heess', 'Leonard Hasenclever', 'Sandy H. Huang', 'Martin Riedmiller', 'Abbas Abdolmaleki', 'Murilo F. Martins', 'Raia Hadsell', 'Michael Neunert', 'H. Francis Song', 'Martina Zambelli'] | 2020-05-15 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [ 1.82599932e-01 -1.72262222e-01 -2.32029364e-01 -2.06290156e-01
-8.09881330e-01 -8.53935421e-01 3.30373257e-01 6.97175562e-02
-8.07394445e-01 1.08185089e+00 3.75933796e-02 -9.62104946e-02
-7.39923298e-01 -5.57557344e-01 -6.02231205e-01 -7.04563141e-01
-1.66170642e-01 8.60208392e-01 1.11097872e-01 -2.01161385e-01
6.06034696e-01 4.98927712e-01 -1.43674672e+00 -1.29283473e-01
1.13219726e+00 8.03235352e-01 4.58702713e-01 8.39699268e-01
1.70556784e-01 2.76741982e-01 -7.64315605e-01 -2.02568639e-02
6.55904412e-01 -1.65376633e-01 -6.96656883e-01 2.38179982e-01
1.26818061e-01 -2.47368023e-01 2.29210541e-01 1.17543185e+00
4.84373987e-01 5.40235698e-01 8.41290057e-01 -1.50833535e+00
-3.91494453e-01 4.48373646e-01 -5.79757094e-01 -1.22272983e-01
2.17323616e-01 4.46146280e-01 1.15378559e+00 -1.61003157e-01
3.88815820e-01 1.46502256e+00 3.35978836e-01 3.23809326e-01
-1.47130084e+00 -2.45002046e-01 3.40354860e-01 -2.19829455e-01
-1.03661239e+00 -1.04569763e-01 6.18512511e-01 -3.28932196e-01
6.29944801e-01 2.90326513e-02 5.74444652e-01 7.97368646e-01
4.91907358e-01 6.39163971e-01 1.44923604e+00 -3.01757812e-01
5.97607791e-01 -4.47912887e-02 -4.67853040e-01 3.66429090e-01
2.72574514e-01 2.91944772e-01 -2.02899113e-01 -3.08786780e-01
1.04441059e+00 -5.41160218e-02 -5.73685169e-02 -9.40055907e-01
-1.24603629e+00 1.03717256e+00 2.13735178e-01 -1.11117601e-01
-5.35829902e-01 4.05530959e-01 9.83212069e-02 5.38734615e-01
-2.16664076e-02 1.14251196e+00 -6.22083426e-01 -2.28719085e-01
-4.02355909e-01 6.21185064e-01 7.65099704e-01 6.20866656e-01
7.26687551e-01 -1.83104932e-01 -2.68552601e-01 9.60343361e-01
2.19256788e-01 3.01259339e-01 3.98270547e-01 -1.55590296e+00
4.02581692e-01 3.02599967e-01 9.42392826e-01 -7.15040326e-01
-4.33966696e-01 -3.12637866e-01 -3.04591954e-01 9.64978814e-01
7.23768890e-01 -4.95209605e-01 -9.51690018e-01 1.89073431e+00
4.42243636e-01 -2.67329276e-01 1.16234824e-01 1.05794621e+00
-3.43969107e-01 4.59087759e-01 1.07652424e-02 -1.77457690e-01
8.30089569e-01 -9.71082568e-01 -4.82179761e-01 -5.27812600e-01
3.46168995e-01 -6.53085530e-01 1.49481177e+00 5.00734687e-01
-1.19277585e+00 -1.09131321e-01 -1.01690924e+00 3.57938796e-01
-9.21465978e-02 -9.46190208e-02 4.43320632e-01 3.59450281e-01
-9.82477486e-01 9.86224294e-01 -8.03152621e-01 -5.58711067e-02
2.35166773e-01 7.05354393e-01 7.63420537e-02 2.19235972e-01
-8.73720348e-01 1.26613128e+00 6.86719179e-01 -1.85697272e-01
-9.54020679e-01 -5.04612982e-01 -7.14566290e-01 -1.84732899e-02
8.91106069e-01 -5.41502893e-01 1.46981347e+00 -1.01546681e+00
-1.88764429e+00 4.21937585e-01 3.68489504e-01 -2.97904640e-01
7.34844506e-01 -1.27953529e-01 -2.52122171e-02 3.93509567e-02
5.58058582e-02 7.34798074e-01 9.79279339e-01 -1.11297202e+00
-7.82505691e-01 -2.33141810e-01 5.25670826e-01 6.58702195e-01
-4.46501613e-01 -6.55389503e-02 -4.21100482e-02 -4.84364003e-01
-2.50367135e-01 -9.77730036e-01 -7.76344717e-01 2.39825219e-01
-5.58406338e-02 -9.23616737e-02 3.65359634e-01 -9.14841518e-02
9.20026004e-01 -1.89752412e+00 6.66442215e-01 2.34702319e-01
4.34910432e-02 -1.82944968e-01 -3.55054080e-01 1.11355178e-01
2.94950753e-01 -2.62322240e-02 -3.73788208e-01 -1.32675052e-01
4.53119248e-01 6.29383802e-01 -6.14246763e-02 4.79956776e-01
1.43679053e-01 5.77219486e-01 -1.17345321e+00 -3.81752610e-01
1.11984216e-01 1.41900703e-01 -8.06293428e-01 2.78410167e-01
-5.04922628e-01 3.37731332e-01 -7.17558861e-01 3.84619236e-01
3.91443014e-01 -3.98723632e-02 2.97677636e-01 3.30254078e-01
-1.46537036e-01 4.84574176e-02 -1.67867613e+00 1.50552511e+00
-5.90812862e-01 6.59321770e-02 3.24209571e-01 -1.14642465e+00
8.29254270e-01 -5.47030717e-02 5.83800018e-01 -3.74386191e-01
2.59889185e-01 3.25886548e-01 3.53321545e-02 -2.18092561e-01
5.15769005e-01 -4.61644053e-01 -3.01433444e-01 5.26370406e-01
3.25894989e-02 -7.00663567e-01 4.21828389e-01 -4.93373632e-01
8.90235722e-01 1.38169453e-01 4.44372177e-01 -5.58056056e-01
2.38895550e-01 -1.06474057e-01 6.25677347e-01 8.96719992e-01
-3.36760044e-01 4.40176815e-01 8.24543893e-01 -2.65953302e-01
-1.01281309e+00 -1.13166094e+00 9.60380360e-02 1.16383004e+00
1.62727267e-01 1.51106194e-01 -4.36083019e-01 -7.51133859e-01
4.11750138e-01 7.22055793e-01 -5.66627860e-01 -1.80100024e-01
-6.13202691e-01 -6.30862832e-01 9.05639678e-02 5.46404004e-01
-1.78332478e-02 -1.02603936e+00 -1.13387692e+00 4.13271070e-01
2.55529553e-01 -8.01449001e-01 -8.13409567e-01 6.33698702e-01
-8.74040782e-01 -1.01044726e+00 -8.08528423e-01 -5.74164212e-01
6.52281463e-01 -6.67159855e-02 9.85778391e-01 -3.57288569e-01
-6.66604713e-02 3.63976300e-01 2.25925539e-03 -4.03603733e-01
-4.40941840e-01 -6.83033168e-02 3.78893524e-01 -8.39409307e-02
-3.55308324e-01 -4.58344907e-01 -4.49618459e-01 4.95157838e-01
-1.00131118e+00 -2.66624033e-01 5.46519220e-01 8.51548314e-01
6.02882564e-01 1.76503107e-01 4.80648428e-01 -5.04598975e-01
1.12536860e+00 -3.10426980e-01 -1.28971362e+00 3.48860562e-01
-5.78214288e-01 7.40129471e-01 7.78261960e-01 -9.47477460e-01
-7.10503638e-01 1.01993129e-01 4.14972007e-01 -5.26477098e-01
1.09784134e-01 4.51047778e-01 -1.39638856e-01 -2.10924134e-01
5.53062916e-01 -3.10043752e-01 2.53904760e-01 -1.98299453e-01
5.46072125e-01 3.54008675e-01 4.22131032e-01 -1.22989047e+00
6.81464672e-01 1.05826989e-01 -3.15161273e-02 -3.32662076e-01
-7.10420072e-01 -1.17844269e-02 -4.17121291e-01 -1.19247846e-01
6.01452768e-01 -4.14598525e-01 -8.41246367e-01 2.99624771e-01
-7.59052753e-01 -6.89498901e-01 -5.67625582e-01 4.71740186e-01
-1.26148534e+00 5.78023866e-02 -4.73553389e-02 -8.29089105e-01
2.00550020e-01 -1.46436489e+00 7.81623244e-01 4.10212845e-01
-2.94642687e-01 -1.09571731e+00 2.81383812e-01 -2.19251234e-02
5.15709698e-01 2.38609314e-01 6.88634217e-01 -4.27328378e-01
-3.54810327e-01 2.80902237e-01 9.17668641e-03 2.36400038e-01
2.51407892e-01 -6.93546608e-02 -3.28322858e-01 -6.63741112e-01
-3.08282766e-02 -7.69761205e-01 4.93084133e-01 5.77274382e-01
1.28165436e+00 -5.62299192e-01 -8.55573341e-02 6.26439750e-01
1.46381640e+00 4.15922642e-01 2.75603712e-01 7.30281830e-01
2.32317463e-01 4.00053024e-01 1.02161717e+00 6.84100688e-01
1.33123443e-01 7.17676699e-01 6.75135672e-01 2.85818487e-01
6.09628975e-01 2.25545336e-02 3.09000194e-01 2.53064856e-02
-4.74622138e-02 -2.34891951e-01 -7.97126770e-01 6.19272053e-01
-2.09364152e+00 -6.81424677e-01 8.45150113e-01 2.32083178e+00
1.07143569e+00 4.17518288e-01 4.25956458e-01 -2.32675031e-01
7.10907757e-01 4.83296392e-03 -1.02877724e+00 -6.26830816e-01
1.44074395e-01 3.11655492e-01 7.58514702e-01 6.86698735e-01
-1.26553071e+00 6.02964818e-01 7.41443968e+00 6.72828913e-01
-1.16014433e+00 -2.67308980e-01 3.71346176e-01 -4.39337075e-01
-3.64557981e-01 1.32548222e-02 -5.09122431e-01 3.61894995e-01
6.83887303e-01 -5.41858792e-01 8.82714927e-01 7.37861335e-01
2.73840457e-01 -9.63762030e-02 -1.07013500e+00 7.36679614e-01
-4.20596778e-01 -1.02134550e+00 -2.19716087e-01 1.51636377e-01
8.69809926e-01 -1.58347994e-01 2.64524400e-01 1.55951530e-01
1.02312016e+00 -1.14660418e+00 7.84065247e-01 2.68481970e-01
4.01521415e-01 -1.00123847e+00 3.14591736e-01 4.09756154e-01
-7.86877334e-01 -6.27979159e-01 -3.33476067e-01 -2.33793199e-01
7.02681914e-02 1.11161001e-01 -7.60055602e-01 1.44303709e-01
3.75402600e-01 2.60019273e-01 -2.67299768e-02 1.28970420e+00
-1.37895435e-01 3.87072400e-03 -5.27263999e-01 -4.09135431e-01
5.26986480e-01 -4.72654223e-01 5.64227104e-01 6.72787011e-01
4.95789111e-01 -1.22836933e-01 6.83447778e-01 9.10169899e-01
7.02439621e-02 -9.27398801e-02 -4.47150201e-01 -3.20742846e-01
4.02725339e-01 1.15722883e+00 -7.94924736e-01 7.05560893e-02
-7.58970529e-02 7.77844489e-01 6.22650027e-01 3.65012586e-01
-9.57474411e-01 -4.08385545e-01 1.14578652e+00 -2.42145732e-01
3.91505212e-01 -4.64352489e-01 -2.78816342e-01 -1.00297797e+00
3.62618864e-02 -1.01870954e+00 6.36337101e-01 -4.00660843e-01
-1.28887284e+00 1.25497013e-01 1.72427550e-01 -1.21519828e+00
-3.88907433e-01 -7.48082519e-01 -4.59593713e-01 7.49071956e-01
-1.44244683e+00 -4.62879956e-01 2.98599035e-01 4.50409055e-01
4.03605342e-01 -9.25672576e-02 5.75346291e-01 -1.48148224e-01
-2.98070461e-01 3.73908579e-01 4.77128714e-01 -4.20200229e-01
8.50233793e-01 -1.51239252e+00 -6.38305023e-02 3.55565310e-01
-1.25481695e-01 3.27541709e-01 1.10197389e+00 -2.51025528e-01
-1.32916176e+00 -6.76654279e-01 1.56970143e-01 -2.07104191e-01
1.04060769e+00 -4.01731301e-03 -7.03124464e-01 5.27907968e-01
2.43249640e-01 -1.16936222e-01 1.73758626e-01 4.49880436e-02
-2.29493733e-02 -9.97245014e-02 -1.35025132e+00 8.54377568e-01
7.10862398e-01 -1.06429629e-01 -6.37059212e-01 2.49544159e-01
6.52787507e-01 -6.19310856e-01 -7.97063172e-01 2.86749363e-01
4.72756833e-01 -5.87579370e-01 1.07508075e+00 -8.58726740e-01
3.68451506e-01 -3.26844305e-01 -2.15355322e-01 -2.07197165e+00
-2.94320434e-01 -7.07165539e-01 -2.99646310e-03 6.77140474e-01
2.64687002e-01 -8.83083761e-01 5.34136832e-01 6.40752494e-01
5.65493945e-03 -8.63077641e-01 -9.50666964e-01 -9.61809278e-01
3.35887194e-01 -3.05052679e-02 5.76709211e-01 6.59707665e-01
-1.46272928e-01 1.49572015e-01 -3.21400791e-01 2.47048914e-01
7.70776689e-01 4.12349284e-01 4.77636486e-01 -1.03846025e+00
-7.51874685e-01 -7.92334437e-01 -9.49518085e-02 -9.67415094e-01
1.61473706e-01 -4.38470572e-01 3.17648709e-01 -1.27607167e+00
3.22030820e-02 -7.68512547e-01 -5.44598818e-01 5.05286217e-01
-1.81979880e-01 -2.26206794e-01 3.35152626e-01 -1.13472588e-01
-6.85516357e-01 7.07820415e-01 1.44233561e+00 -2.75611997e-01
-4.60128665e-01 8.67743492e-02 -9.51281786e-01 8.16602707e-01
9.39834058e-01 -3.42187881e-01 -5.28325915e-01 -6.02896094e-01
1.67166024e-01 2.64698595e-01 9.64018852e-02 -7.69824445e-01
-7.72249624e-02 -1.13760054e+00 2.16691315e-01 -9.25122052e-02
3.13674659e-01 -7.36371458e-01 -9.86018404e-02 4.30802524e-01
-4.61304784e-01 1.45420626e-01 2.34504268e-01 4.30310488e-01
1.28503174e-01 -4.35909212e-01 1.09258604e+00 -1.67694807e-01
-4.41319734e-01 1.62677988e-01 -2.90629506e-01 3.67188096e-01
1.11265850e+00 2.05663070e-02 -1.33432811e-02 -3.30996990e-01
-6.94763780e-01 9.09692705e-01 6.80821896e-01 4.04613137e-01
4.78442430e-01 -1.24727476e+00 -5.13271987e-01 -7.00823143e-02
-6.16049469e-02 -1.41098842e-01 -2.67027557e-01 3.87286156e-01
-3.59001905e-01 1.18910849e-01 -6.92934155e-01 -4.12802607e-01
-1.02016532e+00 5.57195723e-01 5.11044741e-01 -4.56644565e-01
-1.00846544e-01 5.41761100e-01 1.33263078e-02 -7.64925778e-01
1.76035166e-01 -4.89037424e-01 -7.39307404e-02 4.94830944e-02
1.98733509e-01 3.96027625e-01 -3.49014431e-01 -3.31889614e-02
-1.65410772e-01 5.23401558e-01 6.33680746e-02 -5.33545613e-01
1.43076849e+00 2.82154623e-02 2.18595698e-01 3.52288693e-01
9.33314085e-01 -2.11553559e-01 -1.87905800e+00 5.65884002e-02
5.63264787e-02 -7.33600378e-01 -9.89916176e-02 -7.73841798e-01
-7.30538547e-01 4.06692386e-01 3.91350448e-01 1.54369801e-01
1.18807709e+00 -1.88881487e-01 3.51798892e-01 5.95319092e-01
5.91049135e-01 -1.47041667e+00 3.33814770e-01 5.86834192e-01
9.48625922e-01 -1.15710497e+00 -1.24611408e-01 6.11437820e-02
-7.56454051e-01 1.17003179e+00 6.31625354e-01 -3.38098705e-01
3.85395914e-01 4.81926024e-01 1.82985350e-01 1.17796876e-01
-7.89978981e-01 -1.61379829e-01 2.44896244e-02 5.74332416e-01
-7.62220006e-04 2.40174562e-01 -4.66405809e-01 1.15763068e-01
-5.45448549e-02 -1.22628942e-01 6.26164317e-01 1.26831591e+00
-7.33080983e-01 -1.51047075e+00 -6.89214706e-01 4.69564050e-01
-2.87266165e-01 3.66070837e-01 -1.38758808e-01 6.27618670e-01
-2.58490354e-01 7.20420659e-01 -1.06149085e-01 1.79716833e-02
3.62996548e-01 -1.09619141e-01 8.20853293e-01 -4.99892861e-01
-3.03512543e-01 1.75062835e-01 -7.86923394e-02 -5.94136536e-01
-2.64970154e-01 -7.67576277e-01 -1.31736076e+00 -2.47091949e-02
-3.38005126e-02 2.19249323e-01 3.44169080e-01 9.32574511e-01
6.90024644e-02 5.30598581e-01 7.36211061e-01 -8.85950863e-01
-1.40790951e+00 -4.46184099e-01 -6.51955903e-01 2.96093524e-01
5.54882050e-01 -1.17238772e+00 -2.75037378e-01 -4.10381436e-01] | [4.3005266189575195, 2.2826032638549805] |
afae981d-b2a8-4d40-85ea-94ccab2fe7dc | a-unified-software-hardware-scalable | 2201.02262 | null | https://arxiv.org/abs/2201.02262v1 | https://arxiv.org/pdf/2201.02262v1.pdf | A unified software/hardware scalable architecture for brain-inspired computing based on self-organizing neural models | The field of artificial intelligence has significantly advanced over the past decades, inspired by discoveries from the fields of biology and neuroscience. The idea of this work is inspired by the process of self-organization of cortical areas in the human brain from both afferent and lateral/internal connections. In this work, we develop an original brain-inspired neural model associating Self-Organizing Maps (SOM) and Hebbian learning in the Reentrant SOM (ReSOM) model. The framework is applied to multimodal classification problems. Compared to existing methods based on unsupervised learning with post-labeling, the model enhances the state-of-the-art results. This work also demonstrates the distributed and scalable nature of the model through both simulation results and hardware execution on a dedicated FPGA-based platform named SCALP (Self-configurable 3D Cellular Adaptive Platform). SCALP boards can be interconnected in a modular way to support the structure of the neural model. Such a unified software and hardware approach enables the processing to be scaled and allows information from several modalities to be merged dynamically. The deployment on hardware boards provides performance results of parallel execution on several devices, with the communication between each board through dedicated serial links. The proposed unified architecture, composed of the ReSOM model and the SCALP hardware platform, demonstrates a significant increase in accuracy thanks to multimodal association, and a good trade-off between latency and power consumption compared to a centralized GPU implementation. | ['Andres Upegui', 'Quentin Berthet', 'Joachim Schmidt', 'Lyes Khacef', 'Benoit Miramond', 'Laurent Rodriguez', 'Artem R. Muliukov'] | 2022-01-06 | null | null | null | null | ['multimodal-association'] | ['time-series'] | [-1.68149862e-02 2.64916658e-01 3.53191346e-01 -2.86455490e-02
4.15247560e-01 -3.72192800e-01 6.34230554e-01 5.59772372e-01
-6.55436635e-01 5.54302096e-01 -4.37442623e-02 1.79829493e-01
-4.05565321e-01 -9.54486966e-01 -4.62429762e-01 -8.25333118e-01
-3.07035148e-01 7.29584932e-01 4.66660351e-01 -3.55494767e-01
2.67314911e-01 6.68522358e-01 -2.36808205e+00 2.69760013e-01
8.10224056e-01 1.05032718e+00 5.49295723e-01 3.63691360e-01
-1.44127265e-01 2.11098388e-01 -3.91652048e-01 3.43393058e-01
9.69365537e-02 -1.14180446e-01 -3.61488640e-01 -2.41135374e-01
-6.56458810e-02 1.18126266e-01 8.40256810e-02 7.15077817e-01
6.90807998e-01 -1.18880957e-01 8.08181524e-01 -1.02770078e+00
5.83258383e-02 5.43029845e-01 -2.67299592e-01 1.91520661e-01
3.54063690e-01 -3.38206291e-01 3.43597829e-01 -7.73428619e-01
6.50335133e-01 7.23634660e-01 5.16782284e-01 2.38546759e-01
-1.37681234e+00 -4.71306741e-01 -2.84547299e-01 4.84722257e-01
-1.56171978e+00 5.11210319e-03 4.42134142e-01 -5.52685142e-01
1.12051809e+00 2.16995314e-01 1.50663376e+00 8.80446672e-01
6.42981827e-01 2.55275458e-01 1.50055611e+00 -7.07232535e-01
9.32738185e-01 4.06216681e-01 3.43645722e-01 2.36036971e-01
6.38116539e-01 -1.66069880e-01 -9.03778911e-01 6.79523945e-02
9.09998119e-01 -5.93095608e-02 2.66093999e-01 -2.98502237e-01
-9.43803430e-01 3.04123789e-01 4.45689470e-01 8.83033156e-01
-6.70466959e-01 1.23575509e-01 5.14742792e-01 -3.32232341e-02
1.46815926e-01 3.44286621e-01 -2.47917011e-01 2.00026259e-01
-1.14885592e+00 2.17600702e-03 9.31335747e-01 7.26443946e-01
6.60293758e-01 1.33803010e-01 1.40177026e-01 5.88388205e-01
3.29021424e-01 3.20997030e-01 1.03431571e+00 -3.90667230e-01
-5.54332554e-01 7.46507943e-01 -3.64938617e-01 -8.94364297e-01
-1.11218476e+00 -8.11469913e-01 -9.74203885e-01 6.40019417e-01
-4.80746068e-02 -5.78837804e-02 -7.69697785e-01 1.23832178e+00
4.49295908e-01 -2.60793623e-02 -6.42492771e-02 8.36859405e-01
7.48757243e-01 5.94595969e-01 -6.02296405e-02 -7.47712404e-02
1.71844327e+00 -7.20343113e-01 -6.42823339e-01 7.81856850e-02
1.74031079e-01 -4.63716477e-01 4.53564763e-01 7.42787123e-01
-1.03485155e+00 -7.02798188e-01 -1.44173157e+00 1.71347901e-01
-7.98306108e-01 -4.31993231e-03 6.61149502e-01 8.01544130e-01
-1.69999659e+00 4.34780091e-01 -9.51891541e-01 -9.52527940e-01
2.38323927e-01 8.43123972e-01 -4.48873073e-01 5.90667188e-01
-7.93795109e-01 9.47161138e-01 8.22351336e-01 -4.23048176e-02
-6.26969159e-01 -3.66171211e-01 -1.49367869e-01 4.29482788e-01
-4.62668389e-01 -1.13641179e+00 5.87141097e-01 -1.18646860e+00
-1.85578930e+00 7.90170789e-01 9.69820693e-02 -7.45969951e-01
1.79175779e-01 4.15108763e-02 -1.74669370e-01 3.47115457e-01
-2.31461987e-01 9.58413303e-01 6.69209898e-01 -1.01988804e+00
-5.93280673e-01 -4.93859798e-01 -3.62332851e-01 -2.02284008e-02
-1.02373970e+00 -3.22468460e-01 -4.78187427e-02 -5.02401173e-01
3.10855210e-01 -7.89272606e-01 -9.14517604e-03 -4.78206486e-01
2.26937737e-02 4.27572392e-02 3.79498690e-01 -2.87170202e-01
8.52190554e-01 -2.12365413e+00 2.45739534e-01 5.30518115e-01
1.14433557e-01 -1.29679337e-01 2.35904410e-01 6.42662406e-01
-1.16663270e-01 -5.77906549e-01 -5.16938008e-02 -6.90181330e-02
-1.48449123e-01 1.64504543e-01 5.25730886e-02 3.84534776e-01
-3.55865985e-01 3.77687782e-01 -5.12341201e-01 -2.63415843e-01
1.78964153e-01 5.70514560e-01 -4.69202906e-01 -9.99951959e-02
1.76179260e-01 4.12557244e-01 1.48494784e-02 6.10128701e-01
7.38816381e-01 8.86601731e-02 3.12325984e-01 9.76151451e-02
-6.77605331e-01 -2.05112696e-01 -1.24087298e+00 1.94074488e+00
-4.68163997e-01 5.28453529e-01 8.53003561e-02 -9.85476434e-01
1.29867995e+00 5.97378194e-01 4.65815604e-01 -7.97584832e-01
6.26290500e-01 5.38994431e-01 1.36973038e-01 -3.41262460e-01
3.20144236e-01 1.10687003e-01 4.15225089e-01 4.60443228e-01
7.60414720e-01 9.65934992e-02 2.41224691e-01 -2.14978635e-01
1.11882389e+00 1.19909845e-01 4.57924813e-01 -1.14704585e+00
5.61760426e-01 6.26906157e-02 5.08703617e-03 4.90554482e-01
3.92109275e-01 2.21471444e-01 8.75281915e-02 -4.55195487e-01
-9.37236488e-01 -1.07108307e+00 -5.52020907e-01 1.00801039e+00
7.60593712e-02 -1.98415101e-01 -1.24704206e+00 3.14994276e-01
-1.78953424e-01 4.26874548e-01 -6.90177739e-01 2.35353813e-01
-1.04297914e-01 -7.61742353e-01 3.60443711e-01 1.94719464e-01
5.76771617e-01 -1.15702868e+00 -1.43202114e+00 4.29176807e-01
6.38410449e-01 -7.32221186e-01 7.71758199e-01 6.19162977e-01
-1.33968222e+00 -4.31332886e-01 -5.07985413e-01 -9.93400514e-01
8.41142952e-01 -6.87535033e-02 6.06949449e-01 -1.15853094e-01
-5.97519457e-01 5.11366904e-01 -4.09172654e-01 -7.14027345e-01
6.97425306e-02 4.56207812e-01 3.62197101e-01 1.90186217e-01
1.09089538e-01 -1.49193442e+00 -7.57207155e-01 4.66463529e-02
-1.04425514e+00 2.62301356e-01 7.12619662e-01 6.68072522e-01
6.00075305e-01 4.36556824e-02 6.52574480e-01 -6.51544333e-01
4.45731074e-01 -6.07178390e-01 -6.84618175e-01 -2.52849281e-01
-7.91411579e-01 -3.01029142e-02 7.43122518e-01 -2.28261173e-01
-8.56736660e-01 3.65839243e-01 -2.20633864e-01 2.02590004e-01
-5.79647303e-01 3.62396389e-01 6.53693080e-02 -2.87343085e-01
7.37071812e-01 4.56756055e-01 -7.85829350e-02 -3.81226361e-01
1.38300717e-01 6.51793599e-01 6.13462985e-01 -3.98062199e-01
2.91135043e-01 7.26892889e-01 2.22298369e-01 -1.00831389e+00
2.64221817e-01 -2.59552270e-01 -7.60055602e-01 -6.64864302e-01
8.07461917e-01 -6.78835928e-01 -8.53771627e-01 4.84610677e-01
-1.32680786e+00 -5.98533303e-02 -3.26355994e-01 6.72434568e-01
-6.16823196e-01 -1.81427076e-01 -5.09953678e-01 -8.65159929e-01
-7.46496558e-01 -7.81937957e-01 6.30126476e-01 6.39168441e-01
-3.91693443e-01 -8.94982159e-01 2.44383931e-01 -2.64799058e-01
5.10514021e-01 1.65199116e-01 8.59696746e-01 -6.38614893e-01
-4.31700081e-01 -2.53415585e-01 1.89326733e-01 -5.24779484e-02
-3.76265258e-01 -9.23482776e-02 -1.27540052e+00 -1.91973791e-01
3.25200796e-01 1.57518193e-01 7.31276035e-01 4.90218312e-01
5.42561948e-01 8.43032002e-02 -4.50908899e-01 4.47141051e-01
1.73175120e+00 3.08096886e-01 6.77366674e-01 7.37174451e-01
-2.93414928e-02 8.45979393e-01 1.41654909e-01 6.15251243e-01
1.72950342e-01 4.97468799e-01 6.61128461e-01 -2.39674240e-01
4.88271825e-02 3.07720423e-01 1.92673936e-01 1.14103723e+00
-6.24818206e-01 3.39137428e-02 -1.01734126e+00 4.23216313e-01
-1.91242850e+00 -7.46404588e-01 -2.14858070e-01 2.31105089e+00
2.66574770e-01 1.01017915e-01 2.24645942e-01 4.16838109e-01
6.69760108e-01 -6.23097181e-01 -6.01455830e-02 -8.07664812e-01
-2.88397610e-01 6.90358937e-01 3.41185331e-01 2.56524384e-01
-7.47436881e-01 5.55223167e-01 6.26967192e+00 7.10643589e-01
-1.34759724e+00 5.59427679e-01 1.86568588e-01 -2.35711932e-01
1.26184985e-01 -1.50868520e-01 -6.48912072e-01 5.10978103e-01
1.09761977e+00 -1.89731568e-01 5.55535436e-01 7.43978977e-01
5.41957542e-02 -4.07016069e-01 -8.61965477e-01 9.75192726e-01
1.82335049e-01 -1.29903197e+00 -9.19876918e-02 2.64212370e-01
5.86044133e-01 1.98252037e-01 -4.41118190e-03 -1.11573204e-01
-4.50409889e-01 -7.00267375e-01 9.65729713e-01 6.10414743e-01
4.21527177e-01 -8.12289953e-01 9.45709765e-01 3.78266066e-01
-1.08322799e+00 -3.48347694e-01 -3.60807598e-01 -5.51448941e-01
-7.59829730e-02 6.53111160e-01 -8.03668559e-01 4.96210635e-01
9.08973157e-01 1.37010843e-01 -5.66517770e-01 1.35256588e+00
-2.04117447e-02 4.09395993e-01 -5.00826120e-01 -5.47906280e-01
-3.00049633e-02 -2.95219064e-01 5.66748202e-01 1.40074766e+00
6.36764288e-01 -1.04347602e-01 -4.89377648e-01 6.80142462e-01
4.59191561e-01 5.77980340e-01 -6.18269145e-01 5.37525117e-01
3.99378717e-01 1.85085237e+00 -1.46804655e+00 -3.14898729e-01
-9.33033356e-04 7.28321075e-01 6.97000995e-02 -5.54692335e-02
-6.44327164e-01 -4.35788780e-01 8.86085257e-02 4.70117986e-01
1.24248140e-01 -3.28598320e-01 -8.06126177e-01 -5.82745910e-01
-2.55958170e-01 -2.62308747e-01 3.62186059e-02 -7.96386242e-01
-9.09561098e-01 1.18327904e+00 -1.87443510e-01 -1.22168589e+00
-1.11815482e-01 -7.42982924e-01 -3.29342604e-01 5.56472480e-01
-9.00576711e-01 -1.12603092e+00 -5.18594384e-01 4.02598232e-01
1.75160989e-01 -5.36604345e-01 1.30586863e+00 2.44214028e-01
-2.86959141e-01 2.58211374e-01 3.23229134e-01 -5.40274978e-01
2.90380657e-01 -1.08092153e+00 -1.36197865e-01 5.22442937e-01
6.52928427e-02 6.06786549e-01 6.51262283e-01 -3.12536508e-01
-1.30809128e+00 -4.92403209e-01 5.03026903e-01 2.42579967e-01
5.14896512e-01 -7.10837781e-01 -5.14227271e-01 2.54732296e-02
7.46329010e-01 -3.35668087e-01 8.73929143e-01 1.10359818e-01
2.00330503e-02 -5.32273054e-01 -1.14553154e+00 5.30640662e-01
8.25654745e-01 -5.04983626e-02 -5.50303161e-01 2.54304171e-01
8.12251046e-02 -3.78160700e-02 -7.88899302e-01 4.47075814e-02
9.06428397e-01 -1.32071400e+00 5.23170233e-01 2.87048221e-01
1.38576508e-01 -4.03588325e-01 -7.85996951e-03 -1.16181946e+00
-5.53021193e-01 -3.60790104e-01 9.76449773e-02 1.15051937e+00
1.82726488e-01 -1.07262886e+00 6.40612066e-01 1.28798902e-01
-3.51833254e-01 -6.00864232e-01 -1.26039994e+00 -7.24301279e-01
-4.46051508e-01 -1.33385479e-01 4.47588503e-01 6.04049683e-01
7.34516025e-01 4.46389049e-01 3.90567809e-01 7.82118067e-02
4.52138901e-01 -1.24136604e-01 4.86372560e-01 -1.61665678e+00
-4.62282151e-01 -5.61445892e-01 -1.31281483e+00 -3.11787456e-01
-1.78926215e-01 -1.15188849e+00 -1.44791752e-01 -1.48338342e+00
9.95302424e-02 -2.76729643e-01 -5.40024042e-01 5.39337635e-01
6.67955279e-01 5.37632287e-01 3.61850820e-02 2.23497912e-01
-3.11579645e-01 1.42062321e-01 5.54699183e-01 1.30129755e-01
-3.90911371e-01 -5.46561122e-01 -2.57892698e-01 8.08269382e-01
7.65273094e-01 -5.11164308e-01 -1.81746155e-01 -4.72092897e-01
3.60534251e-01 -2.32759714e-01 3.25642735e-01 -1.88037431e+00
8.48790884e-01 7.57220745e-01 5.87446272e-01 -3.79812300e-01
3.80002052e-01 -1.19376731e+00 4.89029676e-01 9.83472943e-01
2.43960008e-01 7.02296868e-02 5.51120460e-01 2.04988658e-01
-2.77126282e-01 -2.93445528e-01 6.36133730e-01 1.93991944e-01
-4.91982877e-01 -3.43248725e-01 -1.08466494e+00 -9.15859401e-01
1.30999494e+00 -5.88161409e-01 -3.27193499e-01 2.06074938e-01
-1.03358090e+00 -3.17882806e-01 4.49432909e-01 -7.08583370e-02
3.43413919e-01 -1.08280337e+00 -4.43731934e-01 4.34470087e-01
-1.04505226e-01 -3.43610615e-01 6.09171748e-01 1.06285942e+00
-1.08728313e+00 4.86711800e-01 -1.32783651e+00 -8.20404470e-01
-9.98557448e-01 4.20812935e-01 1.07506000e-01 1.24729812e-01
-4.56645429e-01 5.53549170e-01 1.40137626e-02 -5.28098643e-02
3.82890552e-02 -4.39796299e-02 -5.86864591e-01 3.74932349e-01
4.77929801e-01 7.37667024e-01 5.30420363e-01 -1.37925044e-01
-4.05182034e-01 5.23145676e-01 3.11395705e-01 -5.34223855e-01
1.64649284e+00 1.38123468e-01 -7.92850018e-01 7.27009356e-01
4.92280185e-01 3.75724360e-02 -5.77265680e-01 4.27619159e-01
-1.95959523e-01 1.57783747e-01 1.29951552e-01 -8.68642569e-01
-7.14191318e-01 7.96181738e-01 1.08747578e+00 4.03588861e-01
1.44191170e+00 -3.30481052e-01 2.06746206e-01 2.56648391e-01
8.98056090e-01 -1.22642493e+00 -2.35765100e-01 3.60154808e-01
7.60661662e-01 -2.74240136e-01 -2.97643989e-02 -2.72730559e-01
-9.92722362e-02 1.51284719e+00 4.07054603e-01 -5.59893191e-01
8.17638814e-01 8.65701795e-01 -2.55326122e-01 -1.55428067e-01
-6.75571024e-01 -1.69635713e-01 -4.57914136e-02 7.27778196e-01
4.39416885e-01 2.12614596e-01 -8.93599749e-01 1.09629297e+00
-3.81863564e-01 3.00506890e-01 4.73488778e-01 9.06530559e-01
-6.59793139e-01 -1.09623826e+00 -6.90159261e-01 2.73683459e-01
-5.33269420e-02 -6.12908565e-02 -1.51024982e-01 6.06279671e-01
7.36521721e-01 6.67106211e-01 4.34978753e-01 -5.25770307e-01
6.55181631e-02 -3.84875797e-02 8.93758237e-01 -5.80180585e-01
-1.11869073e+00 4.12555486e-01 -2.92727858e-01 -1.45872161e-01
-2.77809620e-01 -5.82420766e-01 -1.51360416e+00 -8.32901821e-02
1.53090628e-02 1.93307459e-01 1.26345241e+00 6.97291851e-01
7.38288462e-01 8.51518810e-01 3.50680143e-01 -1.47518587e+00
3.53273839e-01 -9.81687069e-01 -1.11468410e+00 -3.68453294e-01
-3.86399895e-01 -7.87795126e-01 1.33783862e-01 4.29188907e-02] | [8.107510566711426, 2.722280263900757] |
6d95c742-ff22-41d4-a096-8e5b9d70eb26 | visual-scene-graphs-for-audio-source | 2109.11955 | null | https://arxiv.org/abs/2109.11955v1 | https://arxiv.org/pdf/2109.11955v1.pdf | Visual Scene Graphs for Audio Source Separation | State-of-the-art approaches for visually-guided audio source separation typically assume sources that have characteristic sounds, such as musical instruments. These approaches often ignore the visual context of these sound sources or avoid modeling object interactions that may be useful to better characterize the sources, especially when the same object class may produce varied sounds from distinct interactions. To address this challenging problem, we propose Audio Visual Scene Graph Segmenter (AVSGS), a novel deep learning model that embeds the visual structure of the scene as a graph and segments this graph into subgraphs, each subgraph being associated with a unique sound obtained by co-segmenting the audio spectrogram. At its core, AVSGS uses a recursive neural network that emits mutually-orthogonal sub-graph embeddings of the visual graph using multi-head attention. These embeddings are used for conditioning an audio encoder-decoder towards source separation. Our pipeline is trained end-to-end via a self-supervised task consisting of separating audio sources using the visual graph from artificially mixed sounds. In this paper, we also introduce an "in the wild'' video dataset for sound source separation that contains multiple non-musical sources, which we call Audio Separation in the Wild (ASIW). This dataset is adapted from the AudioCaps dataset, and provides a challenging, natural, and daily-life setting for source separation. Thorough experiments on the proposed ASIW and the standard MUSIC datasets demonstrate state-of-the-art sound separation performance of our method against recent prior approaches. | ['Anoop Cherian', 'Narendra Ahuja', 'Jonathan Le Roux', 'Moitreya Chatterjee'] | 2021-09-24 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Chatterjee_Visual_Scene_Graphs_for_Audio_Source_Separation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Chatterjee_Visual_Scene_Graphs_for_Audio_Source_Separation_ICCV_2021_paper.pdf | iccv-2021-1 | ['audio-source-separation'] | ['audio'] | [ 3.44511390e-01 -3.26762706e-01 2.24404544e-01 1.31353771e-03
-1.05407453e+00 -8.52001607e-01 2.38966063e-01 1.22489311e-01
9.49226022e-02 5.10431081e-02 6.72193646e-01 1.54029444e-01
-6.35765493e-02 -2.24390998e-01 -7.77170300e-01 -7.06034064e-01
-1.58097729e-01 1.12759555e-02 1.86565772e-01 5.76683693e-03
1.57484021e-02 2.46879831e-02 -1.91803932e+00 6.56296372e-01
2.91594654e-01 1.01727760e+00 2.94415385e-01 1.25474238e+00
-1.14063673e-01 9.96024311e-01 -6.77495003e-01 3.39742750e-02
1.24469981e-01 -9.04965699e-01 -6.25788569e-01 -1.29796425e-02
9.59281206e-01 1.73781179e-02 -3.89034480e-01 1.10198236e+00
8.71924460e-01 8.31357315e-02 5.08685589e-01 -1.62120032e+00
-6.84491992e-01 1.15352428e+00 -4.86924112e-01 2.97624350e-01
5.91926754e-01 1.12993516e-01 1.51787305e+00 -8.27389479e-01
2.53814250e-01 1.34416986e+00 6.18636966e-01 4.57858652e-01
-1.61499000e+00 -7.92902470e-01 9.87350643e-02 2.42914870e-01
-1.29660928e+00 -8.49717557e-01 1.37798631e+00 -7.17162251e-01
7.55358696e-01 4.11425591e-01 8.21634471e-01 1.32329059e+00
-2.25298911e-01 9.65362549e-01 5.68190038e-01 -4.56692874e-01
3.91857535e-01 -2.02052727e-01 2.20764264e-01 3.27718616e-01
-1.23803720e-01 3.19987722e-03 -1.09011209e+00 -2.47917473e-01
5.57374120e-01 -3.12056154e-01 -7.05089152e-01 -4.80794251e-01
-1.22276795e+00 4.84776944e-01 4.37912077e-01 2.39569113e-01
1.53542496e-03 5.80666900e-01 4.51745778e-01 1.71327218e-01
2.98872352e-01 2.24896178e-01 3.21743544e-04 -9.18227211e-02
-1.03744721e+00 2.33266354e-01 7.00624645e-01 8.61956298e-01
3.05612445e-01 6.82542384e-01 -2.38652408e-01 1.02246773e+00
5.86347401e-01 4.43425328e-01 5.75978279e-01 -8.14974606e-01
3.31755698e-01 2.15232626e-01 -2.39295065e-01 -9.15925145e-01
-3.44904572e-01 -6.72276318e-01 -6.34125948e-01 3.37261528e-01
3.65365595e-01 9.76865552e-03 -1.08777475e+00 1.93613267e+00
2.23568931e-01 7.07317710e-01 -2.45224640e-01 1.26536882e+00
1.40897393e+00 6.36210382e-01 -5.70684299e-02 1.17972858e-01
1.34353089e+00 -1.10940993e+00 -6.40524924e-01 -2.54555732e-01
-1.66372195e-01 -6.67953014e-01 1.38270593e+00 3.93603683e-01
-1.04927158e+00 -9.43616807e-01 -9.94487643e-01 -8.75541121e-02
-4.04492728e-02 1.01828612e-01 1.52577803e-01 4.94459957e-01
-9.18806314e-01 4.18235868e-01 -7.70704806e-01 7.26618385e-03
4.33316767e-01 -4.54690075e-03 -1.36287510e-01 5.13377905e-01
-9.47705626e-01 -1.03640817e-01 3.17345448e-02 -2.94241495e-02
-1.61311436e+00 -1.05203450e+00 -1.16481340e+00 3.82223219e-01
3.88049871e-01 -6.84374809e-01 1.26407540e+00 -1.33895910e+00
-1.43980312e+00 5.85000157e-01 -1.67503968e-01 -2.79074639e-01
-4.86600325e-02 -3.71033669e-01 -6.35412574e-01 3.38635683e-01
1.92266449e-01 3.53349209e-01 1.53988147e+00 -1.53164589e+00
-3.05939406e-01 -1.46760717e-01 -1.55026078e-01 1.82744026e-01
-1.75971568e-01 3.35752130e-01 -4.97151017e-01 -1.04467511e+00
-1.22891709e-01 -7.14625239e-01 2.45064080e-01 -1.54333383e-01
-8.40554595e-01 2.15783015e-01 6.61511064e-01 -5.63735962e-01
1.27627528e+00 -2.67967224e+00 4.99446720e-01 -3.93462414e-03
5.96290052e-01 -2.33089235e-02 -5.54912686e-01 3.95042360e-01
-5.16850173e-01 -1.40934950e-02 -2.57656097e-01 -6.48738027e-01
1.93591774e-01 -3.45348954e-01 -6.88702464e-01 3.56171459e-01
7.01567754e-02 5.33225894e-01 -1.14998162e+00 -3.55853230e-01
-4.32238616e-02 6.60528183e-01 -6.28761590e-01 4.89300579e-01
-1.84095114e-01 4.55146432e-01 1.62292585e-01 6.00954235e-01
4.97724742e-01 -2.17867475e-02 -8.64176229e-02 -3.13033640e-01
1.76043972e-01 5.26208401e-01 -1.53905535e+00 2.16487074e+00
-2.31610015e-01 9.11794305e-01 4.55955148e-01 -6.18336558e-01
5.32707751e-01 4.33400899e-01 3.51821959e-01 -2.56670475e-01
1.26650721e-01 -1.11717680e-04 8.10302496e-02 -4.59999591e-01
1.46960616e-01 -1.45929456e-01 1.77172236e-02 3.84865314e-01
6.68505609e-01 -2.05478311e-01 -2.58010317e-04 3.88731480e-01
1.11412799e+00 4.35536616e-02 7.79794110e-03 1.48189766e-02
2.74939030e-01 -4.97728854e-01 4.12703037e-01 5.87198079e-01
-2.41411239e-01 1.21925771e+00 5.23189962e-01 1.73830181e-01
-5.43585658e-01 -1.75015533e+00 2.28206784e-01 1.45839405e+00
1.52483404e-01 -9.21606898e-01 -6.57457232e-01 -5.48461258e-01
-1.44131510e-02 5.26187539e-01 -4.39069957e-01 -2.54502833e-01
-2.36181229e-01 -1.83839858e-01 8.04949760e-01 4.94029194e-01
-1.86356381e-02 -1.13782942e+00 -5.27621210e-01 1.69370458e-01
-3.63775998e-01 -9.15453017e-01 -7.36341417e-01 3.11941832e-01
-7.57367760e-02 -1.09667373e+00 -6.39910340e-01 -8.50930333e-01
1.23796590e-01 5.31479657e-01 1.25969887e+00 -3.86234939e-01
-4.17486012e-01 7.21086204e-01 -3.87873322e-01 -6.87086403e-01
-4.23722893e-01 -2.72791713e-01 1.58940107e-01 6.68909967e-01
-6.88187853e-02 -1.10827744e+00 -5.82474768e-01 -1.58203244e-01
-8.95112097e-01 -2.87033394e-02 4.76238504e-02 5.52897871e-01
6.98959768e-01 8.30341280e-02 5.93524873e-01 -2.75709897e-01
7.46624351e-01 -6.14022315e-01 -2.03482032e-01 -1.37591332e-01
2.83720195e-01 -9.87356007e-02 9.97831881e-01 -6.95082247e-01
-6.34000003e-01 1.06411278e-01 7.32307583e-02 -1.09762073e+00
-3.95972103e-01 1.81294799e-01 -5.07539928e-01 1.70676962e-01
7.53310740e-01 8.24493244e-02 -4.08685744e-01 -8.19612801e-01
4.87526923e-01 7.04979300e-01 8.42239380e-01 -2.37724081e-01
7.42960811e-01 4.27202582e-01 -1.34694219e-01 -1.13497686e+00
-7.59441316e-01 -6.34241998e-01 -3.94725323e-01 -4.11073714e-01
9.27399457e-01 -1.20142698e+00 -5.68267584e-01 4.21404243e-01
-1.07225800e+00 -5.11159301e-01 -5.27193844e-01 5.06847501e-01
-4.47714329e-01 2.67871886e-01 -3.66848826e-01 -1.06077635e+00
-7.63406903e-02 -1.01357901e+00 1.52030504e+00 9.66150239e-02
-4.35097337e-01 -7.81340480e-01 6.01043940e-01 6.85242116e-02
2.03320533e-02 2.97235757e-01 7.70522058e-01 -5.98940492e-01
-2.39721701e-01 3.06688458e-01 1.03896700e-01 4.65966851e-01
3.17090094e-01 1.33644203e-02 -1.74979138e+00 -1.92384273e-01
-1.45501539e-01 -4.66762334e-01 1.20455527e+00 4.50151712e-01
1.03477097e+00 -1.58153281e-01 -6.60300180e-02 8.01002681e-01
1.19234729e+00 9.47363377e-02 2.77896702e-01 -1.83161646e-01
1.39207137e+00 4.65851426e-01 -9.20865089e-02 3.57807249e-01
1.33225664e-01 6.66432083e-01 7.78898895e-01 -2.73031890e-01
-8.09095502e-01 -5.85946143e-01 7.51810431e-01 8.87055039e-01
2.46987626e-01 -3.12191755e-01 -6.20342791e-01 8.68133664e-01
-1.58710158e+00 -1.11378479e+00 -1.46388844e-01 2.07255220e+00
9.52255428e-01 -1.12446226e-01 5.45027971e-01 7.67165840e-01
7.79227674e-01 6.59091234e-01 -2.86008209e-01 -2.53108770e-01
-1.16949357e-01 4.81318384e-01 -2.59855241e-01 5.65122783e-01
-1.36191177e+00 6.88251615e-01 5.43992615e+00 8.84682775e-01
-1.26758838e+00 5.30118495e-02 -2.91470308e-02 -6.70104206e-01
-4.27000523e-01 -4.21303883e-02 -2.45090082e-01 4.98324305e-01
9.69003975e-01 1.25092909e-01 7.64119446e-01 6.89839721e-01
3.18147056e-02 2.09608629e-01 -1.33184052e+00 1.36707234e+00
3.92808884e-01 -1.01078105e+00 1.09916180e-01 -3.79434645e-01
2.24233001e-01 -1.01719014e-01 1.45874545e-01 5.00012003e-02
2.79171973e-01 -1.04274368e+00 1.42402434e+00 2.63573438e-01
8.09104502e-01 -6.26085639e-01 -1.53479129e-01 2.83247363e-02
-1.72862756e+00 -9.02469456e-02 -2.83036083e-02 4.22065333e-02
2.38576099e-01 4.61859673e-01 -6.61255479e-01 5.96186459e-01
1.02743495e+00 8.53636622e-01 -6.15011215e-01 1.18303072e+00
-2.23140821e-01 1.14804995e+00 -1.52103916e-01 3.68860126e-01
-1.35507047e-01 1.30951747e-01 1.26645279e+00 1.47386563e+00
1.68020919e-01 -5.13877034e-01 9.40997973e-02 1.17171073e+00
-1.38978913e-01 2.32213885e-02 -7.09630668e-01 -2.22862393e-01
2.63013721e-01 1.21977675e+00 -6.68778360e-01 -5.17335124e-02
-2.52922267e-01 9.61050868e-01 1.77482858e-01 5.31563699e-01
-1.02544081e+00 -7.40860641e-01 9.91517127e-01 5.02601042e-02
4.51408386e-01 -6.75133020e-02 6.15588725e-02 -1.12517011e+00
-1.63635220e-02 -9.72151399e-01 3.43153924e-01 -1.13624716e+00
-1.31348670e+00 8.35052371e-01 -1.21291757e-01 -1.62407482e+00
-1.10661522e-01 -4.39905941e-01 -8.18715394e-01 7.37967849e-01
-1.29650939e+00 -1.25985301e+00 -4.07282382e-01 8.43650520e-01
5.93887508e-01 -1.91038206e-01 7.04598665e-01 2.44825691e-01
-5.13113976e-01 6.09350145e-01 -1.36069074e-01 3.77819330e-01
9.14188623e-01 -1.45713067e+00 3.78785938e-01 1.01923621e+00
1.15033555e+00 2.57199287e-01 7.01717436e-01 -4.57491279e-01
-1.30525768e+00 -1.29601812e+00 4.92381394e-01 -3.90093327e-01
8.41032028e-01 -1.01448607e+00 -8.80397439e-01 5.53987265e-01
4.20899630e-01 2.06966847e-01 1.09062922e+00 -1.69342346e-02
-8.50449026e-01 -1.98793992e-01 -4.15168464e-01 6.22669935e-01
1.24292517e+00 -1.07307708e+00 -6.65746391e-01 -2.74875499e-02
9.81362700e-01 -2.80752778e-01 -1.78908750e-01 -2.78202910e-03
5.30819535e-01 -1.09702146e+00 1.25374889e+00 -7.26930678e-01
4.50041831e-01 -6.03251755e-01 -2.15102583e-01 -1.71447992e+00
-5.35000801e-01 -1.00138152e+00 -3.02851230e-01 1.70537376e+00
1.57084897e-01 6.24199025e-02 1.13008849e-01 -1.79405317e-01
-3.25981945e-01 -9.55717042e-02 -8.82684886e-01 -8.32653105e-01
-3.86379600e-01 -1.09490860e+00 4.89133120e-01 9.70835984e-01
5.32306768e-02 7.47128248e-01 -4.00847882e-01 4.50915545e-01
7.58992553e-01 3.75225931e-01 7.88282037e-01 -1.22778916e+00
-8.18655729e-01 -5.01331687e-01 -5.22364974e-01 -7.77467430e-01
4.21371460e-01 -1.20440376e+00 2.52135307e-01 -1.43651211e+00
5.03672706e-03 1.56791762e-01 -5.30832767e-01 3.86304379e-01
-1.56353205e-01 3.86518836e-01 4.24415201e-01 -3.78701650e-02
-6.25169754e-01 6.86895013e-01 9.09930527e-01 -5.75933397e-01
-4.74468291e-01 -2.01835066e-01 -8.21453929e-01 9.10874188e-01
4.48412806e-01 -5.50012529e-01 -7.09020555e-01 -6.40095174e-01
2.29340732e-01 4.59472127e-02 8.97734582e-01 -1.19346941e+00
1.98134810e-01 8.81019309e-02 6.43847212e-02 -2.40013525e-01
5.25230527e-01 -7.39744723e-01 9.13008973e-02 -1.89688340e-01
-5.93758821e-01 -5.70996642e-01 4.45191771e-01 8.41233909e-01
-4.76193458e-01 1.13273583e-01 5.84661126e-01 1.70142204e-01
-3.33829910e-01 9.18837413e-02 -3.23361665e-01 5.31523645e-01
4.99415398e-01 -2.58202963e-02 -1.48964420e-01 -6.70772433e-01
-8.43034029e-01 -1.78595096e-01 -9.26148146e-02 5.40925086e-01
8.02895606e-01 -1.59877801e+00 -8.10347438e-01 3.89266372e-01
2.77539581e-01 -1.51425347e-01 4.69857603e-01 5.49618900e-01
-1.01668082e-01 -2.68503964e-01 -1.10443749e-01 -7.02903152e-01
-1.54860580e+00 7.13469863e-01 3.16678762e-01 3.45414162e-01
-7.39810169e-01 1.25293541e+00 7.88133323e-01 1.53299287e-01
5.61619997e-01 -6.63685262e-01 -3.46205533e-01 2.60393560e-01
6.33092880e-01 4.56601679e-01 -1.79826364e-01 -8.65130365e-01
-5.28091609e-01 6.44722223e-01 6.13093972e-01 -4.91643637e-01
1.12233639e+00 -9.74231288e-02 -3.40445084e-04 1.14502311e+00
1.38329625e+00 7.43039489e-01 -1.28380787e+00 -1.14140123e-01
-5.37191212e-01 -4.94127333e-01 1.80451706e-01 -6.56554222e-01
-1.20769131e+00 1.24995816e+00 5.58775127e-01 5.51904559e-01
1.28032982e+00 2.11265415e-01 6.19409442e-01 -1.73840463e-01
-1.19203642e-01 -6.66862547e-01 3.43598366e-01 2.76567966e-01
1.33740890e+00 -8.95767331e-01 -4.84497011e-01 -3.31954867e-01
-7.16591597e-01 1.05518138e+00 2.85640240e-01 -2.39012554e-01
8.20312142e-01 5.42096257e-01 2.23183170e-01 -2.32511595e-01
-5.97405910e-01 -6.59760118e-01 8.06924760e-01 7.39520669e-01
3.34631145e-01 -3.79089033e-03 7.55947232e-01 1.31025314e+00
-5.28192818e-01 -4.42637950e-01 3.42124581e-01 6.09790921e-01
-2.29302317e-01 -6.60807133e-01 -5.49251676e-01 -1.12470910e-01
-4.45625395e-01 -3.75248015e-01 -9.19066966e-01 3.03956419e-01
2.70660341e-01 1.34051096e+00 1.80191264e-01 -6.96316242e-01
4.58093762e-01 2.29617104e-01 5.00694036e-01 -7.72572815e-01
-7.96450138e-01 8.11383724e-01 -7.15825474e-03 -6.11321509e-01
-4.70533878e-01 -4.65244651e-01 -1.30451870e+00 4.07747477e-01
1.61906071e-02 9.37467739e-02 2.02860847e-01 4.43225563e-01
4.57385480e-01 1.23077333e+00 5.60476363e-01 -1.28698850e+00
-5.16950488e-02 -7.98328519e-01 -9.50001895e-01 7.49394178e-01
9.59169447e-01 -6.79684222e-01 -7.45157838e-01 5.51468968e-01] | [14.8845796585083, 4.983665943145752] |
828aa182-e48c-40da-8566-864ddd2d4fb3 | robust-contact-state-estimation-in-humanoid | 2208.00278 | null | https://arxiv.org/abs/2208.00278v1 | https://arxiv.org/pdf/2208.00278v1.pdf | Robust Contact State Estimation in Humanoid Walking Gaits | In this article, we propose a deep learning framework that provides a unified approach to the problem of leg contact detection in humanoid robot walking gaits. Our formulation accomplishes to accurately and robustly estimate the contact state probability for each leg (i.e., stable or slip/no contact). The proposed framework employs solely proprioceptive sensing and although it relies on simulated ground-truth contact data for the classification process, we demonstrate that it generalizes across varying friction surfaces and different legged robotic platforms and, at the same time, is readily transferred from simulation to practice. The framework is quantitatively and qualitatively assessed in simulation via the use of ground-truth contact data and is contrasted against state of-the-art methods with an ATLAS, a NAO, and a TALOS humanoid robot. Furthermore, its efficacy is demonstrated in base estimation with a real TALOS humanoid. To reinforce further research endeavors, our implementation is offered as an open-source ROS/Python package, coined Legged Contact Detection (LCD). | ['Panos Trahanias', 'Dimitrios Kanoulas', 'Michael Maravgakis', 'Stylianos Piperakis'] | 2022-07-30 | null | null | null | null | ['contact-detection'] | ['robots'] | [-9.83062834e-02 1.43742725e-01 -3.57541591e-01 1.36183664e-01
-3.77753645e-01 -1.18466839e-01 3.24273437e-01 -3.00728589e-01
-4.45044041e-01 9.88577425e-01 -3.72240096e-01 1.68154851e-01
-1.53452620e-01 -7.65036225e-01 -8.58797133e-01 -4.59426522e-01
-5.25387287e-01 7.77453244e-01 5.63212752e-01 -7.40321219e-01
6.63322583e-02 2.34307051e-01 -1.76640022e+00 -4.57538486e-01
8.01432550e-01 9.11300957e-01 4.02416050e-01 3.70856732e-01
1.10806155e+00 4.87105399e-01 -2.63946652e-01 2.00049337e-02
1.60747319e-01 -1.03359438e-01 -8.14181268e-01 -2.07078662e-02
1.29822940e-01 -3.19083869e-01 -3.37206542e-01 5.13241053e-01
7.82953680e-01 7.58834407e-02 8.37844551e-01 -1.53240609e+00
1.68884229e-02 1.97392955e-01 -2.86629289e-01 -1.13687053e-01
6.55677021e-01 4.24715698e-01 8.91678452e-01 -8.92926812e-01
7.82697439e-01 1.10229909e+00 1.13922060e+00 3.45233351e-01
-9.04155970e-01 -3.23735386e-01 -5.11052728e-01 2.82088190e-01
-1.22690248e+00 -2.57178605e-01 6.19323313e-01 -5.58872879e-01
1.10340071e+00 -1.55306414e-01 1.06658173e+00 1.52540171e+00
9.60325062e-01 8.51625204e-01 9.69641984e-01 -3.01026046e-01
3.99129778e-01 -6.37283325e-01 -9.36071724e-02 8.56505692e-01
7.36599267e-01 2.69733816e-01 -7.12606668e-01 -5.74751422e-02
8.97007763e-01 -4.51159477e-01 -1.98936090e-01 -9.87798810e-01
-1.25601363e+00 3.12053651e-01 4.11469430e-01 -1.98008165e-01
-7.32792556e-01 4.07863021e-01 4.13468271e-01 9.25030410e-02
-1.37227438e-02 1.17068216e-01 -3.81392390e-01 -5.29974461e-01
-4.65056628e-01 9.82378542e-01 1.17602587e+00 9.27987933e-01
3.01982522e-01 2.21385464e-01 3.72779340e-01 8.41276348e-01
3.40801001e-01 5.95866203e-01 2.41034612e-01 -1.26271391e+00
4.91529733e-01 3.14031094e-01 7.66277492e-01 -8.37159336e-01
-7.32632935e-01 -8.75935182e-02 -4.22165692e-01 8.35024416e-01
4.08075809e-01 -3.21260184e-01 -7.04683065e-01 1.62267482e+00
2.47804657e-01 -2.90220201e-01 1.89000696e-01 1.18459022e+00
2.80327767e-01 -9.05878842e-03 -2.47266912e-03 3.74667734e-01
1.27765954e+00 -7.74170280e-01 -6.00902140e-01 -6.70314133e-01
2.98736811e-01 -3.16912919e-01 1.18169272e+00 5.60168028e-01
-1.15790188e+00 -5.32556057e-01 -1.68787086e+00 -1.73566304e-02
-1.95860997e-01 2.83570439e-01 5.52401483e-01 2.19750836e-01
-7.16659963e-01 1.10909021e+00 -1.44264603e+00 -8.84917378e-01
3.29283462e-03 3.28692049e-01 -2.66131341e-01 3.38754386e-01
-1.53516591e+00 1.59867644e+00 1.12495460e-01 1.86466187e-01
-9.38241839e-01 -2.77693003e-01 -8.96627665e-01 -3.56546074e-01
2.94343561e-01 -1.05262160e+00 1.37668991e+00 -1.96942478e-01
-1.82184803e+00 9.26705718e-01 3.74696523e-01 -6.92504466e-01
9.97093678e-01 -8.16088974e-01 7.93775246e-02 2.78641254e-01
3.92732710e-01 4.90205109e-01 6.06339335e-01 -1.11952710e+00
-3.63528281e-01 -1.24124691e-01 -1.67438865e-01 5.65769970e-01
1.53997540e-01 -6.24242127e-01 -1.23983680e-03 -5.35731375e-01
2.25383192e-01 -1.26395357e+00 -1.92790955e-01 4.98390317e-01
-3.69350135e-01 2.70367432e-02 3.66692692e-01 -5.21013200e-01
5.81994176e-01 -1.72806954e+00 3.84070486e-01 1.74600542e-01
-8.57222304e-02 1.03777394e-01 5.48805118e-01 8.65244150e-01
4.42178577e-01 -4.04680699e-01 -4.39036816e-01 9.40258242e-03
1.46345481e-01 3.94339949e-01 7.20611662e-02 7.01119900e-01
1.19850732e-01 5.92183471e-01 -1.00441265e+00 -5.76681972e-01
3.28543901e-01 4.24088448e-01 -2.88608313e-01 -2.53402954e-03
8.90045017e-02 2.25051448e-01 -4.27656442e-01 9.06453192e-01
4.38883692e-01 1.17269456e-01 3.30176264e-01 -3.62227827e-01
-1.67434335e-01 1.55866712e-01 -1.27366948e+00 1.65659225e+00
-4.89933968e-01 3.26048672e-01 3.05142909e-01 -1.04233718e+00
1.12316430e+00 3.90591145e-01 6.28969967e-01 -3.53951365e-01
2.34312430e-01 7.62386143e-01 -4.30740342e-02 -5.65021157e-01
3.93881828e-01 2.58645505e-01 -1.73546910e-01 1.72541678e-01
4.19252694e-01 -4.22835380e-01 8.27351864e-03 -4.31097716e-01
1.02377963e+00 1.08993840e+00 5.10660589e-01 -7.07000434e-01
1.62344918e-01 2.35426098e-01 3.51782769e-01 6.25066996e-01
-5.28997242e-01 5.64198256e-01 1.21955998e-01 -1.86635613e-01
-1.05719626e+00 -1.69399202e+00 -2.76912570e-01 6.81103647e-01
6.73675239e-01 -2.69376747e-02 -1.10940301e+00 3.23322713e-01
5.44393539e-01 2.87472382e-02 -4.39668268e-01 -4.00089592e-01
-8.05766284e-01 -5.04610419e-01 7.29297221e-01 1.00756919e+00
5.61627269e-01 -1.35808384e+00 -1.56640828e+00 5.64325273e-01
-3.75888705e-01 -1.17087281e+00 3.07492793e-01 4.15710747e-01
-7.62676120e-01 -1.42267847e+00 -7.69888699e-01 -8.73219490e-01
1.92021427e-03 -1.02280408e-01 9.35734332e-01 -7.45900050e-02
-4.49654490e-01 4.14965302e-01 -3.15268397e-01 -1.66817918e-01
-7.09023178e-02 6.08705170e-02 6.27549529e-01 -7.09062338e-01
-7.85629973e-02 -8.21465135e-01 -7.52155125e-01 6.74483180e-01
-5.57703301e-02 -6.40564319e-03 4.56395090e-01 1.01622260e+00
5.48916459e-01 -3.49309325e-01 5.11814177e-01 -1.13693751e-01
5.52441597e-01 -3.73876065e-01 -3.14409852e-01 -2.22873256e-01
-4.31001246e-01 -4.43715602e-01 3.45480651e-01 -1.73229605e-01
-8.54187012e-01 1.57397225e-01 -2.87902117e-01 1.19811974e-01
6.70579867e-03 5.54548502e-01 2.24647075e-02 -3.04740131e-01
8.86802077e-01 4.43966463e-02 4.56549823e-01 -3.04540843e-01
1.15599751e-01 6.32776856e-01 9.61303294e-01 -8.19429755e-01
3.28464031e-01 6.93631887e-01 2.66325444e-01 -8.79701853e-01
-2.42926925e-03 -8.88789892e-02 -1.11830080e+00 -5.40012300e-01
6.64403498e-01 -8.57009053e-01 -1.03726220e+00 1.12349403e+00
-7.55737603e-01 -8.58624041e-01 2.09168959e-02 4.63600487e-01
-1.62418532e+00 4.39101607e-01 -8.85723054e-01 -7.77564049e-01
-4.24684793e-01 -1.14390111e+00 1.20628488e+00 -1.72513556e-02
-7.17550755e-01 -7.57033587e-01 1.49408565e-03 5.51957497e-03
-1.92520246e-02 8.53790879e-01 3.63256365e-01 8.82504061e-02
3.53440978e-02 -4.92664814e-01 3.00915122e-01 3.24423909e-02
1.81523368e-01 -9.95199233e-02 -6.69490755e-01 -4.99646753e-01
-1.71563670e-01 -7.50819385e-01 3.80649596e-01 4.01007891e-01
4.49279785e-01 1.26311600e-01 -5.60035586e-01 1.60161093e-01
1.18645442e+00 -1.92415476e-01 6.29341722e-01 8.02282929e-01
5.29266715e-01 7.58932650e-01 1.28975153e+00 4.94542420e-01
4.23056126e-01 9.75934625e-01 6.44949436e-01 4.40819412e-02
-1.08014315e-01 -2.45982513e-01 4.08674657e-01 6.21602833e-01
-5.43411016e-01 -1.19877629e-01 -1.13281024e+00 6.06366873e-01
-1.98154795e+00 -5.33006012e-01 -1.01350009e-01 2.04949975e+00
7.22129464e-01 5.24718225e-01 2.73145765e-01 6.40049279e-01
5.48009932e-01 -1.48549780e-01 -7.27729380e-01 -6.31436557e-02
1.00470893e-01 3.90100151e-01 4.84318018e-01 4.61763889e-01
-1.03088260e+00 9.57112134e-01 6.56858635e+00 4.53145713e-01
-1.04619825e+00 -1.61979094e-01 -3.55797321e-01 5.31563759e-01
6.23760164e-01 -3.60437185e-01 -3.24240118e-01 3.78167719e-01
5.60737610e-01 -3.29180062e-01 2.25867346e-01 1.18716133e+00
2.51236826e-01 -5.40229559e-01 -1.15340459e+00 5.17786801e-01
-2.77610898e-01 -7.83359289e-01 -3.98265451e-01 -2.58000374e-01
1.57866895e-01 1.97190151e-01 -2.66043514e-01 5.90738188e-03
2.53736854e-01 -6.84079885e-01 1.06387782e+00 5.57076097e-01
6.14703298e-01 -4.48979467e-01 7.72938907e-01 4.09530640e-01
-1.58676171e+00 -1.01770081e-01 -1.41237155e-01 -5.39623439e-01
3.95677179e-01 1.42243132e-01 -5.00343204e-01 7.52925217e-01
9.04828846e-01 9.23516870e-01 -2.78375655e-01 1.05746961e+00
-3.56860608e-01 2.98573375e-01 -5.49413502e-01 -2.70667583e-01
-7.00308830e-02 -3.97031195e-02 7.45170534e-01 8.74291301e-01
1.22317761e-01 -3.02267134e-01 4.03369427e-01 7.55870342e-01
7.66732097e-01 -2.63077676e-01 -5.84247649e-01 5.02885818e-01
5.29197216e-01 1.00670850e+00 -6.27895057e-01 -1.06559806e-02
1.41494632e-01 1.06450105e+00 1.50230780e-01 3.66855502e-01
-8.02453458e-01 -6.57170951e-01 7.22804844e-01 3.25950563e-01
-4.24792580e-02 -5.68436801e-01 -3.82652223e-01 -8.65735769e-01
4.56306338e-01 -4.14579272e-01 -7.16623589e-02 -9.47710574e-01
-1.21717453e+00 1.67246982e-01 4.34560835e-01 -1.58527493e+00
-6.30345047e-01 -9.49811637e-01 -4.22552675e-01 7.61768639e-01
-1.06400979e+00 -8.82025003e-01 -6.04270995e-01 2.95188397e-01
3.05130333e-01 1.77215129e-01 8.90401661e-01 1.41768441e-01
-4.15094756e-02 4.12199587e-01 -3.05881530e-01 -6.60466328e-02
6.52873933e-01 -1.24466026e+00 8.43836546e-01 3.04570347e-01
-8.29615951e-01 5.26887715e-01 1.14999080e+00 -8.05896819e-01
-1.45443416e+00 -6.54728949e-01 3.18375766e-01 -9.63910744e-02
8.48604620e-01 -2.74477065e-01 -8.28730285e-01 5.94464660e-01
-1.53844029e-01 1.33065143e-04 -4.72729132e-02 -3.31800580e-01
4.27301288e-01 1.27620772e-01 -1.29064810e+00 8.13307703e-01
1.37639129e+00 -1.43855885e-01 -8.71915638e-01 1.87439248e-01
1.84946343e-01 -9.09927666e-01 -9.91001964e-01 9.12642777e-01
1.11782086e+00 -6.82167768e-01 1.21023953e+00 -2.13467002e-01
3.84637207e-01 -2.22172320e-01 -1.62487522e-01 -1.17913902e+00
-1.11792274e-01 -4.62462664e-01 -3.23136240e-01 5.40496826e-01
-1.65789127e-01 -6.88953817e-01 9.07623291e-01 -2.45727420e-01
-2.85829931e-01 -1.14636016e+00 -1.37571108e+00 -1.11759233e+00
2.29012594e-01 -1.52910113e-01 -1.29893180e-02 5.91885030e-01
4.19030994e-01 -9.59240422e-02 -6.19880021e-01 2.62706488e-01
8.69876564e-01 -2.83055037e-01 9.54051673e-01 -1.37774229e+00
-7.89110214e-02 -6.39249384e-02 -8.15148771e-01 -7.67800570e-01
1.52532637e-01 -2.91303307e-01 5.76032639e-01 -1.64245903e+00
-4.13797259e-01 -3.53955358e-01 1.72406565e-02 5.64239979e-01
8.93742591e-02 3.36204022e-01 -1.80061325e-01 3.27074081e-01
-2.07469046e-01 5.00186861e-01 1.11719286e+00 1.88677460e-01
-7.08833486e-02 2.57101238e-01 1.82657167e-01 9.20867145e-01
9.22858715e-01 -5.18366322e-02 -1.17347412e-01 1.08257011e-01
-6.44191280e-02 2.10125223e-01 9.51413929e-01 -1.78442872e+00
2.52612121e-03 1.29435793e-01 1.76885054e-01 -4.77806091e-01
5.48466325e-01 -5.18366277e-01 2.36853287e-01 1.10629797e+00
5.55021800e-02 -1.46163270e-01 1.21092416e-01 4.96844411e-01
7.25885406e-02 2.69878283e-02 8.14269900e-01 -8.86793584e-02
-8.32661331e-01 -1.90098554e-01 -8.62338543e-01 -8.76709702e-04
9.24203336e-01 -6.43619657e-01 -1.68230355e-01 -2.41124585e-01
-9.53235686e-01 4.02560353e-01 7.40291417e-01 4.36972082e-01
4.18221444e-01 -1.34447253e+00 -3.33580136e-01 3.36569659e-02
2.40106344e-01 -2.36536846e-01 -9.95393097e-03 7.53306329e-01
-9.85051811e-01 1.22063264e-01 -9.98007953e-01 -9.20958459e-01
-8.07166398e-01 -1.51783735e-01 6.05482578e-01 6.26056865e-02
-9.36713517e-01 2.46756360e-01 -4.59709555e-01 -7.41425574e-01
1.51788861e-01 -4.79723990e-01 -7.01982602e-02 -4.23827916e-01
-2.70201147e-01 7.92967975e-01 2.11852595e-01 -6.79261088e-01
-4.54231590e-01 9.05107319e-01 5.87227881e-01 -2.77662575e-01
1.07067883e+00 -1.16089396e-01 3.40341598e-01 7.13744700e-01
4.99114931e-01 -5.06434739e-01 -1.83576584e+00 2.94878036e-01
1.65515974e-01 -2.17841621e-02 -6.02601230e-01 -7.14522481e-01
-5.09488344e-01 7.18490303e-01 9.40616906e-01 -2.33339414e-01
5.23864388e-01 -2.63510764e-01 8.52169573e-01 7.82758176e-01
1.10133564e+00 -1.45265436e+00 2.26760045e-01 4.79127973e-01
1.09452891e+00 -1.07923901e+00 3.25533509e-01 -8.04949224e-01
-4.89305884e-01 1.04008019e+00 7.77205110e-01 -8.14136267e-01
6.51023030e-01 6.10129356e-01 7.76593387e-02 -8.99434164e-02
-3.74852747e-01 -5.04748523e-02 -5.40804565e-02 1.06394458e+00
1.90788165e-01 2.07664326e-01 -5.64244270e-01 5.65164983e-01
-4.49274093e-01 4.44835991e-01 5.82393944e-01 1.72557592e+00
-6.31021798e-01 -7.79659688e-01 -1.90052941e-01 6.11527339e-02
-3.49922746e-01 4.88002092e-01 -2.22844601e-01 1.50616157e+00
-1.90191209e-01 6.95516706e-01 -9.55758616e-02 -4.86658782e-01
6.99905455e-01 -2.67748326e-01 8.00160289e-01 -4.40171421e-01
-2.41018027e-01 -2.87638634e-01 2.43620440e-01 -8.41711640e-01
-3.68342042e-01 -7.20046341e-01 -1.56875324e+00 1.40278563e-02
-1.24937154e-01 -1.90247014e-01 3.95444483e-01 7.59199917e-01
6.61467910e-02 4.55256730e-01 1.23869017e-01 -1.80582380e+00
-7.14493632e-01 -1.01601076e+00 -7.09111750e-01 3.04188102e-01
1.79661617e-01 -1.82793331e+00 -2.66052127e-01 2.26034336e-02] | [4.820591926574707, 1.0969507694244385] |
79bc318f-354e-4825-baaa-397940cdb88d | metropolis-hastings-algorithm-in-joint | 2305.19936 | null | https://arxiv.org/abs/2305.19936v1 | https://arxiv.org/pdf/2305.19936v1.pdf | Metropolis-Hastings algorithm in joint-attention naming game: Experimental semiotics study | In this study, we explore the emergence of symbols during interactions between individuals through an experimental semiotic study. Previous studies investigate how humans organize symbol systems through communication using artificially designed subjective experiments. In this study, we have focused on a joint attention-naming game (JA-NG) in which participants independently categorize objects and assign names while assuming their joint attention. In the theory of the Metropolis-Hastings naming game (MHNG), listeners accept provided names according to the acceptance probability computed using the Metropolis-Hastings (MH) algorithm. The theory of MHNG suggests that symbols emerge as an approximate decentralized Bayesian inference of signs, which is represented as a shared prior variable if the conditions of MHNG are satisfied. This study examines whether human participants exhibit behavior consistent with MHNG theory when playing JA-NG. By comparing human acceptance decisions of a partner's naming with acceptance probabilities computed in the MHNG, we tested whether human behavior is consistent with the MHNG theory. The main contributions of this study are twofold. First, we reject the null hypothesis that humans make acceptance judgments with a constant probability, regardless of the acceptance probability calculated by the MH algorithm. This result suggests that people followed the acceptance probability computed by the MH algorithm to some extent. Second, the MH-based model predicted human acceptance/rejection behavior more accurately than the other four models: Constant, Numerator, Subtraction, and Binary. This result indicates that symbol emergence in JA-NG can be explained using MHNG and is considered an approximate decentralized Bayesian inference. | ['Akira Taniguchi', 'Yosinobu Hagiwara', 'Tadahiro Taniguchi', 'Ryota Okumura'] | 2023-05-31 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [-1.27817959e-01 3.45327646e-01 -5.40418401e-02 -2.48532206e-01
4.81689662e-01 -3.67169410e-01 8.71494651e-01 2.60384288e-02
-9.71771359e-01 6.39411807e-01 -4.38577048e-02 -4.21514601e-01
-2.32571974e-01 -7.92683840e-01 -3.82987171e-01 -3.87887836e-01
7.69960657e-02 8.93937647e-01 -1.19374819e-01 -2.62777824e-02
4.79920119e-01 2.33301625e-01 -1.63200545e+00 -3.58807504e-01
7.46621192e-01 3.34633738e-01 1.66580930e-01 7.76904404e-01
3.67578678e-02 6.84098184e-01 -8.93017650e-01 -7.05064654e-01
2.32471019e-01 -9.63499725e-01 -6.69448197e-01 -4.35007513e-01
1.61833670e-02 -6.22686207e-01 -2.27033123e-02 1.22909772e+00
3.75614017e-01 3.01104814e-01 9.52892780e-01 -1.52017069e+00
-1.13809025e+00 1.10523725e+00 -1.59843206e-01 -9.22305286e-02
6.96884573e-01 9.66388434e-02 9.66003239e-01 -6.15912199e-01
7.13345110e-01 1.61301196e+00 3.68041664e-01 4.98624891e-01
-1.46861959e+00 -7.74868727e-01 7.65837356e-02 2.29081422e-01
-2.11656833e+00 -2.12676078e-01 5.18322468e-01 -4.22776878e-01
5.67468345e-01 4.91456240e-01 1.36630261e+00 9.03232396e-01
3.05456579e-01 3.94467145e-01 1.40520966e+00 -6.96092665e-01
7.43680537e-01 1.56654403e-01 1.02517441e-01 3.30819041e-01
8.48687589e-01 1.61225215e-01 -7.56662369e-01 -5.22904992e-01
1.14437449e+00 -5.78219652e-01 6.87649548e-02 -4.40408438e-02
-1.43067706e+00 7.46237814e-01 9.15318057e-02 6.42686665e-01
-6.39682710e-01 3.75791728e-01 -3.75692844e-01 3.29431266e-01
2.96178851e-02 5.08684635e-01 -1.06894717e-01 -2.96759546e-01
-8.51657569e-01 4.16494727e-01 1.20489371e+00 6.00884616e-01
2.56230414e-01 -8.98316875e-02 -5.24378158e-02 6.76673055e-01
9.86273348e-01 9.32618737e-01 4.81937259e-01 -1.20664299e+00
-4.58635122e-01 -5.69298044e-02 4.29484785e-01 -1.09594977e+00
-2.83476502e-01 -3.39963734e-01 -5.87781131e-01 2.69723088e-01
8.22492778e-01 1.25469685e-01 -4.82163101e-01 2.51716590e+00
1.79468557e-01 -7.80189037e-02 -1.49439499e-01 8.79892051e-01
4.06363875e-01 4.28345472e-01 4.69549835e-01 -5.24102807e-01
1.47699749e+00 -7.65280649e-02 -1.12386978e+00 2.86459755e-02
4.10259545e-01 -6.36623085e-01 9.98391449e-01 6.22520626e-01
-9.67234075e-01 -4.78842050e-01 -7.97676206e-01 2.05227226e-01
-1.37535080e-01 -2.79166609e-01 8.47484171e-01 1.17122984e+00
-1.34649718e+00 2.89212197e-01 -6.58662498e-01 -9.62738514e-01
-1.72943085e-01 2.89131790e-01 5.52184992e-02 3.48592609e-01
-1.34268093e+00 1.02256656e+00 5.68924882e-02 2.75655568e-01
-5.54681599e-01 2.41576850e-01 -5.08618116e-01 6.11119121e-02
-1.98181942e-02 -9.80591118e-01 1.44290173e+00 -1.33650446e+00
-1.94210494e+00 9.45198238e-01 -9.57581177e-02 -1.40516847e-01
4.54088509e-01 -5.23244310e-03 -3.63138795e-01 9.71129313e-02
3.62246007e-01 9.54776943e-01 4.89010125e-01 -1.31767249e+00
4.87373173e-02 -1.78508192e-01 -1.18135080e-01 3.94886076e-01
3.45659703e-01 1.94224358e-01 2.10264891e-01 -6.95046782e-01
3.97148341e-01 -9.40847158e-01 1.35355920e-01 1.77701339e-01
-2.07082465e-01 -5.52010894e-01 -2.08129928e-01 -3.95532131e-01
1.29342914e+00 -2.10197711e+00 -1.67005938e-02 6.84745550e-01
4.10001040e-01 -1.55471951e-01 6.59502894e-02 5.54012775e-01
1.43903106e-01 4.14587826e-01 -5.85325547e-02 -1.73491478e-01
7.09061921e-01 2.45304585e-01 -1.05754428e-01 6.85197949e-01
-7.48000741e-01 7.57635355e-01 -8.12993228e-01 -6.49738848e-01
5.79603389e-02 1.16209961e-01 -5.63422561e-01 1.89526781e-01
2.87428647e-01 7.50260502e-02 4.03151289e-02 5.57200134e-01
6.99506760e-01 -2.26976335e-01 5.85688174e-01 3.45120281e-01
-2.86862224e-01 -4.94437031e-02 -1.19265294e+00 1.00088513e+00
2.33467519e-01 5.23465872e-01 7.95761868e-02 -5.97040653e-01
8.58125269e-01 2.57470131e-01 -2.41736993e-01 -6.64633334e-01
5.30201554e-01 3.68365824e-01 7.15971768e-01 -1.00855969e-01
4.50656086e-01 -4.19293791e-01 -3.30413491e-01 8.83444130e-01
-1.08128801e-01 -4.41687942e-01 6.71089068e-02 5.88122368e-01
6.39068127e-01 -9.77145582e-02 8.79703343e-01 -5.29673517e-01
6.04421794e-02 -6.33357465e-01 6.18238151e-01 1.65619111e+00
-5.65563381e-01 8.56598318e-02 4.80140477e-01 -7.37608075e-02
-8.25157404e-01 -1.57097077e+00 -2.86135435e-01 1.08148503e+00
5.79520941e-01 -2.18437001e-01 -1.05531108e+00 7.97463134e-02
3.13676223e-02 1.55707681e+00 -7.04324663e-01 -3.53795230e-01
1.42455436e-02 -5.06438017e-01 9.32091296e-01 5.08222692e-02
5.18992960e-01 -1.20295286e+00 -1.02677405e+00 3.67794707e-02
-3.97169799e-01 -3.17559183e-01 -1.33357733e-01 -2.92259008e-01
-4.71232057e-01 -4.98945385e-01 -6.70409143e-01 -3.16680998e-01
6.21398032e-01 1.73532981e-02 6.77400172e-01 4.71750557e-01
1.77906062e-02 7.45478213e-01 -2.25111336e-01 -3.45118225e-01
-3.13495100e-01 -4.20674711e-01 5.05773067e-01 -9.07424316e-02
5.54224253e-01 -6.71807408e-01 -4.41372037e-01 4.23977971e-01
-7.33132064e-01 -6.32365886e-03 5.83549261e-01 3.76076609e-01
-1.56346053e-01 -4.23097461e-01 5.28144300e-01 -1.75099075e-01
9.38045919e-01 -4.98533010e-01 -3.57729197e-01 1.92684188e-01
-6.43638074e-01 -1.02019273e-01 7.33656110e-03 -8.02483141e-01
-1.04280376e+00 -5.47316849e-01 2.03467220e-01 2.43670285e-01
-3.80122066e-01 5.17373443e-01 -1.47249997e-02 2.44848087e-01
5.36181271e-01 1.85558334e-01 1.41908184e-01 -1.67695343e-01
4.11359102e-01 8.38886321e-01 3.16968471e-01 -8.03952813e-01
4.89288539e-01 1.97560102e-01 -2.61820883e-01 -1.01937425e+00
-2.92544395e-01 2.56695062e-01 -3.95353943e-01 -6.91092789e-01
9.81153190e-01 -5.92382669e-01 -1.42440104e+00 8.35218132e-01
-1.04430079e+00 -4.78636146e-01 -2.96280771e-01 1.16464019e+00
-7.09858835e-01 5.35084188e-01 -6.00311697e-01 -1.55501485e+00
2.56207585e-01 -6.68533862e-01 4.46041495e-01 2.65807778e-01
-1.26807642e+00 -8.22814524e-01 7.15841502e-02 1.42901674e-01
5.47026813e-01 -2.28896052e-01 6.90496564e-01 -9.55966949e-01
-4.63199019e-01 8.00338760e-02 2.50266612e-01 -4.02039945e-01
-2.93736439e-02 3.03490549e-01 -5.54747581e-01 -9.04965028e-02
2.77576625e-01 -1.84328735e-01 2.49868467e-01 5.55807710e-01
3.38846624e-01 -3.31085980e-01 -1.12271078e-01 5.46167344e-02
8.11070144e-01 4.88596261e-01 5.04896462e-01 2.26334840e-01
3.61765213e-02 6.61707699e-01 3.23264867e-01 5.66494763e-01
7.43831217e-01 6.98249638e-01 -2.95380061e-03 4.88309175e-01
2.58202285e-01 -3.87008607e-01 4.34013724e-01 8.39087725e-01
-5.80433309e-02 -4.96574879e-01 -8.13151479e-01 3.11485976e-01
-1.78680015e+00 -1.08540738e+00 -2.58286208e-01 2.45991850e+00
8.66408885e-01 1.40243605e-01 -7.32098380e-03 -4.15629968e-02
1.08395982e+00 -2.25387573e-01 -1.39804542e-01 -5.66956282e-01
-1.71636701e-01 5.91360033e-02 2.44997695e-01 8.49839747e-01
-2.19157353e-01 8.28063786e-01 7.30699062e+00 4.85179573e-01
-6.29967928e-01 3.51695083e-02 3.58481497e-01 1.51248112e-01
-3.55483681e-01 3.82071584e-01 -3.66244972e-01 5.97062588e-01
1.00090182e+00 -4.83481258e-01 6.28757000e-01 3.75124365e-01
1.83696106e-01 -9.55083787e-01 -1.02571213e+00 8.86389673e-01
2.71574557e-01 -3.22375208e-01 1.21152215e-01 3.74214441e-01
2.66929060e-01 -7.49987721e-01 1.35798201e-01 1.60939693e-01
7.89638937e-01 -7.95934200e-01 1.51694834e+00 9.31716263e-01
4.66065764e-01 -1.81474715e-01 7.92440534e-01 4.32143837e-01
-7.39147484e-01 7.33167082e-02 -2.59815395e-01 -9.12921965e-01
5.45753241e-01 3.95058334e-01 -8.37688208e-01 -4.02674340e-02
4.79002535e-01 -9.45846289e-02 -4.23867762e-01 1.00762188e+00
-5.46805501e-01 7.37558961e-01 -7.60768533e-01 -3.88298661e-01
-1.78412363e-01 -4.40280706e-01 5.26687682e-01 6.38338327e-01
4.24650639e-01 3.86702627e-01 -2.89649457e-01 1.46697450e+00
3.20439935e-01 2.19290197e-01 -3.89449865e-01 -1.18820630e-01
1.05015159e+00 6.28816426e-01 -1.43204427e+00 -6.82044387e-01
-2.24891547e-02 9.96066332e-01 -9.92970355e-03 3.36465716e-01
-9.00725722e-01 -2.02156484e-01 1.97672635e-01 5.67983985e-02
-1.55878530e-04 -2.21613601e-01 -5.18559277e-01 -1.02826238e+00
-5.78467131e-01 -5.85461140e-01 -1.48668913e-02 -9.98785138e-01
-1.18980837e+00 9.98374373e-02 6.26523376e-01 -7.03121424e-01
-1.36002511e-01 -1.06663555e-01 -2.88849086e-01 7.46841311e-01
-1.98123723e-01 -7.57451832e-01 -1.08325079e-01 3.35595906e-01
3.88528430e-03 3.11871141e-01 7.11113095e-01 -2.02687681e-01
-3.77639920e-01 5.19156098e-01 -2.28835821e-01 -8.28996226e-02
8.11462998e-01 -1.11679196e+00 1.90096900e-01 4.99900550e-01
6.87710419e-02 1.48799992e+00 9.14889097e-01 -1.15761364e+00
-6.08990669e-01 6.37454689e-02 1.37680137e+00 -3.04834604e-01
3.35219681e-01 -4.65759277e-01 -7.05568969e-01 6.71797037e-01
3.61293852e-01 -5.46895027e-01 1.06681168e+00 -9.24613420e-03
-1.38769835e-01 2.87068933e-01 -1.27724624e+00 9.11581516e-01
1.15109277e+00 -5.80334723e-01 -1.30790341e+00 -6.82588108e-03
2.18733713e-01 2.95487463e-01 -6.75982058e-01 -8.24981332e-02
1.08942318e+00 -9.42510545e-01 5.94177961e-01 -5.35719804e-02
-2.89715260e-01 -5.04782379e-01 -2.58425564e-01 -9.94906664e-01
-7.83015907e-01 -4.08446163e-01 4.81062531e-01 1.12374246e+00
1.41877994e-01 -1.10656929e+00 1.19240634e-01 1.03531492e+00
4.43202794e-01 1.98764816e-01 -1.20423436e+00 -6.12628400e-01
-4.44954038e-02 -2.91687995e-01 4.98960465e-01 1.22221768e+00
4.41697180e-01 1.48715958e-01 -4.52509582e-01 -1.46982685e-01
9.35940862e-01 -4.31022376e-01 6.94798291e-01 -1.64019048e+00
-4.61785376e-01 -6.99301004e-01 -3.55833143e-01 -1.06239104e+00
1.62731424e-01 -6.37542903e-01 2.37530053e-01 -1.47953665e+00
2.38022581e-01 -6.96586147e-02 -3.12289651e-02 2.81016290e-01
2.88988296e-02 4.18318287e-02 5.22479653e-01 5.17242670e-01
-6.03076994e-01 5.82169712e-01 7.29444385e-01 4.10868347e-01
-3.01677167e-01 -2.63769478e-01 -8.48840654e-01 7.77926207e-01
7.02266157e-01 -5.79857588e-01 -3.03736925e-01 1.08053796e-01
6.87078774e-01 4.80221286e-02 5.09238124e-01 -7.91768909e-01
4.99148935e-01 -2.63005912e-01 3.40921968e-01 -2.71412820e-01
2.85599351e-01 -6.04905665e-01 8.10335994e-01 7.74047911e-01
-3.24031591e-01 -3.11537087e-02 -5.58734611e-02 3.93176764e-01
3.30664843e-01 -2.56973267e-01 6.65752947e-01 -3.85649428e-02
-2.38478050e-01 -8.98075163e-01 -1.52164495e+00 -4.09184605e-01
1.03534949e+00 -5.83647490e-01 -1.93329468e-01 -9.34430718e-01
-9.37562764e-01 -1.94615081e-01 6.80284381e-01 1.80753563e-02
5.74944615e-01 -1.34363127e+00 -3.91829789e-01 2.50697970e-01
-1.99978039e-01 -7.50437260e-01 -4.98943450e-03 1.09141743e+00
-7.00926602e-01 2.40205929e-01 -3.48289162e-01 -3.47437650e-01
-8.63710225e-01 3.78976107e-01 1.52368173e-01 4.19969797e-01
1.15537710e-01 9.48689103e-01 3.13142627e-01 -5.21780610e-01
-1.47296339e-01 -2.89857656e-01 -4.01180089e-02 1.88904349e-02
2.84744054e-01 5.51095426e-01 -8.41366291e-01 -8.98394585e-01
-4.45348024e-01 4.19581503e-01 1.53650224e-01 -9.31642771e-01
6.43631458e-01 -5.27640224e-01 -5.57440817e-01 1.04333389e+00
2.85236597e-01 2.94784755e-01 -4.95992929e-01 9.31057557e-02
-3.05207610e-01 -6.68972313e-01 -4.39029485e-01 -9.09803212e-01
-1.40144959e-01 1.89206079e-01 4.49333906e-01 3.50299895e-01
5.07104814e-01 2.08366707e-01 -2.05602527e-01 3.58833820e-01
5.82302392e-01 -1.08605337e+00 -1.31914273e-01 2.12609485e-01
6.55114114e-01 -5.48615456e-01 -9.16468203e-02 -3.12306076e-01
-5.74926794e-01 6.18826985e-01 6.32264495e-01 5.72385639e-03
6.90356076e-01 -1.71108067e-01 -1.20412953e-01 -1.14801474e-01
-7.47743189e-01 4.82583530e-02 -2.08386898e-01 5.13374329e-01
4.72461343e-01 4.48870778e-01 -1.27903926e+00 8.41811001e-01
-7.43346274e-01 1.43044859e-01 8.63931894e-01 8.74386132e-01
-6.56364858e-01 -9.89485562e-01 -8.07391644e-01 2.72883594e-01
-8.13360587e-02 -1.54938787e-01 -9.48806524e-01 8.86129975e-01
2.08342239e-01 1.24740005e+00 3.73677641e-01 -2.34022021e-01
2.74174176e-02 2.44368389e-01 4.09630775e-01 -4.21117127e-01
-4.11780596e-01 1.27755612e-01 -1.53745458e-01 -2.05310181e-01
-4.07939941e-01 -9.76779342e-01 -1.12343359e+00 -6.78094268e-01
-5.31388462e-01 6.85677111e-01 2.79405892e-01 9.41720366e-01
-2.07996164e-02 -5.01051359e-02 -1.90072842e-02 -5.00196099e-01
-6.38501227e-01 -1.23972881e+00 -1.00338984e+00 1.71977863e-01
-2.53143847e-01 -7.79873610e-01 -9.56928790e-01 -1.31012127e-01] | [9.73060417175293, 7.580016136169434] |
3b6ae408-4f66-4fcc-a8df-dfdd1d6d376a | a-constraints-fusion-induced-symmetric | 2302.12114 | null | https://arxiv.org/abs/2302.12114v1 | https://arxiv.org/pdf/2302.12114v1.pdf | A Constraints Fusion-induced Symmetric Nonnegative Matrix Factorization Approach for Community Detection | Community is a fundamental and critical characteristic of an undirected social network, making community detection be a vital yet thorny issue in network representation learning. A symmetric and non-negative matrix factorization (SNMF) model is frequently adopted to address this issue owing to its great interpretability and scalability. However, it adopts a single latent factor matrix to represent an undirected network for precisely representing its symmetry, which leads to loss of representation learning ability due to the reduced latent space. Motivated by this discovery, this paper proposes a novel Constraints Fusion-induced Symmetric Nonnegative Matrix Factorization (CFS) model that adopts three-fold ideas: a) Representing a target undirected network with multiple latent factor matrices, thus preserving its representation learning capacity; b) Incorporating a symmetry-regularizer that preserves the symmetry of the learnt low-rank approximation to the adjacency matrix into the loss function, thus making the resultant detector well-aware of the target network's symmetry; and c) Introducing a graph-regularizer that preserves local invariance of the network's intrinsic geometry, thus making the achieved detector well-aware of community structure within the target network. Extensively empirical studies on eight real-world social networks from industrial applications demonstrate that the proposed CFS model significantly outperforms state-of-the-art models in achieving highly-accurate community detection results. | ['Xin Luo', 'ZhiGang Liu'] | 2023-02-23 | null | null | null | null | ['community-detection'] | ['graphs'] | [ 2.42604852e-01 -1.33120403e-01 -2.55767405e-01 1.52266055e-01
1.70476399e-02 -5.16415119e-01 3.49725544e-01 -8.71148407e-02
1.97301626e-01 2.26539060e-01 1.27489626e-01 -2.17351824e-01
-4.72325146e-01 -7.97064722e-01 -2.13013515e-01 -8.53816152e-01
-3.82053167e-01 3.62534106e-01 1.36621416e-01 -1.85973555e-01
2.30719134e-01 4.76613343e-01 -1.11626899e+00 1.83272734e-01
9.32711005e-01 5.94534516e-01 1.22981735e-01 3.98246378e-01
8.73466805e-02 8.29156935e-01 -3.25530201e-01 -2.23108977e-01
3.57352138e-01 -2.32659608e-01 -4.69662279e-01 6.12189710e-01
3.59806120e-02 1.97669834e-01 -7.62034595e-01 1.36748683e+00
1.62243485e-01 -1.73454974e-02 7.91430593e-01 -1.58999062e+00
-6.79229200e-01 5.72271883e-01 -1.17130053e+00 1.65127829e-01
3.61071855e-01 -1.58030257e-01 1.39657497e+00 -7.58974254e-01
6.26508832e-01 1.40385652e+00 5.96244454e-01 5.75015321e-02
-1.53841209e+00 -1.03092647e+00 3.43793988e-01 9.76851135e-02
-1.78528774e+00 -4.72068861e-02 1.11418796e+00 -7.79667258e-01
5.67502201e-01 2.31169298e-01 7.29512930e-01 8.15144897e-01
2.29431868e-01 6.31949961e-01 6.83230817e-01 -2.90453762e-01
-3.46252359e-02 -9.94970649e-02 1.26114758e-02 8.85277152e-01
6.10691786e-01 -2.16058299e-01 -3.34532499e-01 -4.81736869e-01
1.05085170e+00 4.80037838e-01 -3.10535878e-01 -1.07043481e+00
-1.28307104e+00 1.02340484e+00 7.65544116e-01 3.69620264e-01
-3.64779562e-01 -1.17652953e-01 4.35788810e-01 3.02216053e-01
1.57128870e-01 1.47377357e-01 3.47295493e-01 4.05448586e-01
-6.01627588e-01 -1.10632867e-01 6.61124706e-01 9.64805603e-01
7.10039139e-01 2.53586084e-01 1.38886899e-01 6.15344167e-01
6.12579286e-01 4.11274225e-01 2.18373537e-01 -6.04061604e-01
7.10928798e-01 1.29187405e+00 -2.56173998e-01 -1.92338467e+00
-4.31206763e-01 -8.28393281e-01 -1.64736605e+00 1.17052859e-02
6.35098666e-02 1.06073812e-01 -3.85946840e-01 1.61289656e+00
1.46478683e-01 3.40838432e-01 5.30371703e-02 8.48403573e-01
3.98607671e-01 4.51564014e-01 -3.64687264e-01 -4.57282305e-01
1.22997677e+00 -7.18954027e-01 -4.82410192e-01 -2.05907121e-01
2.68561214e-01 -7.63178885e-01 5.01111865e-01 3.15213770e-01
-2.87504882e-01 -4.56332952e-01 -1.39846396e+00 4.75469112e-01
7.47600347e-02 1.13446422e-01 9.89354610e-01 5.93125165e-01
-6.72769070e-01 2.79765248e-01 -5.67554533e-01 -3.85985762e-01
1.30012393e-01 3.95848781e-01 -6.37284935e-01 -4.09379333e-01
-1.00370181e+00 1.49708882e-01 3.97708863e-01 2.77292252e-01
-5.42474270e-01 -1.95278049e-01 -6.19667828e-01 1.13563217e-01
6.01259291e-01 -7.08820641e-01 3.58259261e-01 -9.15359735e-01
-9.45495963e-01 5.16801000e-01 -4.38909195e-02 -1.28045142e-01
4.47632164e-01 4.00199473e-01 -6.65834904e-01 2.87187189e-01
3.18053037e-01 -5.41478992e-02 1.28167188e+00 -1.37216401e+00
-2.94090033e-01 -4.07395810e-01 1.80395871e-01 1.70423552e-01
-5.34885228e-01 -3.41471344e-01 -3.47999305e-01 -8.23048472e-01
9.05016065e-01 -1.11214733e+00 -4.02184069e-01 -7.76616484e-02
-7.41118908e-01 -8.24954957e-02 1.13366795e+00 -5.00164986e-01
1.51908529e+00 -2.16881633e+00 3.57776821e-01 9.20291007e-01
8.81909788e-01 1.50112107e-01 -3.09103489e-01 7.07889438e-01
-4.53606158e-01 3.95994894e-02 -1.03392832e-01 -8.82629827e-02
-4.38035280e-01 9.45213661e-02 -1.21941946e-01 6.70512736e-01
2.36112297e-01 4.05902922e-01 -1.21904075e+00 -2.89540470e-01
2.39164904e-01 5.36848187e-01 -6.14707828e-01 -7.67823532e-02
4.43328768e-01 3.66301894e-01 -7.60107875e-01 5.02561331e-01
7.73906887e-01 -5.86685121e-01 7.46825576e-01 -5.29595077e-01
1.23735145e-01 -3.99979323e-01 -1.93020523e+00 1.31217325e+00
3.32741551e-02 2.25499317e-01 3.96571487e-01 -1.07156277e+00
1.08228147e+00 3.11500520e-01 9.33889151e-01 -2.62074381e-01
8.77318084e-02 -2.13449430e-02 2.82036424e-01 -1.03505313e-01
1.96011916e-01 5.85009381e-02 2.26777464e-01 5.32574534e-01
-1.27661154e-01 2.86788613e-01 3.92856717e-01 7.92754292e-01
1.23313498e+00 -6.19634032e-01 3.59960228e-01 -4.46594685e-01
9.18026447e-01 -5.40751338e-01 8.67787898e-01 3.47921908e-01
-1.43884793e-01 4.75556165e-01 6.55617356e-01 -2.50984251e-01
-7.64723778e-01 -1.25993109e+00 1.41823798e-01 6.16412640e-01
2.78729200e-01 -5.74815810e-01 -4.24656034e-01 -5.40316522e-01
2.43513852e-01 -1.76197991e-01 -4.92420733e-01 -2.56778240e-01
-2.55542666e-01 -6.96626961e-01 1.25220641e-01 3.61230344e-01
3.74828041e-01 -5.42363048e-01 2.72777319e-01 1.61659509e-01
-4.18951243e-01 -9.22004938e-01 -6.12290084e-01 -3.07797883e-02
-9.29796040e-01 -1.60668159e+00 -4.58595216e-01 -8.21398139e-01
1.18289828e+00 1.00955081e+00 6.56059563e-01 5.25106668e-01
-2.54714400e-01 2.16433764e-01 -3.65907162e-01 2.92973101e-01
-3.30944270e-01 -7.83830136e-03 5.33040702e-01 5.44948161e-01
1.67901248e-01 -8.33651841e-01 -3.94035608e-01 4.91064280e-01
-8.58794928e-01 6.02730401e-02 6.39323890e-01 9.19189215e-01
3.04840386e-01 6.37135088e-01 5.17638087e-01 -9.03987229e-01
6.90450609e-01 -5.30577540e-01 -5.35284340e-01 2.23790482e-01
-6.64760470e-01 -8.47007036e-02 5.12863696e-01 -4.74029243e-01
-5.66570818e-01 1.85562134e-01 5.18779099e-01 -7.32943714e-01
5.79637349e-01 5.91323614e-01 -2.50395745e-01 -3.32526892e-01
5.72577596e-01 2.40637422e-01 2.82995760e-01 -2.73468047e-01
2.51109451e-01 5.26415050e-01 2.14914665e-01 -4.93397832e-01
1.50321519e+00 5.66009164e-01 4.54224437e-01 -1.01062572e+00
-5.15250504e-01 -1.15641427e+00 -9.27277207e-01 -2.30888054e-01
4.88117635e-01 -1.17685318e+00 -9.08062160e-01 3.47071201e-01
-8.43827367e-01 7.32893705e-01 1.22010782e-01 4.59515989e-01
-2.35441089e-01 9.58638728e-01 -6.65587544e-01 -9.26846445e-01
-2.29704484e-01 -9.38185811e-01 6.49351895e-01 -1.22538783e-01
4.60505486e-02 -8.40385318e-01 -1.17423333e-01 4.03293490e-01
1.55073076e-01 4.66650724e-01 9.62920904e-01 -3.30170721e-01
-6.70637846e-01 -4.87660348e-01 -6.55164182e-01 5.39399564e-01
4.19901580e-01 2.55621910e-01 -6.71006382e-01 -8.80721629e-01
-1.47740647e-01 2.69775897e-01 7.08631396e-01 9.25678462e-02
6.27863824e-01 -2.65708685e-01 -5.27409494e-01 2.83785790e-01
1.25514901e+00 -1.15257457e-01 4.31901932e-01 5.00795208e-02
1.15430403e+00 5.09440124e-01 3.63885671e-01 4.56246853e-01
1.79069400e-01 4.00319159e-01 6.52302682e-01 -8.74838904e-02
1.88332036e-01 -3.20689857e-01 2.41372198e-01 1.31755757e+00
-3.28063309e-01 -6.39211386e-02 -8.02592993e-01 2.84421206e-01
-2.16033792e+00 -1.21655321e+00 -2.70457119e-01 2.12368536e+00
2.46154875e-01 4.21158880e-01 2.45932221e-01 5.13054430e-01
1.25286257e+00 3.75445098e-01 -4.77161288e-01 1.11580305e-01
-1.50471777e-01 -4.83502179e-01 2.58007258e-01 2.55805701e-01
-1.10524476e+00 4.11310226e-01 5.35961485e+00 9.40762818e-01
-9.16452706e-01 -1.97144091e-01 1.16593115e-01 4.24302518e-01
-1.88492060e-01 1.05502836e-01 -4.53922659e-01 2.06341118e-01
2.13265046e-01 -5.98426640e-01 6.17323995e-01 1.00832462e+00
1.97709411e-01 5.02822936e-01 -8.47158074e-01 1.02997148e+00
2.27377992e-02 -1.19860899e+00 5.09049475e-01 4.79928106e-01
8.27595174e-01 -1.75162718e-01 -5.06535396e-02 1.81667566e-01
1.04701824e-01 -7.02970743e-01 3.99086505e-01 3.28659296e-01
7.37756073e-01 -9.77766395e-01 6.71431065e-01 3.36119354e-01
-1.81113768e+00 -4.00851101e-01 -5.57281137e-01 -2.21911207e-01
-2.14361623e-02 8.36188853e-01 -8.64643037e-01 7.70761430e-01
2.56318510e-01 1.18667674e+00 -7.05974042e-01 9.74557042e-01
-1.04493901e-01 4.56853032e-01 -2.68977523e-01 3.85636419e-01
1.56377867e-01 -6.39481604e-01 9.71120536e-01 8.27040851e-01
3.40983532e-02 -1.72258466e-01 8.02570403e-01 7.52402246e-01
-2.04365611e-01 1.35246336e-01 -8.09026599e-01 -3.61426026e-01
4.57113534e-01 1.44889283e+00 -9.36978042e-01 1.50086924e-01
-4.01570082e-01 7.65923738e-01 2.30756432e-01 4.66510981e-01
-4.49708492e-01 -2.42855340e-01 7.28716612e-01 3.51694286e-01
1.87990084e-01 -2.81810611e-01 9.67364684e-02 -1.33014798e+00
6.73433095e-02 -8.84010792e-01 4.57358003e-01 -3.05778056e-01
-1.60712755e+00 5.49656868e-01 -3.26639235e-01 -1.68278062e+00
-1.24506675e-01 -4.83198375e-01 -6.41576767e-01 7.71919608e-01
-1.16435242e+00 -1.46540654e+00 -3.01332206e-01 7.70962298e-01
1.04587555e-01 -4.47020262e-01 6.77695453e-01 5.69960296e-01
-8.10990393e-01 4.80949312e-01 2.61513889e-01 4.62682635e-01
4.14645910e-01 -1.05662680e+00 1.98492005e-01 1.02769673e+00
1.42257482e-01 1.22030067e+00 4.23050672e-01 -7.38124609e-01
-1.60697114e+00 -1.06579614e+00 4.91902292e-01 -1.31440938e-01
1.08389342e+00 -4.52161491e-01 -7.52761781e-01 5.14632285e-01
-3.39116305e-01 8.13154057e-02 6.37989104e-01 3.59867305e-01
-8.33189964e-01 -1.70836359e-01 -9.67227399e-01 4.45783377e-01
1.13042414e+00 -8.86770844e-01 -2.21486136e-01 4.29005772e-01
5.04742146e-01 1.29955560e-01 -8.58826339e-01 4.93076921e-01
5.48430502e-01 -8.95145476e-01 1.16984439e+00 -2.84938872e-01
7.35216439e-02 -7.65083432e-01 -1.48190901e-01 -1.01175356e+00
-1.06477451e+00 -6.19410932e-01 -3.50985676e-01 1.28719378e+00
7.66818691e-03 -8.11712563e-01 9.62544024e-01 -8.39847550e-02
3.30267102e-01 -6.04467332e-01 -9.37507570e-01 -7.89170742e-01
-4.27430421e-01 5.59156314e-02 2.66091764e-01 1.45895922e+00
4.90274243e-02 6.70000553e-01 -4.63413537e-01 3.69123787e-01
8.60062420e-01 1.23410620e-01 6.82474673e-01 -1.74527478e+00
-2.21061036e-01 -3.66392046e-01 -8.95198584e-01 -7.97724068e-01
-2.93209893e-03 -9.39534843e-01 -4.83658254e-01 -1.28215456e+00
5.69398582e-01 -5.15607238e-01 -3.92837048e-01 1.14826657e-01
1.30538940e-01 3.16407651e-01 2.82178342e-01 4.75530684e-01
-6.10464454e-01 3.67078245e-01 1.34670866e+00 -3.92940283e-01
-3.18811946e-02 3.49109381e-01 -7.93559432e-01 7.24479675e-01
3.45085621e-01 -3.63076925e-01 -6.47408962e-01 8.23368728e-02
4.52694744e-01 1.17424987e-01 1.78319603e-01 -1.07703292e+00
3.63970459e-01 2.59001963e-02 2.30498165e-01 -5.03293395e-01
2.33066887e-01 -1.01814890e+00 2.94892877e-01 7.26625323e-01
1.86324731e-01 1.45416930e-01 -1.67036012e-01 1.25367045e+00
-3.78710687e-01 2.28637867e-02 6.63363397e-01 1.11036062e-01
-6.20651901e-01 4.25136566e-01 -2.45558769e-01 -1.18460380e-01
9.50966597e-01 -5.79618812e-01 -1.34732619e-01 -3.77528429e-01
-6.50376022e-01 2.87669778e-01 5.96187115e-01 7.66275704e-01
5.32817066e-01 -1.51414931e+00 -6.95551395e-01 4.48942363e-01
3.67129713e-01 -2.62425333e-01 3.72884482e-01 8.14401388e-01
-3.39748770e-01 1.22410998e-01 -1.13669671e-01 -8.29946160e-01
-1.43280864e+00 7.44361520e-01 1.56448334e-02 -6.50502384e-01
-6.42801166e-01 5.41156352e-01 3.44789892e-01 -5.16361594e-01
1.36775136e-01 7.93951824e-02 -3.75253588e-01 2.56321251e-01
4.59320962e-01 5.12458146e-01 -1.74449593e-01 -1.13320160e+00
-4.72484082e-01 7.02619672e-01 -1.53231788e-02 6.60562754e-01
1.06301057e+00 -1.96197093e-01 -4.16509300e-01 2.06759349e-01
1.17363524e+00 1.23153917e-01 -8.82191837e-01 -4.07693595e-01
-7.49636590e-02 -5.49651980e-01 3.77152227e-02 -1.02080330e-01
-1.18936276e+00 6.53076947e-01 4.03111517e-01 2.84592956e-01
9.06782031e-01 -5.75287461e-01 3.64014536e-01 5.09639800e-01
7.38545120e-01 -7.61974275e-01 4.88021106e-01 4.94915158e-01
8.08292150e-01 -1.02141750e+00 4.25563574e-01 -1.07073104e+00
-4.77370232e-01 1.12155831e+00 3.97540838e-01 -3.84573340e-01
9.16157007e-01 -2.90944159e-01 -3.94898742e-01 -3.89172673e-01
-2.52382219e-01 1.16418125e-02 4.85191464e-01 7.35424757e-01
2.21935004e-01 2.77719170e-01 1.28569737e-01 4.38409954e-01
1.63186908e-01 -4.57396299e-01 5.13516188e-01 6.70281112e-01
-3.48648250e-01 -1.04870117e+00 -3.43031853e-01 5.91533661e-01
-1.30662881e-02 1.43061534e-01 -4.29923356e-01 6.23967707e-01
-7.90245757e-02 1.10635686e+00 -2.76558518e-01 -6.87155008e-01
2.94626504e-01 -4.20827389e-01 1.71117395e-01 -9.81499135e-01
-3.04888129e-01 1.57408535e-01 -2.14236528e-01 -5.50568342e-01
-3.48546892e-01 -6.72016084e-01 -1.13648927e+00 -5.16040266e-01
-7.88320422e-01 3.98362905e-01 2.18826741e-01 7.31973410e-01
3.55439872e-01 5.59512317e-01 9.45542991e-01 -7.71912992e-01
-6.65215313e-01 -7.93092787e-01 -1.19000411e+00 5.35797477e-01
2.83304065e-01 -1.05542266e+00 -7.14977801e-01 -3.46775025e-01] | [7.302908420562744, 5.684795379638672] |
edc09155-b3a5-4f3d-bf36-6c7a20e55e63 | adversarial-attacks-on-binary-image-1 | 2010.11782 | null | https://arxiv.org/abs/2010.11782v1 | https://arxiv.org/pdf/2010.11782v1.pdf | Adversarial Attacks on Binary Image Recognition Systems | We initiate the study of adversarial attacks on models for binary (i.e. black and white) image classification. Although there has been a great deal of work on attacking models for colored and grayscale images, little is known about attacks on models for binary images. Models trained to classify binary images are used in text recognition applications such as check processing, license plate recognition, invoice processing, and many others. In contrast to colored and grayscale images, the search space of attacks on binary images is extremely restricted and noise cannot be hidden with minor perturbations in each pixel. Thus, the optimization landscape of attacks on binary images introduces new fundamental challenges. In this paper we introduce a new attack algorithm called SCAR, designed to fool classifiers of binary images. We show that SCAR significantly outperforms existing $L_0$ attacks applied to the binary setting and use it to demonstrate the vulnerability of real-world text recognition systems. SCAR's strong performance in practice contrasts with the existence of classifiers that are provably robust to large perturbations. In many cases, altering a single pixel is sufficient to trick Tesseract, a popular open-source text recognition system, to misclassify a word as a different word in the English dictionary. We also license software from providers of check processing systems to most of the major US banks and demonstrate the vulnerability of check recognitions for mobile deposits. These systems are substantially harder to fool since they classify both the handwritten amounts in digits and letters, independently. Nevertheless, we generalize SCAR to design attacks that fool state-of-the-art check processing systems using unnoticeable perturbations that lead to misclassification of deposit amounts. Consequently, this is a powerful method to perform financial fraud. | ['Richard Wang', 'Yaron Singer', 'Alexander Rilee', 'Kojin Oshiba', 'Harrison Chase', 'Eric Balkanski'] | 2020-10-22 | adversarial-attacks-on-binary-image | https://openreview.net/forum?id=xCm8kiWRiBT | https://openreview.net/pdf?id=xCm8kiWRiBT | null | ['license-plate-recognition'] | ['computer-vision'] | [ 4.14524257e-01 -2.56934106e-01 -8.96152705e-02 -3.19472492e-01
-7.32602954e-01 -1.28096402e+00 5.66608071e-01 -1.75991744e-01
-3.64721835e-01 4.86676574e-01 -7.40404546e-01 -1.14212179e+00
3.14238161e-01 -1.10392141e+00 -7.11560190e-01 -6.31408691e-01
1.69238418e-01 1.73377231e-01 1.82343632e-01 -3.55027884e-01
5.04846573e-01 7.36128747e-01 -1.08251691e+00 5.65648258e-01
5.60752392e-01 8.68893087e-01 -6.22398078e-01 1.08671737e+00
7.89581537e-02 8.83469701e-01 -8.62942338e-01 -1.11404550e+00
8.11279297e-01 -4.31673557e-01 -5.16965151e-01 4.32570837e-02
7.46873021e-01 -3.04079682e-01 -5.10205567e-01 1.73988461e+00
2.46638209e-01 -4.89300042e-01 6.29911959e-01 -1.27223384e+00
-1.04377270e+00 7.46873200e-01 -4.40589845e-01 2.85655469e-01
2.17994630e-01 4.56172913e-01 7.71509409e-01 -5.63995540e-01
3.05189908e-01 1.18976140e+00 6.85898244e-01 6.49860442e-01
-1.19796586e+00 -8.10155034e-01 -1.66320235e-01 4.44301851e-02
-1.33235407e+00 -4.08649296e-01 3.61411273e-01 -3.16839844e-01
6.79523051e-01 7.98442245e-01 2.01969832e-01 1.12443352e+00
4.10664678e-01 6.55054390e-01 1.61676896e+00 -7.71476686e-01
1.86583549e-01 4.89284128e-01 2.49890581e-01 8.94168735e-01
5.53783059e-01 2.57535994e-01 -1.45500660e-01 -5.00894785e-01
5.70566952e-01 -1.82210475e-01 -2.93346971e-01 1.86500940e-02
-8.46340597e-01 1.01154685e+00 -1.39903212e-02 2.13769302e-01
3.28192145e-01 3.32749069e-01 2.32148305e-01 7.64279664e-01
1.64033112e-03 5.61769545e-01 -3.25346708e-01 -4.24293801e-02
-7.67268419e-01 8.16109255e-02 1.24794865e+00 5.49349785e-01
4.48560864e-01 4.02410120e-01 3.17602962e-01 6.00197792e-01
2.98195809e-01 9.75938678e-01 5.19754052e-01 -3.56102943e-01
6.61388993e-01 2.30782017e-01 -2.97617055e-02 -1.17830968e+00
3.83451991e-02 -4.16769609e-02 -6.64122283e-01 9.83794749e-01
6.76963925e-01 -1.49788991e-01 -1.11835027e+00 1.14928269e+00
-2.15697199e-01 -1.85251012e-01 9.62814093e-02 5.08264840e-01
3.70895155e-02 4.50857431e-01 -1.89031065e-01 3.01293194e-01
1.34881556e+00 -5.34472823e-01 -5.24839401e-01 -3.41700822e-01
4.43911880e-01 -8.86431932e-01 9.87101376e-01 6.31620646e-01
-1.06370378e+00 -2.37852141e-01 -1.55270290e+00 2.63175249e-01
-9.88080680e-01 -2.09466442e-01 6.07107937e-01 1.94813466e+00
-6.77148342e-01 4.88441795e-01 -7.41914868e-01 1.06435813e-01
5.97480893e-01 5.32699764e-01 -3.48290056e-01 -1.63839623e-01
-1.23393643e+00 1.09016657e+00 -1.07547671e-01 -1.55213224e-02
-6.41647458e-01 -3.55345100e-01 -8.46660972e-01 1.15681544e-03
-4.12054546e-03 5.82007095e-02 9.82471168e-01 -1.31331527e+00
-1.29437375e+00 1.31355596e+00 1.51595265e-01 -8.16156626e-01
9.49670672e-01 3.93306345e-01 -9.63124812e-01 3.17685217e-01
-2.17180446e-01 9.96272489e-02 1.41764224e+00 -1.07070792e+00
-5.69356561e-01 -1.82599649e-01 1.17664739e-01 -5.54120183e-01
-4.33302492e-01 5.94604254e-01 -1.92893390e-02 -1.06788504e+00
-1.31827995e-01 -8.59062910e-01 -5.81097268e-02 3.03205084e-02
-5.35177469e-01 5.66219270e-01 9.39670265e-01 -5.95383942e-01
1.16355419e+00 -2.34706450e+00 -6.66327953e-01 8.19653988e-01
-6.80876970e-02 5.76700926e-01 -3.66347618e-02 7.38276690e-02
-3.46019238e-01 7.83210754e-01 -5.41418850e-01 4.99528181e-03
4.10424173e-01 4.02411073e-02 -9.50622797e-01 9.75502670e-01
1.70861900e-01 1.02156651e+00 -3.91453266e-01 -1.30901739e-01
2.98088063e-02 5.60575239e-02 -3.47951323e-01 -4.71200287e-01
-4.82258834e-02 -3.90599549e-01 -3.12572390e-01 1.00819230e+00
9.77694333e-01 -2.09236607e-01 1.40499949e-01 3.73211443e-01
2.79131860e-01 1.36223901e-03 -1.28866363e+00 6.52282476e-01
-1.60204142e-01 7.06225812e-01 1.68479264e-01 -9.02495801e-01
7.93499172e-01 4.04110514e-02 -1.36110514e-01 -4.47489053e-01
2.10982099e-01 3.28213155e-01 2.54809588e-01 -1.87353536e-01
4.67230082e-01 -1.83650017e-01 -2.97352821e-01 9.22354162e-01
-4.32865143e-01 -4.89243388e-01 2.75459941e-02 1.43480748e-01
1.31965411e+00 -4.85555559e-01 -1.69445917e-01 -7.03911558e-02
5.23653746e-01 -4.91140038e-02 7.87522420e-02 1.29369295e+00
-3.39525372e-01 5.74082434e-01 5.71752608e-01 -3.76601696e-01
-1.05737388e+00 -9.88568842e-01 -3.19725811e-01 5.57984114e-01
5.23009077e-02 -1.01088032e-01 -7.87714839e-01 -8.92230690e-01
5.00481844e-01 4.12096947e-01 -5.64056218e-01 -2.97673047e-01
-5.00040114e-01 -9.37385261e-01 1.50153100e+00 4.80958045e-01
7.25612104e-01 -7.54927754e-01 -2.55672067e-01 -7.83956051e-03
5.27844548e-01 -1.10378718e+00 -4.89563257e-01 3.19516540e-01
-4.99822408e-01 -1.39892769e+00 -3.56659383e-01 -7.68154204e-01
8.91434550e-01 -2.50263978e-02 7.82907486e-01 5.50943315e-01
-5.37057459e-01 4.01849091e-01 -2.38852754e-01 -4.69348192e-01
-1.21307361e+00 -2.53146589e-01 7.01561151e-03 2.90867895e-01
6.51028275e-01 5.76924309e-02 -4.99799624e-02 5.42187870e-01
-1.32788217e+00 -6.46827519e-01 3.30385715e-01 8.79296422e-01
1.22478325e-02 5.33972085e-01 2.39745960e-01 -1.16241920e+00
7.01401830e-01 -5.07194698e-02 -9.83739436e-01 5.36142766e-01
-6.18875861e-01 1.65788084e-03 7.21310794e-01 -6.66662574e-01
-6.98965669e-01 -4.06468324e-02 -1.33521393e-01 -2.25613505e-01
-9.04342607e-02 3.23144406e-01 -1.38560385e-01 -8.44462395e-01
8.28524947e-01 2.10135251e-01 5.98462708e-02 -2.16567427e-01
9.84156951e-02 9.02552485e-01 7.11781263e-01 -5.31827807e-01
1.42337918e+00 5.92259347e-01 8.18821881e-03 -5.39723694e-01
-6.23520017e-02 4.84335832e-02 -1.94726676e-01 1.77309573e-01
5.60403466e-01 -5.15399814e-01 -8.80160809e-01 1.08233941e+00
-8.36924374e-01 -3.60881925e-01 7.11820126e-02 -1.40119800e-02
-3.12489718e-01 8.18928361e-01 -9.52384949e-01 -8.42573702e-01
2.98106242e-02 -1.17612362e+00 4.74547505e-01 -8.46845731e-02
9.05919522e-02 -1.00920665e+00 -3.59844595e-01 3.15673351e-01
4.30510432e-01 5.82420826e-02 1.02563727e+00 -9.78778303e-01
-6.19520545e-01 -1.02252209e+00 8.11534468e-03 8.30054104e-01
4.43086810e-02 3.43802392e-01 -1.01267803e+00 -2.77557135e-01
2.72348583e-01 -3.89472514e-01 8.52521300e-01 -1.89273909e-01
1.14925408e+00 -4.66360271e-01 1.05953356e-02 6.01437449e-01
1.28579700e+00 4.41924602e-01 1.03775454e+00 6.76285684e-01
2.96341687e-01 8.98256823e-02 2.48570502e-01 3.12510669e-01
-4.07640338e-01 4.75187123e-01 4.04307485e-01 -1.75239220e-01
3.58035862e-01 -5.15485890e-02 5.94519675e-01 1.31522804e-01
2.69514531e-01 -1.81270286e-01 -8.82894754e-01 6.48069754e-02
-1.21252108e+00 -1.31156790e+00 -1.75007224e-01 2.18527293e+00
1.10471940e+00 5.64685047e-01 -2.85858274e-01 5.94340682e-01
9.81651127e-01 1.33858789e-02 -2.14233086e-01 -6.90077960e-01
-5.88807046e-01 7.87509143e-01 1.20315301e+00 6.43678665e-01
-1.40385580e+00 9.91695344e-01 7.05327845e+00 9.48309720e-01
-1.13876712e+00 -1.02452941e-01 9.91601527e-01 1.79400519e-01
-1.74596116e-01 4.94842306e-02 -8.82000744e-01 6.88760757e-01
8.56444776e-01 1.98890688e-03 6.06920600e-01 8.95492852e-01
-2.60302752e-01 5.26537746e-02 -1.05090189e+00 8.52744043e-01
3.30197453e-01 -1.38596749e+00 1.56593323e-02 7.20810145e-02
7.08372355e-01 -3.72471154e-01 6.43126309e-01 2.16849387e-01
6.83886826e-01 -1.50186956e+00 8.78695011e-01 1.67120174e-01
8.88693213e-01 -7.70641625e-01 5.80965757e-01 1.41256735e-01
-4.89795536e-01 -1.82903886e-01 -5.56239665e-01 3.63132246e-02
-3.44316602e-01 2.67288446e-01 -3.72437745e-01 1.31346080e-02
4.25782651e-01 1.80934504e-01 -8.70717406e-01 7.93132961e-01
-4.75764483e-01 7.25517750e-01 -3.94973278e-01 4.60044146e-02
2.96867013e-01 -1.70747921e-01 1.12147860e-01 1.32559669e+00
5.52576780e-02 9.75824296e-02 -1.32844269e-01 8.04468989e-01
-3.66671771e-01 -2.54934549e-01 -6.97771072e-01 -2.04598561e-01
3.69548053e-01 7.79839873e-01 -7.37833738e-01 -3.81476343e-01
-4.92977291e-01 1.07390392e+00 -3.19600910e-01 2.79219776e-01
-1.03069913e+00 -7.56100655e-01 5.72099805e-01 -2.03675389e-01
5.76126337e-01 -2.30766758e-01 -6.32349253e-01 -1.31191349e+00
1.25866547e-01 -1.81148410e+00 4.19636726e-01 -4.76003736e-01
-1.29754019e+00 4.02635753e-01 -4.90274042e-01 -1.15523851e+00
-1.07321061e-01 -1.16567409e+00 -5.33192515e-01 1.07025993e+00
-1.51160789e+00 -8.04835260e-01 2.49695197e-01 8.76148045e-01
2.59662032e-01 -6.41419232e-01 8.99762630e-01 1.38824880e-01
-4.97575462e-01 1.23460782e+00 3.85999143e-01 8.73359203e-01
5.86738586e-01 -1.12214887e+00 8.70600581e-01 1.41329312e+00
5.48849523e-01 7.28545725e-01 3.81620049e-01 -7.01175690e-01
-1.53906882e+00 -7.09771514e-01 6.30976021e-01 -6.30684555e-01
9.91583765e-01 -4.88257378e-01 -7.98359334e-01 7.64499962e-01
-6.82707056e-02 2.82529801e-01 5.57327509e-01 -5.03572583e-01
-9.02455866e-01 1.01348393e-01 -1.45236349e+00 7.36151159e-01
2.71640152e-01 -1.02979445e+00 -3.60311508e-01 4.38542485e-01
8.25450197e-02 -3.20075691e-01 -5.83454728e-01 -8.99598300e-02
6.22416615e-01 -8.49883080e-01 1.03217041e+00 -8.51994216e-01
3.78777564e-01 -2.20731571e-01 -2.55717993e-01 -7.69738138e-01
4.56073284e-02 -9.05974448e-01 1.37376055e-01 9.65809107e-01
6.91910803e-01 -1.05517697e+00 8.22537243e-01 9.80033457e-01
2.95016170e-01 -1.33473411e-01 -9.59125876e-01 -9.83636796e-01
5.30433834e-01 -6.40897453e-01 5.26599646e-01 1.32866645e+00
1.33717611e-01 -6.55301571e-01 -3.68414283e-01 4.96870726e-01
5.29155195e-01 -1.49011731e-01 7.95123160e-01 -6.24388874e-01
-7.25185752e-01 -7.56741345e-01 -8.02229285e-01 -7.66661048e-01
1.80345207e-01 -8.02886546e-01 -1.20957151e-01 -3.40670764e-01
-2.50439197e-01 -6.73507154e-01 -1.52603254e-01 6.35601640e-01
-1.86470792e-01 6.67903543e-01 4.20819342e-01 2.31975719e-01
2.11964950e-01 -3.84819329e-01 6.23926163e-01 -9.02695596e-01
3.97733927e-01 1.74996451e-01 -5.45134962e-01 7.67004490e-01
8.67504299e-01 -7.25967646e-01 1.60314262e-01 -2.73251534e-01
4.13175255e-01 -3.95700157e-01 6.15141451e-01 -8.92412007e-01
2.03971535e-01 -1.32746607e-01 3.95146579e-01 -5.27187400e-02
3.32163721e-02 -1.01786828e+00 -8.76834542e-02 7.76781142e-01
-2.60205984e-01 1.88301702e-03 1.56712398e-01 3.74282956e-01
-1.28911600e-01 -8.66273344e-01 1.06296051e+00 -3.23246300e-01
-2.71817565e-01 2.35934630e-02 -6.79048002e-01 -1.79075301e-01
1.04808855e+00 -4.65617090e-01 -9.46564496e-01 -3.07632804e-01
-5.33850908e-01 -2.55524307e-01 8.75599861e-01 1.60032481e-01
5.44366360e-01 -9.35558915e-01 -6.74077570e-01 7.07127571e-01
-1.94284678e-01 -7.58759081e-01 -2.25817829e-01 4.32896405e-01
-1.08763778e+00 2.58556217e-01 -1.01803012e-01 -1.04971334e-01
-1.51307452e+00 7.52062142e-01 7.10794687e-01 -3.08195442e-01
-3.30669582e-01 7.72806346e-01 -3.00008506e-01 -9.60927457e-03
2.70071924e-01 -3.90945464e-01 3.82818401e-01 -3.08693349e-01
7.12189496e-01 1.39218792e-01 3.47836256e-01 -2.39154965e-01
-2.20531955e-01 4.50122327e-01 -2.78565496e-01 -2.71213412e-01
9.01517570e-01 3.78642738e-01 -1.12812623e-01 -1.00132085e-01
1.03291702e+00 5.33854008e-01 -8.79749179e-01 8.83598477e-02
-7.30327680e-04 -9.07663405e-01 -2.39014104e-01 -9.05678749e-01
-1.16537070e+00 8.59179854e-01 5.41346371e-01 7.07482517e-01
7.93263197e-01 -6.23908639e-01 5.47130167e-01 7.74325192e-01
4.18579668e-01 -1.11602414e+00 -9.24278870e-02 5.97108960e-01
4.55338657e-01 -1.10887849e+00 3.36630456e-02 -3.38521004e-01
-5.20946443e-01 1.26587367e+00 1.71636119e-01 -1.44471854e-01
5.75900912e-01 8.04131091e-01 4.97891575e-01 1.21899188e-01
-3.57065409e-01 3.75268757e-01 -2.14778692e-01 7.39655852e-01
-2.04588220e-01 -1.14732258e-01 -4.90512280e-03 4.01307851e-01
-1.12033419e-01 -3.48306000e-01 1.15285349e+00 1.42379439e+00
-2.75458187e-01 -1.43975365e+00 -1.18744111e+00 4.04917151e-01
-1.01591933e+00 -4.95593220e-01 -6.04441643e-01 8.17445755e-01
-1.47725075e-01 1.03637350e+00 -8.91391635e-02 -3.47911209e-01
2.27400847e-02 3.49012703e-01 3.78512502e-01 -4.09895509e-01
-1.01516140e+00 -4.47653443e-01 3.03482125e-03 -1.39756784e-01
-1.01128802e-01 -6.66281819e-01 -8.64816427e-01 -8.23616624e-01
-4.55999523e-01 -1.19375847e-01 7.81276047e-01 6.64266825e-01
-1.04357153e-01 -5.29389605e-02 9.70897973e-01 -3.85327578e-01
-1.30873716e+00 -5.40244758e-01 -1.06464887e+00 4.94374812e-01
1.99751452e-01 -1.74224168e-01 -7.64596820e-01 4.34772313e-01] | [5.689977645874023, 7.893373012542725] |
06bc89e9-4938-4b33-8a2b-f11155f778cf | multimodal-emotion-recognition-for-one-minute | 1805.01060 | null | http://arxiv.org/abs/1805.01060v1 | http://arxiv.org/pdf/1805.01060v1.pdf | Multimodal Emotion Recognition for One-Minute-Gradual Emotion Challenge | The continuous dimensional emotion modelled by arousal and valence can depict
complex changes of emotions. In this paper, we present our works on arousal and
valence predictions for One-Minute-Gradual (OMG) Emotion Challenge. Multimodal
representations are first extracted from videos using a variety of acoustic,
video and textual models and support vector machine (SVM) is then used for
fusion of multimodal signals to make final predictions. Our solution achieves
Concordant Correlation Coefficient (CCC) scores of 0.397 and 0.520 on arousal
and valence respectively for the validation dataset, which outperforms the
baseline systems with the best CCC scores of 0.15 and 0.23 on arousal and
valence by a large margin. | ['Chenjie Cao', 'Ziqi Zheng', 'Xingwei Chen', 'Guoqiang Xu'] | 2018-05-03 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [-0.08267318 -0.06546824 0.03238434 -0.8378961 -0.7631837 -0.6141438
0.63907504 0.17994802 -0.22246373 0.6405761 0.42627928 0.5692184
0.17172682 -0.08262662 -0.25538564 -0.5394707 -0.4206246 -0.2639759
-0.4146924 -0.47473273 -0.05885255 0.18294613 -1.9048896 0.58836734
0.363532 1.8060606 -0.43669298 1.2355714 -0.07159887 0.5970178
-0.62612474 -0.551731 -0.09231493 -0.2812099 -0.4668938 -0.1764705
0.15422331 0.17308559 -0.08072887 0.7851242 0.5846015 0.32761633
0.85712403 -1.5414685 -0.291544 0.18921997 -0.6754673 0.09396619
0.7466767 -0.20734592 0.98596114 -1.1084392 0.23541379 1.1453624
0.5582351 0.7383677 -0.9762541 -0.64276385 0.06456371 0.27389833
-1.1196506 -0.4239164 1.0095949 -0.20744984 1.1794251 0.76154447
0.8611594 1.4155728 0.23910151 0.61453915 1.1299127 -0.20711587
0.1538195 0.6551024 0.1038592 0.23352006 -0.58185285 -0.29751953
-0.8597635 -0.21574597 0.02526326 -0.43746242 0.02642244 0.2713785
-1.0393598 0.72562283 0.14569499 -0.03616133 -0.7156118 0.09581516
0.8448562 0.5856813 0.37480882 0.38874325 -0.50523114 -0.70636296
-0.69635564 -0.11818781 0.66821927 0.691327 0.49645507 0.3047012
0.06850769 1.3075062 0.465087 0.49558118 0.5788875 -1.0777556
-0.01237446 0.24529703 0.15910988 -1.1702178 -0.6799516 0.06171652
-0.75145197 -0.2462717 -0.3111903 -0.49832255 -0.69113874 1.7629485
0.34854826 0.31749853 0.38260576 1.0450203 1.4471188 1.2337009
0.463117 -0.59859514 1.3394243 -0.65573555 -1.1827594 -0.24314733
0.16924956 -0.7701506 1.1051695 0.89921683 -1.222751 -0.60390466
-1.1760802 0.27722585 -0.40284452 -0.02884884 0.78814286 0.8556885
-0.9600487 0.34865782 -0.39831397 -0.25524545 0.04711398 0.5329327
-0.5084264 0.56558293 -1.4618726 0.8238229 0.0683521 0.22386432
-0.69135827 -0.397165 -0.90295756 -0.1890161 -0.4378204 -0.05998929
0.95658773 -1.4052596 -1.8251668 0.8247651 -0.03059907 -0.13561098
-0.05404817 -0.42788896 -1.1929263 0.3528829 -0.68082684 1.0235465
0.84853125 -1.1953642 -0.2556383 -0.34371346 -0.38603148 0.47197363
-0.4571842 0.37534794 -0.28797412 -0.1438537 0.13789487 -0.8797196
-0.11110227 -0.77925056 -0.03628811 -0.16605884 0.7138823 -0.7032076
1.3730799 -2.2458112 0.25824144 0.3503518 -0.05501946 -0.09199739
-0.29743534 0.35855487 -0.23376626 -0.25621 0.3177317 -0.49117818
0.25419858 0.19018432 -0.4305795 0.30189225 0.47253764 0.730407
-0.4874633 -0.475701 0.281889 0.78136975 -0.46383205 0.68790984
0.0941736 0.27822068 0.10038835 0.8684018 0.3848889 0.2364308
0.28196433 -0.43395907 0.15881659 -0.12592646 -1.0523131 1.4337784
-0.34198922 0.84651333 0.02756379 -0.8829884 1.5139134 0.5988414
0.5020625 -0.5410639 0.5531644 -0.15063868 -0.18303049 -0.7506973
0.829247 -0.37938774 -0.85123384 0.084292 0.48975596 -0.1924632
-0.15878206 0.230214 0.4114015 -0.19157517 0.1704349 0.06246421
0.6036681 -0.46821204 0.47224915 0.12016591 -0.5889478 0.5184371
0.8313297 -0.22841339 -0.79446113 -1.020762 -0.03697284 1.4215208
0.06309824 -0.6337509 -0.6485039 -0.17475167 -0.3095784 0.5567368
-0.75986767 -0.42283642 0.04327414 -0.77227324 0.4940099 0.60038245
-0.05289637 -1.0129012 -0.42127082 -0.07104048 -0.32084814 -1.1874534
0.16392165 0.30499634 -0.6037636 -0.342108 -0.19336762 -0.34994528
0.04709993 -0.41998327 0.8834713 -0.5909221 -0.17710342 0.8177362
-0.39504215 -0.67689276 0.01203987 -0.4291894 0.51354057 0.4025934
0.54580444 -0.6151767 -0.48972994 -0.02922282 -0.6061462 -0.18104358
0.2442828 0.695975 0.53787076 -0.6624908 1.1110253 -0.26999408
0.99210954 -0.69048 -0.06911813 0.01269403 -0.5612576 -0.57290053
0.3671112 -0.8769858 -1.0116465 0.11507318 -0.16030666 -0.7352904
-0.39512867 0.39750013 -0.12221915 0.2920879 0.5205685 0.03359323
-0.25584957 -0.09382527 0.5116672 1.080512 0.8749479 -0.28740403
-0.0236242 0.01177191 -0.13928849 -0.9822459 -0.6156137 -0.38264793
-0.4434645 -0.89042294 0.95492107 -1.2191594 -0.91339034 0.24368888
-0.82651657 0.12072528 0.21623994 0.8655157 -0.71587 0.21153615
-0.8763933 -1.2185818 -0.6046908 -0.68928695 0.9735464 0.4862304
-0.80279624 -0.8345672 0.24512921 0.40821552 0.47013745 0.5725021
0.45447752 -0.55537647 0.55875355 -0.46667406 0.06822318 0.5865019
-0.2464865 0.28369322 -1.389245 0.00971213 -0.07825113 -0.8497896
0.61238194 0.36968783 1.072345 -0.09404322 0.23953289 0.43751204
1.1424164 0.23567075 0.7564708 0.01211699 0.3359938 0.6628657
0.98963016 1.0457603 0.3698141 0.37113428 0.71437114 -0.0178893
0.6859839 0.16682936 0.6627301 1.0523287 -0.12251806 -0.00765915
-0.6760646 0.33242625 -1.6787922 -0.9906801 -0.34318623 1.7390162
0.87413925 -0.20039983 0.3301695 0.204969 0.4755078 0.31702954
-0.1931337 -1.5407876 -0.27291054 0.01475207 -0.20707595 0.36214152
-1.3069869 0.81269807 7.317264 0.24597357 -1.3434232 -0.10702863
0.74733686 -0.75169605 -0.05741481 -0.6009424 -0.39444143 0.48774263
1.6851707 -0.02995225 0.49330544 0.8795444 0.07156203 -0.04492155
-0.85535234 1.6319585 0.46888342 -0.5308583 -0.3703421 -0.5457014
0.81432676 -0.11131387 0.2687862 0.5597568 -0.32508945 -1.2058601
0.4772344 0.9580781 0.64483577 -1.1833824 0.777414 -0.13596854
-0.90603 -0.12215541 -0.3431068 -0.13487063 0.21115108 0.3934467
-0.5001294 0.13194315 0.93731123 0.57783043 -0.33967984 0.45710385
0.0850131 0.7194841 -0.1837991 -0.28707942 0.2594402 -0.10158298
0.29243025 1.8493319 0.26492128 0.5267563 -0.20725212 0.3187398
0.07980057 0.34011307 -0.31689203 -0.22109051 0.27569905 1.8898556
-0.20648238 -0.6327145 -0.33052477 0.9545182 -0.10246812 0.26928505
-1.3271979 -0.50447416 0.92380446 -0.7040794 0.04531577 0.01815763
-0.39345396 -1.0685027 -0.32305518 -0.69689184 0.45348132 -1.1746829
-1.274817 0.87739265 -0.22446147 -1.2685674 -0.4345589 -0.5914713
-0.63487834 0.7866215 -1.0503796 -0.8006469 -0.5768924 0.7692549
0.28607818 -0.18369281 1.1020328 0.3048944 -0.4380707 0.5561544
-0.305421 -0.22087401 0.9070521 -1.1387705 -0.48273975 0.20727333
-0.10055555 0.3612579 0.96741813 -0.11232419 -1.7191049 -0.78077096
0.8439326 -0.41863924 0.57781947 -0.3200047 -0.82125926 0.42256582
0.6662231 -0.11382682 1.4254481 0.31333423 -0.40830538 -0.17671251
-1.1826872 0.26650304 0.25480947 -0.6061074 -0.6183448 -0.11989549
0.46129218 -0.20082292 -1.4437712 0.61250705 1.0945624 -0.88803595
0.82126313 -0.9196835 0.5475172 0.0082901 -0.8337605 -1.3215789
-0.09482201 -0.34486914 -0.310747 1.2409763 0.4008853 -0.15264122
0.29542303 0.69961286 -0.03469242 -0.8221474 -0.7436977 -0.16232459
-0.4584701 -1.106499 0.4735241 1.1453321 0.8332366 0.72656894
-0.8852458 0.10193506 0.17932686 0.05445978 0.72721875 -1.0703758
0.13196772 -0.2777153 -0.6819593 -0.32854164 0.11094025 -0.39629585
-0.19087604 -1.1858367 0.16638303 0.39708644 -0.87292886 0.41369057
0.13869108 0.70160717 0.08644278 -0.34767398 -0.9696555 0.87593377
0.5016602 0.04448776 -0.42397627 -0.28865746 -0.6905834 0.6998558
1.0237052 0.03659362 -0.30662173 0.1930335 0.45544225 0.389553
-0.08618002 -0.7419699 -0.2809261 -0.10036276 0.82610327 -0.6845262
1.1657622 -0.5481356 0.22685613 -0.06896735 -0.5601754 0.19047037
0.48174176 0.39934844 -0.56123275 0.3085199 0.76342565 0.3332138
-0.77483344 -0.09100467 -0.5487887 -0.48862174 0.95718676 -0.24712718
-0.06531307 -0.83947605 -1.287144 0.18034755 0.0323607 0.88335514
0.9567599 -1.7675451 -0.510709 0.18538012 0.21696913 -0.9664122
0.5427615 1.2832315 0.16043247 0.19545375 -0.54632497 -0.70217425
-1.7871817 0.20939726 0.32517093 0.35219842 0.21892801 0.83292526
-0.10332052 -0.33115208 0.21666078 0.11706057 -0.71280855 0.5542369
0.7247794 0.42114422 -0.12079854 -1.1023785 -0.51367056 0.43513167
0.30228776 -0.25629315 1.4518467 -0.30859882 -0.16097519 1.0108482
1.453105 -0.25354132 -0.8106864 0.21455656 -0.15588588 -0.2632013
0.05959112 -1.0069234 -0.7300708 0.9432374 1.1123483 0.23248294
1.4551499 0.18096401 0.51239324 0.25011572 -0.15271349 -1.6364151
0.34903827 0.49034342 0.9389424 -1.3729905 -0.25042063 0.03754893
-1.5821587 1.0538768 0.6283178 -0.07663873 0.51047456 0.2535098
0.4388006 -0.24163829 -1.4948146 0.06680933 0.6663359 0.41586202
0.76550686 0.4471826 -0.42002594 1.2714952 -0.44250563 -0.41253534
0.4189072 0.40592575 -0.64580417 -0.500157 -0.40861133 0.35379654
-0.8771112 0.27387923 -0.61576986 0.25109833 -0.20475024 1.2160999
0.15446036 -1.0114038 0.25671256 0.48332158 0.3852343 -0.11157041
-0.62018967 0.56288916 0.43858075 -0.89163744 -0.6017995 -0.58164
-1.3658987 0.06704792 -0.02492366 0.24134885 1.0480901 0.43741214
0.31283656 0.2688245 1.1009755 -0.95028466 -0.08020014 -1.2173277
-0.8871261 0.59422666 0.11524758 -0.3971525 -0.5438593 0.08112962] | [13.347736358642578, 5.095846176147461] |
96473f1f-5819-4246-a348-7d55b6d0a44c | an-unsupervised-domain-adaptive-approach-for | 2203.03568 | null | https://arxiv.org/abs/2203.03568v1 | https://arxiv.org/pdf/2203.03568v1.pdf | An Unsupervised Domain Adaptive Approach for Multimodal 2D Object Detection in Adverse Weather Conditions | Integrating different representations from complementary sensing modalities is crucial for robust scene interpretation in autonomous driving. While deep learning architectures that fuse vision and range data for 2D object detection have thrived in recent years, the corresponding modalities can degrade in adverse weather or lighting conditions, ultimately leading to a drop in performance. Although domain adaptation methods attempt to bridge the domain gap between source and target domains, they do not readily extend to heterogeneous data distributions. In this work, we propose an unsupervised domain adaptation framework, which adapts a 2D object detector for RGB and lidar sensors to one or more target domains featuring adverse weather conditions. Our proposed approach consists of three components. First, a data augmentation scheme that simulates weather distortions is devised to add domain confusion and prevent overfitting on the source data. Second, to promote cross-domain foreground object alignment, we leverage the complementary features of multiple modalities through a multi-scale entropy-weighted domain discriminator. Finally, we use carefully designed pretext tasks to learn a more robust representation of the target domain data. Experiments performed on the DENSE dataset show that our method can substantially alleviate the domain gap under the single-target domain adaptation (STDA) setting and the less explored yet more general multi-target domain adaptation (MTDA) setting. | ['Bin Yang', 'Karim Guirguis', 'Mario Döbler', 'Pavithran Pandiyan', 'Robert A. Marsden', 'George Eskandar'] | 2022-03-07 | null | null | null | null | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 7.45973229e-01 -3.05911124e-01 -2.15332896e-01 -6.38165057e-01
-8.28148723e-01 -7.85681307e-01 7.32902348e-01 -1.15493998e-01
-4.70573723e-01 6.65336728e-01 1.19660255e-02 -1.90165550e-01
-6.24817498e-02 -5.89989960e-01 -6.27050698e-01 -8.42371941e-01
4.65060860e-01 2.39173889e-01 5.49051225e-01 -2.11009473e-01
4.62737605e-02 7.02753007e-01 -1.66398966e+00 -7.39617571e-02
1.04211974e+00 1.32630539e+00 3.52876157e-01 5.16195238e-01
-9.91215780e-02 5.36562145e-01 -5.16116381e-01 -5.00604697e-02
7.16960192e-01 -2.68189847e-01 -3.58925015e-01 1.47531241e-01
6.43922031e-01 -3.22254598e-01 -3.96206915e-01 9.78489161e-01
6.64811432e-01 3.34504604e-01 8.66637945e-01 -1.48093343e+00
-4.46888655e-01 -2.79703945e-01 -6.66783750e-01 2.51583695e-01
-9.14828759e-03 3.76698285e-01 5.46038032e-01 -5.64029217e-01
4.34850365e-01 1.18368196e+00 7.17398643e-01 5.94600081e-01
-1.26120281e+00 -8.30583811e-01 3.37511837e-01 2.03752905e-01
-1.07687938e+00 -4.31303293e-01 1.05782449e+00 -5.52792728e-01
7.95664132e-01 -7.17152730e-02 3.84114176e-01 1.33055151e+00
-2.03428436e-02 4.83195871e-01 1.46890163e+00 -4.90719788e-02
3.57229799e-01 1.76849112e-01 -1.70560524e-01 1.64188027e-01
3.67731720e-01 3.32037956e-01 -5.56465626e-01 -5.85421175e-02
4.41591829e-01 -7.42881447e-02 -2.76280548e-02 -8.47118258e-01
-9.93423641e-01 7.02766299e-01 4.28953707e-01 -8.33454803e-02
-2.49506325e-01 -2.62744248e-01 3.98064882e-01 3.39787036e-01
4.80439544e-01 2.89307714e-01 -4.84746903e-01 4.50868495e-02
-6.48914158e-01 2.55572081e-01 4.05345708e-01 8.25447559e-01
9.01446760e-01 1.17635660e-01 -7.73399994e-02 7.73978293e-01
2.95542687e-01 1.07190979e+00 3.13956052e-01 -8.97006392e-01
5.90521336e-01 5.83775401e-01 1.44576564e-01 -7.47604549e-01
-4.48995918e-01 -3.53680164e-01 -7.02305794e-01 5.65532863e-01
5.67465901e-01 -6.12596311e-02 -1.34786570e+00 1.91159523e+00
6.32540405e-01 1.40606329e-01 3.78155887e-01 1.00545418e+00
7.39972711e-01 3.04990351e-01 9.83475447e-02 1.82547569e-01
1.07942665e+00 -4.55271095e-01 -3.84279191e-01 -8.58419776e-01
3.06939512e-01 -5.97663581e-01 9.38361824e-01 3.93193848e-02
-5.44328630e-01 -7.03294456e-01 -1.23743606e+00 -9.74950939e-02
-6.33008182e-01 -1.90741628e-01 1.99456751e-01 6.88496649e-01
-5.42990029e-01 -1.24908015e-02 -5.78370750e-01 -5.62428713e-01
5.86027443e-01 2.08478615e-01 -4.28567857e-01 -2.93514729e-01
-1.14337254e+00 1.16064346e+00 5.68012655e-01 -2.55349874e-01
-9.26241636e-01 -8.40383232e-01 -1.07987630e+00 -4.67205882e-01
4.16241109e-01 -6.12033248e-01 9.95369375e-01 -9.57404137e-01
-1.46526539e+00 1.02116120e+00 -1.20581292e-01 -5.67403376e-01
4.98556256e-01 -1.98020369e-01 -5.47071040e-01 9.88190807e-03
7.79636502e-02 7.38386691e-01 1.25721407e+00 -1.45763564e+00
-7.73759365e-01 -4.82463986e-01 -3.81638631e-02 5.55352807e-01
-1.25229657e-01 -4.18166190e-01 -2.27045268e-01 -5.46554863e-01
4.03197482e-02 -9.95319724e-01 -1.36939436e-01 2.20825195e-01
-1.02721594e-01 9.98365059e-02 1.20420420e+00 -4.84464467e-01
4.93353337e-01 -2.39718556e+00 1.64881125e-01 1.77417621e-01
2.42776833e-02 2.39564911e-01 -2.15804756e-01 -3.07459887e-02
7.26460367e-02 -3.89166802e-01 -5.94371140e-01 -2.62597710e-01
2.07402539e-02 6.16792440e-01 -4.38222766e-01 6.65818155e-01
7.82199025e-01 5.50256193e-01 -7.51902938e-01 -3.24573040e-01
4.20137405e-01 4.37448770e-01 -2.99021661e-01 2.62959510e-01
-2.17656806e-01 7.63330221e-01 -3.63252997e-01 6.86109006e-01
1.05309725e+00 1.07362211e-01 -1.28240719e-01 -8.44535232e-02
1.08480543e-01 3.46429080e-01 -1.22834027e+00 1.71297503e+00
-3.53935719e-01 6.60405874e-01 2.67565995e-01 -1.11370206e+00
1.25807405e+00 -1.56217471e-01 5.11058629e-01 -1.04613137e+00
3.02267335e-02 2.91873485e-01 -4.97676209e-02 -4.16094899e-01
5.52180886e-01 -5.05122066e-01 -3.52492213e-01 8.95424038e-02
-7.17308298e-02 -5.31264126e-01 -3.11715990e-01 -6.99694753e-02
9.18683648e-01 3.43783438e-01 3.59001338e-01 2.76364898e-03
4.60467130e-01 2.25219011e-01 8.87548923e-01 6.81636453e-01
-6.87168181e-01 7.26739168e-01 1.49012640e-01 -1.47290409e-01
-1.07274461e+00 -1.25198781e+00 -1.64540306e-01 9.89813864e-01
6.49418950e-01 3.82140666e-01 -2.55926400e-01 -7.75379121e-01
4.22688901e-01 6.77720070e-01 -5.21819949e-01 -5.80593467e-01
-4.60750192e-01 -6.64477527e-01 6.84007704e-01 7.62656450e-01
8.77934575e-01 -5.15784144e-01 -9.94395554e-01 -1.71292976e-01
-2.27794662e-01 -1.49959755e+00 -1.75073743e-01 5.20658493e-01
-6.85629427e-01 -9.50548530e-01 -7.24669814e-01 -4.96375650e-01
3.02748233e-01 4.74248409e-01 9.59841847e-01 -7.17887998e-01
-1.07087821e-01 5.79192102e-01 -2.99835622e-01 -6.70068979e-01
-4.24557328e-01 -2.43325122e-02 2.43313625e-01 1.92149505e-01
6.37865126e-01 -5.02050221e-01 -3.93198013e-01 3.75334412e-01
-9.25678015e-01 -1.98451862e-01 7.02330828e-01 7.28702068e-01
6.14110231e-01 -1.82589874e-01 4.74249214e-01 -3.08432400e-01
1.73402369e-01 -5.36223054e-01 -6.92072034e-01 -1.40000001e-01
-5.43917596e-01 2.25139588e-01 4.41206276e-01 -6.70156300e-01
-1.24594438e+00 3.93280566e-01 1.23039134e-01 -6.11908913e-01
-6.09137535e-01 1.14190601e-01 -4.87638980e-01 -1.68539196e-01
8.25147390e-01 2.05672145e-01 1.37834564e-01 -4.09595728e-01
3.31382453e-01 6.56603515e-01 7.76389182e-01 -5.34301043e-01
1.36691654e+00 7.28701830e-01 1.17706627e-01 -7.32260644e-01
-7.07267463e-01 -6.52193427e-01 -6.68621957e-01 -1.11006312e-01
9.97563243e-01 -1.21557117e+00 -3.62847932e-02 5.03835618e-01
-8.07525635e-01 -4.08877641e-01 -4.92310554e-01 5.42173147e-01
-5.39392591e-01 2.84952253e-01 3.07138473e-01 -8.29926610e-01
2.30505720e-01 -9.94294584e-01 1.19201660e+00 4.14989889e-01
1.16331585e-01 -7.31422663e-01 1.56159192e-01 3.35110366e-01
4.31937665e-01 5.44483721e-01 6.37956977e-01 -7.11469650e-01
-3.15812141e-01 -6.56266436e-02 -4.42648321e-01 4.91053849e-01
3.76901716e-01 -3.61989349e-01 -1.21791089e+00 -3.03660959e-01
-4.98408303e-02 -4.86510128e-01 1.06089222e+00 2.30102256e-01
8.57270062e-01 2.73101449e-01 -3.15189183e-01 7.25937128e-01
1.19583535e+00 2.31387198e-01 3.97421271e-01 6.12717390e-01
6.82019234e-01 6.68766439e-01 8.58753979e-01 3.15033942e-01
5.07273138e-01 6.28557384e-01 7.29591727e-01 -1.60076514e-01
-4.00563329e-01 -1.19902231e-01 5.19543469e-01 5.04662916e-02
2.61902690e-01 -7.18698651e-02 -1.06489396e+00 7.24459112e-01
-1.63710511e+00 -8.12177956e-01 2.44782731e-01 2.32087994e+00
6.34042501e-01 2.98990816e-01 3.57816219e-01 6.13787323e-02
5.95810473e-01 2.70832002e-01 -1.24940574e+00 -1.81643009e-01
-4.93179202e-01 -1.79252438e-02 9.57032800e-01 2.04356000e-01
-1.34473383e+00 6.48652434e-01 5.51421499e+00 5.18692195e-01
-1.34828544e+00 7.26063922e-02 2.32002944e-01 -1.07914589e-01
-2.49668702e-01 -1.48321107e-01 -7.46894658e-01 2.94945627e-01
8.02022040e-01 1.30129233e-03 2.21328616e-01 8.51263165e-01
9.24141146e-03 -7.21760616e-02 -8.97744596e-01 1.09917593e+00
6.07204176e-02 -7.75284231e-01 -3.52160662e-01 1.37897015e-01
6.91403627e-01 3.95281255e-01 3.13925534e-01 4.03613567e-01
3.12170148e-01 -6.67238235e-01 8.32956135e-01 2.68383980e-01
8.79291177e-01 -5.59540272e-01 4.52436507e-01 3.54065031e-01
-1.13823223e+00 -3.49530518e-01 -2.94703454e-01 -2.53989603e-02
-1.31355943e-02 4.49541688e-01 -9.19126034e-01 6.88283622e-01
9.51271236e-01 7.30664194e-01 -5.20950496e-01 9.03924823e-01
1.41225025e-01 3.16778451e-01 -5.95899880e-01 4.14539784e-01
1.30688041e-01 5.14813438e-02 8.56762052e-01 9.97295201e-01
2.94452071e-01 -2.97640599e-02 1.31840363e-01 7.46160209e-01
3.77523713e-02 -4.11256611e-01 -1.04688954e+00 1.10570319e-01
5.64293802e-01 9.93539572e-01 -3.77700210e-01 -2.53551036e-01
-4.84785438e-01 9.25722241e-01 1.57169923e-02 5.65503418e-01
-1.01959538e+00 -2.57178694e-01 1.33729374e+00 2.01235823e-02
5.61882257e-01 -3.87624055e-01 -6.38301671e-01 -1.14621079e+00
1.01617619e-01 -8.09017420e-01 3.75750542e-01 -6.50416851e-01
-1.52836096e+00 2.57214904e-01 2.59082556e-01 -1.59027100e+00
-1.82917982e-01 -7.02092469e-01 -2.49393895e-01 1.04425883e+00
-2.14815831e+00 -1.21598577e+00 -5.38294077e-01 8.75449359e-01
4.77525234e-01 -1.54018953e-01 4.91594434e-01 2.72879690e-01
-3.10611010e-01 5.88429093e-01 1.89915538e-01 -9.59142298e-02
1.15669262e+00 -1.19698882e+00 2.78829962e-01 1.04295242e+00
-2.59648651e-01 1.15837090e-01 8.01602125e-01 -5.68535984e-01
-1.24399674e+00 -1.42314410e+00 3.39266002e-01 -4.86409277e-01
4.32079583e-01 -4.05356199e-01 -1.03000617e+00 3.54366899e-01
-6.77022114e-02 3.27679694e-01 5.90164363e-01 -2.28060231e-01
-8.35090160e-01 -4.58833218e-01 -1.43830979e+00 2.68658757e-01
9.12804246e-01 -7.68686473e-01 -6.94805682e-01 -4.62132320e-02
6.47868693e-01 -6.04293644e-01 -7.72337556e-01 7.99347997e-01
4.71844137e-01 -8.76040816e-01 1.11029196e+00 -4.46628630e-01
2.03412756e-01 -6.04777217e-01 -6.58575058e-01 -1.28731644e+00
-1.12001657e-01 -2.73873210e-01 -2.12049022e-01 1.09937799e+00
1.60887703e-01 -9.05183434e-01 6.72882020e-01 5.39112151e-01
-2.40371495e-01 -1.25969946e-01 -1.25591516e+00 -1.15125942e+00
3.20556968e-01 -5.94411016e-01 7.09056079e-01 8.44543934e-01
-6.48438573e-01 2.57824719e-01 -2.14462131e-01 6.27715588e-01
6.83510005e-01 2.18200058e-01 9.68143940e-01 -1.38815272e+00
-1.60233602e-01 -3.94586086e-01 -4.77948934e-01 -1.09590769e+00
1.07152022e-01 -5.95474064e-01 4.15449440e-01 -1.26022935e+00
-8.78225863e-02 -5.27886212e-01 -5.24741888e-01 4.83194619e-01
-2.22997934e-01 3.72810781e-01 2.67338902e-01 1.55234426e-01
-3.26387078e-01 6.88003063e-01 9.73453045e-01 -3.44201714e-01
-2.28760958e-01 -9.69848633e-02 -6.96012139e-01 5.92898011e-01
8.50311518e-01 -4.16945875e-01 -5.61640441e-01 -5.60938120e-01
-8.21750909e-02 -4.46330816e-01 7.19260633e-01 -1.20982552e+00
-1.29834920e-01 -4.55539495e-01 4.85103339e-01 -5.89992046e-01
5.96679509e-01 -1.09692144e+00 -1.77970469e-01 1.93504363e-01
-1.58255607e-01 -3.65161926e-01 6.14095867e-01 8.65762472e-01
-1.22735649e-01 2.65967488e-01 1.16089022e+00 2.88571924e-01
-1.14720786e+00 1.40182421e-01 -2.02324092e-01 1.04231410e-01
1.10732520e+00 -6.04773641e-01 -3.35745066e-01 -1.64832503e-01
-3.65120977e-01 2.55197912e-01 7.78122962e-01 7.14130104e-01
5.11688471e-01 -1.34876835e+00 -5.74442267e-01 4.06068623e-01
6.14983499e-01 2.90429682e-01 1.64151996e-01 7.70729303e-01
5.21867312e-02 1.97003528e-01 -3.32573682e-01 -9.76695478e-01
-1.05991900e+00 3.29996616e-01 5.00081778e-01 -4.80630901e-03
-5.05508721e-01 8.64265740e-01 3.42996210e-01 -6.92418754e-01
7.12677538e-02 -3.92600328e-01 4.54439176e-03 7.28154033e-02
3.20094019e-01 1.26510710e-01 9.23643261e-02 -9.66623545e-01
-6.32316828e-01 5.92858613e-01 1.07205316e-01 5.82899190e-02
1.10104609e+00 -4.40702558e-01 5.28383374e-01 5.11144280e-01
1.02421284e+00 -2.68893659e-01 -1.93392503e+00 -5.16427100e-01
-2.47395560e-01 -5.03999591e-01 1.07097641e-01 -9.42960680e-01
-8.35276663e-01 1.01702833e+00 1.09841692e+00 3.25409435e-02
1.54395342e+00 8.58419761e-03 7.96087801e-01 3.03583264e-01
3.32736596e-02 -1.11927807e+00 1.00248195e-01 7.20130384e-01
5.17952144e-01 -1.73301411e+00 -1.85044020e-01 -2.22238507e-02
-9.52759922e-01 8.32096577e-01 7.19900310e-01 1.16015553e-01
3.20856899e-01 2.46266231e-01 2.11901769e-01 2.14991951e-03
-3.54570895e-01 -6.67384267e-01 3.69501412e-01 1.27063274e+00
-2.02326849e-01 -1.10902227e-01 4.09863710e-01 3.58630478e-01
1.95469692e-01 -1.14650264e-01 1.20711528e-01 1.02262402e+00
-4.99486774e-01 -8.70123029e-01 -7.21038699e-01 2.49741077e-01
4.46882844e-02 3.08494031e-01 -5.94505250e-01 8.55783403e-01
4.71006811e-01 8.77180994e-01 1.20230444e-01 -5.57240963e-01
5.21881998e-01 1.57646567e-01 4.39666510e-01 -3.92295629e-01
-3.70860770e-02 -1.09636970e-01 -1.27144232e-01 -5.25931776e-01
-6.46480143e-01 -8.69339228e-01 -1.10778618e+00 3.16469967e-02
-4.25316654e-02 -4.52235729e-01 7.13405728e-01 1.07538986e+00
4.58206266e-01 4.37125474e-01 5.47383964e-01 -1.11326814e+00
-6.26976371e-01 -7.88123369e-01 -3.92956376e-01 4.28796083e-01
8.94524276e-01 -1.17097640e+00 -3.36258471e-01 1.41582102e-01] | [8.297542572021484, -2.19661283493042] |
97a0b32b-e192-4a7c-a8f1-10e89944c6f6 | star-boosting-low-resource-event-extraction | 2305.15090 | null | https://arxiv.org/abs/2305.15090v1 | https://arxiv.org/pdf/2305.15090v1.pdf | STAR: Boosting Low-Resource Event Extraction by Structure-to-Text Data Generation with Large Language Models | Structure prediction tasks such as event extraction require an in-depth understanding of the output structure and sub-task dependencies, thus they still heavily rely on task-specific training data to obtain reasonable performance. Due to the high cost of human annotation, low-resource event extraction, which requires minimal human cost, is urgently needed in real-world information extraction applications. We propose to synthesize data instances given limited seed demonstrations to boost low-resource event extraction performance. We propose STAR, a structure-to-text data generation method that first generates complicated event structures (Y) and then generates input passages (X), all with Large Language Models. We design fine-grained step-by-step instructions and the error cases and quality issues identified through self-reflection can be self-refined. Our experiments indicate that data generated by STAR can significantly improve the low-resource event extraction performance and they are even more effective than human-curated data points in some cases. | ['Wei Wang', 'Nanyun Peng', 'P. Jeffrey Brantingham', 'Po-Nien Kung', 'Xiaoxuan Wang', 'Mingyu Derek Ma'] | 2023-05-24 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 2.38390282e-01 2.75194377e-01 -6.57064468e-02 -3.91831994e-01
-1.32004082e+00 -7.25321770e-01 5.73396981e-01 5.34545124e-01
-5.10006666e-01 8.42518270e-01 5.31942725e-01 -2.28421345e-01
1.16411313e-01 -8.39806139e-01 -7.71670520e-01 4.94113900e-02
1.39749840e-01 5.17268240e-01 2.80362546e-01 -1.46585479e-01
1.85701195e-02 8.43510032e-02 -1.70763040e+00 7.23484755e-01
1.20206654e+00 6.72692657e-01 3.43594134e-01 6.35152161e-01
-3.48346323e-01 9.48319912e-01 -8.54469776e-01 -3.47904205e-01
1.10780997e-02 -7.04986095e-01 -7.65548170e-01 -2.08546579e-01
-1.79102063e-01 -2.94334263e-01 5.21642715e-02 5.36877513e-01
5.50886095e-01 1.01165399e-01 5.40769339e-01 -1.09722233e+00
-2.43521586e-01 1.26614499e+00 -1.08124480e-01 2.89402068e-01
6.09602034e-01 2.58198410e-01 1.11491954e+00 -9.79958653e-01
9.05381083e-01 8.43543828e-01 5.23645699e-01 4.37838882e-01
-1.21570432e+00 -7.62807548e-01 1.72859967e-01 1.61857247e-01
-1.21867633e+00 -5.24008274e-01 5.82310140e-01 -2.46586040e-01
1.51572585e+00 3.34552109e-01 5.32032669e-01 1.44842243e+00
-5.35562411e-02 8.80898297e-01 7.30889559e-01 -5.85317433e-01
1.25641525e-01 4.53432053e-02 5.22360764e-02 5.97231567e-01
3.99807513e-01 7.99739361e-03 -8.77775013e-01 -4.41147247e-03
6.13724113e-01 -3.11845064e-01 -1.26621351e-01 3.53353232e-01
-1.43754423e+00 6.02014065e-01 1.96275767e-02 3.21820289e-01
-5.38267553e-01 -2.02125400e-01 5.84177732e-01 1.02655031e-01
3.44377875e-01 1.04359734e+00 -9.51007545e-01 -6.15078747e-01
-1.07196033e+00 5.22410154e-01 8.90137494e-01 1.48759210e+00
5.57943463e-01 -7.07101356e-03 -6.51496053e-01 6.49661601e-01
-1.63518459e-01 2.61644661e-01 4.65574056e-01 -6.20618641e-01
9.63066280e-01 8.25769365e-01 3.28877777e-01 -4.72850412e-01
-5.84005773e-01 -4.94993955e-01 -4.38691944e-01 -2.17161372e-01
4.90139723e-01 -4.09494430e-01 -6.68807983e-01 1.67778015e+00
2.62770325e-01 5.93063496e-02 8.01578239e-02 5.77882409e-01
9.20995295e-01 7.80140758e-01 4.43989903e-01 -3.60350817e-01
1.72533250e+00 -8.01360488e-01 -8.83283019e-01 -3.26735049e-01
8.26955199e-01 -6.92569554e-01 1.53297126e+00 2.95007944e-01
-1.18272448e+00 -4.79350626e-01 -9.75452185e-01 -1.86805651e-01
-2.29408622e-01 3.35941821e-01 6.82611525e-01 2.63889611e-01
-1.04202643e-01 6.85209334e-01 -9.85871136e-01 -1.15378574e-01
4.25862908e-01 8.01552832e-02 -2.27485135e-01 2.97941238e-01
-1.22013462e+00 8.03358257e-01 7.47909546e-01 -1.80339158e-01
-8.48920524e-01 -9.03070807e-01 -9.89726186e-01 2.38494217e-01
9.34825897e-01 -5.45251489e-01 1.60203958e+00 -4.59002733e-01
-1.29830289e+00 4.37361658e-01 -2.96965688e-01 -2.67740458e-01
3.11633587e-01 -6.16927683e-01 -5.48577011e-01 1.07404456e-01
2.75204301e-01 5.15787363e-01 4.44007337e-01 -8.43008876e-01
-8.90058219e-01 6.85610175e-02 -1.17903031e-01 1.18745349e-01
-5.02632201e-01 3.70839536e-01 -4.47509855e-01 -8.44662189e-01
-2.64279127e-01 -6.38951540e-01 -3.00097167e-01 -6.40749693e-01
-6.41151726e-01 -4.44155782e-01 4.34288025e-01 -8.00200164e-01
1.78118908e+00 -1.92016971e+00 -2.53647804e-01 -4.56813611e-02
7.96888769e-03 6.40382916e-02 -2.38231704e-01 4.79071409e-01
1.40088156e-01 2.40232706e-01 2.57298462e-02 -2.15247482e-01
-2.27045137e-02 4.08708863e-02 -3.86938810e-01 -3.62158567e-01
7.45622277e-01 1.00643921e+00 -1.19125259e+00 -9.21790302e-01
-1.75463175e-03 9.81848836e-02 -8.30700517e-01 6.64763451e-01
-6.32289290e-01 4.26757574e-01 -6.86334431e-01 5.09043634e-01
-1.11636259e-01 -5.56568742e-01 -1.04979672e-01 -4.17054236e-01
-2.00852647e-01 9.53484178e-01 -1.21359837e+00 1.75362384e+00
-7.83046126e-01 4.58953857e-01 -5.50579786e-01 -5.96685112e-01
7.61021674e-01 4.51737821e-01 2.43420273e-01 -7.11398244e-01
-3.38240750e-02 1.44020095e-02 -6.11084439e-02 -7.23447084e-01
8.32094133e-01 8.48204941e-02 -4.34486628e-01 6.13668561e-01
1.89633161e-01 -9.31826383e-02 8.12648058e-01 3.79152030e-01
1.38472104e+00 4.08555478e-01 4.94375080e-01 1.34107769e-01
-5.24445027e-02 4.18166578e-01 9.35938776e-01 7.11557686e-01
3.23953778e-01 8.08427393e-01 5.31296015e-01 -1.09677084e-01
-1.27304447e+00 -8.23109686e-01 -1.46481320e-02 1.10561562e+00
-2.14947805e-01 -1.10905206e+00 -8.87457788e-01 -9.27919507e-01
-4.19497013e-01 1.30038655e+00 -2.69117117e-01 1.40497822e-03
-8.78178060e-01 -7.60738313e-01 7.45147526e-01 8.90412688e-01
2.65776664e-01 -1.45842540e+00 -9.70175683e-01 7.88470685e-01
-7.08033502e-01 -1.32819462e+00 -5.30319571e-01 4.44243252e-01
-7.23097324e-01 -9.99103367e-01 -1.89186051e-01 -4.06011224e-01
5.91484785e-01 -2.62610883e-01 1.58634400e+00 -4.41599376e-02
-3.23871225e-01 -2.64520735e-01 -8.45501482e-01 -7.09950984e-01
-6.74289763e-01 4.16892141e-01 -1.93350598e-01 -4.97362792e-01
5.26255846e-01 -4.29953516e-01 -4.20624882e-01 3.27395797e-01
-8.92691135e-01 4.77512360e-01 6.47217453e-01 7.26651549e-01
5.95834076e-01 2.67660350e-01 8.84172559e-01 -1.16114545e+00
7.47592211e-01 -2.18787178e-01 -5.37434340e-01 2.97655642e-01
-6.26218677e-01 3.85601252e-01 9.17502284e-01 -6.78694129e-01
-1.36877286e+00 3.34745273e-02 -8.87395293e-02 4.13290672e-02
-2.51879156e-01 7.12301135e-01 -4.27252024e-01 8.69458795e-01
1.08758712e+00 -9.27835852e-02 -5.87945163e-01 -5.93012035e-01
4.78865862e-01 6.15087330e-01 5.04128993e-01 -9.28124189e-01
6.66318953e-01 -3.65680419e-02 -4.76072669e-01 -3.97797674e-01
-1.10222721e+00 -1.54056987e-02 -4.30221528e-01 3.99420150e-02
8.08586717e-01 -1.06936395e+00 -3.69188040e-01 -1.97248235e-02
-1.26200581e+00 -4.95703280e-01 -8.09809685e-01 4.78315234e-01
-3.73214453e-01 -8.99489820e-02 -6.09628201e-01 -6.03953898e-01
-5.34018159e-01 -8.35110247e-01 1.13059020e+00 2.17404887e-01
-8.94368470e-01 -4.12696540e-01 -9.29362923e-02 1.66067824e-01
1.37134954e-01 1.43811211e-01 9.96080041e-01 -1.03620589e+00
-7.49203801e-01 -1.23890616e-01 -8.09070840e-02 2.13508997e-02
2.07371250e-01 1.35881439e-01 -7.59893298e-01 2.29680493e-01
-2.85068452e-01 -3.50848436e-01 3.32470924e-01 5.44587560e-02
1.35172439e+00 -4.71926868e-01 -4.66771036e-01 2.13180289e-01
8.56264234e-01 1.39262006e-01 4.71496880e-01 2.30617717e-01
6.99854195e-01 5.95777452e-01 8.93224895e-01 7.88927972e-01
4.33557630e-01 5.79869628e-01 -1.69268712e-01 3.94449197e-02
-2.63976842e-01 -8.36042881e-01 3.71307611e-01 8.33912969e-01
2.01953435e-03 -3.74240100e-01 -7.45012164e-01 7.76452124e-01
-1.75901115e+00 -1.07990968e+00 -1.89621165e-01 1.86550534e+00
1.50234938e+00 5.09418011e-01 1.97483912e-01 3.35301042e-01
4.06524688e-01 -1.97305992e-01 -3.86328161e-01 1.84169109e-03
-2.28032377e-02 4.09410715e-01 2.04242975e-01 3.20835933e-02
-8.02927136e-01 9.96634483e-01 6.47721910e+00 1.02891731e+00
-6.56584263e-01 7.64242746e-03 4.05882299e-01 -3.46924871e-01
-6.21655822e-01 1.64757982e-01 -1.25686741e+00 5.36771476e-01
1.20756114e+00 -2.95092046e-01 1.96077809e-01 8.79218459e-01
5.27510285e-01 -6.29778951e-02 -1.58003640e+00 8.98596168e-01
-2.68021941e-01 -1.62915885e+00 1.63044840e-01 -1.87753156e-01
4.84104931e-01 -4.12391275e-01 -5.95041752e-01 6.50527954e-01
5.53015411e-01 -8.60531390e-01 9.22659993e-01 3.33125442e-01
1.00271153e+00 -6.27185643e-01 4.03852105e-01 5.78398705e-01
-1.36589515e+00 -3.64499055e-02 -1.15695052e-01 1.69446692e-02
7.19463170e-01 1.05749166e+00 -1.10909939e+00 5.31490028e-01
5.98036528e-01 5.07423937e-01 -6.84724450e-01 7.32546270e-01
-4.23301935e-01 1.02451277e+00 -5.12211025e-01 -3.07150781e-01
-2.55919904e-01 2.76878983e-01 4.25577402e-01 1.52008212e+00
4.33258295e-01 3.36668611e-01 2.74561197e-01 1.01047051e+00
-1.21937059e-01 2.27630049e-01 -4.55174893e-01 -2.94935495e-01
7.31534660e-01 1.36708236e+00 -7.97670662e-01 -6.04906023e-01
-4.30068970e-01 7.17621028e-01 3.48082483e-01 1.53287813e-01
-9.61025655e-01 -6.41583920e-01 3.17226261e-01 2.54905075e-01
2.34979898e-01 -7.95925781e-02 -5.79581141e-01 -1.32178640e+00
1.59655750e-01 -1.09674382e+00 4.59675997e-01 -8.40747237e-01
-1.16481471e+00 7.76741743e-01 2.43341044e-01 -1.39149666e+00
-7.65164137e-01 -2.04376087e-01 -4.98439163e-01 6.85687780e-01
-1.06644905e+00 -9.83359516e-01 -3.71820688e-01 3.73496711e-01
8.63138914e-01 -7.87385181e-02 9.37524140e-01 3.93674046e-01
-4.77249563e-01 7.41232991e-01 -6.84732378e-01 3.70985627e-01
6.10455096e-01 -1.18597472e+00 7.17790723e-01 9.71964955e-01
4.60746408e-01 5.46007574e-01 7.64225543e-01 -9.55349207e-01
-1.19753468e+00 -1.23830664e+00 1.29764664e+00 -8.16673458e-01
5.94938457e-01 -5.95477819e-01 -9.26903367e-01 6.22270942e-01
3.53414426e-03 -2.59139508e-01 7.43676960e-01 3.81862730e-01
-2.83627152e-01 -6.44196058e-03 -8.15124512e-01 8.12551558e-01
1.39118814e+00 -6.00265563e-01 -8.34519386e-01 2.80171990e-01
9.87029076e-01 -6.09643221e-01 -8.76901269e-01 5.11667728e-01
1.87219724e-01 -4.96156722e-01 6.74807966e-01 -8.12761128e-01
6.38824046e-01 -4.15131778e-01 1.94573969e-01 -1.23910797e+00
-7.87020996e-02 -7.17661023e-01 -2.27330700e-01 1.63805866e+00
1.06921935e+00 7.13998452e-03 5.73879719e-01 9.67739046e-01
-1.81746379e-01 -4.23857212e-01 -4.90982771e-01 -7.76755512e-01
-3.76547098e-01 -7.80555964e-01 8.21939826e-01 7.53172576e-01
2.47207016e-01 8.48774374e-01 -1.75032631e-01 1.38229374e-02
2.55531162e-01 1.75455213e-01 8.75811219e-01 -1.19234908e+00
-3.30452591e-01 -2.93228254e-02 4.32196170e-01 -8.76795292e-01
1.77708287e-02 -8.42533588e-01 2.74024934e-01 -1.48913324e+00
1.54497072e-01 -5.87038755e-01 -8.19312930e-02 7.74550915e-01
-5.23454726e-01 -2.64472425e-01 5.85610531e-02 4.14230563e-02
-7.32198119e-01 5.33908248e-01 1.12101269e+00 2.07203731e-01
-6.15646899e-01 -3.10809948e-02 -8.80800068e-01 5.58560371e-01
6.18048787e-01 -7.51468360e-01 -4.94818866e-01 -3.76862586e-01
4.97921735e-01 2.55906850e-01 1.31662013e-02 -9.28819776e-01
2.80876793e-02 -2.98979819e-01 4.83808100e-01 -7.68871725e-01
1.30320825e-02 -6.08609319e-01 1.83705688e-01 -2.66073570e-02
-6.42729759e-01 2.48782903e-01 2.38708630e-01 2.38554046e-01
-2.02842087e-01 -3.87324661e-01 3.10512483e-01 -2.51531452e-01
-4.13998663e-01 1.43739387e-01 -2.95079589e-01 5.06597996e-01
8.33983123e-01 -6.01777632e-04 -3.53995085e-01 -2.18598187e-01
-5.85938334e-01 6.49660975e-02 1.64708585e-01 5.04556537e-01
3.76397818e-01 -1.18685901e+00 -8.97485077e-01 1.95121139e-01
3.58531475e-01 3.77322912e-01 -4.64492962e-02 2.84966677e-01
-9.17963237e-02 8.96632522e-02 6.35903254e-02 -3.67552608e-01
-9.96963978e-01 4.81297702e-01 -2.99922794e-01 -7.18118787e-01
-8.68146181e-01 5.26471555e-01 -1.15703233e-01 -6.91765398e-02
1.42444819e-01 -6.42493904e-01 -2.24329203e-01 1.39264569e-01
8.35336804e-01 1.22233681e-01 2.29183465e-01 -4.67983680e-03
-1.69076979e-01 -5.18933460e-02 -1.62465408e-01 -3.74027789e-01
1.45401287e+00 2.24662930e-01 3.83408993e-01 2.84145594e-01
7.09058344e-01 1.64062545e-01 -1.44349110e+00 -2.15904191e-01
4.09795016e-01 -2.52421141e-01 -2.17970788e-01 -1.03963566e+00
-5.87191343e-01 5.74637890e-01 -9.17171538e-02 2.15596884e-01
1.14645064e+00 1.70422763e-01 1.01887786e+00 5.32605410e-01
4.08717901e-01 -1.40676761e+00 2.27140933e-01 4.62459803e-01
8.40748250e-01 -1.05179417e+00 -4.20935266e-02 -5.69877207e-01
-8.70532751e-01 7.45384514e-01 8.11585307e-01 1.87561080e-01
2.95546442e-01 7.20612288e-01 -1.93382949e-01 -1.03223789e-02
-1.10091257e+00 -2.50635654e-01 3.36553097e-01 5.11206388e-01
5.45268297e-01 -5.76966628e-02 -2.08659232e-01 1.30393744e+00
-5.28431356e-01 1.37804016e-01 4.19237345e-01 8.85331511e-01
-2.40528405e-01 -1.33155215e+00 -1.25551552e-01 7.22286403e-01
-5.62375844e-01 -3.48178118e-01 -2.03802083e-02 6.55911326e-01
1.74491797e-02 1.00539231e+00 -7.37215430e-02 -1.55739740e-01
7.31160522e-01 2.80850053e-01 4.32340503e-01 -8.80969286e-01
-8.55564415e-01 1.94782123e-01 8.06501269e-01 -5.17573595e-01
-8.67531672e-02 -7.26084888e-01 -1.66765809e+00 2.47775950e-02
-4.00742054e-01 2.73983806e-01 5.44947922e-01 1.01722300e+00
7.06295550e-01 8.07527483e-01 2.93264151e-01 -6.12167835e-01
-3.33268464e-01 -1.20962286e+00 -5.11859767e-02 7.15773821e-01
-7.61143789e-02 -5.33986866e-01 9.06156562e-03 5.06454289e-01] | [9.439888954162598, 9.016218185424805] |
dd870455-40c9-482f-aaa7-0a52512da21c | learning-to-agree-on-vision-attention-for | 2302.02117 | null | https://arxiv.org/abs/2302.02117v2 | https://arxiv.org/pdf/2302.02117v2.pdf | Learning to Agree on Vision Attention for Visual Commonsense Reasoning | Visual Commonsense Reasoning (VCR) remains a significant yet challenging research problem in the realm of visual reasoning. A VCR model generally aims at answering a textual question regarding an image, followed by the rationale prediction for the preceding answering process. Though these two processes are sequential and intertwined, existing methods always consider them as two independent matching-based instances. They, therefore, ignore the pivotal relationship between the two processes, leading to sub-optimal model performance. This paper presents a novel visual attention alignment method to efficaciously handle these two processes in a unified framework. To achieve this, we first design a re-attention module for aggregating the vision attention map produced in each process. Thereafter, the resultant two sets of attention maps are carefully aligned to guide the two processes to make decisions based on the same image regions. We apply this method to both conventional attention and the recent Transformer models and carry out extensive experiments on the VCR benchmark dataset. The results demonstrate that with the attention alignment module, our method achieves a considerable improvement over the baseline methods, evidently revealing the feasibility of the coupling of the two processes as well as the effectiveness of the proposed method. | ['Kejie Wang', 'Mohan Kankanhalli', 'Liqiang Nie', 'Fan Liu', 'Yangyang Guo', 'Zhenyang Li'] | 2023-02-04 | null | null | null | null | ['visual-reasoning', 'visual-commonsense-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning', 'reasoning'] | [ 4.53244746e-01 1.10481717e-01 -6.88017625e-03 -2.21111789e-01
-6.67696357e-01 -4.71303374e-01 9.38124716e-01 -7.41634071e-02
-1.81552678e-01 2.65452713e-01 4.75148201e-01 -5.86317778e-01
3.53260823e-02 -5.77094793e-01 -5.27386069e-01 -5.64859629e-01
7.39764392e-01 2.76879728e-01 3.25811535e-01 -2.48936161e-01
5.55293679e-01 1.96489573e-01 -1.31963122e+00 3.60741347e-01
9.10866082e-01 8.60746264e-01 3.51407081e-01 2.85474926e-01
-3.77491742e-01 1.43912494e+00 -4.83847886e-01 -9.00659919e-01
-4.16748822e-02 -6.34658039e-01 -1.00167513e+00 2.77414739e-01
2.35820606e-01 -3.87448817e-01 -4.55200970e-01 1.28214240e+00
2.92157859e-01 1.49568543e-01 6.55720592e-01 -1.19222713e+00
-1.35752094e+00 5.86737096e-01 -9.75734770e-01 4.63291198e-01
3.18167925e-01 2.12927416e-01 1.33326221e+00 -9.12153006e-01
3.15431178e-01 1.34295392e+00 2.31878430e-01 4.16219682e-01
-1.02969778e+00 -4.66768652e-01 6.33769870e-01 7.25445867e-01
-1.09627616e+00 -4.68299270e-01 1.11368585e+00 -4.41153884e-01
7.69739628e-01 2.13528737e-01 6.30950689e-01 9.67815697e-01
5.18298522e-02 9.93882298e-01 1.13989520e+00 -2.76531965e-01
2.44029686e-02 1.34564668e-01 2.56116569e-01 4.32199448e-01
5.59801757e-02 -2.37850785e-01 -3.80613267e-01 1.18854538e-01
6.84323430e-01 1.70746267e-01 -4.35127020e-01 -2.43328229e-01
-1.33196592e+00 8.12056720e-01 7.44433403e-01 3.54019076e-01
-5.56057572e-01 1.75978929e-01 2.99760938e-01 -6.99243546e-02
1.41726956e-01 1.98124811e-01 1.93079844e-01 3.34362656e-01
-7.14607537e-01 9.98394042e-02 3.38522553e-01 8.96503866e-01
4.05350983e-01 -1.84942484e-01 -8.27020705e-01 7.69545972e-01
5.20570099e-01 1.08850636e-01 2.69450963e-01 -6.93603933e-01
6.40016913e-01 7.75185883e-01 2.69696712e-01 -1.54481077e+00
4.88670170e-02 -4.53510970e-01 -8.18808794e-01 2.34385440e-03
2.70572543e-01 3.08277279e-01 -8.60404253e-01 1.70409560e+00
3.44636589e-01 -4.38217446e-03 1.24096617e-01 1.14362442e+00
9.63981688e-01 6.71360970e-01 3.73098344e-01 -1.13621987e-01
1.67871344e+00 -1.45528936e+00 -1.00335479e+00 -4.08686697e-01
-3.83320637e-02 -7.87875056e-01 1.24948692e+00 9.97109860e-02
-1.23350096e+00 -5.57107925e-01 -1.18429816e+00 -6.82123840e-01
-5.07160500e-02 1.12174124e-01 4.89117116e-01 5.54704294e-02
-9.49471772e-01 2.27493554e-01 -3.53984296e-01 -1.93261713e-01
5.38645625e-01 -1.14466012e-01 2.48488057e-02 -1.20712332e-01
-1.18492830e+00 9.98980165e-01 4.55302261e-02 6.46061182e-01
-8.16847324e-01 -3.58568519e-01 -4.89888549e-01 2.72880554e-01
4.50442612e-01 -7.24801958e-01 1.26919281e+00 -9.22280967e-01
-1.16147411e+00 1.09533930e+00 -5.12206972e-01 -3.11430126e-01
8.43463778e-01 -2.81236321e-01 -2.41860554e-01 2.10450813e-01
2.72231638e-01 4.62970585e-01 8.63465250e-01 -1.55873752e+00
-7.70175517e-01 -3.72359186e-01 4.42741841e-01 4.71212268e-01
-1.06137559e-01 -2.97989906e-03 -8.09088528e-01 -6.08825684e-01
2.08381772e-01 -4.68546420e-01 -1.58265773e-02 -6.39356002e-02
-4.19475973e-01 -4.97081399e-01 6.24320984e-01 -7.93256819e-01
1.29107308e+00 -2.13680649e+00 4.78116602e-01 -1.38039991e-01
5.54216444e-01 1.30706817e-01 -5.85683100e-02 2.99369693e-01
-2.33667910e-01 3.91072920e-03 -8.64710510e-02 -2.35034466e-01
-2.27125250e-02 -4.84417826e-02 -8.06682289e-01 3.98966581e-01
3.06809723e-01 1.19961989e+00 -1.04399025e+00 -6.31734729e-01
1.64288133e-01 3.47559452e-01 -2.47256249e-01 4.21526670e-01
-1.75161958e-01 4.23395485e-01 -4.57161278e-01 5.80263555e-01
4.91784871e-01 -6.12879813e-01 2.86018163e-01 -5.15254021e-01
2.78480239e-02 1.61507249e-01 -6.71080649e-01 1.25111330e+00
-1.88358724e-01 6.91388428e-01 -2.22660229e-01 -1.13925278e+00
1.08621347e+00 3.02808285e-01 5.21469973e-02 -1.02645350e+00
2.50848234e-01 -3.79604474e-02 1.16636246e-01 -7.52482116e-01
4.06339794e-01 -4.12936270e-01 1.33882105e-01 4.97219652e-01
-1.72655135e-01 1.34082824e-01 8.32676813e-02 3.77056688e-01
7.09506035e-01 2.17366278e-01 5.99583805e-01 7.57677946e-04
8.56119335e-01 4.46682312e-02 3.99992973e-01 7.18544304e-01
-6.06555641e-01 5.95456481e-01 7.49435842e-01 -4.88753706e-01
-1.03954649e+00 -8.33216012e-01 2.79720277e-01 1.00576794e+00
7.50986218e-01 -2.31080860e-01 -5.73603451e-01 -7.28283346e-01
-1.97720617e-01 9.57197130e-01 -9.52395082e-01 -1.80211872e-01
-4.72787738e-01 -5.89346945e-01 2.00132146e-01 7.56988049e-01
6.39568985e-01 -1.23999393e+00 -6.67934597e-01 7.67465606e-02
-5.55998504e-01 -1.21772158e+00 -3.60749185e-01 -1.75436169e-01
-4.90362674e-01 -1.18340600e+00 -7.05255210e-01 -8.25706363e-01
5.91989577e-01 8.23036790e-01 1.08121061e+00 5.47343194e-01
8.35060999e-02 2.15401605e-01 -3.68657559e-01 -1.80665106e-01
-2.87066221e-01 -2.96443999e-01 -4.75671738e-01 3.51696461e-01
5.15950441e-01 -4.65892226e-01 -7.65984893e-01 1.54910251e-01
-7.57558703e-01 5.26340246e-01 6.77832663e-01 7.92443812e-01
5.75726867e-01 -1.87808886e-01 5.87826431e-01 -6.73236668e-01
7.31342673e-01 -4.68651295e-01 -3.95415843e-01 7.20538020e-01
-4.01256531e-01 1.14325978e-01 5.84033310e-01 -3.64148319e-01
-1.27503884e+00 -2.73015440e-01 5.65448869e-03 -5.93270183e-01
7.34425485e-02 3.16689461e-01 -3.42417240e-01 3.26603562e-01
1.45247757e-01 4.67906117e-01 -1.66536301e-01 -1.19510934e-01
5.48992991e-01 5.41945994e-01 7.13199496e-01 -4.83390927e-01
8.67656350e-01 5.71161866e-01 -4.23274547e-01 -2.69084930e-01
-1.14618075e+00 -3.10862035e-01 -4.60386306e-01 -4.31130528e-01
1.12567425e+00 -7.85350025e-01 -9.36985910e-01 2.15568379e-01
-1.43313575e+00 1.30057354e-02 -1.96013674e-02 -7.26238638e-02
-4.63352352e-01 4.77232635e-01 -4.92077500e-01 -8.80305648e-01
-3.82040441e-01 -1.33365667e+00 9.12602663e-01 3.31511736e-01
-9.96797383e-02 -8.00912857e-01 -8.43160823e-02 8.22282553e-01
2.53483504e-01 4.46288707e-03 1.17214561e+00 -6.40130997e-01
-7.33198404e-01 9.87527668e-02 -8.72965753e-01 3.14854048e-02
1.06834933e-01 3.04216519e-02 -9.54440296e-01 -7.98213179e-04
2.11833075e-01 -2.78551489e-01 9.59925413e-01 -4.22290750e-02
1.17375422e+00 -1.14798404e-01 -1.71363622e-01 2.43945658e-01
1.40552354e+00 4.77463305e-01 8.46129596e-01 4.77159411e-01
9.35297489e-01 7.31012344e-01 5.56917131e-01 -5.02994396e-02
6.87741518e-01 6.18072689e-01 6.86894715e-01 -3.07457507e-01
-1.78307846e-01 -4.59376663e-01 7.31102228e-02 8.70617986e-01
-3.30356956e-01 -1.45941198e-01 -9.42644358e-01 5.83363414e-01
-2.09145808e+00 -1.19186401e+00 -2.13506684e-01 1.86406589e+00
6.19051635e-01 3.62758115e-02 -2.10625023e-01 9.41675603e-02
1.01129389e+00 4.10833240e-01 -5.84239185e-01 -3.48880231e-01
8.69314596e-02 -1.99404463e-01 -1.91915184e-01 3.78297865e-01
-9.34258401e-01 9.18605685e-01 6.27774286e+00 7.18569934e-01
-9.20556784e-01 1.54092118e-01 7.61297584e-01 9.60054547e-02
-6.36089027e-01 2.11895093e-01 -5.09897232e-01 3.19902152e-01
2.48927906e-01 -2.51934022e-01 5.49398839e-01 6.29432976e-01
1.29233852e-01 -1.38783092e-02 -9.77996826e-01 1.05806983e+00
2.92265415e-01 -1.29585159e+00 4.93776649e-01 -8.99828523e-02
5.39123237e-01 -4.63572711e-01 8.32441747e-02 2.64452964e-01
2.49755502e-01 -1.11096323e+00 1.15420687e+00 4.50967729e-01
4.22358096e-01 -6.49398267e-01 6.30713999e-01 2.51064509e-01
-1.26409948e+00 -2.10718602e-01 -2.67129153e-01 -2.71417499e-02
3.51787478e-01 3.02727014e-01 -3.24228793e-01 7.43536949e-01
6.03696227e-01 7.25685060e-01 -6.61974013e-01 6.37601852e-01
-6.37500823e-01 2.64138669e-01 4.66553509e-01 -4.60047089e-03
3.04174125e-01 -3.16856802e-01 4.88386661e-01 7.79748321e-01
-6.33522198e-02 2.39105225e-01 -1.92588061e-01 1.20099580e+00
-1.10469654e-01 1.18713841e-01 -3.46015900e-01 -4.69646156e-02
4.08695579e-01 1.32958519e+00 -8.77088130e-01 -5.67424834e-01
-6.22709811e-01 1.02953744e+00 8.91243041e-01 5.10479808e-01
-1.32418108e+00 -1.96325064e-01 3.17335010e-01 -1.21423461e-01
4.01901901e-01 5.68660349e-02 -3.99554521e-01 -1.31543982e+00
1.51842013e-01 -1.03595412e+00 3.34370822e-01 -1.16655397e+00
-1.35320985e+00 7.41392791e-01 -1.88916698e-01 -1.13216031e+00
1.44846171e-01 -5.49723268e-01 -7.71102130e-01 9.02696192e-01
-1.76882148e+00 -1.19006097e+00 -5.29026151e-01 5.42376041e-01
7.45730162e-01 2.01274380e-01 4.13955003e-01 1.01830803e-01
-7.39704370e-01 2.81298965e-01 -3.57066035e-01 1.49734303e-01
4.49105799e-01 -1.21449578e+00 3.16741168e-01 1.17293632e+00
2.29865789e-01 8.17395926e-01 6.06799185e-01 -5.02740026e-01
-1.13123357e+00 -8.83555353e-01 8.80248249e-01 -4.88273859e-01
7.87579477e-01 -2.08034635e-01 -1.13786495e+00 7.66125500e-01
6.62644982e-01 -2.89156139e-01 5.96425772e-01 4.51648049e-02
-6.46045983e-01 -3.20779346e-02 -6.61392093e-01 9.96479809e-01
1.05600548e+00 -5.85584700e-01 -1.21238303e+00 7.45005757e-02
7.62812674e-01 -2.49909982e-01 -5.31535804e-01 3.42775196e-01
4.41678941e-01 -1.10468650e+00 8.50433588e-01 -6.52615666e-01
1.12228286e+00 -5.20896494e-01 -2.03204840e-01 -8.35393846e-01
-5.63451469e-01 -3.87299627e-01 -1.82646364e-01 1.48907471e+00
1.30381078e-01 -4.70336914e-01 6.71238527e-02 4.58533764e-01
1.56316072e-01 -8.86307657e-01 -5.32675743e-01 -2.00473636e-01
-1.19768113e-01 -1.74547151e-01 6.25253141e-01 1.01132357e+00
-5.77292442e-02 8.77419591e-01 -4.16455001e-01 1.60482436e-01
5.80522954e-01 6.24264836e-01 6.84137464e-01 -9.65235174e-01
-3.21401209e-01 -6.80712521e-01 -4.48225588e-02 -1.22030973e+00
1.15817405e-01 -7.66344726e-01 2.08997354e-01 -1.92454863e+00
8.59733462e-01 -1.58881724e-01 -3.42644423e-01 4.17008460e-01
-7.76496649e-01 1.61229908e-01 3.80815834e-01 5.62260866e-01
-6.61044896e-01 6.77800536e-01 1.47738171e+00 -1.52624846e-01
7.28755519e-02 -4.49240863e-01 -1.30293608e+00 7.44140685e-01
5.66851735e-01 -6.04096018e-02 -6.74153864e-01 -6.10467613e-01
1.22190118e-01 5.40934317e-02 7.27134109e-01 -5.58340073e-01
2.98253864e-01 -2.81005859e-01 3.51367295e-01 -6.27954841e-01
3.28038894e-02 -8.18517029e-01 -6.47580251e-02 1.79950356e-01
-5.49066901e-01 1.70332983e-01 -1.00131586e-01 7.73782790e-01
-2.96844006e-01 5.98415965e-03 7.34985888e-01 -1.10054672e-01
-8.52943897e-01 1.40204400e-01 1.66473966e-02 7.57386386e-02
1.18190432e+00 -2.46613130e-01 -5.33830702e-01 -3.82588714e-01
-6.35655761e-01 3.95600170e-01 2.26272687e-01 5.50682247e-01
6.48737133e-01 -1.31128013e+00 -6.36250734e-01 -1.39864281e-01
2.28596255e-01 -1.46777444e-02 5.07726789e-01 9.43635881e-01
-2.81899124e-01 4.03306425e-01 -1.94259256e-01 -4.84621435e-01
-1.05149519e+00 1.08037472e+00 3.10371727e-01 -4.32849854e-01
-7.71003783e-01 5.77637970e-01 7.00183213e-01 5.02694808e-02
2.29242593e-01 -8.00517127e-02 -6.84535861e-01 1.74547434e-01
7.13963151e-01 9.34345275e-03 -3.04755628e-01 -7.65061915e-01
-3.25562209e-01 5.13083518e-01 -2.27462322e-01 -5.77821918e-02
1.14391279e+00 -4.22635317e-01 -2.56704271e-01 4.91735846e-01
8.24087679e-01 -9.95312557e-02 -1.26524889e+00 -2.39485323e-01
-1.32053152e-01 -4.96768802e-01 -1.73070714e-01 -5.93149304e-01
-1.18217766e+00 1.08036983e+00 -6.08844683e-02 3.61897856e-01
1.27272964e+00 2.78599471e-01 4.86888736e-01 -3.65798697e-02
5.32740215e-03 -6.53459489e-01 3.22494298e-01 2.13315740e-01
1.14984000e+00 -1.10220182e+00 3.46163735e-02 -4.97909248e-01
-9.59847748e-01 9.63648856e-01 8.13994944e-01 -6.64456263e-02
5.16966954e-02 -1.88223645e-01 5.55952713e-02 -3.52502018e-01
-9.45129931e-01 -2.99952775e-01 3.76690328e-01 3.88114154e-01
5.00804484e-01 -1.54761463e-01 -4.15261954e-01 6.52077377e-01
1.84298635e-01 -1.08639644e-02 2.79720068e-01 7.02347219e-01
-2.51506776e-01 -6.32765651e-01 -3.41529191e-01 1.45951197e-01
-3.52312148e-01 -2.31473133e-01 -4.30759430e-01 7.60862291e-01
-2.26249471e-01 9.49678183e-01 3.50002535e-02 -3.49188119e-01
4.58603024e-01 -1.34190004e-02 3.64595979e-01 -2.39748403e-01
-5.05358040e-01 7.76299974e-03 -1.70183301e-01 -5.37618518e-01
-6.42416477e-01 -4.40835148e-01 -1.06412721e+00 -2.54236072e-01
-2.63547748e-01 -6.50583431e-02 9.51260924e-02 1.22921002e+00
2.26828188e-01 8.88493359e-01 4.72409099e-01 -5.61488867e-01
-6.08662248e-01 -7.09028065e-01 -2.19448239e-01 6.83603704e-01
2.24600762e-01 -7.37654209e-01 -3.72654527e-01 1.66653588e-01] | [10.67404842376709, 1.7343604564666748] |
f13dc381-27ec-443c-a1bb-7889974f11a6 | iterative-greedy-matching-for-3d-human-pose | 2101.09745 | null | https://arxiv.org/abs/2101.09745v1 | https://arxiv.org/pdf/2101.09745v1.pdf | Iterative Greedy Matching for 3D Human Pose Tracking from Multiple Views | In this work we propose an approach for estimating 3D human poses of multiple people from a set of calibrated cameras. Estimating 3D human poses from multiple views has several compelling properties: human poses are estimated within a global coordinate space and multiple cameras provide an extended field of view which helps in resolving ambiguities, occlusions and motion blur. Our approach builds upon a real-time 2D multi-person pose estimation system and greedily solves the association problem between multiple views. We utilize bipartite matching to track multiple people over multiple frames. This proofs to be especially efficient as problems associated with greedy matching such as occlusion can be easily resolved in 3D. Our approach achieves state-of-the-art results on popular benchmarks and may serve as a baseline for future work. | ['Juergen Gall', 'Julian Tanke'] | 2021-01-24 | null | null | null | null | ['3d-human-pose-tracking'] | ['computer-vision'] | [-3.27591628e-01 -2.62512594e-01 -2.86655314e-03 -3.00258577e-01
-7.80268848e-01 -7.27739573e-01 4.33624059e-01 -2.26664618e-01
-5.28223932e-01 6.00564480e-01 4.50878918e-01 4.39403623e-01
1.57922417e-01 -3.35102886e-01 -6.30202830e-01 -2.27599591e-01
-4.29284610e-02 9.24776435e-01 3.34722877e-01 3.32345488e-03
-8.88581499e-02 5.68811178e-01 -1.33422983e+00 -1.50981992e-02
2.02076342e-02 3.40066016e-01 -2.86873758e-01 9.30790126e-01
4.77982938e-01 3.29184830e-01 -5.36632359e-01 -6.25432014e-01
7.42574811e-01 -2.12413341e-01 -6.42631114e-01 7.03580797e-01
1.31708539e+00 -7.04760253e-01 -5.38281977e-01 8.13494563e-01
7.32882559e-01 2.96619207e-01 2.35212773e-01 -1.43218780e+00
9.53171495e-03 -2.71195620e-01 -1.09304667e+00 1.79593056e-01
1.33630252e+00 -1.82496104e-02 7.91155279e-01 -9.91658926e-01
8.64588201e-01 1.70925391e+00 9.25609171e-01 4.83062267e-01
-1.27433765e+00 -2.99277544e-01 1.60265610e-01 -2.60910932e-02
-1.55900872e+00 -4.87086654e-01 3.54110062e-01 -3.82152557e-01
8.08011293e-01 3.43427867e-01 1.12098789e+00 1.13470685e+00
5.17217554e-02 7.04301536e-01 8.53932321e-01 -4.39255238e-01
-2.84354836e-01 -1.52330905e-01 1.08266778e-01 8.92918587e-01
6.47624016e-01 5.27711399e-02 -8.21602821e-01 -4.58294362e-01
1.19878852e+00 3.70893240e-01 -2.05143124e-01 -9.57819045e-01
-1.62653863e+00 6.17998838e-01 2.58854210e-01 -3.83990020e-01
-2.45008275e-01 3.48892003e-01 2.07891017e-01 1.68122903e-01
2.24938363e-01 1.29792050e-01 -1.55635148e-01 -1.13016907e-02
-6.62785947e-01 8.96330237e-01 7.45304704e-01 1.29579103e+00
5.98781168e-01 -5.80424368e-01 7.09983939e-03 4.87855285e-01
3.63287717e-01 7.20784962e-01 -2.22564921e-01 -1.44909334e+00
6.22630596e-01 5.67940950e-01 4.44076747e-01 -1.17467916e+00
-5.40496826e-01 4.36853059e-02 -6.23116434e-01 1.78294748e-01
7.94766128e-01 -2.17050061e-01 -4.34260249e-01 1.64900613e+00
8.22937787e-01 1.11762680e-01 -2.91719764e-01 1.26259077e+00
6.14933133e-01 1.31646886e-01 -4.39470619e-01 1.62230171e-02
1.75240111e+00 -1.12770355e+00 -5.72329104e-01 -6.71167076e-01
-6.59738854e-02 -8.03868830e-01 3.06854367e-01 2.85933852e-01
-1.28004205e+00 -5.62242031e-01 -7.03473628e-01 -1.63741991e-01
6.44718260e-02 -3.16403806e-02 5.41799247e-01 7.15649545e-01
-1.10381281e+00 1.78457618e-01 -9.77181256e-01 -8.74007225e-01
-1.10264998e-02 5.09868622e-01 -9.17111337e-01 -1.97878808e-01
-8.02207828e-01 1.13550162e+00 1.82519313e-02 4.33948152e-02
-5.74504256e-01 -2.52874702e-01 -9.70603645e-01 -4.64427859e-01
6.16078734e-01 -1.37447417e+00 1.18382728e+00 -3.60918939e-01
-1.06217253e+00 1.30329359e+00 -4.52212393e-01 -1.29622653e-01
8.63308072e-01 -7.89595962e-01 -6.96836486e-02 3.78759176e-01
2.86888480e-01 5.96543729e-01 5.61575234e-01 -1.05989015e+00
-5.73343515e-01 -8.09620082e-01 2.55475551e-01 6.67080581e-01
1.25701609e-03 2.52303809e-01 -1.25791645e+00 -5.42118669e-01
4.79540884e-01 -1.43257558e+00 -3.81804258e-01 2.93220907e-01
-4.76775795e-01 1.49963617e-01 4.11231399e-01 -6.09847903e-01
7.92576492e-01 -1.58400238e+00 6.43276989e-01 1.49319097e-01
5.32243848e-01 -2.40482137e-01 2.64574885e-01 4.05991226e-01
1.68105170e-01 -4.26788867e-01 5.62616229e-01 -7.14109838e-01
-2.54104231e-02 -3.67050506e-02 3.90524268e-01 9.90143120e-01
-3.74365151e-01 7.89840698e-01 -7.62366593e-01 -7.26510942e-01
3.60198736e-01 4.68760967e-01 -4.92926866e-01 2.89635748e-01
4.54742819e-01 5.48208058e-01 -2.08956644e-01 7.36903429e-01
6.30893469e-01 -4.32988107e-01 3.45845371e-01 -2.37781286e-01
1.22914851e-01 -1.12011150e-01 -1.87511957e+00 2.00529671e+00
2.16775745e-01 3.30633372e-01 1.88847885e-01 -3.73466969e-01
4.49444145e-01 5.56311488e-01 7.26016581e-01 4.33295555e-02
7.82636460e-03 -3.01773608e-01 -6.78919852e-01 -2.20328569e-01
5.69021463e-01 1.34339839e-01 -3.24866623e-01 5.20541430e-01
2.58946698e-02 2.14794651e-01 1.37844473e-01 3.56099367e-01
1.05722106e+00 3.72572184e-01 7.87536860e-01 -5.35265356e-02
3.14129025e-01 -1.11549534e-01 6.99458897e-01 7.70506561e-01
-4.80443627e-01 8.14369798e-01 1.63008466e-01 -9.03271794e-01
-1.17768860e+00 -1.50042093e+00 2.92504042e-01 8.00897181e-01
4.18356419e-01 -7.64537454e-01 -5.45034528e-01 -7.01281309e-01
2.33639106e-01 -4.10888046e-01 -5.41839778e-01 3.16543788e-01
-7.78941870e-01 -4.17001814e-01 3.85648042e-01 7.02364326e-01
3.31128478e-01 -1.62289903e-01 -8.10118973e-01 -5.60273193e-02
-7.03853786e-01 -1.55910432e+00 -9.62223351e-01 -3.43386471e-01
-7.54514694e-01 -1.49233425e+00 -1.08213615e+00 -6.03292346e-01
8.86237919e-01 9.43202317e-01 1.51136351e+00 3.00828367e-02
-3.58757913e-01 1.02616096e+00 -2.93535814e-02 -1.85922638e-01
2.34782249e-01 -3.29555035e-01 6.53986096e-01 -1.22616746e-01
6.52193487e-01 -3.10301572e-01 -7.49502063e-01 7.67587364e-01
-1.14914417e-01 5.07950112e-02 7.25556612e-02 5.42144001e-01
4.93413985e-01 -3.75533670e-01 -2.23726496e-01 -7.98652649e-01
1.14738502e-01 9.18626506e-03 -6.80076957e-01 3.12199533e-01
6.95342422e-02 -2.93618768e-01 -2.10165814e-01 -3.16100180e-01
-9.35373366e-01 5.95505238e-01 3.63285542e-01 -5.46411395e-01
-3.20574373e-01 -2.13581890e-01 -4.97222953e-02 -1.65296867e-01
6.07969344e-01 -3.37221980e-01 8.50247219e-02 -3.86422038e-01
2.24917248e-01 1.96576715e-01 8.01017761e-01 -6.34642959e-01
1.00687099e+00 7.90067673e-01 2.88485914e-01 -6.15019500e-01
-9.37976956e-01 -1.10372615e+00 -1.12413430e+00 -4.33762074e-01
9.26685631e-01 -1.56549668e+00 -1.09451604e+00 4.56642240e-01
-1.35182309e+00 3.98661494e-01 1.60254166e-01 5.84261358e-01
-5.08647919e-01 6.80126309e-01 -6.89184189e-01 -8.61942410e-01
-4.04798090e-02 -1.01657188e+00 1.52060330e+00 1.38441324e-01
-7.01064408e-01 -9.28829789e-01 3.47348154e-01 7.15280652e-01
-2.54534602e-01 6.10549450e-01 -3.59633774e-01 -1.61066353e-01
-7.92347729e-01 -5.33962786e-01 1.29463494e-01 -3.55637342e-01
-5.25569916e-02 -3.98852617e-01 -7.40834653e-01 -6.97624803e-01
-2.49584854e-01 -3.51931006e-01 5.29168427e-01 4.83738780e-01
2.56458074e-01 -2.49086283e-02 -6.14463687e-01 4.67994899e-01
1.11907554e+00 -4.25729901e-01 3.07572991e-01 3.46583843e-01
8.87514353e-01 6.00312352e-01 5.10113418e-01 5.82630575e-01
8.10517609e-01 1.27597797e+00 1.13382362e-01 -5.12304902e-02
5.08349128e-02 -1.45956576e-01 2.56588280e-01 4.15752739e-01
-4.62785274e-01 -3.91194560e-02 -8.49843919e-01 2.15752989e-01
-2.12953496e+00 -1.24697101e+00 -2.84161538e-01 2.39481068e+00
3.92603308e-01 -3.21685486e-02 8.67752671e-01 -1.51813537e-01
1.06680882e+00 7.35618398e-02 -1.56471491e-01 5.16513407e-01
-5.67475148e-02 -2.92245805e-01 6.25361502e-01 5.45052052e-01
-1.34206271e+00 6.24797165e-01 7.35632658e+00 -1.26557544e-01
-8.41792747e-02 9.30485427e-02 1.65525511e-01 -6.05542064e-01
2.86398888e-01 -7.84786567e-02 -1.31098223e+00 6.94480985e-02
2.35110402e-01 1.60677746e-01 3.63787591e-01 5.11496246e-01
-4.81956266e-02 -3.22799683e-01 -1.39164364e+00 1.51313317e+00
4.25952435e-01 -1.00369811e+00 -3.63161981e-01 4.81855541e-01
6.98911786e-01 -1.96064577e-01 -3.54240417e-01 -1.96164235e-01
4.75668967e-01 -5.56302905e-01 6.93622530e-01 5.77128947e-01
4.87030327e-01 -7.25041687e-01 5.75411856e-01 3.80024552e-01
-1.53838563e+00 2.58047432e-01 -4.15728539e-01 -3.02672803e-01
5.03655851e-01 3.98958594e-01 -5.49691796e-01 5.35886884e-01
9.50586259e-01 7.54540205e-01 -6.37171805e-01 1.05044484e+00
-2.00686585e-02 -3.42934102e-01 -6.02138400e-01 3.26598614e-01
-3.08915496e-01 -1.52988255e-01 6.91032708e-01 1.00539029e+00
2.46684104e-01 2.16719359e-01 7.75579333e-01 2.58801550e-01
6.73306882e-02 -2.93694466e-01 -7.36412466e-01 5.81667066e-01
4.08861667e-01 1.20195675e+00 -7.80518353e-01 -3.70042890e-01
-7.26040125e-01 1.27847195e+00 4.32510853e-01 2.62627780e-01
-9.06190932e-01 2.48713389e-01 7.26813138e-01 2.34816790e-01
2.00706065e-01 -5.31754315e-01 1.32135347e-01 -1.81667292e+00
2.29390264e-01 -9.93389249e-01 7.89188445e-01 -6.72725201e-01
-1.26605690e+00 3.05869162e-01 2.52101541e-01 -1.38778746e+00
-3.70542467e-01 -4.50398475e-01 -1.46541744e-01 6.82615876e-01
-7.81568110e-01 -1.18459690e+00 -5.75912595e-01 9.30739880e-01
4.33733135e-01 -9.38408524e-02 8.70795429e-01 3.85496110e-01
-3.77468884e-01 4.22768652e-01 -3.96519691e-01 3.22578430e-01
1.16663110e+00 -1.38683712e+00 6.83347285e-01 1.04323733e+00
3.85467768e-01 7.90773273e-01 7.29266465e-01 -7.18598783e-01
-1.69038916e+00 -6.31480217e-01 9.56832111e-01 -1.20035553e+00
5.07416166e-02 -4.95681137e-01 -1.95293218e-01 1.24026918e+00
3.18255238e-02 1.90138161e-01 7.58348346e-01 5.12166381e-01
-4.53840196e-01 1.73682277e-03 -1.02542603e+00 4.31380153e-01
1.39560568e+00 -3.62187177e-01 -6.31573439e-01 4.74674165e-01
2.03302577e-01 -1.13700461e+00 -7.13139832e-01 -1.31487632e-02
8.76534462e-01 -1.22292721e+00 1.63883996e+00 -4.62820262e-01
-2.61488914e-01 -3.53553116e-01 -2.04120398e-01 -8.63548577e-01
-5.90672731e-01 -7.57026017e-01 -3.20578784e-01 9.34925258e-01
-3.11240733e-01 -3.52690130e-01 1.08275819e+00 1.00965381e+00
5.00173271e-01 -1.56275913e-01 -8.19911540e-01 -7.45622098e-01
-6.76074326e-01 6.27872953e-03 2.51906484e-01 6.28611624e-01
-4.05924693e-02 4.07350838e-01 -9.39482272e-01 4.47097093e-01
1.34668911e+00 7.73756206e-02 1.45182335e+00 -1.29201329e+00
-5.41315138e-01 1.31638423e-01 -7.00465679e-01 -1.28096712e+00
-2.51029097e-02 -2.10415468e-01 -1.79722473e-01 -1.27307868e+00
6.80812538e-01 6.29752427e-02 1.60767570e-01 2.11969569e-01
-3.57410342e-01 6.68576300e-01 5.17674148e-01 2.58297026e-01
-1.17924380e+00 -6.40837625e-02 9.45691943e-01 4.68050353e-02
1.11575276e-01 2.87252367e-01 -5.48947752e-01 9.90440071e-01
3.44844073e-01 -4.58294541e-01 -4.94831167e-02 -6.09266877e-01
2.42346004e-01 4.71705496e-01 7.67594337e-01 -1.16117287e+00
5.77787161e-01 -1.10407509e-01 1.00088418e+00 -8.09388340e-01
8.97131920e-01 -9.29867446e-01 6.69748843e-01 4.10351247e-01
1.18125133e-01 6.48183644e-01 -1.04646794e-01 8.30690920e-01
3.40841785e-02 1.68726951e-01 7.09425807e-01 -8.00326645e-01
-6.43746376e-01 4.37193841e-01 -3.90341729e-02 2.49017671e-01
1.14132226e+00 -3.93898785e-01 -5.19747809e-02 -6.23981714e-01
-7.58122385e-01 4.65896815e-01 9.26618338e-01 4.60650921e-01
5.99824488e-01 -1.67654657e+00 -6.82941020e-01 7.09340200e-02
7.54362941e-02 -1.16941772e-01 1.44667864e-01 7.08364964e-01
-6.67522252e-01 3.37746650e-01 -2.79374182e-01 -9.38598752e-01
-1.84343851e+00 4.25252676e-01 3.11821282e-01 -2.37854540e-01
-7.41115987e-01 8.51327181e-01 3.16281803e-03 -4.86912906e-01
4.00877506e-01 3.30270141e-01 1.49644166e-01 -4.58059795e-02
8.26281905e-01 7.86049783e-01 -2.64375210e-01 -1.03121924e+00
-7.15482295e-01 9.81311321e-01 1.61807403e-01 -3.71566564e-01
1.13927400e+00 -6.90250516e-01 -2.15808097e-02 3.17750156e-01
9.64447141e-01 1.97672814e-01 -1.37259543e+00 -4.46313709e-01
-2.85950243e-01 -1.01015997e+00 -4.46050882e-01 -4.60888207e-01
-8.68098319e-01 4.15193409e-01 3.67240280e-01 -9.56708938e-02
7.62918413e-01 1.51484787e-01 7.38200188e-01 5.11964858e-01
7.17111945e-01 -9.54653502e-01 2.77107716e-01 3.17137659e-01
5.88344991e-01 -1.39867139e+00 5.64696014e-01 -5.64702034e-01
-5.07465422e-01 1.03888619e+00 7.69254327e-01 -1.86957553e-01
2.76836574e-01 3.36921215e-01 1.51162460e-01 -3.01186085e-01
-5.10036469e-01 -1.47725970e-01 5.11017919e-01 6.18216813e-01
3.60038221e-01 -1.00361936e-01 1.69490233e-01 5.47627285e-02
6.98197410e-02 -1.83667034e-01 3.77199054e-01 1.06388283e+00
-2.69008487e-01 -1.22203469e+00 -1.10523593e+00 -4.12779301e-02
-4.35828865e-01 4.24587637e-01 -4.43139344e-01 8.00507486e-01
-3.19395848e-02 9.46987927e-01 -7.14369307e-05 -9.73956510e-02
6.00302577e-01 -3.02338451e-02 1.08877432e+00 -5.95232189e-01
-4.17233139e-01 4.91376191e-01 1.85066536e-01 -9.89033222e-01
-9.09912586e-01 -1.12190652e+00 -6.45911574e-01 -5.86953819e-01
-2.43520930e-01 -1.09884925e-01 4.86474894e-02 7.41858423e-01
2.36557692e-01 6.84480295e-02 2.41718486e-01 -1.36974621e+00
-4.40413266e-01 -4.87791985e-01 -5.50714672e-01 7.69724846e-01
4.25171822e-01 -8.80085945e-01 -2.39883848e-02 2.67718226e-01] | [7.048717021942139, -0.9976739883422852] |
0d13779f-a3e2-41b8-911f-5b4e5340429e | multi-task-text-classification-using-graph | 2205.01204 | null | https://arxiv.org/abs/2205.01204v1 | https://arxiv.org/pdf/2205.01204v1.pdf | Multi-Task Text Classification using Graph Convolutional Networks for Large-Scale Low Resource Language | Graph Convolutional Networks (GCN) have achieved state-of-art results on single text classification tasks like sentiment analysis, emotion detection, etc. However, the performance is achieved by testing and reporting on resource-rich languages like English. Applying GCN for multi-task text classification is an unexplored area. Moreover, training a GCN or adopting an English GCN for Indian languages is often limited by data availability, rich morphological variation, syntax, and semantic differences. In this paper, we study the use of GCN for the Telugu language in single and multi-task settings for four natural language processing (NLP) tasks, viz. sentiment analysis (SA), emotion identification (EI), hate-speech (HS), and sarcasm detection (SAR). In order to evaluate the performance of GCN with one of the Indian languages, Telugu, we analyze the GCN based models with extensive experiments on four downstream tasks. In addition, we created an annotated Telugu dataset, TEL-NLP, for the four NLP tasks. Further, we propose a supervised graph reconstruction method, Multi-Task Text GCN (MT-Text GCN) on the Telugu that leverages to simultaneously (i) learn the low-dimensional word and sentence graph embeddings from word-sentence graph reconstruction using graph autoencoder (GAE) and (ii) perform multi-task text classification using these latent sentence graph embeddings. We argue that our proposed MT-Text GCN achieves significant improvements on TEL-NLP over existing Telugu pretrained word embeddings, and multilingual pretrained Transformer models: mBERT, and XLM-R. On TEL-NLP, we achieve a high F1-score for four NLP tasks: SA (0.84), EI (0.55), HS (0.83) and SAR (0.66). Finally, we show our model's quantitative and qualitative analysis on the four NLP tasks in Telugu. | ['Radhika Mamidi', 'Venkata Charan Chinni', 'Lakshmi Sireesha Vakada', 'Subba Reddy Oota', 'Mounika Marreddy'] | 2022-05-02 | null | null | null | null | ['graph-reconstruction', 'xlm-r'] | ['graphs', 'natural-language-processing'] | [ 6.78367242e-02 1.02781951e-01 6.31676316e-02 -3.11894089e-01
-7.27874279e-01 -6.48410559e-01 3.65912557e-01 2.96661407e-01
-4.00097817e-01 3.19380224e-01 3.72783840e-01 -6.91282034e-01
2.90213168e-01 -6.60569310e-01 -4.87420857e-01 -4.90940988e-01
7.37158209e-02 4.66021955e-01 -4.31325048e-01 -5.33378363e-01
-2.24160030e-02 1.51419327e-01 -8.78423750e-01 3.58882308e-01
9.88924563e-01 7.42799044e-01 2.94642933e-02 9.55728889e-01
-1.50222629e-01 8.60409677e-01 -5.75654745e-01 -9.60946500e-01
-2.56300986e-01 -2.87671268e-01 -9.26776052e-01 4.04988788e-03
3.14202696e-01 -1.77527741e-02 -1.06013671e-01 1.02251208e+00
5.10806799e-01 2.73052454e-01 5.95090747e-01 -1.23770130e+00
-1.19333375e+00 7.78358281e-01 -5.58265030e-01 -5.63339628e-02
2.96401024e-01 -9.31972861e-02 1.55945659e+00 -1.00440609e+00
6.88902617e-01 1.35071111e+00 8.54640424e-01 4.71114814e-01
-9.35957372e-01 -4.54427600e-01 8.75453204e-02 -5.84644005e-02
-1.11038637e+00 -1.19187109e-01 9.84817684e-01 -1.57853067e-01
1.57512605e+00 9.89190340e-02 4.09029037e-01 1.46534789e+00
5.07978916e-01 1.04092872e+00 8.99486363e-01 -4.84470904e-01
-8.97880420e-02 -6.61971048e-03 3.02069724e-01 1.07721543e+00
-1.78518638e-01 -4.80853647e-01 -5.86657941e-01 8.63394663e-02
1.45004496e-01 -1.93779990e-01 -3.83754581e-01 2.49754697e-01
-8.43029976e-01 1.27992666e+00 2.67193079e-01 4.28812593e-01
-7.51394257e-02 -1.14603927e-02 9.97326374e-01 5.38079083e-01
1.12051725e+00 2.50749558e-01 -8.23027790e-01 -1.91930279e-01
-5.89801848e-01 -1.88154146e-01 9.31143463e-01 7.84708858e-01
7.12504804e-01 4.95232314e-01 2.03004293e-03 1.35736549e+00
3.82325351e-01 5.78292131e-01 7.65885174e-01 -1.15323998e-01
9.40781534e-01 6.99349105e-01 -8.41510236e-01 -1.09716082e+00
-6.83883846e-01 -2.10420460e-01 -1.09734738e+00 -3.38924825e-01
4.32685018e-02 -5.66198826e-01 -8.75446498e-01 1.70764589e+00
4.73693907e-02 -2.95805663e-01 6.44152522e-01 6.06839836e-01
1.22556198e+00 1.03124130e+00 -1.51662184e-02 -4.89820354e-03
1.51273239e+00 -1.28238750e+00 -7.68380344e-01 -6.15711391e-01
1.24226046e+00 -7.12788165e-01 1.48554349e+00 2.54300654e-01
-5.16290545e-01 -3.36070836e-01 -7.21201658e-01 -3.89677346e-01
-7.95801282e-01 4.29142773e-01 6.34340465e-01 6.62253737e-01
-1.14216924e+00 2.36326858e-01 -5.94566107e-01 -6.81326449e-01
2.12765500e-01 2.48914272e-01 -6.48034871e-01 -1.60195336e-01
-1.35264421e+00 8.35078418e-01 4.22408462e-01 3.58540416e-01
-5.43149292e-01 -4.20058489e-01 -1.46462607e+00 1.13356784e-01
2.31736615e-01 -3.71077567e-01 7.77335346e-01 -9.13219512e-01
-1.62752593e+00 1.06425369e+00 4.93392013e-02 -2.76189983e-01
-6.17981292e-02 -2.49854207e-01 -5.99915922e-01 9.68394503e-02
-6.30308874e-03 3.69964212e-01 7.78643608e-01 -8.81649494e-01
-1.05022639e-01 -5.49826562e-01 -9.11969841e-02 3.37556601e-01
-5.82849145e-01 3.50343227e-01 -2.02904820e-01 -7.06865370e-01
-3.35073858e-01 -9.10106659e-01 5.39726838e-02 -6.39797151e-01
-6.89543903e-01 -5.62160015e-01 1.15321159e+00 -1.09749389e+00
1.06523287e+00 -2.27618647e+00 3.54973763e-01 -1.28993154e-01
2.29680881e-01 3.27088296e-01 -5.71072936e-01 6.93239689e-01
-1.52772531e-01 4.67193633e-01 -2.10197598e-01 -9.10633624e-01
7.66318804e-03 4.71980602e-01 -2.14164466e-01 3.40856522e-01
4.14876461e-01 1.28764224e+00 -6.43637419e-01 -4.09987360e-01
2.30868697e-01 6.27226532e-01 -4.71284896e-01 1.06570274e-01
-4.32311520e-02 1.87854692e-02 -2.63094991e-01 7.00769663e-01
4.71491933e-01 -1.06066629e-01 4.64105874e-01 -3.66782844e-01
1.11994527e-01 2.33847186e-01 -5.32444596e-01 1.61750555e+00
-1.02329409e+00 7.75214314e-01 1.62510291e-01 -1.12170839e+00
1.05269051e+00 3.94398719e-01 9.40602720e-02 -7.85325348e-01
3.44097316e-01 1.77431032e-01 -1.42370760e-01 -4.37721968e-01
7.12102532e-01 -2.77505279e-01 -4.71992493e-01 4.14901018e-01
7.29700565e-01 -3.70249242e-01 2.10161269e-01 4.39952999e-01
1.07428455e+00 -1.70065418e-01 3.13491940e-01 -2.03106076e-01
4.95676786e-01 -1.98800966e-01 2.41136894e-01 2.86591738e-01
-1.86670452e-01 6.27291620e-01 8.71542335e-01 -2.37672240e-01
-7.97420740e-01 -7.67912924e-01 2.14510575e-01 1.33551300e+00
-4.39270139e-01 -7.24872887e-01 -6.14784598e-01 -1.08423603e+00
-1.79132581e-01 8.64582300e-01 -6.49613321e-01 -1.59570575e-01
-7.27699280e-01 -1.11945415e+00 8.70760322e-01 4.10042107e-01
4.49740797e-01 -1.38906109e+00 1.41636312e-01 9.60046947e-02
-3.47706735e-01 -1.53490555e+00 -7.18944252e-01 3.92297268e-01
-5.89607477e-01 -8.67700994e-01 -3.19680631e-01 -1.14418089e+00
4.25033599e-01 -9.20398757e-02 1.14433706e+00 -6.28028959e-02
7.57779256e-02 4.77836996e-01 -8.00502419e-01 -1.75580293e-01
-5.37149549e-01 3.39682698e-01 -1.06525958e-01 6.21645078e-02
2.76620001e-01 -3.02042246e-01 5.83713017e-02 -1.93865046e-01
-9.86989498e-01 2.37576254e-02 3.50584358e-01 1.03462636e+00
3.88826758e-01 -3.56036946e-02 6.62849247e-01 -1.12998676e+00
1.06951261e+00 -3.23436290e-01 -2.39294872e-01 3.60765994e-01
-4.10203218e-01 -9.38903689e-02 1.10903895e+00 -3.62250000e-01
-9.84113812e-01 -3.40131283e-01 -5.93319118e-01 -3.38375241e-01
5.31568229e-02 1.02137661e+00 -1.49539262e-01 1.10433713e-01
3.48693401e-01 1.59141049e-01 -6.45996183e-02 -2.91273862e-01
6.19351685e-01 8.78617764e-01 2.05132619e-01 -2.74593621e-01
4.10839379e-01 1.71722360e-02 -2.92665005e-01 -1.25004387e+00
-1.11121917e+00 -4.28637326e-01 -4.78055537e-01 -7.58396238e-02
1.28416491e+00 -8.74822378e-01 -6.64084017e-01 6.24786556e-01
-1.30209565e+00 -6.06668532e-01 2.29925290e-02 2.80250847e-01
-2.19287708e-01 6.16126776e-01 -1.04320312e+00 -7.10790515e-01
-8.32385778e-01 -1.12192702e+00 1.44057715e+00 -1.25679731e-01
-1.00100406e-01 -1.74340665e+00 1.51832312e-01 7.06330121e-01
2.49257073e-01 9.70339105e-02 1.25772631e+00 -1.10430217e+00
2.69272238e-01 -2.88684033e-02 -3.47556740e-01 7.38842368e-01
1.31683066e-01 -7.87845328e-02 -9.94740188e-01 -4.26638186e-01
-1.69451916e-04 -8.15368831e-01 8.35598052e-01 3.40605110e-01
8.00305963e-01 -3.56061965e-01 1.22551084e-01 6.19300008e-01
1.50598800e+00 -2.70348161e-01 4.92381573e-01 1.04773000e-01
1.32419419e+00 6.78623736e-01 2.79484987e-01 1.03846230e-01
6.59952641e-01 2.80741215e-01 3.90558273e-01 -1.96084559e-01
-1.70718849e-01 -2.16349944e-01 9.06553984e-01 1.67268157e+00
2.89430380e-01 -9.40996706e-01 -1.10076714e+00 6.32993281e-01
-1.78686130e+00 -4.12479103e-01 -3.83631319e-01 1.48150301e+00
6.20922863e-01 -8.60693981e-04 -2.88544118e-01 -6.93367347e-02
4.82258409e-01 5.62287152e-01 -9.05777588e-02 -1.14000392e+00
-5.26317477e-01 3.69546324e-01 2.65439451e-01 5.41838169e-01
-1.20840108e+00 1.45486403e+00 4.76441908e+00 8.31484497e-01
-1.18552566e+00 2.78397441e-01 6.75240457e-01 4.33820128e-01
-3.76546532e-01 -1.96235776e-01 -7.68730462e-01 4.92701195e-02
1.18621325e+00 5.94757088e-02 4.77402657e-01 6.41654074e-01
8.81078392e-02 3.18144619e-01 -7.08742321e-01 1.06250250e+00
6.91801727e-01 -1.08409846e+00 3.59908670e-01 -1.91490114e-01
5.66459060e-01 5.34907818e-01 -3.63687016e-02 8.70262802e-01
4.31407183e-01 -1.08840108e+00 3.88807207e-01 -2.57209510e-01
8.87670338e-01 -7.98556328e-01 1.08806908e+00 1.77085429e-01
-1.28393030e+00 8.50285366e-02 -2.49161929e-01 8.08787793e-02
2.61451274e-01 6.84678018e-01 -8.14898610e-01 9.30654109e-01
6.62983775e-01 1.07094562e+00 -6.53247058e-01 -1.36756180e-02
-5.17084777e-01 1.04527330e+00 -1.71947435e-01 -3.06439519e-01
4.80572343e-01 -5.04385710e-01 4.13798660e-01 1.52277899e+00
8.06458965e-02 -2.96036184e-01 3.40772897e-01 6.37809157e-01
-5.89123011e-01 6.71944857e-01 -7.94659555e-01 -5.37678897e-01
-2.17982173e-01 1.62444317e+00 -7.65952170e-01 -1.47238165e-01
-6.10794127e-01 1.25899279e+00 7.58122444e-01 5.10088086e-01
-7.61255205e-01 -6.09234154e-01 3.62249911e-01 -7.42041051e-01
2.47238576e-01 -4.29314852e-01 -1.09582789e-01 -1.52003002e+00
-1.91138145e-02 -9.87651229e-01 4.64143783e-01 -7.20155001e-01
-1.54684627e+00 9.36293662e-01 -4.79854673e-01 -5.75329185e-01
-9.26901624e-02 -1.06780887e+00 -6.23382092e-01 8.04070175e-01
-1.60536039e+00 -1.83427942e+00 1.84711888e-01 5.83733678e-01
7.72246420e-01 -2.44498685e-01 9.65616584e-01 2.58104324e-01
-9.52399611e-01 6.48924291e-01 -1.77011658e-02 6.37203872e-01
5.65713763e-01 -1.52817488e+00 5.02446473e-01 8.13830674e-01
4.33015853e-01 1.70938849e-01 2.55851060e-01 -6.22471213e-01
-1.56628823e+00 -1.45655870e+00 1.23250163e+00 -2.95743674e-01
1.18145585e+00 -9.89238501e-01 -9.49901938e-01 1.04753232e+00
4.33451980e-01 -6.63461909e-02 7.86754966e-01 4.84708160e-01
-3.63354117e-01 1.84394002e-01 -8.69497657e-01 6.32286310e-01
5.99851429e-01 -9.45727468e-01 -4.88233149e-01 6.78855002e-01
1.02866817e+00 -1.99751571e-01 -1.02472138e+00 1.81073546e-01
2.36575708e-01 -6.10001147e-01 5.45740485e-01 -6.05846763e-01
6.07022703e-01 2.99018949e-01 -3.77650917e-01 -1.72163236e+00
6.13843426e-02 -4.40243930e-01 2.43087679e-01 1.47088850e+00
6.94323540e-01 -8.27504337e-01 4.52606082e-01 -1.92359276e-02
-5.49945056e-01 -8.32559466e-01 -8.82228196e-01 -5.00538349e-01
3.56287569e-01 -7.28193164e-01 1.23265170e-01 1.22344613e+00
1.50928333e-01 1.12699008e+00 -3.15903008e-01 1.17245696e-01
2.14869961e-01 2.20901240e-02 5.11895478e-01 -1.04856181e+00
-1.30248219e-01 -2.11056769e-01 -2.30521709e-01 -7.15727866e-01
9.27680969e-01 -1.40863872e+00 -5.69321699e-02 -1.76358688e+00
3.21780965e-02 8.58938470e-02 -2.98762489e-02 8.26311350e-01
-1.64317444e-01 1.83730781e-01 2.57289916e-01 -2.39644960e-01
-5.85984409e-01 8.37364554e-01 1.13392961e+00 -2.10049644e-01
-1.50188074e-01 -4.50832754e-01 -5.68634808e-01 5.47575116e-01
9.46096480e-01 -3.16831261e-01 -3.10406446e-01 -6.21244371e-01
6.66677654e-01 -2.96833497e-02 8.36320221e-02 -4.03735071e-01
-1.79307222e-01 2.82560050e-01 -5.93330972e-02 -4.20127690e-01
2.46277571e-01 -4.46471244e-01 -7.10114121e-01 9.14278254e-02
-2.14752574e-02 3.96974027e-01 4.38874006e-01 2.80879378e-01
-4.60944086e-01 -1.50359303e-01 5.17986894e-01 2.57579493e-03
-5.11747777e-01 2.09660664e-01 -5.84677517e-01 3.83421123e-01
4.91120517e-01 7.84066543e-02 -5.64787984e-01 -3.99096757e-01
-5.78559577e-01 2.96641976e-01 -1.28985476e-02 6.91927195e-01
8.76745343e-01 -9.26041424e-01 -8.27311873e-01 2.43788660e-01
2.08531916e-01 -1.41332418e-01 2.56491721e-01 8.67133498e-01
-4.87033576e-01 2.35812321e-01 2.58697003e-01 -3.99277031e-01
-1.40204096e+00 3.22906941e-01 2.56566733e-01 -8.04878235e-01
-5.31461239e-01 7.98459053e-01 2.87392586e-02 -1.31235611e+00
-1.85966760e-01 -5.60974479e-01 -4.71014738e-01 2.59706676e-01
-1.17020316e-01 2.82937121e-02 3.07889193e-01 -8.87954772e-01
-3.82982343e-01 5.86005151e-01 -1.92724675e-01 8.86391848e-02
1.49828684e+00 -8.94754082e-02 -4.51035082e-01 5.62312543e-01
1.68266344e+00 2.17238367e-01 -3.91843170e-01 1.49014458e-01
-3.47149558e-02 2.50403076e-01 2.42631868e-01 -7.24958658e-01
-1.17564082e+00 1.10780871e+00 1.42304108e-01 3.09015989e-01
1.06737065e+00 6.87949359e-02 1.05294025e+00 4.04971033e-01
-1.22699142e-01 -1.16152787e+00 2.18125135e-01 1.07937121e+00
9.23431873e-01 -1.37961268e+00 -2.30309859e-01 -4.31924701e-01
-1.10646999e+00 1.26403499e+00 4.62912619e-01 1.53354838e-01
7.32709229e-01 2.15664327e-01 2.36935273e-01 -6.37779415e-01
-7.40479946e-01 -1.16937563e-01 3.29491556e-01 4.45691884e-01
7.59341717e-01 1.94730014e-01 -7.90758133e-02 4.61103410e-01
-5.27005374e-01 -5.32446563e-01 6.80603981e-01 6.95574284e-01
3.92220728e-02 -9.96449590e-01 9.73602384e-02 5.43809712e-01
-7.12517977e-01 -6.16893589e-01 -5.84162951e-01 7.73220301e-01
-3.64789844e-01 1.17671573e+00 -4.60074190e-03 -3.94404233e-01
2.92853951e-01 2.59241521e-01 1.41895071e-01 -8.31758261e-01
-9.65956807e-01 -7.02927709e-02 5.64486980e-01 -2.77568132e-01
-3.42874706e-01 -3.40490341e-01 -1.30524671e+00 -8.66224989e-02
-5.50620258e-01 1.60416305e-01 7.34159946e-01 1.05619240e+00
1.93895757e-01 7.26095915e-01 5.15762568e-01 -5.55091202e-01
-2.05036759e-01 -1.13502693e+00 -8.02599609e-01 5.50465882e-01
2.46808842e-01 -1.23109244e-01 -5.78967214e-01 -2.25637078e-01] | [10.78431224822998, 9.555343627929688] |
879c5240-8904-4867-b653-7d9576d1da3c | the-kriston-ai-system-for-the-voxceleb | 2209.11433 | null | https://arxiv.org/abs/2209.11433v1 | https://arxiv.org/pdf/2209.11433v1.pdf | The Kriston AI System for the VoxCeleb Speaker Recognition Challenge 2022 | This technical report describes our system for track 1, 2 and 4 of the VoxCeleb Speaker Recognition Challenge 2022 (VoxSRC-22). By combining several ResNet variants, our submission for track 1 attained a minDCF of 0:090 with EER 1:401%. By further incorporating three fine-tuned pre-trained models, our submission for track 2 achieved a minDCF of 0:072 with EER 1:119%. For track 4, our system consisted of voice activity detection (VAD), speaker embedding extraction, agglomerative hierarchical clustering (AHC) followed by a re-clustering step based on a Bayesian hidden Markov model and overlapped speech detection and handling. Our submission for track 4 achieved a diarisation error rate (DER) of 4.86%. The submissions all ranked the 2nd places for the corresponding tracks. | ['Haizhou Li', 'Ximin Li', 'Zhijian Ye', 'Guoqiang Hong', 'Qutang Cai'] | 2022-09-23 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [-2.56358415e-01 4.11242068e-01 5.96687123e-02 -2.98531890e-01
-1.26354909e+00 -3.23312283e-01 7.73223877e-01 -5.00369072e-02
-5.13745904e-01 2.61287719e-01 5.43815613e-01 -2.39405766e-01
3.25389892e-01 3.75852846e-02 -1.23461351e-01 -6.43021703e-01
-2.62914687e-01 1.48993477e-01 1.99507669e-01 1.23263158e-01
-1.52454913e-01 4.43543375e-01 -1.49315858e+00 1.59530416e-01
3.64746094e-01 9.98014331e-01 -2.90062129e-01 1.20456731e+00
-5.21055830e-04 6.86640501e-01 -9.13546443e-01 -2.19656199e-01
-7.30808079e-02 -3.22825342e-01 -5.16435683e-01 -1.13938905e-01
7.10280001e-01 -2.66362298e-02 -4.66168970e-01 7.98198760e-01
1.08361864e+00 4.25989777e-01 5.64634562e-01 -1.34149992e+00
-1.08034462e-01 9.46293056e-01 -4.16577578e-01 5.76477766e-01
2.48407066e-01 -6.19566813e-02 1.03436387e+00 -1.21891153e+00
2.09831655e-01 1.34667206e+00 8.12266707e-01 9.10297811e-01
-1.28451824e+00 -8.85301292e-01 -2.66742911e-02 3.88806283e-01
-1.71509027e+00 -1.31735647e+00 7.18913615e-01 -3.39562893e-01
1.32325637e+00 4.69787627e-01 4.31991667e-01 1.35603130e+00
-4.38036740e-01 9.42264497e-01 9.63278890e-01 -3.38993758e-01
4.03743953e-01 4.12075818e-01 2.84338117e-01 2.65742838e-01
-4.56094235e-01 1.24106050e-01 -7.39951134e-01 -1.80478007e-01
3.07060689e-01 -7.71128356e-01 -8.65189210e-02 5.08365571e-01
-1.11933303e+00 7.12710738e-01 2.77700108e-02 4.40346181e-01
-5.40459037e-01 6.49340823e-02 5.78038335e-01 2.30698019e-01
3.78537834e-01 7.05001354e-02 -2.06002891e-01 -3.79871339e-01
-1.57529020e+00 -8.01695418e-03 8.58084619e-01 1.00470865e+00
5.23714488e-03 7.52694726e-01 -4.05658692e-01 1.26826715e+00
7.03632832e-01 4.83570486e-01 5.37494779e-01 -8.79429460e-01
2.09892169e-01 -1.95587203e-01 -2.60022849e-01 -4.85208184e-01
-2.90356904e-01 -5.97384453e-01 -8.54385555e-01 -1.94885910e-01
1.26949340e-01 -3.32367718e-01 -1.03188896e+00 1.63903522e+00
2.96317190e-01 2.83383340e-01 1.75930470e-01 6.17773473e-01
1.33086288e+00 9.92259920e-01 2.49339104e-01 -3.20750922e-01
1.37359250e+00 -1.15200531e+00 -9.41428602e-01 2.54737049e-01
8.22508633e-02 -1.00015247e+00 6.29309833e-01 5.85626245e-01
-1.24222779e+00 -7.48024404e-01 -1.03920853e+00 2.88324475e-01
1.38003483e-01 8.65943544e-03 -9.90139842e-02 1.08487833e+00
-1.49807942e+00 1.60106599e-01 -8.31343055e-01 -4.12343353e-01
1.42681807e-01 3.64224195e-01 -1.32301837e-01 5.33492625e-01
-1.05676639e+00 5.77008665e-01 -1.11885019e-01 8.95114914e-02
-1.19559968e+00 -7.52858579e-01 -6.78098202e-01 -7.48007223e-02
-1.69714138e-01 -1.64368227e-01 1.37898088e+00 -4.66041952e-01
-2.01678038e+00 8.13645780e-01 -4.10042495e-01 -6.87653959e-01
6.49085224e-01 -2.36607894e-01 -1.10509241e+00 9.54453796e-02
-1.23535529e-01 7.45812058e-01 9.20533597e-01 -8.77215564e-01
-6.57221496e-01 -1.10222846e-01 -6.18958950e-01 8.73176530e-02
-2.69261330e-01 5.88810503e-01 -5.93257070e-01 -6.95321202e-01
8.62210393e-02 -7.57654428e-01 9.35610756e-02 -7.06290960e-01
-9.33185816e-01 -5.39211094e-01 6.37747765e-01 -1.16504669e+00
1.39313078e+00 -2.54293513e+00 2.03959960e-02 1.96231022e-01
4.15202320e-01 3.86833280e-01 -9.19316709e-02 2.36392885e-01
-2.34541893e-01 1.56277090e-01 9.73001868e-02 -8.88927698e-01
3.36008549e-01 -4.61865157e-01 -2.08243564e-01 4.72165436e-01
-2.51175854e-02 3.92521441e-01 -5.33649206e-01 -3.99677873e-01
3.94315332e-01 9.60520029e-01 -4.97477889e-01 3.79770815e-01
3.71605307e-01 5.04577644e-02 2.53139019e-01 4.83741462e-01
6.47771835e-01 3.55957925e-01 -1.12586934e-02 -2.11871967e-01
-2.95795768e-01 6.78096235e-01 -1.41403472e+00 1.09991539e+00
-2.66461462e-01 9.99076426e-01 4.11426038e-01 -4.15562570e-01
1.03791881e+00 8.62982035e-01 4.94405448e-01 -2.65756220e-01
8.75584781e-02 6.88386932e-02 -8.54686275e-02 -3.17270964e-01
3.12100023e-01 -1.44327894e-01 -1.53703149e-02 1.29935995e-01
3.81593198e-01 9.46677551e-02 -3.68762583e-01 2.39880025e-01
1.08669555e+00 -5.76624334e-01 -1.35405199e-03 -2.94513106e-01
5.35652339e-01 -5.06264508e-01 5.69990516e-01 8.17314982e-01
-8.72189820e-01 7.12535501e-01 8.51719528e-02 -2.27165725e-02
-9.34089839e-01 -1.40444148e+00 -3.20050776e-01 9.18247461e-01
-5.05075991e-01 -4.93179500e-01 -1.02879643e+00 -4.57469106e-01
-3.56383771e-01 8.21733415e-01 -3.40565801e-01 2.39877805e-01
-6.09084249e-01 -5.53308666e-01 1.26166523e+00 3.84521902e-01
5.84221661e-01 -1.04812431e+00 4.15185168e-02 2.77621448e-01
-2.93457568e-01 -1.27552986e+00 -7.86274612e-01 2.98365027e-01
-5.08987010e-01 -4.36344415e-01 -8.37854147e-01 -7.77110040e-01
-3.00741401e-02 -6.82461113e-02 7.75943816e-01 -2.30500624e-01
-8.59287679e-02 2.79595345e-01 -1.70833677e-01 -3.22475523e-01
-5.80122709e-01 3.35368186e-01 4.97886389e-01 1.42885029e-01
5.52803159e-01 -4.24100101e-01 -4.40659970e-01 2.45153368e-01
-2.86524653e-01 -3.04663420e-01 1.42096639e-01 7.50943303e-01
2.51670003e-01 -1.34505466e-01 6.61916196e-01 -3.74579370e-01
4.14958954e-01 -1.63967267e-01 -5.58439791e-01 -1.22476369e-01
-4.45803195e-01 -3.74056876e-01 2.89038807e-01 -4.67913300e-01
-9.06476915e-01 4.82142493e-02 -7.53882349e-01 -5.17504811e-01
-5.18379986e-01 -2.60575801e-01 -3.50081980e-01 3.87903988e-01
5.98877192e-01 2.26391256e-01 -2.57229269e-01 -6.54030621e-01
3.16245556e-01 1.25882387e+00 6.05098665e-01 2.56526358e-02
6.46735787e-01 -5.35657257e-02 -7.65995443e-01 -1.57589471e+00
-4.16614920e-01 -7.11416006e-01 -4.60917056e-01 -2.36286402e-01
1.06512105e+00 -1.34208798e+00 -8.69308650e-01 6.10607326e-01
-1.12372983e+00 -1.48055926e-01 -3.77438486e-01 8.04804683e-01
-1.65296420e-01 2.61323512e-01 -8.22789013e-01 -1.15680456e+00
-6.44131720e-01 -1.11474454e+00 9.00329053e-01 -5.82761355e-02
-7.11001337e-01 -8.09205949e-01 3.48607123e-01 7.75686502e-01
6.14860177e-01 -1.86351866e-01 3.67712617e-01 -1.11679637e+00
-4.53688353e-02 -3.83528620e-02 1.20369509e-01 6.42906725e-01
-1.21234342e-01 1.31513238e-01 -1.82308626e+00 -3.55641603e-01
-3.44251469e-02 1.76201239e-01 8.82893562e-01 4.72902417e-01
8.36115837e-01 -3.57522577e-01 -6.14414588e-02 4.01022851e-01
7.85240948e-01 3.07993442e-01 6.21154189e-01 -9.40408707e-02
5.29009819e-01 6.06106400e-01 -1.44239590e-01 2.88289517e-01
4.09812540e-01 1.10102451e+00 -9.03832465e-02 1.41880199e-01
-7.22125232e-01 -1.76229596e-01 7.83482611e-01 1.44122362e+00
1.86092079e-01 -2.65768588e-01 -8.82100403e-01 9.50524151e-01
-1.27467453e+00 -1.26580560e+00 -3.16842675e-01 2.17557740e+00
7.82690167e-01 5.03799319e-02 7.75923133e-01 5.88958085e-01
1.04028010e+00 2.80120343e-01 -2.61262864e-01 -5.94922841e-01
-6.20491132e-02 2.14973599e-01 1.69273555e-01 8.88774931e-01
-1.22147095e+00 1.04094207e+00 7.15118265e+00 9.24753308e-01
-1.05281961e+00 3.12262237e-01 4.21545923e-01 -3.69395107e-01
4.42357026e-02 -5.85784554e-01 -1.09457612e+00 4.65848744e-01
1.72428966e+00 -1.44484174e-02 3.55383724e-01 7.37798393e-01
2.13960841e-01 2.59009242e-01 -9.17170405e-01 1.28233016e+00
3.41722369e-01 -9.59549904e-01 -3.53076965e-01 2.20879644e-01
4.96746063e-01 4.98436689e-01 2.39849836e-02 5.24379373e-01
4.38560992e-01 -1.06176722e+00 8.35264921e-01 1.72595575e-01
7.02129126e-01 -7.76106954e-01 6.58796608e-01 -1.71675850e-02
-1.26700604e+00 1.15973748e-01 -4.02021743e-02 5.06587744e-01
2.33277425e-01 6.52689815e-01 -1.24056256e+00 2.12354675e-01
8.99143159e-01 3.39063019e-01 -1.07453175e-01 1.15922213e+00
-2.81864345e-01 1.26284468e+00 -5.03236353e-01 -1.34027883e-01
-1.66428804e-01 5.48164725e-01 1.08067834e+00 1.78155112e+00
3.05020250e-02 -1.44328177e-01 -2.11879924e-01 3.21336776e-01
-2.74664998e-01 1.42715648e-01 -5.10825440e-02 1.81818411e-01
1.03792131e+00 1.15346742e+00 -3.23430747e-01 -3.83074224e-01
1.12008899e-01 9.09288287e-01 -7.92263150e-02 4.10138190e-01
-9.30816472e-01 -7.22425938e-01 8.66858244e-01 -5.21777272e-02
4.06318188e-01 2.36858428e-02 -7.28186220e-02 -1.12499130e+00
-1.44444555e-01 -8.42369318e-01 3.02941829e-01 -2.30186448e-01
-1.21422791e+00 1.17851901e+00 -1.03465654e-01 -8.50134611e-01
-3.47314149e-01 -7.93882534e-02 -7.35130966e-01 9.34773088e-01
-1.22077942e+00 -8.42133820e-01 2.83305198e-02 5.01009405e-01
7.20978796e-01 -6.14836276e-01 9.63122070e-01 7.24417567e-01
-9.91311491e-01 1.30327618e+00 -1.17056118e-02 2.92969435e-01
5.98172009e-01 -1.30619454e+00 7.86521852e-01 8.04662406e-01
3.19841206e-01 4.11956787e-01 8.65763187e-01 -1.73103228e-01
-1.08409023e+00 -1.18346918e+00 1.44312000e+00 -4.78443146e-01
5.83848596e-01 -6.74028337e-01 -7.33480752e-01 5.35283566e-01
5.20202100e-01 -1.87445879e-01 8.87616456e-01 3.40282470e-01
-5.28433383e-01 -1.92693308e-01 -1.24223316e+00 2.77744353e-01
6.96280599e-01 -7.94460893e-01 -7.13449240e-01 -9.61123593e-03
7.86285460e-01 -9.78628993e-02 -1.01796031e+00 1.89865813e-01
5.56656182e-01 -8.13349843e-01 8.70981693e-01 -3.91067177e-01
-3.59485835e-01 -2.49482140e-01 -5.38271666e-01 -1.11975849e+00
-3.06899607e-01 -1.00386214e+00 -2.22912341e-01 1.64790213e+00
5.67101359e-01 -3.61049056e-01 6.76289380e-01 1.30116478e-01
-2.55926400e-01 -2.19395995e-01 -1.40232086e+00 -9.67806518e-01
-7.23393634e-02 -8.39647830e-01 2.67338693e-01 8.08850765e-01
-2.05312595e-02 4.69137311e-01 -3.92780304e-01 3.46329570e-01
8.78117859e-01 -4.91783410e-01 7.23968148e-01 -1.17768013e+00
-2.24605814e-01 -5.85756063e-01 -4.93268669e-01 -9.66924906e-01
4.57466505e-02 -7.47241199e-01 2.28570193e-01 -1.26108623e+00
4.36535338e-03 -1.86487497e-03 -4.85043257e-01 3.47216249e-01
9.20099169e-02 3.31918806e-01 3.58242214e-01 1.61430836e-01
-6.10652268e-01 5.84790230e-01 2.51673937e-01 -1.86372459e-01
-4.96382177e-01 9.25270095e-02 -4.68902916e-01 5.84301949e-01
7.36564279e-01 -5.34576774e-01 7.56273717e-02 -1.88364610e-02
-6.38706446e-01 -2.70572454e-02 1.60734251e-01 -1.14396453e+00
4.34085518e-01 4.67111409e-01 1.71416298e-01 -7.61227429e-01
7.32166052e-01 -3.93218726e-01 9.58678052e-02 3.52489740e-01
-3.51451039e-01 -2.79781491e-01 5.00267565e-01 1.26679614e-01
-3.36734653e-01 2.04505712e-01 9.42697465e-01 4.48392183e-01
-2.92188913e-01 -5.96571248e-03 -9.63297367e-01 1.26022294e-01
7.38538086e-01 -1.54888794e-01 4.03426178e-02 -4.29027528e-01
-1.28105128e+00 7.34163299e-02 -3.17029923e-01 6.15218878e-01
6.07797027e-01 -1.24776411e+00 -1.04968464e+00 3.73288304e-01
-9.60149541e-02 -5.09476066e-01 4.65864718e-01 7.53247142e-01
-6.47306144e-02 4.50593352e-01 3.37803394e-01 -7.33706772e-01
-1.77081847e+00 -2.33567595e-01 4.02752697e-01 -3.17205722e-03
-4.64928508e-01 1.22776067e+00 -2.86101907e-01 -4.66122359e-01
8.76641512e-01 1.04911849e-01 -2.52659947e-01 3.42156112e-01
8.41447949e-01 6.84068739e-01 3.65072131e-01 -1.06377220e+00
-7.51519203e-01 1.48175940e-01 -1.90987393e-01 -7.30954170e-01
1.29419112e+00 -8.05205777e-02 3.32998246e-01 6.04089200e-01
1.24140120e+00 2.63003677e-01 -9.66278851e-01 -1.48159176e-01
7.16821179e-02 3.68088633e-02 5.73121548e-01 -8.50329101e-01
-9.30571079e-01 1.10438144e+00 9.10789251e-01 1.38919145e-01
9.79155838e-01 5.10012321e-02 8.62818718e-01 7.29300454e-02
-1.51497602e-01 -1.05253947e+00 -2.07367450e-01 6.76708400e-01
8.12078118e-01 -9.34306562e-01 -4.43758398e-01 -1.36693344e-01
-5.69368303e-01 8.12455058e-01 3.72177750e-01 1.76724181e-01
8.87139976e-01 3.83989930e-01 3.85734826e-01 2.89065316e-02
-8.57817590e-01 7.09175915e-02 4.30431664e-01 4.95942861e-01
5.94946444e-01 2.69714236e-01 3.39414120e-01 6.02372408e-01
-6.22998893e-01 -5.74889362e-01 1.11299917e-01 3.50500405e-01
-4.55207795e-01 -7.24474072e-01 -4.48201805e-01 -3.47134955e-02
-5.70038021e-01 -2.37689763e-01 -4.03797239e-01 3.82794052e-01
-1.88341856e-01 1.39320850e+00 1.87911853e-01 -7.32650399e-01
4.97456312e-01 2.90241867e-01 1.45984337e-01 -5.81464827e-01
-1.05746245e+00 7.58921981e-01 3.35797489e-01 -4.32629138e-01
-1.76520526e-01 -9.90252018e-01 -1.08237040e+00 -4.32847083e-01
-2.60944337e-01 3.97390842e-01 1.04063666e+00 6.06410325e-01
4.15216327e-01 6.92661703e-01 8.43277991e-01 -4.85653847e-01
-4.24076349e-01 -1.32920146e+00 -5.77457368e-01 -2.09185392e-01
8.08133960e-01 -1.74641579e-01 -7.50109196e-01 8.87983665e-03] | [14.443395614624023, 6.025880813598633] |
5a4fa934-d0ac-479b-bc15-a6ca3a24299a | introduction-to-core-sets-an-updated-survey | 2011.09384 | null | https://arxiv.org/abs/2011.09384v1 | https://arxiv.org/pdf/2011.09384v1.pdf | Introduction to Core-sets: an Updated Survey | In optimization or machine learning problems we are given a set of items, usually points in some metric space, and the goal is to minimize or maximize an objective function over some space of candidate solutions. For example, in clustering problems, the input is a set of points in some metric space, and a common goal is to compute a set of centers in some other space (points, lines) that will minimize the sum of distances to these points. In database queries, we may need to compute such a some for a specific query set of $k$ centers. However, traditional algorithms cannot handle modern systems that require parallel real-time computations of infinite distributed streams from sensors such as GPS, audio or video that arrive to a cloud, or networks of weaker devices such as smartphones or robots. Core-set is a "small data" summarization of the input "big data", where every possible query has approximately the same answer on both data sets. Generic techniques enable efficient coreset \changed{maintenance} of streaming, distributed and dynamic data. Traditional algorithms can then be applied on these coresets to maintain the approximated optimal solutions. The challenge is to design coresets with provable tradeoff between their size and approximation error. This survey summarizes such constructions in a retrospective way, that aims to unified and simplify the state-of-the-art. | ['Dan Feldman'] | 2020-11-18 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [-1.26188472e-01 -2.49088630e-01 -2.48598486e-01 -4.48021829e-01
-7.94200420e-01 -6.24836862e-01 -3.64645422e-01 8.99642766e-01
-5.20583570e-01 3.50312233e-01 -2.63467789e-01 6.93463087e-02
-7.90043831e-01 -1.18025076e+00 -9.83303428e-01 -6.08223915e-01
-5.96268177e-01 1.22428060e+00 2.07422987e-01 -1.58933043e-01
3.88770729e-01 4.02949154e-01 -1.90036118e+00 2.95565516e-01
6.08734608e-01 1.52001882e+00 4.37217027e-01 8.83389950e-01
-2.11590260e-01 2.05662176e-01 -7.50193417e-01 -3.43446672e-01
4.62802082e-01 5.12069464e-02 -9.06191409e-01 1.66899011e-01
-3.30992341e-02 -6.55003041e-02 -3.12222354e-02 1.11332428e+00
4.68871236e-01 5.34778424e-02 2.80046970e-01 -1.60008013e+00
-2.89604902e-01 7.77436972e-01 -7.84912407e-01 9.82376933e-02
5.32110333e-01 -2.24969760e-01 1.14467704e+00 -8.06353986e-01
3.88847351e-01 1.04922628e+00 3.71531576e-01 1.33139908e-01
-1.23379874e+00 -1.25050008e-01 -1.03644095e-01 4.90758330e-01
-1.64281976e+00 -3.59706193e-01 2.23090664e-01 1.91069245e-01
9.24994230e-01 8.01887989e-01 6.00133538e-01 -1.08516298e-01
-1.39657214e-01 6.80561423e-01 2.96947360e-02 -2.91815877e-01
9.43651199e-01 1.60027906e-01 1.79011058e-02 2.17151508e-01
5.64268351e-01 -4.55561638e-01 -8.16892445e-01 -7.11827457e-01
3.91301811e-02 5.21640658e-01 -2.63404045e-02 -4.19484913e-01
-1.24168432e+00 9.75905359e-01 3.21064353e-01 9.90428776e-02
-4.17643815e-01 2.24339798e-01 4.42691743e-01 5.86187303e-01
1.61858678e-01 2.70858258e-01 -6.36497259e-01 -1.71632677e-01
-9.67804193e-01 5.85332632e-01 1.24269760e+00 1.49301445e+00
9.87826943e-01 -6.69951260e-01 1.85308993e-01 5.65297604e-01
5.48308231e-02 7.37965107e-01 1.42841101e-01 -1.16920853e+00
7.40486383e-01 6.87968850e-01 3.23273033e-01 -1.16043007e+00
-4.04094875e-01 -1.98430978e-02 -9.02661383e-01 -3.02863926e-01
2.37664759e-01 -1.50129095e-01 -3.14254649e-02 1.40188360e+00
7.97391653e-01 -1.32589191e-01 -1.19272456e-01 1.17710686e+00
4.51250449e-02 1.03218031e+00 -7.24694848e-01 -5.72392702e-01
1.23416436e+00 -7.15839565e-01 -2.58315474e-01 5.71558513e-02
8.32753003e-01 -7.80465066e-01 8.21597397e-01 7.66786337e-01
-1.76742327e+00 -1.30017847e-01 -9.17468548e-01 2.51393672e-02
-3.53696048e-01 -4.62036073e-01 4.33215946e-01 5.51215112e-01
-1.24152124e+00 5.19436121e-01 -7.01791286e-01 -2.49887824e-01
1.47341013e-01 6.85820639e-01 7.76560903e-02 -3.90123606e-01
-6.37536108e-01 2.55009383e-01 2.59731472e-01 -1.66989386e-01
-5.58842063e-01 -8.13890457e-01 -3.48381519e-01 1.91475809e-01
4.58408296e-01 -8.37252736e-01 1.20501804e+00 -8.38058472e-01
-8.10525119e-01 7.43705809e-01 -2.74589986e-01 -5.36019146e-01
1.07793152e-01 -7.78341293e-02 -5.12478985e-02 3.69877577e-01
1.07636042e-01 3.60549420e-01 5.50399184e-01 -6.78224087e-01
-1.10010481e+00 -8.12644482e-01 1.43421220e-03 2.45994568e-01
-7.16709971e-01 2.01655984e-01 -7.24223077e-01 -5.04191555e-02
2.22868696e-01 -9.10654902e-01 -7.31384754e-01 1.24203078e-01
-2.85310835e-01 -4.59968209e-01 7.66909897e-01 1.14474535e-01
1.37703454e+00 -2.05380297e+00 2.40167931e-01 6.99118614e-01
2.81000584e-01 -3.42084527e-01 2.00010259e-02 7.42912233e-01
3.09396416e-01 -1.13544073e-02 -1.67094588e-01 -4.28616583e-01
1.26052245e-01 4.77381319e-01 -2.85824656e-01 7.81072199e-01
-3.58646333e-01 2.68271923e-01 -8.21186721e-01 -4.99860436e-01
-1.19122677e-01 -1.94151461e-01 -8.99404585e-01 2.59686023e-01
-6.24612331e-01 -4.48031247e-01 -3.69600922e-01 4.82331663e-01
8.78909111e-01 -4.73358929e-01 6.15888787e-03 3.42577219e-01
4.58998093e-03 1.09588891e-01 -1.93210530e+00 1.83930790e+00
-1.58803701e-01 1.45657569e-01 5.75624585e-01 -1.41112494e+00
8.47247243e-01 1.65805906e-01 1.18639731e+00 -3.24374765e-01
-2.12210879e-01 4.62510169e-01 -4.11812931e-01 -4.05069292e-01
7.19414651e-01 2.94051796e-01 -3.47637028e-01 8.53364706e-01
-5.55448890e-01 -1.56374946e-01 4.16426539e-01 4.44083095e-01
1.39774692e+00 -1.04939568e+00 -2.60113180e-01 -2.07167715e-01
2.81259030e-01 3.59274834e-01 4.57176834e-01 7.53098607e-01
2.33368620e-01 6.57711387e-01 4.58189905e-01 -5.20035803e-01
-1.35107470e+00 -9.84469891e-01 -3.13259102e-02 1.31368220e+00
3.57312322e-01 -7.66355276e-01 -8.08154285e-01 -6.34057820e-03
3.41867208e-01 2.15898126e-01 -7.66161606e-02 1.40508264e-01
-4.68773782e-01 -5.87147117e-01 7.47454390e-02 1.99725702e-01
-5.84143661e-02 -5.70609152e-01 -1.05475557e+00 5.08204639e-01
-1.92790076e-01 -8.43423069e-01 -8.31667185e-01 1.48907021e-01
-1.13702917e+00 -1.04265165e+00 -3.11078966e-01 -8.35526645e-01
6.55228496e-01 7.65137732e-01 1.29695153e+00 2.60481298e-01
-3.48513722e-01 4.51603115e-01 -3.39401960e-01 -6.09311461e-01
3.29829872e-01 4.08018567e-02 2.88393945e-01 1.02394456e-02
6.96864009e-01 -5.49958169e-01 -9.06363189e-01 5.44460833e-01
-1.21082318e+00 -4.82042432e-01 1.35354221e-01 2.83655733e-01
1.08335173e+00 5.04361749e-01 5.27594090e-01 -5.13443291e-01
6.54179752e-01 -1.03555608e+00 -7.86205888e-01 3.05556059e-01
-5.82200885e-01 -1.83557514e-02 7.88118839e-01 -2.44873554e-01
-1.44110480e-02 1.68976143e-01 2.67501861e-01 -5.45052052e-01
1.40424132e-01 4.36151415e-01 -2.46482790e-01 3.60897064e-01
6.11097574e-01 4.81566489e-02 6.58409894e-02 -3.41731519e-01
5.41960537e-01 8.42318475e-01 4.79621381e-01 -9.20523584e-01
2.43921846e-01 7.68771112e-01 1.24556623e-01 -8.32758307e-01
-4.92586136e-01 -1.01561975e+00 -2.98933864e-01 7.85048231e-02
1.93140894e-01 -5.47435880e-01 -1.21541905e+00 -2.81987824e-02
-1.14209616e+00 4.84960489e-02 -7.92371571e-01 1.85866013e-01
-9.20121253e-01 1.37451962e-01 -3.33624333e-01 -7.44118214e-01
-6.22149050e-01 -8.51342678e-01 1.04418337e+00 1.38972193e-01
-1.40401963e-02 -6.08078003e-01 1.12525895e-01 4.58305441e-02
4.47104275e-01 1.48828894e-01 5.92482865e-01 -7.76093423e-01
-9.18106556e-01 -5.42457342e-01 2.29302398e-03 -5.17538488e-02
-6.01538643e-02 -2.88080629e-02 -2.76489258e-01 -5.33897519e-01
2.91957855e-01 -2.69499868e-01 2.60042340e-01 5.43058157e-01
1.76729655e+00 -9.01004255e-01 -5.92523217e-01 6.18848979e-01
1.53636098e+00 -6.21019378e-02 2.92431444e-01 1.73359867e-02
5.33502698e-02 6.58869207e-01 9.29191351e-01 1.33790755e+00
4.75096643e-01 4.98929709e-01 6.36988223e-01 2.83902258e-01
9.20788169e-01 2.06478283e-01 1.20813370e-01 8.27838421e-01
4.12452132e-01 -3.54664564e-01 -7.00825930e-01 8.85033369e-01
-2.20325327e+00 -9.54590201e-01 -1.32901609e-01 2.57552695e+00
7.18036532e-01 7.69810081e-02 6.16433740e-01 4.99122292e-01
7.53755569e-01 -4.10371661e-01 -8.72494996e-01 -6.15727365e-01
-1.04011677e-01 1.28556088e-01 8.20555449e-01 1.88870225e-02
-6.43272102e-01 1.92657113e-01 6.16574383e+00 7.54457176e-01
-8.86307359e-01 -4.93195690e-02 7.42793918e-01 -9.50376272e-01
-4.43802804e-01 -1.62569836e-01 -7.63178229e-01 6.31484151e-01
1.51578355e+00 -7.28036106e-01 8.30882251e-01 1.07167506e+00
3.04540455e-01 -3.20588887e-01 -1.65956736e+00 1.34181237e+00
-1.66069001e-01 -1.62247348e+00 -4.64492261e-01 3.27518463e-01
7.76371479e-01 2.15056166e-01 1.10210851e-01 -2.30368197e-01
2.99743086e-01 -9.14574206e-01 5.90296566e-01 4.78526890e-01
5.48189700e-01 -1.10211408e+00 3.60076904e-01 7.70937324e-01
-1.20938313e+00 -3.48695606e-01 -8.09431970e-01 -7.98793510e-02
1.97073907e-01 9.08460498e-01 -5.97587168e-01 3.22511464e-01
1.29159784e+00 2.04218894e-01 9.85245965e-03 1.41337717e+00
8.25465918e-01 2.57377058e-01 -1.28888857e+00 -5.57813048e-01
1.68019727e-01 -4.22715873e-01 3.08316678e-01 1.08931017e+00
6.13880992e-01 3.76819849e-01 3.03242326e-01 5.40846527e-01
-2.99547911e-01 3.72588843e-01 -6.75354302e-01 3.30124140e-01
1.01482117e+00 1.22716522e+00 -6.20617807e-01 -5.24585307e-01
-4.56202209e-01 5.48434198e-01 2.30118215e-01 4.57170419e-02
-7.14954138e-01 -6.89023077e-01 9.89851475e-01 5.65869212e-01
2.19430313e-01 -3.06425363e-01 -5.59559226e-01 -7.08508790e-01
4.15362567e-01 -8.13621163e-01 9.88946557e-01 -4.67238963e-01
-1.33559954e+00 1.13093734e-01 -3.32541615e-02 -1.04777575e+00
-2.10734203e-01 -1.43888131e-01 -5.85813284e-01 4.75280315e-01
-7.93061018e-01 -1.56085581e-01 -1.55351281e-01 1.13402259e+00
4.73942995e-01 1.15775362e-01 5.27653396e-01 4.05773610e-01
-2.00121969e-01 5.05402029e-01 5.74171066e-01 -5.72105885e-01
3.67244452e-01 -1.28557050e+00 -1.03111817e-02 4.31507558e-01
1.14475012e-01 3.37862253e-01 7.77879715e-01 -1.72960415e-01
-2.13477349e+00 -9.79770660e-01 8.66063297e-01 -3.40245754e-01
5.05861402e-01 -2.87758589e-01 -6.17827475e-01 3.60806078e-01
-1.21037088e-01 3.56439739e-01 7.28066146e-01 -3.29304934e-02
7.26770386e-02 -7.13123441e-01 -1.37826443e+00 2.55981952e-01
8.79213631e-01 -3.46595794e-02 -1.43126264e-01 8.19036722e-01
9.82554018e-01 -2.65902400e-01 -1.03480697e+00 -6.78818077e-02
1.27427787e-01 -1.03524494e+00 9.83041227e-01 -7.78641105e-01
-1.68494843e-02 -3.45641345e-01 -6.02870286e-01 -9.34579253e-01
-4.88747936e-03 -9.48381305e-01 -4.27203149e-01 8.43061209e-01
3.67019892e-01 -2.81575978e-01 1.28170884e+00 1.02988434e+00
1.06528345e-02 -1.07652950e+00 -1.17881322e+00 -6.78632855e-01
-2.08284423e-01 -6.43372536e-01 1.08344066e+00 4.76937294e-01
5.27492642e-01 2.09874004e-01 5.62676322e-03 3.46566528e-01
9.46760654e-01 4.87218529e-01 9.36916828e-01 -1.17853463e+00
-1.63034603e-01 -1.75977930e-01 -2.13023856e-01 -1.30339360e+00
-5.58928907e-01 -7.16435254e-01 -1.25931859e-01 -1.35415649e+00
1.14395365e-01 -1.03408694e+00 9.73121971e-02 -1.23720199e-01
4.75185901e-01 -3.91296178e-01 1.24203078e-01 4.06767339e-01
-1.40889597e+00 2.65359819e-01 5.86320460e-01 1.02067823e-02
-4.26786959e-01 5.34307361e-01 -5.73037744e-01 4.69030291e-01
7.37731516e-01 -7.35023975e-01 -3.80953193e-01 -7.45805323e-01
7.68373609e-01 5.56312084e-01 -1.71959341e-01 -9.02167082e-01
1.13417840e+00 -3.65486205e-01 -2.54587997e-02 -1.05819416e+00
2.57732004e-01 -1.16409779e+00 2.06869036e-01 4.80334461e-01
-2.54919320e-01 7.18344629e-01 -3.05355757e-01 6.86665058e-01
-2.20926046e-01 -3.59302580e-01 5.08646131e-01 -1.00724764e-01
-5.53351864e-02 7.49012232e-01 -1.07942075e-01 3.57104331e-01
1.44498539e+00 -2.60160089e-01 -1.70419738e-01 -4.26936626e-01
-4.23636228e-01 8.91068518e-01 5.94878316e-01 1.41437709e-01
8.48870397e-01 -1.21741486e+00 -8.07114959e-01 -2.16381643e-02
-1.06353406e-02 7.92997897e-01 8.13851357e-02 7.74665833e-01
-7.00358093e-01 4.60554838e-01 2.45719910e-01 -9.13967788e-01
-1.01684451e+00 1.20332003e+00 4.72372957e-02 -2.97709256e-02
-2.18909159e-01 8.45059693e-01 -3.49014014e-01 -1.01043507e-01
5.81947863e-01 -3.95587146e-01 5.31857610e-01 -4.25081439e-02
8.11089993e-01 8.65286708e-01 3.50573629e-01 -7.33573288e-02
-4.18488562e-01 2.63159811e-01 9.61944982e-02 -1.88966021e-01
1.56334484e+00 -4.44164127e-01 -5.54239750e-01 2.80429959e-01
1.41197741e+00 -1.81801945e-01 -6.89617336e-01 -5.40315926e-01
3.66190523e-01 -6.17960334e-01 -5.16781747e-01 2.23668888e-01
-1.21930134e+00 5.17601788e-01 4.03673440e-01 8.65532458e-01
1.42657435e+00 3.92030865e-01 1.14431858e+00 7.20448792e-01
7.93575644e-01 -1.62485790e+00 1.80742010e-01 1.62517235e-01
6.13570511e-01 -8.79561782e-01 -9.45744216e-02 6.56595454e-03
-3.38614762e-01 1.28697050e+00 3.66734505e-01 -2.75792092e-01
1.07468140e+00 5.77145278e-01 -6.06429696e-01 -3.41030300e-01
-1.40487218e+00 9.02823806e-02 -2.35263765e-01 3.40742528e-01
-7.04800934e-02 2.25020036e-01 -2.25676700e-01 6.05253816e-01
-5.70729315e-01 -1.80282369e-01 3.73913765e-01 1.09213269e+00
-1.01001036e+00 -9.19768155e-01 -8.32640588e-01 7.76252389e-01
-4.99386072e-01 2.21037745e-01 -2.18419475e-03 8.80887266e-03
7.27625787e-02 1.35961616e+00 6.96792245e-01 -2.05590293e-01
4.36857134e-01 -5.28023124e-01 2.28233654e-02 -4.99557376e-01
-4.84208375e-01 -5.00975624e-02 -3.23569745e-01 -7.25469351e-01
-1.90962449e-01 -7.82476127e-01 -1.42081738e+00 -1.00631189e+00
-2.32663125e-01 6.23575509e-01 9.65426147e-01 5.14397442e-01
6.17688596e-01 -3.23245794e-01 1.17635393e+00 -4.53777075e-01
-8.27523649e-01 -3.90632510e-01 -8.32161427e-01 2.90870428e-01
2.54172951e-01 2.10837632e-01 -1.45338938e-01 9.57646966e-02] | [6.6469502449035645, 4.949954032897949] |
b4f4fbad-950c-43d2-8fa4-3c97afbcd508 | parameter-efficient-deep-probabilistic | 2112.02905 | null | https://arxiv.org/abs/2112.02905v2 | https://arxiv.org/pdf/2112.02905v2.pdf | Parameter Efficient Deep Probabilistic Forecasting | Probabilistic time series forecasting is crucial in many application domains such as retail, ecommerce, finance, or biology. With the increasing availability of large volumes of data, a number of neural architectures have been proposed for this problem. In particular, Transformer-based methods achieve state-of-the-art performance on real-world benchmarks. However, these methods require a large number of parameters to be learned, which imposes high memory requirements on the computational resources for training such models. To address this problem, we introduce a novel Bidirectional Temporal Convolutional Network (BiTCN), which requires an order of magnitude less parameters than a common Transformer-based approach. Our model combines two Temporal Convolutional Networks (TCNs): the first network encodes future covariates of the time series, whereas the second network encodes past observations and covariates. We jointly estimate the parameters of an output distribution via these two networks. Experiments on four real-world datasets show that our method performs on par with four state-of-the-art probabilistic forecasting methods, including a Transformer-based approach and WaveNet, on two point metrics (sMAPE, NRMSE) as well as on a set of range metrics (quantile loss percentiles) in the majority of cases. Secondly, we demonstrate that our method requires significantly less parameters than Transformer-based methods, which means the model can be trained faster with significantly lower memory requirements, which as a consequence reduces the infrastructure cost for deploying these models. | ['Maarten de Rijke', 'Sebastian Schelter', 'Olivier Sprangers'] | 2021-12-06 | null | null | null | null | ['probabilistic-time-series-forecasting'] | ['time-series'] | [-2.39372283e-01 -2.98737556e-01 8.54955167e-02 -4.30515736e-01
-7.44057059e-01 -5.44833839e-01 7.97769129e-01 -1.32548865e-02
-2.72601336e-01 5.40486753e-01 7.67291561e-02 -5.30900955e-01
-3.88940573e-01 -8.20557356e-01 -7.68210530e-01 -9.06181335e-01
-4.79219824e-01 2.58887112e-01 2.77209729e-01 -1.58070147e-01
7.42549896e-02 4.10040230e-01 -1.39239144e+00 4.24679108e-02
4.94788587e-01 1.71286297e+00 -2.47829497e-01 4.03419614e-01
-1.28470674e-01 6.38578117e-01 -2.86421001e-01 -5.26587248e-01
6.53623343e-02 1.25266433e-01 -5.15038371e-01 -5.20602107e-01
4.77528684e-02 -2.90685773e-01 -6.06239140e-01 5.29272199e-01
3.64407986e-01 3.12131286e-01 4.95421946e-01 -1.35225511e+00
-3.38415295e-01 6.94266677e-01 -5.01197934e-01 2.20346555e-01
-2.62096792e-01 -7.76453540e-02 9.31167126e-01 -6.57194674e-01
5.86640686e-02 1.07147014e+00 1.12872899e+00 -7.27458820e-02
-1.31989229e+00 -7.87181377e-01 4.01720017e-01 3.39304000e-01
-1.27126813e+00 -1.93748474e-01 9.07834232e-01 -5.05656242e-01
1.00948417e+00 2.05210913e-02 4.20352966e-01 1.12274790e+00
4.03949946e-01 5.15400529e-01 8.87439847e-01 -1.18317241e-02
3.24263483e-01 -5.62844202e-02 -1.53093398e-01 1.85233444e-01
-2.83133358e-01 2.82164752e-01 -6.50217891e-01 -3.05444360e-01
5.96533656e-01 2.85686672e-01 -9.42786932e-02 -1.69228747e-01
-1.34820938e+00 8.53272557e-01 3.39688569e-01 4.75195080e-01
-5.71209073e-01 6.11797273e-01 5.09367883e-01 1.72458604e-01
1.04790211e+00 -1.05430849e-01 -7.34335780e-01 -4.38750118e-01
-1.20125079e+00 2.69456893e-01 8.04034889e-01 6.35449171e-01
3.29484165e-01 1.17509574e-01 -5.58626801e-02 7.03284383e-01
2.78509557e-01 5.75382888e-01 4.01344121e-01 -6.24122083e-01
5.84385395e-01 1.47748709e-01 1.62280947e-01 -1.08468091e+00
-5.93841255e-01 -6.86049938e-01 -1.26631343e+00 -2.67201602e-01
5.03149807e-01 -3.25266011e-02 -8.61795485e-01 1.92842889e+00
2.50350177e-01 5.62906623e-01 -1.59047350e-01 6.70851648e-01
2.63610899e-01 1.08034897e+00 5.18241860e-02 -2.49565169e-02
1.21144021e+00 -5.91910720e-01 -4.87195164e-01 3.93120758e-02
3.90296429e-01 -7.79245019e-01 5.91758311e-01 4.94965732e-01
-7.43612945e-01 -3.22009861e-01 -8.52231622e-01 7.19982982e-02
-5.56938648e-01 1.02073178e-01 7.73490489e-01 5.60780048e-01
-1.10770440e+00 9.53311026e-01 -1.19133806e+00 -2.08178014e-01
1.87043339e-01 3.52755666e-01 -4.86540422e-02 1.72115862e-01
-1.38737178e+00 6.75336003e-01 1.17052622e-01 5.58052480e-01
-7.26249695e-01 -1.01473331e+00 -6.62582278e-01 5.07136047e-01
-2.32085451e-01 -2.99515665e-01 1.36260283e+00 -3.39851618e-01
-1.56577158e+00 1.34842485e-01 -7.14485720e-02 -7.07730532e-01
5.38790882e-01 -2.16594562e-01 -6.00128233e-01 -2.97185451e-01
-2.88725406e-01 3.85194838e-01 6.46403730e-01 -6.47100508e-01
-6.71680033e-01 -1.73651367e-01 -1.06293581e-01 -3.63287330e-01
-6.08737767e-01 -6.39963076e-02 -2.22968355e-01 -8.35019767e-01
1.23587258e-01 -8.50834250e-01 -2.92894483e-01 9.53119025e-02
-1.89631730e-01 -5.07040799e-01 9.41341043e-01 -7.12239444e-01
1.24003255e+00 -2.27817106e+00 -1.03099659e-01 2.99485654e-01
2.01618418e-01 -1.00327939e-01 1.23813702e-03 6.87482059e-01
-1.18661627e-01 -7.27960933e-03 -2.94795781e-01 -5.60581684e-01
2.34237671e-01 5.80662265e-02 -9.78370249e-01 5.82811177e-01
1.33328050e-01 6.31712854e-01 -6.62657797e-01 8.79585929e-03
2.41879046e-01 8.96745205e-01 -2.66692847e-01 3.42153907e-02
-2.73528486e-01 3.55017185e-01 -2.97828138e-01 2.30870560e-01
5.91896951e-01 -4.11195368e-01 1.96578845e-01 -1.86121970e-01
-2.43850708e-01 5.23158550e-01 -9.65087295e-01 1.30242777e+00
-7.28911817e-01 8.84176314e-01 -4.35400218e-01 -1.22345006e+00
9.72835600e-01 6.15862310e-01 9.18285608e-01 -7.29939044e-01
1.05376564e-01 3.21404278e-01 -3.00664961e-01 3.52254622e-02
3.98347676e-01 -1.51379734e-01 -2.17159852e-01 5.87287903e-01
-4.73060124e-02 2.98183709e-01 7.13224383e-03 -9.03097540e-02
8.93323660e-01 1.19168915e-01 -3.24537814e-01 -1.07976288e-01
2.51783729e-01 -3.54799807e-01 6.30244136e-01 3.99559259e-01
1.48098767e-01 3.38865399e-01 6.15434229e-01 -8.82382989e-01
-1.11941493e+00 -1.03691041e+00 -1.61328852e-01 1.04070020e+00
-3.01527172e-01 -2.55801767e-01 -3.84874433e-01 -2.03141481e-01
1.33713976e-01 7.86977410e-01 -7.23516762e-01 -2.29363814e-02
-7.01211154e-01 -9.12068963e-01 5.32484114e-01 7.65411675e-01
2.64800429e-01 -7.38036692e-01 -5.94806969e-01 6.12980306e-01
-2.75241733e-01 -1.16954207e+00 -3.89030308e-01 1.95135161e-01
-1.02115858e+00 -6.07432365e-01 -7.83859611e-01 -3.41434211e-01
7.92383701e-02 3.44412923e-02 1.03160787e+00 -5.06526053e-01
1.25484452e-01 1.84066921e-01 -1.58223823e-01 -6.25705481e-01
9.90900695e-02 2.13087454e-01 8.95726085e-02 4.20108825e-01
3.19749057e-01 -1.16368926e+00 -6.29297435e-01 3.56817186e-01
-9.11640286e-01 7.24525098e-03 3.47050786e-01 8.13781321e-01
4.67930824e-01 3.48185092e-01 6.68143809e-01 -4.26317245e-01
4.63483214e-01 -8.29087794e-01 -1.06609380e+00 1.36092454e-01
-7.08273590e-01 1.92069039e-01 8.49415541e-01 -7.37529635e-01
-7.68554270e-01 -1.31327942e-01 2.71397866e-02 -6.39155388e-01
3.64172816e-01 9.35469270e-01 4.53999281e-01 2.98423350e-01
2.90342093e-01 4.79668528e-01 -1.91928476e-01 -6.71520948e-01
1.95280656e-01 5.64405978e-01 5.92521489e-01 -4.58137125e-01
6.71900809e-01 6.01307154e-01 1.28877148e-01 -4.47149128e-01
-7.29643703e-01 -3.46677393e-01 -4.17368323e-01 -2.09387749e-01
4.94559437e-01 -8.61496687e-01 -9.76788521e-01 7.65610754e-01
-1.30303001e+00 -2.99703747e-01 4.76328172e-02 7.42209971e-01
-4.19832706e-01 -8.51401538e-02 -5.56494594e-01 -9.72392499e-01
-4.28320467e-01 -8.13491881e-01 1.03405762e+00 -1.53364807e-01
6.37539551e-02 -1.12076652e+00 1.51695134e-02 -2.25832686e-01
9.57338691e-01 4.73029643e-01 9.94918883e-01 -5.14427006e-01
-5.92886984e-01 -3.82562190e-01 -2.77783155e-01 1.67442143e-01
-2.36115277e-01 -7.17226341e-02 -1.08287454e+00 -2.24767849e-01
6.06958941e-02 9.70596075e-02 8.81530583e-01 5.71981668e-01
1.51830721e+00 -2.70471871e-01 -4.61829811e-01 7.46218443e-01
1.25110972e+00 2.30900899e-01 5.53482533e-01 8.21300521e-02
5.85168958e-01 5.52575111e-01 2.90226400e-01 6.83276653e-01
6.54642761e-01 7.58666158e-01 3.46770465e-01 2.66149435e-02
3.91110331e-01 -1.36734948e-01 3.50384086e-01 1.03820240e+00
1.91447213e-02 -4.53839928e-01 -1.09584785e+00 7.60155082e-01
-2.12321973e+00 -8.56482208e-01 -2.69618002e-03 2.21209574e+00
6.39942229e-01 8.86313096e-02 5.93245439e-02 3.49095821e-01
4.08157706e-01 2.42027357e-01 -7.28902519e-01 -2.52521932e-01
1.66477531e-01 2.92368412e-01 6.13412201e-01 3.90028469e-02
-1.31494844e+00 4.58042651e-01 6.17834330e+00 7.93819487e-01
-1.60748911e+00 1.34541616e-01 9.33568537e-01 -1.77462772e-01
-2.87126660e-01 -1.59871176e-01 -6.41363800e-01 7.38830686e-01
1.66952860e+00 -3.23903531e-01 5.27737260e-01 7.99011827e-01
2.45397881e-01 2.15979069e-01 -1.23577631e+00 1.04396057e+00
-4.95851994e-01 -1.43174052e+00 -3.82832170e-01 6.26690239e-02
4.86359984e-01 4.36098337e-01 2.70075202e-01 4.35583621e-01
2.79599220e-01 -1.03026080e+00 9.22930717e-01 7.91955590e-01
8.01622093e-01 -9.01098251e-01 8.86041760e-01 2.53106594e-01
-1.45741236e+00 -1.24868415e-01 -2.40298972e-01 -2.02680901e-02
4.36647922e-01 1.25217807e+00 -4.08626199e-01 4.76265341e-01
1.14462185e+00 7.46589661e-01 -6.97451755e-02 1.09932935e+00
1.66372448e-01 1.03640962e+00 -7.25166500e-01 -3.85085084e-02
3.71485680e-01 8.43803864e-03 2.28674620e-01 1.04327607e+00
7.86459863e-01 -1.05560280e-01 -5.92160933e-02 7.34435201e-01
-7.09145442e-02 -2.28067860e-01 -2.17739090e-01 -2.23233551e-01
4.87541646e-01 1.11775327e+00 -4.64132994e-01 -2.19711870e-01
-3.86459827e-01 5.11426210e-01 2.16521233e-01 4.57612216e-01
-1.12036479e+00 -3.70739400e-01 7.21956372e-01 -5.37502579e-02
6.65757477e-01 -5.24362445e-01 -2.23302126e-01 -9.41595197e-01
2.53846318e-01 -4.58186239e-01 3.94966006e-01 -5.82054913e-01
-1.66659093e+00 7.82834709e-01 4.51403260e-02 -1.39165020e+00
-5.18113256e-01 -5.42255521e-01 -5.43622315e-01 1.07011497e+00
-1.75480199e+00 -1.10280573e+00 -1.46967888e-01 4.84856457e-01
1.12360746e-01 8.73396024e-02 7.44249761e-01 6.07163489e-01
-4.44122881e-01 4.81155664e-01 5.40655851e-01 1.13479756e-02
5.74774504e-01 -1.02378333e+00 8.10425341e-01 6.88433349e-01
3.10968515e-02 4.52971995e-01 5.45379221e-01 -2.17066586e-01
-1.30272293e+00 -1.27808511e+00 1.19346845e+00 -1.92892030e-01
9.55785751e-01 -5.48083782e-01 -9.07183230e-01 5.36870360e-01
-9.49354842e-02 2.05644801e-01 5.57468116e-01 4.15601939e-01
-6.22687042e-01 -6.01122975e-01 -8.50341380e-01 3.04960519e-01
5.89249492e-01 -5.63474417e-01 -3.44584994e-02 5.00283718e-01
8.66901338e-01 -3.98495913e-01 -1.23035944e+00 3.74266535e-01
9.36492622e-01 -7.72008717e-01 8.46393585e-01 -3.24246138e-01
4.36944216e-01 -2.69747674e-01 -1.17257021e-01 -1.29784918e+00
-4.10456061e-01 -6.84689403e-01 -4.64614362e-01 1.16903090e+00
4.46442366e-01 -1.01883280e+00 5.02409875e-01 6.36208594e-01
3.76976058e-02 -1.20130074e+00 -1.15118039e+00 -1.00031412e+00
2.96630770e-01 -8.45508993e-01 1.03250217e+00 8.20960701e-01
-3.03672075e-01 -2.30562426e-02 -5.47058821e-01 2.02215716e-01
4.61425841e-01 2.15180174e-01 3.68431956e-01 -1.45568955e+00
-1.42267659e-01 -6.16708219e-01 -3.79564077e-01 -1.00821483e+00
-8.52997676e-02 -5.29409051e-01 -2.45022196e-02 -1.28616488e+00
-1.82883963e-01 -8.83232892e-01 -7.24656165e-01 6.49020016e-01
2.15059042e-01 1.11016296e-01 -1.67900950e-01 2.90844530e-01
-1.28832757e-01 8.85381699e-01 5.08107781e-01 -1.22607157e-01
-7.09300786e-02 1.69290513e-01 -3.25945795e-01 5.00797927e-01
8.34173322e-01 -5.09025156e-01 -4.66736972e-01 -4.81257349e-01
3.64308953e-01 3.01231414e-01 5.36716104e-01 -1.08964419e+00
3.96923304e-01 -1.14953950e-01 4.46988255e-01 -9.08275425e-01
5.03914297e-01 -7.82174468e-01 3.61227602e-01 2.05063075e-01
-1.81822464e-01 3.88339877e-01 4.04748857e-01 7.19667494e-01
-3.28749210e-01 5.40914476e-01 5.30267417e-01 4.72593546e-01
-3.12347114e-01 5.66539049e-01 -3.63592148e-01 -3.23960871e-01
6.70557618e-01 2.23139375e-01 -2.21971244e-01 -6.22293532e-01
-3.43010873e-01 2.37896532e-01 -2.33966500e-01 5.41395187e-01
2.72143573e-01 -1.37109351e+00 -6.07070744e-01 1.64546259e-02
-1.21207334e-01 -1.15753874e-01 4.72531736e-01 1.12760842e+00
-1.26510218e-01 7.60109425e-01 2.42781609e-01 -8.31996322e-01
-6.61721170e-01 4.89638180e-01 2.88426787e-01 -6.00622892e-01
-6.28935695e-01 8.08243573e-01 1.94992661e-01 -5.10515451e-01
3.37501734e-01 -7.42069721e-01 4.10351111e-03 4.43987697e-02
4.47373837e-01 3.68046999e-01 3.09286386e-01 -3.85015070e-01
-4.12103534e-01 3.52196157e-01 1.51788853e-02 -2.37854049e-01
1.77575147e+00 4.82046828e-02 -7.21087232e-02 6.56153560e-01
1.24385107e+00 -4.37652528e-01 -1.38966000e+00 -3.97817045e-01
7.60745481e-02 -1.13112450e-01 3.41391385e-01 -7.71931708e-01
-1.15704548e+00 1.05435228e+00 6.18332922e-01 4.76310432e-01
1.39813566e+00 -1.87278509e-01 1.17574239e+00 2.77231127e-01
5.05730987e-01 -8.83126974e-01 -3.61008614e-01 6.26830637e-01
8.46145391e-01 -8.95497620e-01 -1.76175401e-01 -1.82654336e-01
-2.24524826e-01 1.13566232e+00 3.96370888e-02 -6.80420175e-03
1.16070056e+00 2.11379334e-01 -5.27334101e-02 -8.40041488e-02
-1.15627789e+00 2.88243383e-01 5.85780799e-01 1.68460414e-01
5.20628035e-01 1.82447016e-01 4.12378693e-03 7.19022632e-01
-4.74481672e-01 1.01125283e-04 1.04103558e-01 5.68058670e-01
1.03621008e-02 -9.78338540e-01 -1.71966314e-01 5.53087533e-01
-4.91517454e-01 -2.01242894e-01 2.89376318e-01 5.62662899e-01
-2.78312981e-01 9.76432621e-01 3.13913614e-01 -4.34393913e-01
1.49001509e-01 6.42529726e-02 -2.58369446e-02 5.67381196e-02
-5.35471439e-01 3.64497080e-02 -5.83770163e-02 -7.13319004e-01
-3.70062828e-01 -7.03340709e-01 -8.17718804e-01 -6.34962499e-01
-2.85856217e-01 7.96402842e-02 1.01279151e+00 1.00365901e+00
6.60357177e-01 5.90644658e-01 8.03482950e-01 -9.41722691e-01
-6.06073081e-01 -1.07039034e+00 -5.30053139e-01 1.80959497e-02
2.56771326e-01 -8.70536447e-01 -2.48428479e-01 -2.96630412e-02] | [6.970136642456055, 3.0271527767181396] |
99adfa11-195c-482e-b01b-be4a371cce66 | points2vec-unsupervised-object-level-feature | 2102.04136 | null | https://arxiv.org/abs/2102.04136v1 | https://arxiv.org/pdf/2102.04136v1.pdf | Points2Vec: Unsupervised Object-level Feature Learning from Point Clouds | Unsupervised representation learning techniques, such as learning word embeddings, have had a significant impact on the field of natural language processing. Similar representation learning techniques have not yet become commonplace in the context of 3D vision. This, despite the fact that the physical 3D spaces have a similar semantic structure to bodies of text: words are surrounded by words that are semantically related, just like objects are surrounded by other objects that are similar in concept and usage. In this work, we exploit this structure in learning semantically meaningful low dimensional vector representations of objects. We learn these vector representations by mining a dataset of scanned 3D spaces using an unsupervised algorithm. We represent objects as point clouds, a flexible and general representation for 3D data, which we encode into a vector representation. We show that using our method to include context increases the ability of a clustering algorithm to distinguish different semantic classes from each other. Furthermore, we show that our algorithm produces continuous and meaningful object embeddings through interpolation experiments. | ['Roland Siegwart', 'Julian Förster', 'Kenneth Blomqvist', 'Joël Bachmann'] | 2021-02-08 | null | null | null | null | ['learning-word-embeddings'] | ['methodology'] | [-7.22633069e-03 9.83834341e-02 -1.99572202e-02 -5.53669572e-01
-1.06925227e-01 -6.30737126e-01 8.40034306e-01 8.93406332e-01
-3.82400513e-01 -5.12055233e-02 6.31490707e-01 -2.58096069e-01
-1.11282200e-01 -9.32908535e-01 -4.94279206e-01 -4.11229730e-01
-2.95477986e-01 4.26968366e-01 2.08770961e-01 -1.13852412e-01
2.46960849e-01 6.90045238e-01 -1.83658397e+00 1.28559515e-01
2.70052940e-01 7.28133142e-01 1.27570003e-01 4.65546191e-01
-8.78163159e-01 4.39668864e-01 -4.23839450e-01 2.18566790e-01
2.36118585e-01 -1.73902243e-01 -7.62123346e-01 3.62225115e-01
2.53765315e-01 1.94740146e-01 -1.69105455e-01 8.36826503e-01
8.48868787e-02 6.11265481e-01 1.20503390e+00 -1.07253015e+00
-1.09149027e+00 1.74062878e-01 -4.34301108e-01 -1.95714012e-01
4.75000173e-01 -4.28091705e-01 1.21372485e+00 -9.67543244e-01
6.92073286e-01 1.54609156e+00 5.99334538e-01 4.97003198e-01
-1.28379178e+00 1.20196857e-01 1.05073579e-01 1.38338357e-01
-1.44988143e+00 -5.10654114e-02 9.59727228e-01 -7.88840175e-01
1.09469903e+00 1.39108256e-01 6.59222066e-01 7.12348759e-01
-4.46408056e-02 5.04768431e-01 6.59315944e-01 -9.29484308e-01
6.09626174e-01 3.45884740e-01 4.00500238e-01 5.97827375e-01
3.49096090e-01 -2.12802619e-01 -1.77279100e-01 -1.00484774e-01
7.13914692e-01 5.07412553e-01 -2.96503082e-02 -1.06126869e+00
-1.18380582e+00 1.19182825e+00 6.84358180e-01 7.41151929e-01
-2.92868018e-01 4.24400330e-01 3.14901769e-01 4.05231901e-02
6.11500502e-01 5.02675951e-01 -1.95811465e-01 7.52518773e-02
-5.79187751e-01 2.06556946e-01 5.81106007e-01 8.77142668e-01
8.92852962e-01 -1.22797556e-01 3.31159830e-01 9.84330773e-01
5.31506360e-01 2.23564550e-01 6.50343776e-01 -8.30770612e-01
-2.41758153e-01 1.05555570e+00 -4.41692071e-03 -1.24691463e+00
-3.51123512e-01 1.01815417e-01 -5.04033208e-01 3.98156881e-01
9.80512500e-02 4.10919487e-01 -9.97402906e-01 1.45344913e+00
3.35332006e-01 7.15365633e-02 2.57095039e-01 9.07874584e-01
7.56281495e-01 7.82822073e-01 3.99529971e-02 2.81821132e-01
1.32688153e+00 -2.87258893e-01 -4.41282034e-01 1.81077030e-02
8.87626588e-01 -5.13541102e-01 1.14899075e+00 3.81737687e-02
-6.61024570e-01 -5.33061147e-01 -1.20008969e+00 -4.34265167e-01
-1.07223749e+00 -6.77252412e-01 8.91068280e-01 6.71997547e-01
-8.32483232e-01 4.86680567e-01 -7.74914622e-01 -8.77484381e-01
5.09996057e-01 5.29914275e-02 -5.57430089e-01 -2.25047320e-02
-7.75496662e-01 1.08126104e+00 6.48273051e-01 -5.73940158e-01
-3.44180346e-01 -5.07989466e-01 -1.33326888e+00 -7.72014558e-02
6.30585700e-02 -5.74131191e-01 9.89904761e-01 -6.66303039e-01
-8.10807288e-01 1.26717520e+00 -2.41799146e-01 -4.65210736e-01
-2.36390069e-01 -1.64599225e-01 -3.82290989e-01 1.07373118e-01
-5.75834289e-02 7.11983681e-01 8.30650806e-01 -1.52657986e+00
-2.36299440e-01 -6.02943480e-01 1.35040388e-01 1.05588153e-01
-5.80469728e-01 -1.45814821e-01 -1.61397129e-01 -6.81811094e-01
5.17123342e-01 -6.88527346e-01 -4.54285532e-01 6.15867853e-01
-8.42189882e-03 -5.71214855e-01 9.89977002e-01 -2.97303516e-02
8.18729818e-01 -2.38115501e+00 2.71789074e-01 4.15116727e-01
4.31692928e-01 -5.69615364e-02 4.29880172e-02 5.02568305e-01
-1.22782283e-01 3.30972701e-01 -5.69467127e-01 -9.10863653e-02
4.08900946e-01 8.12295258e-01 -2.16726243e-01 5.37105739e-01
4.27139789e-01 8.79506409e-01 -1.13357997e+00 -5.19440532e-01
5.12409270e-01 6.78167939e-01 -6.46865785e-01 -8.26688334e-02
-3.47583115e-01 -7.88300484e-02 -5.27923882e-01 3.67488384e-01
2.49807432e-01 -1.66612312e-01 1.07616941e-02 1.41648538e-02
-7.27649732e-03 2.47213125e-01 -1.17196560e+00 2.02548599e+00
-5.62205672e-01 8.66493821e-01 -5.39218605e-01 -1.42283845e+00
1.24760544e+00 1.12937480e-01 5.99736691e-01 -3.65508258e-01
1.31123498e-01 -1.47917852e-01 -2.27163583e-01 -4.80314523e-01
8.72762561e-01 -4.24249917e-01 -2.95727611e-01 7.39956141e-01
8.84538069e-02 -6.82035267e-01 -1.56515285e-01 3.61705124e-01
8.05905104e-01 -1.04766358e-02 3.23473126e-01 -2.89422244e-01
2.25718379e-01 1.93710133e-01 1.18939266e-01 4.01449949e-01
1.34010658e-01 6.03390694e-01 2.15251267e-01 -4.71391886e-01
-1.18199861e+00 -1.57212675e+00 -3.21028560e-01 7.51008868e-01
2.75955915e-01 -6.34235442e-01 -3.24464291e-01 -4.05073643e-01
3.91233683e-01 8.89643073e-01 -7.68771410e-01 -2.37017572e-01
-2.32967690e-01 -2.01288342e-01 -2.81435419e-02 6.85658514e-01
-3.11114520e-01 -7.86186635e-01 -9.64373767e-01 1.03353754e-01
5.17173767e-01 -9.23272789e-01 -7.03036925e-03 4.01468545e-01
-9.96024132e-01 -9.22703564e-01 -5.47827899e-01 -1.05194259e+00
8.41589808e-01 4.42440063e-01 1.35132456e+00 6.63767159e-02
-7.32885063e-01 1.05265164e+00 -7.01960444e-01 -7.83061147e-01
-4.43650216e-01 -4.71215904e-01 3.54613870e-01 -1.65023148e-01
9.38363969e-01 -6.15098238e-01 -9.31739062e-02 -4.15916927e-02
-1.15237212e+00 -2.35827431e-01 3.74243930e-02 4.95906055e-01
7.96212256e-01 2.48229336e-02 1.84908152e-01 -6.94336474e-01
6.03470087e-01 -6.28742993e-01 -2.09044755e-01 -3.78410071e-02
-2.26294652e-01 4.45064843e-01 1.96298361e-01 -4.72454548e-01
-3.94128084e-01 -6.54182024e-03 1.92852065e-01 -6.50610924e-01
-5.34870327e-01 4.49183792e-01 -1.36470320e-02 2.46345416e-01
8.54859471e-01 1.26185998e-01 2.07439736e-01 -6.44974947e-01
1.14227843e+00 8.05221736e-01 3.16804469e-01 -5.70897102e-01
8.95251155e-01 6.35861933e-01 6.87579364e-02 -1.47451437e+00
-4.65251148e-01 -6.42973125e-01 -8.91460955e-01 1.67212278e-01
1.16694736e+00 -6.02077007e-01 -1.31177992e-01 -2.78127819e-01
-1.10915935e+00 1.26819596e-01 -9.19047594e-01 6.51189387e-01
-7.04252362e-01 5.11989534e-01 -1.04463443e-01 -7.91053236e-01
3.03105593e-01 -6.74224675e-01 9.99794781e-01 -2.51082741e-02
-6.94307745e-01 -1.27159441e+00 2.41683364e-01 -1.30677700e-01
5.19706681e-02 4.69035357e-01 1.37882400e+00 -7.81050801e-01
-1.60156354e-01 -1.01108082e-01 -9.81381834e-02 4.19667840e-01
5.84049046e-01 -9.30223688e-02 -8.53547573e-01 5.39214946e-02
-6.72686845e-02 -8.82437155e-02 7.32775688e-01 2.64952302e-01
1.32138216e+00 -4.13836949e-02 -3.53818357e-01 3.11606407e-01
1.55061758e+00 1.04073249e-01 4.19282883e-01 2.03101486e-01
7.30457366e-01 8.67819488e-01 1.62886634e-01 3.11860085e-01
1.44291520e-01 4.52677578e-01 3.83478463e-01 -6.28211871e-02
6.29628194e-04 -2.73572415e-01 7.40068685e-03 8.46241832e-01
1.38187572e-01 6.95409849e-02 -1.12656462e+00 8.92712355e-01
-1.54896760e+00 -8.77098799e-01 -1.46569740e-02 1.98526609e+00
5.05677342e-01 9.70058218e-02 -8.40472244e-03 4.75054413e-01
4.90414292e-01 1.55414119e-01 -3.24329257e-01 -8.37330937e-01
7.54477158e-02 5.59678078e-01 2.47119904e-01 2.87653476e-01
-1.04483533e+00 8.00914705e-01 6.19513321e+00 2.60975927e-01
-8.03534210e-01 -4.07315977e-02 4.67392504e-02 6.20791735e-03
-7.23980784e-01 -1.54802203e-01 -1.34852216e-01 1.29988402e-01
7.55538940e-01 -2.54393637e-01 -5.77778034e-02 8.36234093e-01
1.38580650e-02 1.33887306e-01 -1.51604140e+00 1.28028989e+00
2.65543103e-01 -1.41532493e+00 4.35842544e-01 3.60456482e-02
5.14823318e-01 -2.80027688e-01 -2.91462038e-02 2.98056658e-02
3.44507307e-01 -1.36065793e+00 5.11122167e-01 4.62860525e-01
4.34233218e-01 -7.67722487e-01 2.46484563e-01 7.22059011e-02
-1.05041814e+00 1.19180910e-01 -6.97509944e-01 -2.14960411e-01
1.17295146e-01 6.21553421e-01 -7.95259893e-01 3.99807364e-01
6.42117560e-01 9.54281271e-01 -3.27378094e-01 9.17891383e-01
-4.97303382e-02 2.18218744e-01 -3.30853701e-01 -3.38860601e-01
4.90121037e-01 -3.49102885e-01 4.75856900e-01 1.16181636e+00
2.23454088e-01 8.59723762e-02 5.84822074e-02 1.02410567e+00
-3.77300009e-03 1.56809047e-01 -1.19652045e+00 -3.78815979e-01
3.59235406e-01 7.23465562e-01 -8.52754891e-01 -3.53761077e-01
-6.82796240e-01 8.20934534e-01 2.67440826e-02 1.35202110e-01
-4.63265657e-01 -5.48916757e-01 1.15828753e+00 1.84541658e-01
3.76766592e-01 -1.00496352e+00 -3.16758633e-01 -8.41061413e-01
-4.19743098e-02 -1.58198833e-01 4.46139798e-02 -8.83519232e-01
-1.56317043e+00 2.01942265e-01 1.63591772e-01 -1.32457888e+00
-1.90790161e-01 -9.28392768e-01 -3.38466048e-01 5.82465649e-01
-1.14108944e+00 -8.14944804e-01 -1.22229397e-01 7.03164935e-01
4.95851994e-01 -7.05276355e-02 1.19126451e+00 -1.07894354e-01
2.75265247e-01 8.67211074e-02 2.24138409e-01 2.91753262e-01
3.15104753e-01 -1.45425713e+00 4.65425521e-01 3.63920957e-01
1.10030913e+00 8.21965873e-01 5.67430139e-01 -2.24781811e-01
-1.55509424e+00 -1.02095890e+00 9.31467772e-01 -7.63239086e-01
7.65652537e-01 -7.22634673e-01 -1.06091392e+00 5.78237057e-01
-2.31816739e-01 1.72943756e-01 1.10360110e+00 3.39973032e-01
-7.85722077e-01 3.41198474e-01 -1.11005640e+00 6.37973607e-01
1.19691360e+00 -8.79759550e-01 -1.37402225e+00 4.32846993e-01
1.12040913e+00 4.57622819e-02 -9.03589308e-01 -9.99928564e-02
3.49363685e-01 -6.06358111e-01 1.34041631e+00 -1.01308274e+00
2.60591179e-01 -3.06063116e-01 -7.04083681e-01 -1.36151528e+00
-2.64724255e-01 4.90315780e-02 -1.57015204e-01 8.24348390e-01
1.56730190e-01 -4.85254407e-01 8.13612938e-01 4.07847881e-01
-1.73787773e-01 -5.37549615e-01 -7.94364750e-01 -1.04234481e+00
4.81572509e-01 -8.04492950e-01 6.66261196e-01 1.14558482e+00
1.08779937e-01 2.68912971e-01 3.07841003e-01 -4.43334952e-02
6.03030384e-01 3.80863190e-01 5.41534722e-01 -1.79731464e+00
-2.15446344e-03 -4.46108639e-01 -1.33675420e+00 -1.05322742e+00
2.97601104e-01 -1.34616888e+00 -6.01109341e-02 -1.77607334e+00
-9.68216509e-02 -7.28146911e-01 -3.44258666e-01 3.75569254e-01
3.38497311e-01 1.81201160e-01 2.53530502e-01 3.90823632e-02
-3.48040283e-01 5.93215823e-01 7.86226094e-01 -3.50409448e-01
-2.00133443e-01 -6.47799730e-01 -7.26429820e-01 8.22984040e-01
6.52728736e-01 -4.21487600e-01 -5.69356441e-01 -6.94879651e-01
6.92555532e-02 -5.83961427e-01 4.42586094e-01 -8.49200130e-01
-7.05770925e-02 -9.20137856e-03 4.85649645e-01 -3.87860119e-01
4.57136959e-01 -1.23417819e+00 -1.77367464e-01 1.90753639e-01
-4.03938383e-01 -1.53405830e-01 2.22372428e-01 6.71822608e-01
-8.95591304e-02 -3.46437752e-01 6.31212115e-01 -2.71688193e-01
-1.24854040e+00 1.08549632e-01 -5.09499431e-01 4.45563830e-02
1.11355531e+00 -7.35265791e-01 2.62312740e-01 -1.49134159e-01
-8.25440228e-01 -2.95881052e-02 8.73418212e-01 8.88743579e-01
9.67912078e-01 -1.42504680e+00 -4.32826459e-01 2.04245165e-01
5.67470908e-01 1.46945521e-01 -2.49952048e-01 -8.05810392e-02
-4.61174309e-01 2.48118579e-01 -2.03544155e-01 -9.69802797e-01
-9.95312989e-01 8.17789376e-01 1.13470443e-02 6.03056908e-01
-1.11252809e+00 6.91297591e-01 1.37446642e-01 -5.29757619e-01
1.50740162e-01 -6.75519109e-01 -2.57917255e-01 1.03695832e-01
4.24144506e-01 -5.63026220e-02 -9.87229049e-02 -7.38588750e-01
-5.55421710e-01 9.21044946e-01 1.85801715e-01 -1.65945478e-02
1.58029521e+00 1.27212152e-01 -3.70178260e-02 1.07999921e+00
1.53258109e+00 -2.02869967e-01 -6.22669518e-01 -3.03837925e-01
4.24831867e-01 -6.50975227e-01 2.89239641e-02 -1.41749904e-01
-5.58121860e-01 1.01862013e+00 7.10027099e-01 5.31427324e-01
6.06966615e-01 6.70447052e-01 5.92361927e-01 4.57593054e-01
3.77819926e-01 -8.60669732e-01 2.00349405e-01 5.66858232e-01
6.58386707e-01 -1.13489449e+00 7.61584789e-02 -2.47283489e-01
-3.50334167e-01 1.05275261e+00 -2.55421624e-02 -4.75318819e-01
8.34940553e-01 -7.70596862e-02 -1.25042796e-01 -5.92549205e-01
-3.93432379e-01 -4.05493915e-01 2.34203890e-01 1.16686618e+00
4.42959934e-01 2.55073130e-01 -1.48055684e-02 3.25448155e-01
-2.64619172e-01 -3.55372280e-01 4.32278693e-01 1.30902410e+00
-7.64218986e-01 -1.12809110e+00 -3.89657617e-01 3.78442258e-01
4.35064621e-02 6.32631108e-02 -5.55653214e-01 7.65358925e-01
1.96401164e-01 8.34850073e-01 7.10241973e-01 -2.00405017e-01
2.90309489e-01 3.30619007e-01 6.79701447e-01 -1.08081877e+00
-2.73073697e-03 -5.03303528e-01 -2.81832010e-01 -4.19451386e-01
-6.39345467e-01 -4.62802172e-01 -1.67367101e+00 -7.93981086e-03
5.49680665e-02 3.37311804e-01 1.11206532e+00 7.88579345e-01
2.46343657e-01 3.04296702e-01 4.68280792e-01 -9.47979033e-01
-2.02508122e-01 -5.33932149e-01 -8.00699353e-01 7.68526912e-01
2.90165871e-01 -1.05473340e+00 -2.93631643e-01 2.97733605e-01] | [10.321168899536133, 2.3803248405456543] |
8fff9f12-ac4f-43b9-8614-b9c4764a292d | mia-cov19d-covid-19-detection-through-3-d | 2106.07524 | null | https://arxiv.org/abs/2106.07524v2 | https://arxiv.org/pdf/2106.07524v2.pdf | MIA-COV19D: COVID-19 Detection through 3-D Chest CT Image Analysis | Early and reliable COVID-19 diagnosis based on chest 3-D CT scans can assist medical specialists in vital circumstances. Deep learning methodologies constitute a main approach for chest CT scan analysis and disease prediction. However, large annotated databases are necessary for developing deep learning models that are able to provide COVID-19 diagnosis across various medical environments in different countries. Due to privacy issues, publicly available COVID-19 CT datasets are highly difficult to obtain, which hinders the research and development of AI-enabled diagnosis methods of COVID-19 based on CT scans. In this paper we present the COV19-CT-DB database which is annotated for COVID-19, consisting of about 5,000 3-D CT scans, We have split the database in training, validation and test datasets. The former two datasets can be used for training and validation of machine learning models, while the latter will be used for evaluation of the developed models. We also present a deep learning approach, based on a CNN-RNN network and report its performance on the COVID19-CT-DB database. | ['Stefanos Kollias', 'Levon Soukissian', 'Anastasios Arsenos', 'Dimitrios Kollias'] | 2021-06-14 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [-2.54774064e-01 -2.67307460e-01 -9.29984376e-02 -4.57860053e-01
-6.87931478e-01 -1.31279781e-01 5.97217791e-02 3.83625567e-01
-5.44866621e-01 6.34415030e-01 7.86170065e-02 -7.26100028e-01
-3.24543118e-01 -8.90915275e-01 -2.65286952e-01 -6.44015968e-01
-2.65520841e-01 1.28091633e+00 1.64441437e-01 2.63847142e-01
-4.49766576e-01 7.92976320e-01 -1.03680003e+00 4.12738174e-01
3.18606973e-01 1.15977681e+00 4.20622021e-01 8.95716369e-01
1.29666820e-01 9.21647310e-01 -3.95463973e-01 -3.41537088e-01
2.56190777e-01 -3.42890739e-01 -8.08463097e-01 -2.71874517e-01
-7.58074597e-02 -7.44135082e-01 -4.18535590e-01 4.38003510e-01
8.81256640e-01 -2.88052946e-01 9.88559425e-01 -8.31004739e-01
-2.15792343e-01 3.89301479e-01 -2.53932685e-01 5.32840788e-01
8.11874717e-02 1.47921190e-01 7.38450050e-01 -7.56205499e-01
8.33508432e-01 6.80256963e-01 8.71095002e-01 6.49575353e-01
-4.51405942e-01 -5.14446259e-01 -6.08038664e-01 2.95367569e-01
-1.28806317e+00 4.51454878e-01 2.84282058e-01 -6.96827590e-01
8.94885182e-01 2.95882344e-01 1.05796683e+00 1.27562451e+00
6.21703744e-01 7.36272395e-01 8.42668355e-01 -1.04954436e-01
-3.43807787e-02 -3.33828330e-02 9.93354395e-02 8.67376447e-01
4.47551101e-01 2.17626959e-01 1.36439219e-01 -1.96041062e-01
8.37092638e-01 3.94276202e-01 -1.64580360e-01 -1.95148423e-01
-1.06734467e+00 1.14639854e+00 7.53069043e-01 6.13766074e-01
-6.21035337e-01 -1.01062737e-01 7.78527319e-01 1.29386291e-01
3.26283962e-01 2.02012658e-01 -4.79834914e-01 1.51360840e-01
-8.39701235e-01 4.87380236e-01 5.90479434e-01 5.81151545e-01
-2.08323419e-01 -9.59107245e-04 -2.06302345e-01 8.18432748e-01
4.36054796e-01 6.14672601e-01 8.43112171e-01 -2.00870603e-01
4.68967140e-01 5.02614141e-01 -2.29325995e-01 -8.69066179e-01
-9.98924851e-01 -3.92518610e-01 -1.28738809e+00 -2.66389161e-01
1.32667765e-01 -3.04828495e-01 -1.21655476e+00 1.09301412e+00
2.89408684e-01 -5.24699427e-02 4.68705632e-02 1.05524623e+00
1.37910748e+00 3.18921387e-01 3.80824767e-02 -1.75678924e-01
1.47217631e+00 -6.25312924e-01 -5.32624900e-01 4.61326152e-01
7.21275926e-01 -5.09070516e-01 5.72116256e-01 4.93355393e-01
-9.18190241e-01 -3.58255446e-01 -8.72271478e-01 1.82765186e-01
-4.32888418e-01 2.43032485e-01 4.24637765e-01 5.56486547e-01
-8.83195877e-01 3.28400016e-01 -1.07601154e+00 -3.60470742e-01
7.35961378e-01 4.36850578e-01 -2.27978036e-01 -2.59433150e-01
-1.40786123e+00 1.14640975e+00 5.29127538e-01 1.65231172e-02
-1.40843070e+00 -4.80967700e-01 -5.35716653e-01 -2.28339627e-01
1.61837861e-01 -1.00395322e+00 1.29151499e+00 -2.31955796e-01
-7.52330959e-01 1.11061466e+00 4.04961884e-01 -7.54294574e-01
8.30728769e-01 1.19611673e-01 -6.56422198e-01 3.00912648e-01
9.26793963e-02 3.48430961e-01 7.98755348e-01 -9.46355760e-01
-5.86222768e-01 -3.07432622e-01 -3.51810426e-01 -5.89277185e-02
-7.27371797e-02 9.90420878e-02 -4.93613571e-01 -7.74043143e-01
-1.08929515e-01 -9.78008151e-01 -4.85134602e-01 -7.48441219e-02
-4.14941907e-01 -2.21716717e-01 8.87673855e-01 -8.48087013e-01
1.03085852e+00 -1.78847206e+00 -2.49616101e-01 2.97861457e-01
7.47136116e-01 7.02520013e-01 2.52162248e-01 9.18525681e-02
-3.84199589e-01 2.01499104e-01 -2.80149192e-01 -3.66124101e-02
-4.91737753e-01 4.58996356e-01 2.62664020e-01 4.75793689e-01
3.18101346e-01 7.98893571e-01 -4.90568042e-01 -7.85669923e-01
4.45940107e-01 3.87711942e-01 -3.87362927e-01 5.43874323e-01
-2.28254691e-01 6.11841798e-01 -8.17241132e-01 7.97831416e-01
6.22861445e-01 -4.53968942e-01 -9.84691232e-02 2.11507156e-02
2.51576841e-01 -6.84854165e-02 -5.81753671e-01 1.33251512e+00
-3.97230476e-01 5.07289708e-01 -1.70257583e-01 -1.09481299e+00
6.99636519e-01 7.83334315e-01 9.82751787e-01 -6.33694828e-01
5.73791981e-01 2.50349283e-01 2.98668206e-01 -1.08042717e+00
1.01342790e-01 -3.01154375e-01 3.35352942e-02 7.32675314e-01
-1.67330250e-01 -2.00925305e-01 6.48187175e-02 -8.29364210e-02
1.16458464e+00 -6.75513625e-01 3.35599363e-01 6.39366359e-02
5.24620831e-01 2.92522848e-01 3.46419066e-01 6.72238648e-01
-2.62944728e-01 7.57447124e-01 2.24150732e-01 -1.02333391e+00
-1.12565958e+00 -1.04499185e+00 -8.17436337e-01 4.90274012e-01
-3.99965972e-01 3.78234079e-04 -4.33400482e-01 -9.42631900e-01
-1.45766966e-03 3.77913147e-01 -6.84319794e-01 2.16468811e-01
-6.77893877e-01 -1.01080024e+00 8.95618141e-01 8.47715974e-01
4.09062862e-01 -1.14726532e+00 -9.97962773e-01 3.33071917e-01
-1.79054111e-01 -8.35285723e-01 2.64132529e-01 3.84367079e-01
-8.47804129e-01 -1.41980731e+00 -1.10403168e+00 -7.71428406e-01
3.64420563e-01 -1.13816120e-01 1.24630213e+00 5.14525950e-01
-8.36699486e-01 8.83115828e-02 -4.66210008e-01 -8.91490102e-01
-6.03749275e-01 2.30248868e-01 1.00802131e-01 -6.15053594e-01
2.48550639e-01 -2.99454957e-01 -5.13471663e-01 1.92890212e-01
-9.38090324e-01 9.40410979e-03 7.19234407e-01 1.09551537e+00
8.27965915e-01 3.02343480e-02 4.47353184e-01 -1.09168422e+00
6.77513719e-01 -9.38565075e-01 -3.97274166e-01 1.57248318e-01
-5.81889868e-01 -3.04484099e-01 4.76870984e-01 2.20769979e-02
-8.36294293e-01 -1.72457248e-02 -7.67187536e-01 -7.38273084e-01
-2.66067475e-01 8.27595711e-01 5.63014627e-01 1.95735976e-01
7.03145564e-01 1.23020835e-01 -2.82225847e-01 -5.41322112e-01
-1.76156223e-01 9.73398745e-01 4.09880221e-01 -1.37718871e-01
4.47204560e-01 2.40941867e-01 1.53039411e-01 -6.37985051e-01
-7.48766065e-01 -4.54390734e-01 -8.54858577e-01 -1.38807863e-01
1.34990215e+00 -8.54157865e-01 -4.38386083e-01 3.31191838e-01
-9.69045103e-01 1.41933829e-01 -7.25704879e-02 7.06679165e-01
-2.91080117e-01 -1.01000899e-02 -7.93616951e-01 -3.29648226e-01
-8.36565256e-01 -1.41792500e+00 9.02717829e-01 -8.62689614e-02
-1.58886358e-01 -1.08871496e+00 3.90791357e-01 4.08977151e-01
4.77517247e-01 5.94371974e-01 1.21513426e+00 -1.08274341e+00
-3.75986576e-01 -6.08159900e-01 -3.20701510e-01 2.90990263e-01
1.67950038e-02 -4.96937409e-02 -8.80952656e-01 -1.82543740e-01
1.43150672e-01 -6.40522957e-01 8.41996133e-01 6.48496091e-01
1.73624039e+00 8.31509903e-02 -4.43676502e-01 7.09033132e-01
1.50650191e+00 4.77120340e-01 2.56626815e-01 1.17398821e-01
8.11294496e-01 5.66371577e-03 5.00598252e-01 6.25836432e-01
2.84510672e-01 2.12964371e-01 7.19335377e-01 -4.13890272e-01
4.51469868e-02 2.17614785e-01 -5.27320921e-01 1.01342666e+00
-3.89529377e-01 -4.85879332e-01 -1.58772850e+00 5.95034242e-01
-1.31754494e+00 -5.54880261e-01 -5.52965581e-01 1.57554853e+00
5.49412608e-01 -1.32649094e-02 1.18528552e-01 2.67003655e-01
4.16711450e-01 -1.31625265e-01 -5.44191420e-01 -4.47915941e-01
2.64854342e-01 6.20954275e-01 2.60402888e-01 -3.04195553e-01
-1.29215038e+00 4.04097885e-01 7.03785944e+00 6.09760463e-01
-1.52325535e+00 3.48266393e-01 6.04123414e-01 -1.76009871e-02
2.13112161e-02 -8.72676492e-01 -4.40393358e-01 3.38928908e-01
9.37300742e-01 2.18445718e-01 -2.67751813e-01 1.03003001e+00
1.97643906e-01 1.04952017e-02 -1.03336024e+00 1.00962293e+00
-1.36317939e-01 -1.48754275e+00 -3.42903621e-02 1.18397195e-02
4.93587792e-01 6.54718459e-01 4.34514843e-02 1.67242333e-01
3.18222463e-01 -1.33093321e+00 1.77302673e-01 2.43887752e-01
1.22698629e+00 -7.13665783e-01 1.42155385e+00 3.72766703e-01
-1.02523077e+00 1.08434953e-01 -4.82852012e-01 4.53690261e-01
1.06890211e-02 5.04704416e-01 -1.55966294e+00 8.51084352e-01
9.68705416e-01 7.40827262e-01 -5.64930081e-01 9.60329652e-01
-5.22200987e-02 7.93552577e-01 -1.37838721e-01 -1.35905385e-01
2.40937874e-01 2.22054318e-01 3.02569181e-01 1.28864682e+00
3.85402828e-01 2.40679994e-01 2.37183958e-01 7.40558147e-01
-7.10098892e-02 1.70434475e-01 -8.41776550e-01 -5.36191761e-02
9.36990306e-02 1.10409307e+00 -7.57377088e-01 -4.73422289e-01
-3.80335093e-01 4.57554400e-01 -5.64605556e-02 -1.69427678e-01
-1.08046770e+00 1.95287690e-02 1.32803604e-01 1.16286300e-01
2.48104990e-01 1.44226223e-01 -2.92467088e-01 -9.52910125e-01
-2.91713685e-01 -1.00534320e+00 9.06418264e-01 -8.07486296e-01
-1.51165247e+00 9.74554121e-01 9.51841101e-02 -1.30984092e+00
-6.19384229e-01 -7.96481013e-01 -5.23157477e-01 7.43943214e-01
-1.26887059e+00 -1.19601083e+00 -5.62448144e-01 8.73768687e-01
4.04205054e-01 -4.25632566e-01 1.16709745e+00 4.24588442e-01
-6.60235167e-01 5.20694554e-01 1.64443552e-01 6.21964216e-01
4.15729225e-01 -1.13871610e+00 1.85223132e-01 3.95562440e-01
-1.30128130e-01 1.25099868e-01 1.82498515e-01 -6.93513870e-01
-1.08971393e+00 -1.53064907e+00 7.24515975e-01 -3.75939369e-01
2.66498178e-01 1.20317280e-01 -8.36600482e-01 7.33833611e-01
-2.19200309e-02 2.29784861e-01 8.31446052e-01 -4.09869224e-01
1.63708135e-01 1.53667778e-01 -1.50942874e+00 6.62191734e-02
7.07784951e-01 -2.47913480e-01 -6.68821990e-01 5.12280166e-01
6.29768908e-01 -6.60037637e-01 -1.31453240e+00 6.03626370e-01
4.18201566e-01 -7.78358281e-01 1.05209506e+00 -7.50969172e-01
7.40711868e-01 1.25510097e-01 2.89774667e-02 -1.20692170e+00
-2.25173578e-01 4.08559799e-01 1.35759845e-01 4.64109391e-01
3.57229263e-01 -4.52410668e-01 8.27031910e-01 2.08675727e-01
-7.02550113e-02 -1.19324088e+00 -9.20465231e-01 -4.07102704e-01
3.00606281e-01 -8.15390468e-01 6.64229453e-01 1.07939887e+00
-6.96654975e-01 -7.50729218e-02 -1.59452170e-01 -7.89261833e-02
2.21710861e-01 1.08704790e-01 4.43994939e-01 -1.24104691e+00
-3.02632958e-01 -2.26554081e-01 -4.86407906e-01 -3.46282810e-01
-3.91668111e-01 -8.70224059e-01 -2.80214995e-01 -1.63610244e+00
4.29746091e-01 -8.20738256e-01 -2.66103506e-01 5.35611391e-01
1.60216495e-01 3.74605000e-01 -2.79335547e-02 2.54770607e-01
-8.29038173e-02 1.58648863e-01 1.49411333e+00 -1.82832316e-01
1.49761423e-01 2.95869052e-01 -4.45092916e-02 7.91869760e-01
9.04007316e-01 -7.95158446e-01 -3.58183086e-01 -3.92344058e-01
1.34916017e-02 6.85597003e-01 1.86753094e-01 -1.11970270e+00
-1.08227536e-01 -6.14287443e-02 8.68966997e-01 -1.33475626e+00
1.85472012e-01 -1.08359516e+00 2.21727744e-01 1.12023342e+00
-2.47202262e-01 3.27592909e-01 1.25543505e-01 4.20924217e-01
-3.96653980e-01 -3.14807355e-01 8.31070423e-01 -5.23052573e-01
-4.24393535e-01 7.94543922e-01 -7.79536963e-01 1.14589900e-01
1.20611191e+00 -1.55721642e-02 1.16355233e-01 -2.40311742e-01
-8.64720702e-01 1.28371999e-01 -1.12422211e-02 1.15244038e-01
9.10503268e-01 -1.25141430e+00 -1.05992091e+00 3.69899511e-01
1.86189309e-01 6.12161636e-01 3.39263290e-01 9.61367905e-01
-1.34263146e+00 7.11354375e-01 -4.55011725e-01 -9.97683585e-01
-1.48743391e+00 6.41945660e-01 5.51696837e-01 -8.71165514e-01
-7.32841015e-01 9.28958178e-01 4.90448624e-02 -2.97847688e-01
3.19900036e-01 -6.00212157e-01 -4.71667618e-01 -9.05093774e-02
2.82264084e-01 -3.39556374e-02 4.46024626e-01 -4.91488069e-01
-4.69554991e-01 2.62492806e-01 -2.59068131e-01 2.47386634e-01
1.78048503e+00 2.77803659e-01 4.37276736e-02 1.90104797e-01
1.27685487e+00 -5.34125090e-01 -2.95870781e-01 -1.69846967e-01
-1.81984290e-01 -2.55626291e-01 1.38388976e-01 -9.78768945e-01
-1.38832581e+00 1.12084007e+00 1.11940861e+00 2.20357589e-02
1.25216377e+00 2.68355235e-02 1.00888503e+00 4.21576649e-01
1.95106983e-01 -6.14310145e-01 1.48740351e-01 3.95229965e-01
7.99690366e-01 -1.42093933e+00 3.36682536e-02 -5.51752709e-02
-6.78349793e-01 1.20028353e+00 4.32318270e-01 -8.92772432e-03
8.63653243e-01 3.03352267e-01 2.56362826e-01 -6.71548605e-01
-7.46067226e-01 2.73448462e-03 2.43556827e-01 6.95857525e-01
5.48580766e-01 4.37656373e-01 -1.72117099e-01 8.19534838e-01
-1.92353532e-01 2.11802706e-01 3.52109224e-01 9.30273056e-01
-1.53021529e-01 -1.12032223e+00 -4.40435588e-01 1.07192922e+00
-8.97720158e-01 -2.89680380e-02 -3.90174687e-01 1.33096206e+00
5.61338842e-01 4.44551289e-01 -5.03395647e-02 -4.88218755e-01
1.07449301e-01 -1.63073316e-01 2.78636962e-01 -5.37126422e-01
-7.23395169e-01 -6.63945153e-02 1.32857904e-01 -2.02424705e-01
-3.47510695e-01 -5.93547404e-01 -1.21995473e+00 -3.40192556e-01
-3.17729264e-01 -1.34492382e-01 6.11508310e-01 8.54227543e-01
-1.28174797e-01 8.37260365e-01 5.49140334e-01 -2.81377017e-01
-4.75544870e-01 -1.06292808e+00 -6.05122924e-01 2.92244554e-01
4.50000674e-01 -7.11770236e-01 1.82359830e-01 -1.66962873e-02] | [15.362744331359863, -1.876570701599121] |
dc2d8a21-4a35-48d0-b663-15b6ec82819d | towards-unsupervised-speech-recognition-and | 1910.12729 | null | https://arxiv.org/abs/1910.12729v2 | https://arxiv.org/pdf/1910.12729v2.pdf | Towards Unsupervised Speech Recognition and Synthesis with Quantized Speech Representation Learning | In this paper we propose a Sequential Representation Quantization AutoEncoder (SeqRQ-AE) to learn from primarily unpaired audio data and produce sequences of representations very close to phoneme sequences of speech utterances. This is achieved by proper temporal segmentation to make the representations phoneme-synchronized, and proper phonetic clustering to have total number of distinct representations close to the number of phonemes. Mapping between the distinct representations and phonemes is learned from a small amount of annotated paired data. Preliminary experiments on LJSpeech demonstrated the learned representations for vowels have relative locations in latent space in good parallel to that shown in the IPA vowel chart defined by linguistics experts. With less than 20 minutes of annotated speech, our method outperformed existing methods on phoneme recognition and is able to synthesize intelligible speech that beats our baseline model. | ['Lin-shan Lee', 'Hung-Yi Lee', 'Tao Tu', 'Alexander H. Liu'] | 2019-10-28 | null | null | null | null | ['unsupervised-speech-recognition'] | ['speech'] | [ 2.10003942e-01 5.83831906e-01 -5.78095131e-02 -6.22113705e-01
-1.07760692e+00 -8.17712188e-01 4.75318253e-01 -2.66904272e-02
-1.15462482e-01 5.98214626e-01 6.03042901e-01 -4.16464716e-01
1.75568268e-01 -4.63910103e-01 -5.03240585e-01 -4.17877585e-01
-3.45452093e-02 7.64136672e-01 3.99930403e-02 -2.16719374e-01
2.87439674e-02 2.42390692e-01 -1.84660959e+00 6.20214283e-01
4.89832699e-01 6.28377676e-01 4.49260056e-01 1.22829270e+00
-4.03304338e-01 4.90703344e-01 -8.91804636e-01 2.94309780e-02
2.71953195e-01 -7.33790100e-01 -8.63760114e-01 2.79608220e-01
4.10863519e-01 -2.97927231e-01 -1.69416681e-01 9.27644491e-01
3.71116132e-01 4.85806912e-01 1.03646374e+00 -9.95341539e-01
-7.73127437e-01 1.13897729e+00 -1.91826206e-02 1.08745612e-01
3.87313008e-01 -4.08560514e-01 1.12784684e+00 -8.68207216e-01
3.88253063e-01 1.47789276e+00 4.91614163e-01 5.57237744e-01
-1.22487378e+00 -5.68041384e-01 -1.65043354e-01 3.74089256e-02
-1.42564929e+00 -8.16721439e-01 5.41461647e-01 -4.89395291e-01
1.43111706e+00 3.13053310e-01 3.42637509e-01 1.14568961e+00
-1.88195780e-01 4.10808712e-01 7.00514972e-01 -8.63942206e-01
2.49252513e-01 2.92772919e-01 1.42974174e-02 3.34032178e-01
-5.49609423e-01 9.05143917e-02 -3.83681267e-01 1.05492719e-01
9.82373834e-01 -3.54130745e-01 -2.16683611e-01 6.43974170e-02
-1.07809281e+00 9.59020376e-01 -7.52098858e-02 6.59543812e-01
-4.20235515e-01 5.19015230e-02 4.98131275e-01 3.14798862e-01
4.07517180e-02 2.70196706e-01 -5.25213599e-01 -4.62256163e-01
-1.04421163e+00 -1.36495933e-01 7.59515107e-01 1.02536654e+00
5.67493737e-01 8.16708207e-01 2.58375853e-01 1.06804740e+00
2.24189118e-01 1.53647959e-01 1.05223691e+00 -1.40132570e+00
4.90522325e-01 -1.22482769e-01 -1.01105586e-01 -7.41421521e-01
1.97263524e-01 -1.24958545e-01 -6.83161259e-01 2.08485022e-01
3.53010595e-01 -7.26606622e-02 -9.99467254e-01 1.59518659e+00
-1.48756877e-01 2.53919601e-01 5.92171490e-01 4.87776399e-01
6.57178640e-01 1.40822458e+00 -1.73184872e-01 -4.79513854e-01
1.21735489e+00 -8.28795850e-01 -1.12977350e+00 -1.06486566e-01
3.67091060e-01 -8.71901214e-01 1.19693410e+00 6.07841849e-01
-1.05636311e+00 -1.18483222e+00 -1.16642690e+00 -1.00315049e-01
-2.89406627e-01 2.19857529e-01 3.53889525e-01 7.83542454e-01
-1.12087762e+00 5.60767949e-01 -7.64101088e-01 -1.04375251e-01
-1.09933369e-01 4.38670725e-01 -4.02980149e-01 4.70373333e-01
-1.18208218e+00 5.83569288e-01 9.06179905e-01 -5.15169613e-02
-1.18136871e+00 -3.29009622e-01 -1.09091246e+00 1.71130121e-01
-2.72201478e-01 -4.87747677e-02 1.48693871e+00 -1.16275680e+00
-1.94160974e+00 5.40586233e-01 -2.91244358e-01 -7.46482730e-01
-1.60980925e-01 -1.51501849e-01 -7.88068652e-01 4.27375406e-01
-7.91728497e-02 1.12364721e+00 1.12716603e+00 -1.28301895e+00
-7.16067076e-01 1.70808822e-01 -5.40129185e-01 3.29383105e-01
-3.90051305e-01 4.55785580e-02 -2.24146530e-01 -7.32318282e-01
1.83905959e-01 -6.33121848e-01 -1.53418317e-01 -6.51710629e-01
-2.25292400e-01 -5.02314746e-01 8.05366635e-01 -9.12063956e-01
1.17286849e+00 -2.34503293e+00 8.18722472e-02 4.12966087e-02
-1.86715066e-01 1.58119313e-02 -7.59548992e-02 5.77764988e-01
-3.62822175e-01 3.22598815e-02 -2.95121461e-01 -5.45495331e-01
1.18152341e-02 7.29722202e-01 -8.80662680e-01 1.46140009e-01
5.37123047e-02 4.47540760e-01 -8.69072139e-01 -4.28734094e-01
2.75285751e-01 6.42751038e-01 -5.02342105e-01 4.38385665e-01
-2.52283439e-02 3.31745744e-01 2.87135333e-01 1.88471004e-01
2.35628948e-01 4.07650799e-01 2.26304114e-01 9.90818515e-02
-7.68433064e-02 9.70540345e-01 -1.32123709e+00 1.62869132e+00
-5.40288210e-01 8.39541197e-01 1.66493312e-01 -9.90804315e-01
1.27662849e+00 1.12959361e+00 2.13880956e-01 -2.10428193e-01
-2.07930058e-02 1.99431121e-01 1.04925677e-01 -1.73773795e-01
6.08011186e-01 -4.25712913e-01 -1.38776258e-01 3.73460501e-01
6.16387904e-01 -5.08822620e-01 -1.07335396e-01 -2.84943474e-03
4.34606284e-01 -1.08254634e-01 2.94242948e-01 -7.51754940e-02
2.28652835e-01 -2.20995307e-01 6.53548837e-01 5.36189973e-01
-1.82873085e-01 9.05797303e-01 8.15399960e-02 -5.35296686e-02
-1.07912040e+00 -1.41861689e+00 -2.04869449e-01 1.20955777e+00
-4.35024112e-01 -6.30156159e-01 -9.55071330e-01 -2.62941957e-01
-4.95580167e-01 9.07744229e-01 -3.34188432e-01 8.01173821e-02
-6.64755344e-01 -5.72810275e-03 7.72555709e-01 8.14746261e-01
-1.38161197e-01 -1.45512164e+00 -2.03621477e-01 6.36402726e-01
-1.82126597e-01 -1.02782643e+00 -3.61466259e-01 8.51686835e-01
-8.30394208e-01 -5.65086484e-01 -6.38669133e-01 -1.46407342e+00
2.92343557e-01 -2.29293957e-01 1.07630527e+00 -6.03683293e-01
-2.30781138e-02 -5.82705587e-02 -5.91567636e-01 -3.18881422e-01
-8.31236184e-01 2.83701695e-03 4.60177481e-01 -1.25187844e-01
3.46976399e-01 -8.07571769e-01 -6.59211800e-02 -3.28758694e-02
-7.46988833e-01 -3.86458486e-01 2.03785554e-01 7.95992792e-01
7.74719715e-01 5.19797742e-01 7.63954937e-01 -5.57266355e-01
7.08143353e-01 -4.17100728e-01 -3.40483576e-01 -3.04687351e-01
-1.37575731e-01 -1.96406655e-02 8.77213061e-01 -4.56841737e-01
-9.36026514e-01 4.26429480e-01 -4.93509740e-01 -5.40622234e-01
-6.86225414e-01 2.47667626e-01 -1.96144357e-01 8.35012555e-01
6.74522102e-01 1.91714704e-01 -7.04874517e-03 -5.70189595e-01
8.95255744e-01 1.21186411e+00 1.08485353e+00 -3.47732961e-01
4.01078403e-01 -1.11343428e-01 -8.84069741e-01 -1.17853463e+00
-2.02001333e-01 -4.21908826e-01 -8.06726813e-01 2.38102436e-01
9.71817493e-01 -9.45942163e-01 -3.50264341e-01 -8.13755095e-02
-1.20734811e+00 -1.13275595e-01 -6.68004274e-01 8.98338616e-01
-8.69649231e-01 3.17420185e-01 -8.59332800e-01 -1.04125643e+00
-1.37154594e-01 -1.04768682e+00 9.69641030e-01 -4.36568595e-02
-8.13485444e-01 -9.17418122e-01 3.03130835e-01 1.50623351e-01
5.55719435e-03 -1.40152961e-01 9.02250707e-01 -8.67798626e-01
-8.17287266e-02 2.81106345e-02 4.98656750e-01 9.22855973e-01
6.20752752e-01 1.82366192e-01 -1.41106987e+00 -1.83148205e-01
8.24174061e-02 -4.36138779e-01 7.89905846e-01 6.06017113e-01
1.06159365e+00 -3.76420289e-01 4.84639173e-03 4.24369246e-01
9.75610971e-01 8.63760889e-01 5.43776751e-01 -2.50150293e-01
4.78414834e-01 8.43624532e-01 3.52518916e-01 2.40353242e-01
1.91192031e-02 3.65540832e-01 1.41071528e-01 1.43358395e-01
-2.41420254e-01 -4.74539191e-01 5.98430753e-01 2.01940942e+00
2.95390576e-01 -2.75060385e-01 -7.87505269e-01 9.89301682e-01
-1.45660675e+00 -1.00246453e+00 1.74995348e-01 1.93629944e+00
9.00503278e-01 3.24742258e-01 2.41690457e-01 6.95368588e-01
7.83686817e-01 8.65766704e-02 6.51238114e-02 -1.31766474e+00
5.42338984e-03 5.58891118e-01 5.38516119e-02 9.79173839e-01
-9.08986986e-01 1.14559901e+00 8.07289886e+00 9.16128933e-01
-7.66587555e-01 1.45135015e-01 4.16903168e-01 2.45273218e-01
-5.38221121e-01 -1.34264916e-01 -8.26712310e-01 3.42171907e-01
1.64789069e+00 1.63416266e-01 5.22838533e-01 8.81563365e-01
1.07612245e-01 4.20297593e-01 -1.23995852e+00 1.10025585e+00
-1.46121169e-02 -1.26502025e+00 1.85121730e-01 -1.45539880e-01
7.00191140e-01 -1.40271276e-01 2.17243165e-01 4.75397736e-01
7.04022527e-01 -1.44175422e+00 7.50275910e-01 1.32316351e-01
9.34319854e-01 -1.04385889e+00 4.75907326e-01 3.61581624e-01
-1.52277660e+00 1.52468532e-02 -5.21497428e-01 -2.17306674e-01
1.86493561e-01 7.73314461e-02 -1.62629604e+00 3.09049428e-01
5.62199175e-01 5.91564119e-01 -4.70554158e-02 3.88843775e-01
-1.39863655e-01 1.14380395e+00 -3.08402836e-01 2.11547568e-01
3.71817797e-01 -1.74206212e-01 1.88083768e-01 1.33037913e+00
4.60088104e-01 3.44087444e-02 1.28174633e-01 6.02986455e-01
1.34750053e-01 1.95298001e-01 -8.07928979e-01 -5.12364447e-01
1.02386153e+00 7.18230009e-01 -5.92850804e-01 -5.86519480e-01
-5.88863529e-02 1.18651259e+00 5.56245260e-02 4.29143846e-01
-3.84602696e-01 -7.24145114e-01 7.55865633e-01 -3.63808692e-01
6.20382726e-01 -4.19243604e-01 -1.75890904e-02 -6.24819517e-01
-3.02813739e-01 -9.34499860e-01 8.39357004e-02 -6.75904751e-01
-1.27194107e+00 1.19278729e+00 -1.02354683e-01 -1.35228968e+00
-1.30581892e+00 -5.57185054e-01 -6.32339537e-01 1.07995498e+00
-8.54260862e-01 -6.45773828e-01 5.94572723e-01 5.41782498e-01
1.03561509e+00 -6.09913945e-01 1.57258856e+00 1.08786924e-02
-1.94021866e-01 5.21123528e-01 2.46133149e-01 3.36590290e-01
4.08879548e-01 -1.61866117e+00 6.53741717e-01 6.94357216e-01
9.97263432e-01 6.76497698e-01 6.97414041e-01 -3.33313525e-01
-5.93163073e-01 -8.29092681e-01 1.10697901e+00 -5.62133014e-01
3.79465163e-01 -6.03109479e-01 -1.00098622e+00 8.62213910e-01
6.72562182e-01 -1.56484246e-01 1.04939747e+00 5.66449985e-02
-3.23389441e-01 -6.83079809e-02 -7.15715468e-01 2.90826827e-01
6.21751070e-01 -1.29037535e+00 -1.36025119e+00 1.18512966e-01
1.09885430e+00 -1.40363947e-01 -8.16110134e-01 -5.91823310e-02
4.14380193e-01 -7.54063427e-01 8.70244801e-01 -4.53658879e-01
1.89320758e-01 -3.70410860e-01 -6.32216334e-01 -1.39726424e+00
-1.63975790e-01 -8.24093461e-01 -2.78377831e-02 1.59881747e+00
4.31193799e-01 -1.79900140e-01 6.33409977e-01 -3.00024122e-01
-5.23895264e-01 -2.23431170e-01 -1.00765514e+00 -8.77792537e-01
1.51435122e-01 -7.16853321e-01 4.89453822e-01 9.41328347e-01
1.34521183e-02 4.48021322e-01 -2.55427599e-01 3.96439999e-01
2.97342569e-01 1.19832605e-01 3.11844170e-01 -1.05987883e+00
-7.43921518e-01 -2.34542936e-01 -3.72369528e-01 -1.21000051e+00
4.14247572e-01 -9.44068909e-01 4.92208928e-01 -1.30797267e+00
-6.58197522e-01 -2.14165822e-01 -4.63944763e-01 4.36997592e-01
2.03229576e-01 1.55320928e-01 4.40597394e-03 -9.54635069e-02
4.65473980e-02 6.02127850e-01 7.14446843e-01 -1.53652713e-01
-4.61769938e-01 -1.57001361e-01 -4.34539139e-01 7.33447909e-01
8.05259228e-01 -4.83464599e-01 -8.48905385e-01 -3.04507673e-01
-3.76973927e-01 3.71643543e-01 -3.02365869e-01 -1.08485210e+00
-6.09342987e-03 1.90766796e-01 4.59444165e-01 -9.29034293e-01
8.18073630e-01 -6.66409373e-01 9.00301412e-02 1.07037447e-01
-6.25483394e-01 8.02651644e-02 1.93659723e-01 4.35697854e-01
-8.89275074e-01 -4.18631673e-01 4.40064758e-01 -1.83173165e-01
-8.14733624e-01 -2.26503298e-01 -1.00073850e+00 -1.64226949e-01
5.66799939e-01 -5.07980227e-01 3.62572551e-01 -6.58080220e-01
-1.09886503e+00 -2.37788498e-01 -9.65096895e-03 6.37556493e-01
7.08677530e-01 -1.56402922e+00 -5.58656573e-01 6.16959572e-01
-1.44917339e-01 -1.06428102e-01 -1.27531081e-01 -1.62591398e-01
-4.61194634e-01 6.06727719e-01 -3.93736273e-01 -5.88813245e-01
-1.17017758e+00 4.35760111e-01 1.37881309e-01 2.98825264e-01
-6.50483608e-01 1.17168868e+00 1.07335776e-01 -7.26227045e-01
7.31792033e-01 -7.78395832e-01 -5.29802561e-01 2.76936859e-01
4.97487634e-01 5.86020276e-02 -2.01843809e-02 -8.53572488e-01
-2.88666904e-01 3.56046796e-01 -8.98882523e-02 -8.11177373e-01
1.19032240e+00 -2.30435669e-01 2.47718677e-01 1.12681556e+00
1.39357293e+00 3.48707825e-01 -1.43692470e+00 3.98615152e-01
-1.05642155e-02 -4.20183539e-02 -5.15042990e-02 -3.52468997e-01
-5.92684925e-01 1.29390502e+00 6.65080965e-01 3.93822879e-01
9.45601404e-01 3.67904790e-02 9.56797838e-01 3.03299993e-01
5.94250374e-02 -1.15147114e+00 1.57835692e-01 6.41229510e-01
9.31450367e-01 -8.73730779e-01 -6.78658545e-01 -3.94845068e-01
-8.63602161e-01 1.26952040e+00 2.55222976e-01 -3.18031609e-01
6.20662808e-01 3.35647225e-01 4.29478765e-01 2.47932479e-01
-7.36807108e-01 -1.07873574e-01 2.29452655e-01 1.12033367e+00
5.82065225e-01 3.05379897e-01 3.89170974e-01 7.25936532e-01
-9.62135613e-01 -5.87357044e-01 4.62748826e-01 6.22115850e-01
-7.18810916e-01 -1.28323364e+00 -4.42650676e-01 -3.00059468e-02
-3.80451411e-01 -2.30024248e-01 -3.12812716e-01 5.37606120e-01
2.93596059e-01 1.17647696e+00 7.56940365e-01 -4.43533868e-01
3.14027071e-02 4.73990828e-01 1.71291605e-01 -1.04968333e+00
-4.80756938e-01 6.08635962e-01 1.51813388e-01 -3.15143675e-01
-1.19274937e-01 -4.61332262e-01 -1.66221523e+00 4.09567654e-01
-8.82198513e-02 7.03193069e-01 5.91971517e-01 8.16001594e-01
4.80118841e-02 6.58851624e-01 7.80331254e-01 -9.21308637e-01
-5.54087818e-01 -1.11743689e+00 -8.57764840e-01 1.86879337e-01
5.89195549e-01 -1.59017086e-01 -4.20422971e-01 6.18280590e-01] | [14.629858016967773, 6.644370079040527] |
dac530a2-805f-4833-90ef-dc70b9710e23 | knowledge-acquisition-and-completion-for-long | 2301.06834 | null | https://arxiv.org/abs/2301.06834v1 | https://arxiv.org/pdf/2301.06834v1.pdf | Knowledge Acquisition and Completion for Long-Term Human-Robot Interactions using Knowledge Graph Embedding | In Human-Robot Interaction (HRI) systems, a challenging task is sharing the representation of the operational environment, fusing symbolic knowledge and perceptions, between users and robots. With the existing HRI pipelines, users can teach the robots some concepts to increase their knowledge base. Unfortunately, the data coming from the users are usually not enough dense for building a consistent representation. Furthermore, the existing approaches are not able to incrementally build up their knowledge base, which is very important when robots have to deal with dynamic contexts. To this end, we propose an architecture to gather data from users and environments in long-runs of continual learning. We adopt Knowledge Graph Embedding techniques to generalize the acquired information with the goal of incrementally extending the robot's inner representation of the environment. We evaluate the performance of the overall continual learning architecture by measuring the capabilities of the robot of learning entities and relations coming from unknown contexts through a series of incremental learning sessions. | ['D. Nardi', 'V. Suriani', 'F. Argenziano', 'E. Bartoli'] | 2023-01-17 | null | null | null | null | ['knowledge-graph-embedding'] | ['graphs'] | [-2.18519464e-01 5.86113453e-01 1.53895840e-01 -4.04069424e-01
-8.65616743e-03 -5.99470317e-01 4.69620615e-01 6.57893181e-01
-4.57911819e-01 6.24780059e-01 6.12725616e-02 3.15251164e-02
-1.76369205e-01 -8.59202802e-01 -8.98779929e-01 -6.04868717e-02
-3.91103655e-01 8.69795978e-01 4.91410255e-01 -5.59845924e-01
-1.97015777e-01 4.30599749e-01 -1.69445121e+00 1.45940885e-01
5.29813290e-01 5.19312441e-01 6.97822154e-01 4.29700822e-01
-1.32476643e-01 1.10854840e+00 -1.02576010e-01 -1.95678174e-02
1.07633904e-01 -3.04697491e-02 -1.17853677e+00 -3.17913443e-02
5.65106375e-03 -4.89699841e-01 -4.56235737e-01 8.31266582e-01
-4.51320484e-02 6.18506849e-01 3.86667401e-01 -1.63865066e+00
-7.88517594e-01 1.13483441e+00 1.52237684e-01 -2.97918499e-01
7.47139692e-01 6.45762086e-02 8.88766289e-01 -9.57906425e-01
9.13554013e-01 1.47497940e+00 4.67429489e-01 4.75885719e-01
-9.74035382e-01 -2.71632373e-01 5.15014172e-01 6.64326012e-01
-1.51629388e+00 -2.23307416e-01 6.81249082e-01 -4.19783920e-01
1.06831825e+00 -1.06088690e-01 6.49303079e-01 9.23651874e-01
-4.91544634e-01 8.09197366e-01 4.34323728e-01 -4.10535425e-01
3.55876952e-01 7.25274026e-01 4.90873277e-01 6.26258910e-01
2.44130760e-01 -1.28017172e-01 -6.55666292e-01 2.30596438e-01
6.71581328e-01 3.09909344e-01 -5.51345088e-02 -8.94850194e-01
-1.16320932e+00 3.16851199e-01 9.29433107e-01 7.26284206e-01
-4.44484204e-01 8.09829906e-02 2.44170040e-01 6.16043091e-01
-3.12897325e-01 8.17050755e-01 -5.59081495e-01 -1.59806788e-01
8.36397186e-02 7.40169287e-02 9.66531336e-01 1.50341356e+00
1.26617169e+00 -4.63730961e-01 4.96342361e-01 5.19587159e-01
2.77358770e-01 2.00534821e-01 6.40767574e-01 -9.31387603e-01
2.25064546e-01 9.88892674e-01 3.65939200e-01 -9.47889030e-01
-4.71139818e-01 1.10908724e-01 -2.77695358e-01 1.46752317e-03
3.17177456e-03 -2.07570150e-01 -5.35988390e-01 1.70881522e+00
4.45702076e-01 2.76773423e-01 4.87220228e-01 5.44982612e-01
6.02377713e-01 3.65086347e-01 1.10111848e-01 1.83094501e-01
1.01745510e+00 -1.07256222e+00 -4.21460658e-01 -5.44158757e-01
9.35595870e-01 9.41547230e-02 9.53973651e-01 2.37886384e-01
-5.25545716e-01 -1.03882682e+00 -9.90109146e-01 -1.94263637e-01
-8.11441243e-01 -2.22146809e-01 6.05706096e-01 1.58809442e-02
-1.05163991e+00 7.31258273e-01 -8.28808784e-01 -7.90258646e-01
4.75938544e-02 4.75243866e-01 -7.81547368e-01 -4.15893078e-01
-1.05525053e+00 1.17768800e+00 9.56292510e-01 1.62733898e-01
-1.05090714e+00 -3.41081530e-01 -1.20083189e+00 4.44667749e-02
7.46521235e-01 -4.99838442e-01 1.32255197e+00 -6.36313617e-01
-1.29822373e+00 1.70873225e-01 3.17440718e-01 -2.65194029e-01
1.05470136e-01 -5.13566494e-01 -3.90708335e-02 -2.16266453e-01
-1.37609601e-01 6.79786205e-01 4.93721873e-01 -1.55841315e+00
-8.16343784e-01 -6.82475507e-01 8.13052475e-01 5.81627965e-01
-4.33631539e-01 -4.92220521e-01 -4.81001318e-01 3.25719297e-01
1.53779373e-01 -1.26989067e+00 -4.46255118e-01 -2.70258218e-01
8.08719918e-03 -3.22065681e-01 1.03767705e+00 -4.36326146e-01
6.35238528e-01 -2.36979008e+00 6.45548820e-01 2.73112446e-01
1.63074210e-01 6.32214472e-02 -3.54317486e-01 5.90942562e-01
3.54782231e-02 -1.58591256e-01 2.83382565e-01 -4.65095401e-01
7.49506876e-02 6.54345334e-01 -1.70473665e-01 -1.82479143e-01
-3.72398235e-02 9.41574872e-01 -1.42311645e+00 -1.40656337e-01
4.06968027e-01 3.31989557e-01 -4.88778025e-01 5.82049489e-01
-3.70373964e-01 5.34049451e-01 -3.92432064e-01 6.45685866e-02
9.23656374e-02 -2.26102814e-01 8.00100982e-01 -1.33304492e-01
1.80678010e-01 1.04140706e-01 -1.32505465e+00 2.09867930e+00
-7.52844810e-01 4.89769846e-01 -7.09399059e-02 -8.88577461e-01
7.45776296e-01 3.03872406e-01 3.73365194e-01 -4.21163261e-01
-9.62650850e-02 -1.02811530e-01 -1.41396716e-01 -6.30015373e-01
7.82262206e-01 1.26294538e-01 -2.38063455e-01 3.19998384e-01
4.97904986e-01 -2.50517666e-01 -9.61943343e-02 4.00776654e-01
1.16442275e+00 2.08151177e-01 3.35969269e-01 4.21338260e-01
3.00557941e-01 7.45556653e-02 1.77886963e-01 6.53075933e-01
5.86506038e-04 -1.94852814e-01 2.23869324e-01 -5.15336394e-01
-7.21107423e-01 -8.86716068e-01 4.92140740e-01 1.44566476e+00
5.27212203e-01 -5.45273125e-01 -3.39406192e-01 -9.06734765e-01
-5.66768646e-02 9.80058193e-01 -6.58482850e-01 -4.48381126e-01
-4.00692642e-01 1.07959874e-01 9.12033468e-02 7.20777392e-01
2.43407935e-01 -1.31524646e+00 -1.01728058e+00 2.93972075e-01
-6.62733093e-02 -1.16330063e+00 1.11594073e-01 4.08529222e-01
-5.27128041e-01 -1.04277050e+00 1.24244966e-01 -1.00913835e+00
9.15198386e-01 3.93189341e-01 8.53318214e-01 9.32215750e-02
-2.11954981e-01 1.11065662e+00 -7.03707755e-01 -4.18350577e-01
-5.91305077e-01 1.37792766e-01 3.64266574e-01 -3.73551726e-01
2.69344240e-01 -8.05667520e-01 -1.09683402e-01 3.02493840e-01
-7.30156660e-01 6.94952384e-02 6.12888157e-01 5.47960401e-01
2.83421278e-01 4.87437010e-01 5.31400025e-01 -8.50113809e-01
5.15238762e-01 -7.46413112e-01 -4.67383832e-01 5.06734788e-01
-5.37059009e-01 3.13278794e-01 4.71660584e-01 -6.11008883e-01
-1.18761992e+00 3.68895382e-01 3.85956913e-01 -4.35743392e-01
-3.99862617e-01 9.21981871e-01 -2.65364558e-01 1.24454908e-01
8.44153702e-01 -3.95623706e-02 -1.59331635e-02 -5.26147485e-01
1.08473217e+00 6.34969413e-01 7.28863120e-01 -6.94421172e-01
7.64681935e-01 6.14289455e-02 -3.87852967e-01 -6.19381011e-01
-7.62854815e-01 -8.12147021e-01 -1.14952958e+00 -1.27421230e-01
5.84649920e-01 -9.00632858e-01 -8.32679987e-01 2.23228931e-01
-1.16815102e+00 -8.13916683e-01 -6.82573080e-01 4.90829110e-01
-5.68111956e-01 2.39606678e-01 -3.76914591e-01 -7.59062469e-01
2.27671161e-01 -1.12557054e+00 6.93616748e-01 1.17833562e-01
-2.99555093e-01 -9.73521411e-01 1.88878283e-01 -7.21977204e-02
3.40177149e-01 -1.02356598e-01 8.68124127e-01 -9.35297012e-01
-5.04634500e-01 -1.32372648e-01 -4.52119447e-02 2.03765675e-01
3.55248690e-01 -5.56444049e-01 -8.74124885e-01 -3.63931507e-01
-2.71206975e-01 -6.90285206e-01 8.91813561e-02 -5.99548161e-01
8.51911068e-01 -2.02280372e-01 -6.45835817e-01 -1.73696101e-01
1.23880291e+00 2.07360849e-01 3.87931228e-01 2.15919733e-01
8.32148910e-01 8.59746933e-01 8.49045694e-01 3.82624686e-01
9.41263914e-01 5.49262702e-01 4.21148121e-01 4.45232064e-01
5.92804886e-02 -5.53414226e-01 3.64164203e-01 7.05288768e-01
-6.04188256e-02 2.32867777e-01 -1.07699871e+00 7.94038355e-01
-2.23484421e+00 -8.32390070e-01 5.02587795e-01 2.02901602e+00
7.63582110e-01 -1.16238780e-01 -1.89229280e-01 -1.37291104e-01
2.36501694e-01 -4.05088902e-01 -8.32123697e-01 -5.31786494e-02
6.37811422e-01 -3.54933351e-01 1.73145026e-01 6.14844382e-01
-9.00177002e-01 1.15416217e+00 6.05437183e+00 3.18938233e-02
-7.74047375e-01 7.11031184e-02 -1.01580828e-01 2.71620482e-01
-1.08731568e-01 1.57939643e-01 -6.88311934e-01 -4.73957472e-02
1.08529997e+00 -7.79096335e-02 7.59887874e-01 1.43455422e+00
-5.27796388e-01 -1.45362243e-01 -1.79623818e+00 9.93441999e-01
7.40004405e-02 -8.96001697e-01 -1.61078483e-01 -1.52550936e-01
3.16228598e-01 2.58446693e-01 -1.70776606e-01 1.12279689e+00
8.16765845e-01 -8.04591477e-01 6.94109261e-01 6.55986607e-01
3.93291831e-01 -5.02347469e-01 6.93711340e-01 8.69868040e-01
-1.11866724e+00 -5.54624736e-01 -2.64810711e-01 -1.68273106e-01
1.34838760e-01 -9.25682783e-02 -1.54402900e+00 7.35349536e-01
7.95610547e-01 7.95734107e-01 -8.19808424e-01 5.52590072e-01
-3.65092248e-01 -1.51695907e-01 -4.63281304e-01 6.07883036e-02
-7.80168846e-02 -1.42384723e-01 2.18214214e-01 6.66725934e-01
3.00426692e-01 2.04736099e-01 5.51441610e-01 7.01236427e-01
-1.91073492e-01 -1.19737647e-01 -1.13270497e+00 -7.00503737e-02
8.05988252e-01 1.08035862e+00 -4.07672852e-01 -4.52395499e-01
-6.78875625e-01 1.09880996e+00 9.58899140e-01 2.50697196e-01
-3.86622995e-01 -2.45755911e-01 3.63523871e-01 -2.59726673e-01
2.05347374e-01 -5.62966108e-01 3.09530795e-01 -9.17454183e-01
-3.70960380e-03 -8.14487755e-01 3.01350117e-01 -1.04395628e+00
-9.27719772e-01 4.00169402e-01 3.09011430e-01 -8.83089662e-01
-5.27324557e-01 -6.24394834e-01 -8.99833962e-02 3.47685575e-01
-1.25313699e+00 -1.39867854e+00 -7.15471864e-01 5.90653658e-01
4.46108073e-01 -8.93277004e-02 1.05119908e+00 5.52502573e-02
-2.99045086e-01 2.13910580e-01 -2.93929398e-01 1.68564916e-02
6.26401544e-01 -1.15544164e+00 2.40114704e-01 2.86196828e-01
5.31206727e-01 9.47873950e-01 6.21603012e-01 -6.91748619e-01
-1.71261811e+00 -1.05885863e+00 4.46874946e-01 -9.44539428e-01
7.48929262e-01 -3.97846729e-01 -1.19744384e+00 1.42201567e+00
-7.39172027e-02 3.29109654e-02 6.06139123e-01 6.87033415e-01
-6.03586674e-01 -1.97215099e-02 -8.93971920e-01 5.35920680e-01
1.28977764e+00 -7.08551407e-01 -1.06432450e+00 1.68266132e-01
1.06421280e+00 -3.79061908e-01 -1.07016790e+00 2.69170612e-01
5.04494905e-01 -5.08252084e-01 9.00760114e-01 -5.99233270e-01
-1.24792553e-01 -4.11943287e-01 -3.27119291e-01 -1.57596135e+00
-1.71558455e-01 -3.36267143e-01 -1.65788904e-01 1.04207957e+00
5.33235908e-01 -4.27738935e-01 4.00337577e-01 1.02803934e+00
-1.04350924e-01 -2.16255426e-01 -4.56918925e-01 -7.16207862e-01
-3.80546302e-01 -6.24672890e-01 9.81181800e-01 9.95250404e-01
7.49921679e-01 7.07790792e-01 -9.52665210e-02 5.30258894e-01
1.29230544e-01 -1.38478547e-01 1.25028861e+00 -1.37849569e+00
-4.41767424e-01 2.46046260e-01 -6.11723483e-01 -8.53356719e-01
3.55236888e-01 -7.53821015e-01 4.00707453e-01 -1.70711362e+00
1.62459075e-01 -7.47756541e-01 -4.91043240e-01 9.17523086e-01
-1.91532932e-02 -5.68804979e-01 3.50356907e-01 2.35921249e-01
-1.16572154e+00 5.90956688e-01 7.57926881e-01 -1.58211380e-01
-6.13147557e-01 -4.46429074e-01 -5.33404231e-01 9.17195916e-01
5.99245369e-01 -1.84652433e-01 -1.04009330e+00 -7.11860478e-01
7.02896774e-01 -2.42495731e-01 4.57566649e-01 -1.32026625e+00
6.46204591e-01 -1.65944368e-01 -5.72087653e-02 -3.70986611e-01
5.33126831e-01 -1.47941196e+00 4.11133140e-01 7.95710757e-02
-3.75924975e-01 -2.10158691e-01 3.52482736e-01 7.82446623e-01
-2.68513113e-01 -3.09366703e-01 1.35823891e-01 -3.47092062e-01
-1.50403702e+00 1.74621269e-01 -1.50997534e-01 -4.17341560e-01
1.22273779e+00 5.10155298e-02 -1.63084064e-02 -3.55397254e-01
-1.11858404e+00 6.32798731e-01 6.86849535e-01 7.29952812e-01
7.31219351e-01 -1.12700987e+00 1.60494857e-02 8.58540908e-02
6.97481453e-01 5.31490922e-01 1.82175353e-01 3.53517115e-01
-9.27345827e-02 6.45301342e-02 -2.85800546e-01 -2.90212005e-01
-1.02231085e+00 1.08356011e+00 1.24331914e-01 -6.66477382e-02
-7.39943385e-01 6.46441221e-01 1.51681662e-01 -9.51007485e-01
4.73277599e-01 -4.75437194e-01 -5.37040234e-01 1.30313516e-01
5.17420173e-01 2.56092459e-01 -3.65397818e-02 -5.29048264e-01
-1.02448948e-01 1.37545601e-01 -1.27993912e-01 -1.88352764e-01
1.42159820e+00 -5.15082240e-01 -2.56040901e-01 9.13861394e-01
9.55690026e-01 -3.38900656e-01 -1.01050687e+00 -7.08130956e-01
3.20223391e-01 -8.73369798e-02 -3.65252733e-01 -5.48695028e-01
-3.77938569e-01 6.29243135e-01 5.32976210e-01 4.75047678e-02
5.67724228e-01 4.32061821e-01 4.18875158e-01 1.40224874e+00
1.14855707e+00 -1.25682199e+00 3.52732480e-01 6.96558535e-01
1.00631309e+00 -1.29154921e+00 -1.39523268e-01 -2.72116512e-01
-7.88680255e-01 9.38885570e-01 9.35094297e-01 2.60103613e-01
6.89753056e-01 -1.15460530e-01 -6.83655515e-02 -3.70755225e-01
-7.80187249e-01 -3.18484038e-01 5.61787421e-03 1.04356980e+00
-7.32380748e-02 1.81679741e-01 6.24000788e-01 6.90073311e-01
-2.29012012e-01 -5.04484363e-02 5.00485003e-01 1.08876562e+00
-5.58165371e-01 -1.05575716e+00 4.84113544e-02 -2.85193864e-02
5.95348239e-01 4.14692044e-01 -4.81039703e-01 7.49027848e-01
2.96577662e-01 9.45331037e-01 -1.98090941e-01 -6.20685399e-01
7.74311185e-01 2.88087755e-01 4.27281797e-01 -1.19684517e+00
-1.08687431e-01 -7.72756100e-01 1.45225257e-01 -9.87077713e-01
-3.03980649e-01 -6.23810291e-01 -1.70657873e+00 1.72596171e-01
-2.40571916e-01 1.14772223e-01 7.57651031e-01 9.30354774e-01
4.36737686e-01 5.05215347e-01 3.82526994e-01 -8.96766663e-01
-3.87143791e-01 -9.44827735e-01 -3.49709392e-01 6.99743450e-01
2.72277862e-01 -8.38940859e-01 -3.55329132e-03 2.06132203e-01] | [4.517464637756348, 0.8066277503967285] |
7cc9d6e6-a012-4fb7-a6c1-ed6ca58b898c | reader-aware-multi-document-summarization-via | 1504.07324 | null | http://arxiv.org/abs/1504.07324v1 | http://arxiv.org/pdf/1504.07324v1.pdf | Reader-Aware Multi-Document Summarization via Sparse Coding | We propose a new MDS paradigm called reader-aware multi-document
summarization (RA-MDS). Specifically, a set of reader comments associated with
the news reports are also collected. The generated summaries from the reports
for the event should be salient according to not only the reports but also the
reader comments. To tackle this RA-MDS problem, we propose a
sparse-coding-based method that is able to calculate the salience of the text
units by jointly considering news reports and reader comments. Another
reader-aware characteristic of our framework is to improve linguistic quality
via entity rewriting. The rewriting consideration is jointly assessed together
with other summarization requirements under a unified optimization model. To
support the generation of compressive summaries via optimization, we explore a
finer syntactic unit, namely, noun/verb phrase. In this work, we also generate
a data set for conducting RA-MDS. Extensive experiments on this data set and
some classical data sets demonstrate the effectiveness of our proposed
approach. | ['Piji Li', 'Hang Li', 'Yi Liao', 'Wai Lam', 'Lidong Bing'] | 2015-04-28 | null | null | null | null | ['reader-aware-summarization'] | ['natural-language-processing'] | [ 3.17650586e-01 2.08564833e-01 -1.52347043e-01 -2.57557988e-01
-1.12832749e+00 -4.41959113e-01 7.28349805e-01 7.64503419e-01
-2.31123626e-01 7.36181855e-01 1.45087016e+00 3.19756150e-01
-2.88434267e-01 -7.14539349e-01 -3.48981410e-01 -3.90907764e-01
3.12825620e-01 2.71134347e-01 2.01849323e-02 -3.54816407e-01
7.43938744e-01 1.76991999e-01 -1.53174424e+00 6.21148884e-01
1.47718465e+00 5.43764591e-01 5.67998350e-01 6.46290123e-01
-3.96956921e-01 6.71447933e-01 -8.73847425e-01 -3.52092624e-01
-6.04028516e-02 -6.03912652e-01 -5.68891048e-01 5.74096918e-01
7.38233477e-02 -3.17223787e-01 1.69758126e-02 9.25700963e-01
5.87952554e-01 4.03053403e-01 7.58124292e-01 -7.61465490e-01
-5.66130519e-01 1.10014224e+00 -7.73759723e-01 3.41765344e-01
7.70712137e-01 -4.90724713e-01 1.45529580e+00 -8.39912713e-01
7.29814529e-01 1.25986123e+00 1.03807412e-01 1.47248060e-01
-7.17948377e-01 -1.70964926e-01 4.26998049e-01 8.00933838e-02
-1.12345457e+00 -5.96080303e-01 1.03201509e+00 -1.46377638e-01
5.45794606e-01 6.57837689e-01 3.78937989e-01 6.14054859e-01
1.03164330e-01 9.44555223e-01 5.63153863e-01 -4.10025358e-01
4.09642905e-01 -4.59431410e-02 3.11966866e-01 2.55266488e-01
4.98296559e-01 -7.27764010e-01 -6.70417547e-01 -4.10405725e-01
4.86236624e-03 -6.48178384e-02 -2.98803031e-01 2.16274917e-01
-1.43804550e+00 9.11792338e-01 -3.33508849e-02 4.50867087e-01
-7.80148327e-01 -2.59434104e-01 5.50098181e-01 -1.89041182e-01
7.12070286e-01 3.54781330e-01 -5.35036474e-02 6.11817650e-02
-1.20404208e+00 4.21873182e-01 8.35400283e-01 1.02241206e+00
5.68088710e-01 -3.90804000e-02 -7.79664636e-01 8.54154587e-01
3.19312692e-01 4.30852294e-01 5.38678765e-01 -8.00053596e-01
1.05275226e+00 8.37105334e-01 2.39260137e-01 -1.58874619e+00
-3.31806213e-01 -6.65269494e-01 -1.02996635e+00 -6.11367345e-01
-3.10062170e-01 -3.46227735e-01 -2.75669992e-01 1.49692166e+00
3.57143253e-01 1.03340387e-01 4.66659814e-01 6.55243337e-01
1.07953143e+00 1.05730569e+00 -3.64754081e-01 -8.85207951e-01
1.46602786e+00 -8.66556227e-01 -1.02749157e+00 -7.25932270e-02
6.61515534e-01 -6.92411065e-01 6.62263691e-01 2.53778517e-01
-1.31653214e+00 -3.52693707e-01 -1.13474357e+00 -1.19184762e-01
3.20006281e-01 6.59714878e-01 1.14149719e-01 2.18668789e-01
-7.87485957e-01 2.36045703e-01 -3.83498669e-01 -3.86557102e-01
9.11341384e-02 1.81788094e-02 -1.23631567e-01 4.62586619e-02
-1.00114965e+00 6.41404152e-01 5.28409958e-01 -2.09545732e-01
-4.31002915e-01 -5.18884420e-01 -8.02770376e-01 4.36951220e-01
4.43089783e-01 -8.13592196e-01 9.72883403e-01 -5.47779262e-01
-1.15074849e+00 2.89128929e-01 -6.33368909e-01 -4.07091230e-01
2.13166922e-01 -1.66396707e-01 -4.56210971e-01 4.35413092e-01
3.92268151e-01 1.64307371e-01 6.23796105e-01 -1.45675623e+00
-9.64088976e-01 -1.71180129e-01 1.00295439e-01 5.86614788e-01
-6.75228953e-01 1.86824560e-01 -4.88697559e-01 -9.94507015e-01
8.45848098e-02 -6.32633388e-01 -2.44536728e-01 -6.80756927e-01
-8.45582783e-01 -2.36965120e-01 3.42113763e-01 -8.53934050e-01
1.96514010e+00 -2.13184071e+00 4.51026261e-01 2.42137596e-01
4.16294068e-01 -1.72927473e-02 -2.97190249e-01 8.34575474e-01
2.43458271e-01 1.42869785e-01 -3.77335966e-01 -6.37046516e-01
-7.88907111e-02 -6.47763535e-02 -4.87556696e-01 1.69510484e-01
7.58597404e-02 4.56860840e-01 -8.22123110e-01 -7.33061492e-01
-2.84378201e-01 4.82695177e-02 -8.06153655e-01 9.69723836e-02
-2.54351139e-01 4.39058691e-01 -7.20297098e-01 2.56663054e-01
6.59177959e-01 -2.16126651e-01 7.27885514e-02 -4.33190256e-01
-4.43046957e-01 2.77554333e-01 -1.36880362e+00 1.72626579e+00
-4.08421934e-01 1.85843602e-01 -1.07241757e-01 -9.21093047e-01
1.16633737e+00 2.54047781e-01 6.72964752e-01 -3.89721990e-01
1.10971041e-01 1.14866167e-01 -2.38580763e-01 -4.72092301e-01
1.33399212e+00 1.96671993e-01 -3.73803914e-01 7.58679748e-01
-2.73788661e-01 7.46702701e-02 7.24935949e-01 7.28055596e-01
9.55788910e-01 -5.35288990e-01 6.56132340e-01 -1.87426805e-01
8.81896019e-01 1.03138864e-01 7.23383486e-01 5.73993266e-01
3.95566791e-01 7.77902246e-01 6.15127802e-01 7.16247633e-02
-1.08305693e+00 -3.97009850e-01 2.44704291e-01 8.88644814e-01
2.18667328e-01 -9.53991234e-01 -6.73939049e-01 -4.51660991e-01
-1.39856905e-01 1.09648228e+00 -5.20387173e-01 -1.29152253e-01
-5.04305124e-01 -6.84663117e-01 2.25903660e-01 1.32324383e-01
3.09493184e-01 -6.64916515e-01 -4.00549352e-01 4.94976938e-01
-6.67968094e-01 -1.06156611e+00 -7.81519473e-01 -2.97814071e-01
-5.68682253e-01 -7.41477191e-01 -7.70863533e-01 -6.34770215e-01
7.33073771e-01 4.84173268e-01 6.36675835e-01 6.36229338e-03
3.90557051e-01 3.54274094e-01 -8.97329271e-01 -2.11513087e-01
-7.46686339e-01 3.09458315e-01 6.27983883e-02 4.80932534e-01
-2.05512583e-01 -5.66806436e-01 -3.54327172e-01 -1.35202454e-02
-1.13934362e+00 2.09457293e-01 5.98430812e-01 5.03378332e-01
5.94300389e-01 1.65321618e-01 1.01697624e+00 -7.89154708e-01
1.28327286e+00 -8.12471032e-01 1.51395500e-02 4.98501211e-01
-1.51914179e-01 2.76074529e-01 7.62860656e-01 -2.42208228e-01
-1.35656488e+00 -2.47532502e-01 -2.91492809e-02 1.31657571e-01
1.79137364e-01 9.99499083e-01 -2.65454859e-01 6.52284741e-01
4.17641699e-01 4.25033957e-01 -2.94484735e-01 -3.70153517e-01
4.43206310e-01 9.09663260e-01 4.59983557e-01 -4.53968495e-01
6.93490386e-01 4.95498806e-01 -2.30938390e-01 -8.45359802e-01
-9.58973289e-01 -6.31943405e-01 -3.46592426e-01 -2.17871770e-01
7.35114813e-01 -1.06077886e+00 -4.16872412e-01 -1.59088984e-01
-1.59792888e+00 7.20476449e-01 -3.60676259e-01 5.43405652e-01
-4.33553398e-01 7.49839306e-01 -1.29478306e-01 -7.80975997e-01
-6.24506831e-01 -1.06949031e+00 1.09115934e+00 1.85492381e-01
-4.51484919e-01 -8.00305724e-01 1.73435196e-01 2.11079985e-01
1.72837988e-01 3.24399799e-01 8.30002129e-01 -1.13131618e+00
-2.67298788e-01 -9.71739292e-02 -2.30930075e-01 4.94029447e-02
4.79026914e-01 6.70882836e-02 -4.94297922e-01 -1.47871926e-01
1.58267900e-01 2.35210314e-01 9.85488176e-01 4.39467758e-01
7.57896721e-01 -8.35435271e-01 -1.59437805e-01 1.64045647e-01
1.18557799e+00 -1.83390584e-04 3.28009337e-01 1.16837099e-01
6.82693541e-01 6.67858005e-01 9.67955649e-01 1.37043357e+00
7.76488006e-01 6.04062557e-01 2.08578914e-01 2.60567904e-01
-5.48011996e-03 -1.32005066e-01 4.64324951e-01 1.64185667e+00
1.12788700e-01 -8.28822136e-01 -5.28768778e-01 5.90545237e-01
-1.99160707e+00 -1.07581306e+00 -3.97287816e-01 1.68622625e+00
8.14572036e-01 -7.77870044e-02 4.62726466e-02 2.71640331e-01
9.31254566e-01 5.14477015e-01 -2.79659778e-01 -4.25935417e-01
-2.83924609e-01 -3.14144224e-01 7.80039504e-02 4.00916427e-01
-7.68698096e-01 4.37276006e-01 5.22266150e+00 9.26124156e-01
-6.57994151e-01 5.28254248e-02 2.46392116e-01 -8.92711133e-02
-1.12158942e+00 1.54583557e-02 -9.92150366e-01 5.74172556e-01
7.35783696e-01 -9.60597336e-01 7.20119849e-02 4.86319274e-01
7.50989914e-01 -3.63123357e-01 -8.08067858e-01 7.77038991e-01
6.10414088e-01 -1.74203622e+00 5.23249388e-01 -1.69991087e-02
9.87001896e-01 -5.45011938e-01 -1.78915069e-01 -2.28138223e-01
7.13989511e-02 -2.82521605e-01 9.03028965e-01 6.84144974e-01
6.55970275e-01 -9.71102297e-01 6.56165123e-01 6.34111583e-01
-1.27515078e+00 -1.32449955e-01 -3.28479975e-01 1.82797194e-01
4.68314797e-01 9.40785229e-01 -7.14852929e-01 1.13959563e+00
9.12918001e-02 1.13507462e+00 -5.36903083e-01 9.16147530e-01
-2.91923154e-03 4.33532387e-01 -8.16319659e-02 -1.21033736e-01
1.26298934e-01 -1.87892616e-01 9.68887806e-01 1.32888222e+00
7.80509055e-01 3.56138557e-01 2.93550670e-01 7.16389596e-01
-1.64196029e-01 5.62658787e-01 -3.23737770e-01 2.40535419e-02
6.84570968e-01 1.20077825e+00 -5.57815850e-01 -5.53558946e-01
-2.68314153e-01 7.55855918e-01 1.47867963e-01 1.86417729e-01
-5.99490643e-01 -3.33182096e-01 2.01369002e-01 -1.46279007e-01
3.53615582e-01 -1.59787744e-01 -4.43994492e-01 -1.40396321e+00
2.53220767e-01 -8.70036006e-01 2.85586804e-01 -5.62380314e-01
-9.06907558e-01 5.38727820e-01 1.14456825e-01 -1.63548267e+00
-1.19929679e-01 4.13526386e-01 -8.15800846e-01 5.64053655e-01
-1.55754876e+00 -8.65683496e-01 -2.68435329e-01 3.50730896e-01
8.40039432e-01 -2.68359482e-01 5.63070774e-01 2.43059814e-01
-6.44979596e-01 2.87371516e-01 2.23533705e-01 -3.33985001e-01
6.35658741e-01 -9.17595327e-01 2.40489081e-01 1.15156901e+00
-7.23179337e-03 5.62943518e-01 8.72324705e-01 -9.34254587e-01
-1.29050803e+00 -1.12788928e+00 1.24786162e+00 2.38254443e-01
5.20081460e-01 9.83531326e-02 -8.35658193e-01 3.47679138e-01
4.80794698e-01 -8.42844844e-01 9.01216745e-01 -1.17210634e-01
1.64132223e-01 -1.54881984e-01 -9.91185606e-01 6.70268893e-01
8.78421664e-01 -1.38106778e-01 -1.03782022e+00 3.56070936e-01
1.02881777e+00 -3.20892721e-01 -7.78216898e-01 1.04288690e-01
1.57496873e-02 -7.10504651e-01 6.82339370e-01 -1.57357782e-01
8.44891727e-01 -5.10909259e-01 -2.52750367e-01 -1.59925401e+00
-4.35152203e-01 -6.27491772e-01 -2.65223086e-01 1.80458987e+00
2.83110112e-01 -2.06763417e-01 2.42452070e-01 2.48471886e-01
-4.67270404e-01 -3.48683476e-01 -7.40355670e-01 -3.40755343e-01
-4.96705800e-01 -3.06226552e-01 8.38660181e-01 7.08909512e-01
3.41371208e-01 4.47378695e-01 -6.39247239e-01 3.48602772e-01
4.17455703e-01 2.78512925e-01 6.88513279e-01 -1.15053642e+00
2.53627216e-03 -2.68689781e-01 9.51501727e-03 -1.14453506e+00
1.13964841e-01 -9.41666126e-01 -7.87448734e-02 -2.11422515e+00
3.80678892e-01 -1.85720697e-01 1.21284120e-01 -2.12616939e-02
-3.56858790e-01 -4.83726829e-01 3.63340408e-01 3.84768665e-01
-9.36578870e-01 8.74624252e-01 1.23835886e+00 -1.09229341e-01
-4.63279396e-01 1.49773005e-02 -1.28569210e+00 4.08613861e-01
6.87954307e-01 -5.66765010e-01 -3.87331456e-01 -4.40552771e-01
4.07483160e-01 4.05179769e-01 -1.27034098e-01 -9.11392391e-01
6.19328678e-01 -2.96180844e-01 -1.66718200e-01 -1.16851866e+00
2.58617960e-02 -5.52384615e-01 1.30913764e-01 2.16761515e-01
-7.95806408e-01 2.30524868e-01 -1.56732291e-01 7.95279920e-01
-4.90927160e-01 -3.54462892e-01 2.48554677e-01 6.70321509e-02
-3.49142253e-01 5.98702468e-02 -4.69004750e-01 1.95897877e-01
9.12350237e-01 -3.11654229e-02 -4.73925829e-01 -6.15855038e-01
-4.88146424e-01 5.00371695e-01 2.78707981e-01 1.96145758e-01
8.29572558e-01 -1.41981840e+00 -1.35616231e+00 -1.54089034e-01
3.15847635e-01 1.08757578e-01 3.75649482e-01 7.77478635e-01
-1.32373214e-01 3.12493831e-01 7.61499926e-02 -1.87901095e-01
-1.18755257e+00 4.61275071e-01 -5.37674367e-01 -5.22713482e-01
-6.24524295e-01 3.03137124e-01 3.65412119e-03 1.75052032e-01
6.42785849e-03 -7.19175994e-01 -8.85576010e-01 7.46154487e-01
8.97704780e-01 5.75358510e-01 -1.87023208e-02 -8.59470785e-01
-1.64201409e-01 4.13487583e-01 -6.80229068e-02 -2.32586414e-01
1.54880142e+00 -6.88748002e-01 -2.72511125e-01 3.53446007e-01
1.01951206e+00 6.75644875e-01 -7.07415998e-01 -3.83087486e-01
2.75206000e-01 -2.34324172e-01 -5.43813873e-03 -2.72232920e-01
-8.48671615e-01 2.31932491e-01 -4.00167316e-01 3.38011682e-01
1.19244862e+00 6.25251699e-03 8.50695610e-01 3.19108456e-01
2.26869783e-03 -1.13033724e+00 1.47056088e-01 4.54143643e-01
1.31789207e+00 -8.91381621e-01 3.92633617e-01 -4.34436947e-01
-1.10422945e+00 1.09223604e+00 1.18866101e-01 -4.70445119e-02
3.09544265e-01 2.63170749e-02 -3.85060221e-01 -2.02823132e-02
-8.83408010e-01 -2.48654231e-01 4.92048681e-01 1.70003712e-01
2.50285029e-01 -7.58846998e-02 -8.96664858e-01 9.75592375e-01
-2.18423903e-01 -2.29877830e-01 1.06542361e+00 8.52961421e-01
-9.03344870e-01 -8.99285018e-01 -5.51934481e-01 5.90795457e-01
-3.13722163e-01 -6.97987229e-02 -2.39444837e-01 -5.38537130e-02
-1.30279928e-01 1.44741130e+00 1.58907827e-02 -3.97289783e-01
4.35826242e-01 -3.09291005e-01 -2.34452575e-01 -9.04821336e-01
-7.42765903e-01 3.14290345e-01 4.11653101e-01 -9.56577286e-02
-7.68146396e-01 -9.74918962e-01 -1.46343553e+00 -3.89387608e-01
-3.80407184e-01 4.99146134e-01 5.96213758e-01 1.05964446e+00
7.04940975e-01 9.45374489e-01 1.18628716e+00 -7.15726018e-01
-4.96553570e-01 -9.18741107e-01 -6.29765034e-01 2.76222378e-01
4.97105479e-01 -2.48253420e-01 -4.29838747e-01 1.28117368e-01] | [12.585785865783691, 9.5249662399292] |
4653c094-147f-4ffe-95b5-39b4f7d00661 | hipool-modeling-long-documents-using-graph | 2305.03319 | null | https://arxiv.org/abs/2305.03319v2 | https://arxiv.org/pdf/2305.03319v2.pdf | HiPool: Modeling Long Documents Using Graph Neural Networks | Encoding long sequences in Natural Language Processing (NLP) is a challenging problem. Though recent pretraining language models achieve satisfying performances in many NLP tasks, they are still restricted by a pre-defined maximum length, making them challenging to be extended to longer sequences. So some recent works utilize hierarchies to model long sequences. However, most of them apply sequential models for upper hierarchies, suffering from long dependency issues. In this paper, we alleviate these issues through a graph-based method. We first chunk the sequence with a fixed length to model the sentence-level information. We then leverage graphs to model intra- and cross-sentence correlations with a new attention mechanism. Additionally, due to limited standard benchmarks for long document classification (LDC), we propose a new challenging benchmark, totaling six datasets with up to 53k samples and 4034 average tokens' length. Evaluation shows our model surpasses competitive baselines by 2.6% in F1 score, and 4.8% on the longest sequence dataset. Our method is shown to outperform hierarchical sequential models with better performance and scalability, especially for longer sequences. | ['Rex Ying', 'Dragomir Radev', 'Aosong Feng', 'Irene Li'] | 2023-05-05 | null | null | null | null | ['document-classification'] | ['natural-language-processing'] | [ 2.94490904e-01 -1.70215949e-01 -5.22039115e-01 -3.85211319e-01
-9.64966893e-01 -6.89086199e-01 3.41772079e-01 5.24076462e-01
-6.83855891e-01 7.00311720e-01 4.89947975e-01 -5.11668563e-01
3.13965976e-01 -5.17087162e-01 -8.01877975e-01 -4.26289558e-01
-3.06431532e-01 3.40867490e-01 4.68794554e-01 -1.36352107e-01
2.82209426e-01 1.16433755e-01 -1.11497545e+00 7.99775362e-01
9.27208662e-01 6.30742967e-01 3.73554200e-01 8.74070227e-01
-5.23929060e-01 7.95688689e-01 -6.91290200e-01 -4.07813728e-01
-6.07507024e-03 -3.43389153e-01 -9.55976844e-01 -8.75658989e-02
4.62722182e-01 -5.10971062e-02 -2.98490644e-01 1.01204455e+00
5.48969448e-01 1.27775356e-01 3.45571309e-01 -8.73288155e-01
-6.93045735e-01 1.05716598e+00 -8.16826522e-01 3.24380815e-01
5.53344846e-01 3.18486392e-02 1.54176605e+00 -7.57857144e-01
6.01841867e-01 1.39533734e+00 6.85518980e-01 6.22258663e-01
-1.17866659e+00 -6.40445650e-01 5.61807871e-01 1.73819691e-01
-1.16164923e+00 -3.57204944e-01 3.91361952e-01 -2.96053618e-01
1.62960827e+00 -3.12830918e-02 3.04511309e-01 1.17787433e+00
4.86227661e-01 1.05581963e+00 8.65642071e-01 -4.67105269e-01
2.26977430e-02 -3.52940977e-01 7.59807050e-01 5.93186677e-01
6.48101494e-02 -3.18333060e-01 -4.83697653e-01 -6.99586123e-02
1.47962496e-01 -2.28185043e-01 -2.92274147e-01 1.64666802e-01
-1.14681888e+00 8.91295016e-01 3.16171587e-01 4.81584638e-01
-1.80175275e-01 6.52703345e-02 7.17392564e-01 3.43171656e-01
5.02314866e-01 3.96808505e-01 -5.59632003e-01 -2.80975848e-01
-7.78306067e-01 2.70679235e-01 1.00339174e+00 1.15535033e+00
3.95870388e-01 -1.03746369e-01 -4.82973218e-01 1.09798968e+00
8.81503746e-02 1.89330280e-01 7.09777951e-01 -3.88364822e-01
1.05622065e+00 5.08068383e-01 -2.97132254e-01 -1.05835557e+00
-5.52953303e-01 -6.33456707e-01 -1.06964993e+00 -5.51264226e-01
2.25114658e-01 -1.12800620e-01 -9.25963759e-01 1.64396095e+00
-9.63589475e-02 6.45002648e-02 -3.30267139e-02 7.22815156e-01
6.40584946e-01 1.24021900e+00 2.25074053e-01 -3.75594079e-01
1.36411655e+00 -1.45135045e+00 -6.20569825e-01 -4.93836671e-01
1.16268420e+00 -6.16835296e-01 1.48802102e+00 4.45686519e-01
-9.95184839e-01 -4.42566514e-01 -9.79943037e-01 -2.58527905e-01
-2.16049477e-01 -2.07473755e-01 5.02196431e-01 5.68512261e-01
-1.05704141e+00 5.56108356e-01 -6.28362715e-01 -3.17752242e-01
2.45606527e-01 2.31258348e-01 -9.59449485e-02 -3.35993618e-01
-1.34680521e+00 6.13519430e-01 6.98051929e-01 -5.09396195e-02
-4.79725838e-01 -7.90552378e-01 -9.40790415e-01 2.13007092e-01
3.67433310e-01 -5.31620145e-01 1.18551517e+00 -5.33477187e-01
-1.45187879e+00 5.26339471e-01 -3.31656307e-01 -8.71762991e-01
3.46468717e-01 -4.36817855e-01 -3.80009085e-01 1.00955039e-01
-6.77926093e-02 5.87866366e-01 3.70085597e-01 -9.50509369e-01
-4.93346542e-01 -1.38425529e-01 1.23714425e-01 2.07657576e-01
-5.93101740e-01 2.47653097e-01 -7.36114681e-01 -7.69338608e-01
-3.00668925e-01 -9.18561757e-01 -5.01368344e-01 -7.63970196e-01
-4.54905182e-01 -3.95109713e-01 5.53047776e-01 -6.25247836e-01
1.67554867e+00 -2.02884269e+00 4.82827425e-02 -7.46811628e-02
1.29188016e-01 4.58342493e-01 -6.61923707e-01 6.70062244e-01
1.71117082e-01 4.96642798e-01 -3.48698020e-01 -7.44016528e-01
7.29705999e-03 3.55973244e-01 -3.35779607e-01 1.81615546e-01
1.90862700e-01 9.87546682e-01 -9.92722034e-01 -6.24204457e-01
-3.35587502e-01 2.50394821e-01 -7.25536942e-01 7.85282552e-02
-4.89944935e-01 1.14637382e-01 -2.17903346e-01 2.98109651e-01
5.67496240e-01 -4.51501608e-01 4.48844403e-01 2.44129151e-01
1.09616600e-01 7.15251565e-01 -6.58339739e-01 1.97335291e+00
-5.67165315e-01 4.12580580e-01 -1.89711109e-01 -1.01489663e+00
7.71931708e-01 1.33379579e-01 2.34530032e-01 -6.92793548e-01
-1.71258703e-01 1.03301108e-01 2.38967761e-01 -4.16485846e-01
7.60752797e-01 -9.06719714e-02 -2.32934177e-01 3.59649956e-01
-3.79657075e-02 8.52013156e-02 6.63452983e-01 4.95897770e-01
1.38391674e+00 -2.50378400e-01 3.09116453e-01 -2.75806785e-01
5.43495774e-01 -2.07154036e-01 6.42056644e-01 9.04599249e-01
-6.26356080e-02 7.12185085e-01 8.92883122e-01 -2.94103801e-01
-1.03886628e+00 -5.86689532e-01 1.60347655e-01 1.33584583e+00
-1.66919693e-01 -8.85436594e-01 -6.02030694e-01 -1.06291866e+00
2.49260366e-02 6.35916173e-01 -3.87393266e-01 -7.67940432e-02
-7.98285425e-01 -9.26887870e-01 6.57517493e-01 6.83433235e-01
2.07024187e-01 -1.06848073e+00 -2.79843360e-02 5.16990006e-01
-2.99370944e-01 -1.55023539e+00 -8.56821716e-01 4.04823758e-02
-8.11150908e-01 -7.01967061e-01 -7.14558184e-01 -9.92667198e-01
3.11468661e-01 2.74401277e-01 1.31052232e+00 2.27129072e-01
-1.85507704e-02 -1.24655284e-01 -7.34122097e-01 -1.60384238e-01
-5.12010992e-01 7.10060477e-01 -1.04179837e-01 -2.86718488e-01
3.40899348e-01 -3.66183609e-01 -3.35746258e-01 2.76132766e-02
-8.00614119e-01 5.51725067e-02 7.62471139e-01 1.06458604e+00
3.85999560e-01 -1.29674956e-01 8.44799817e-01 -1.03203154e+00
9.26974952e-01 -4.30876851e-01 -4.48204488e-01 4.27381337e-01
-5.52958965e-01 1.64795682e-01 1.06016529e+00 -3.98903072e-01
-8.28276217e-01 -1.39396310e-01 -3.46648008e-01 -5.72221801e-02
-1.36020407e-01 8.95506680e-01 -1.46828130e-01 2.16900706e-01
3.70091498e-01 3.25053066e-01 -2.15085238e-01 -5.21484315e-01
1.66012734e-01 7.13377714e-01 1.61485240e-01 -6.28952324e-01
3.38742703e-01 -1.12962849e-01 -2.70166337e-01 -9.03793693e-01
-1.31274974e+00 -7.21735597e-01 -6.97997987e-01 3.40476543e-01
7.19091535e-01 -9.32358801e-01 -6.05642736e-01 3.26780081e-01
-1.40926039e+00 -5.89379668e-01 2.13131070e-01 4.37883675e-01
-8.25888664e-02 8.12976182e-01 -1.22975230e+00 -5.89042902e-01
-6.53361499e-01 -1.03134537e+00 1.07597613e+00 -3.56611133e-01
-1.53706804e-01 -1.02858746e+00 6.04490079e-02 3.43365103e-01
1.86401442e-01 -7.03102127e-02 1.06062329e+00 -7.37983584e-01
-2.17228860e-01 -1.09171093e-01 -3.03432643e-01 5.09657621e-01
-2.10608020e-01 -2.15355083e-01 -5.09669363e-01 -5.86795092e-01
-2.69822955e-01 -5.10844707e-01 1.28548157e+00 3.57314362e-03
1.41513455e+00 -3.98552507e-01 -2.89182484e-01 4.67549056e-01
1.39676285e+00 1.35227278e-01 4.72199112e-01 1.96772456e-01
8.68928254e-01 5.93580961e-01 4.64217395e-01 4.27519143e-01
4.08395648e-01 4.45316464e-01 1.27448082e-01 2.77099699e-01
-8.32754597e-02 -3.41506630e-01 6.45517111e-01 1.50968635e+00
3.08559746e-01 -7.87037849e-01 -1.26955128e+00 5.69486976e-01
-1.77860343e+00 -7.46380806e-01 -3.08865100e-01 1.89618635e+00
9.78010952e-01 5.95691025e-01 2.02202741e-02 1.51317984e-01
4.75056589e-01 4.74521130e-01 -2.36333475e-01 -6.42928720e-01
-1.33395329e-01 1.53145239e-01 4.85789239e-01 6.45401716e-01
-1.01420259e+00 1.16822207e+00 6.24721003e+00 9.99324083e-01
-9.20856357e-01 -9.98111144e-02 6.39329970e-01 -8.15638751e-02
-3.61941099e-01 -1.30389020e-01 -1.33349669e+00 6.33077562e-01
1.33472407e+00 -1.98309675e-01 1.17064463e-02 5.34693360e-01
2.41772048e-02 1.46464154e-01 -1.00842094e+00 7.85543859e-01
2.22242385e-01 -1.31669533e+00 4.50642705e-01 2.39617210e-02
8.13932419e-01 2.14316174e-01 -2.66697675e-01 7.30870843e-01
2.56121337e-01 -1.04354727e+00 4.07143623e-01 4.52437811e-03
6.56335771e-01 -9.07586873e-01 8.49558890e-01 6.91613674e-01
-1.23284686e+00 1.03664608e-03 -6.68251991e-01 -2.08037987e-01
2.81266540e-01 7.84299195e-01 -8.66182327e-01 5.93580425e-01
4.38337028e-01 8.90935838e-01 -6.85076356e-01 8.43894780e-01
-2.37546206e-01 1.00310731e+00 -1.66233420e-01 -5.15518904e-01
6.87785685e-01 6.53292760e-02 3.11431020e-01 1.96153331e+00
1.19399741e-01 8.77272163e-04 5.33575594e-01 2.42125586e-01
-3.75779450e-01 4.70506042e-01 -5.21813869e-01 -2.63145179e-01
3.18620592e-01 9.95795727e-01 -5.89590251e-01 -5.67795813e-01
-6.34064376e-01 9.93804634e-01 7.30964184e-01 4.06559169e-01
-8.41298401e-01 -6.63035572e-01 3.30346018e-01 -1.27959028e-01
2.75803983e-01 -5.54213881e-01 -1.85226709e-01 -1.42108214e+00
1.55851796e-01 -9.76664901e-01 6.67377651e-01 -2.75258750e-01
-1.41629040e+00 7.46689141e-01 -2.18000904e-01 -8.95829499e-01
-2.32444629e-01 -6.70626998e-01 -2.25610584e-01 7.61133015e-01
-1.66348195e+00 -9.19935882e-01 4.33788896e-02 2.51467705e-01
9.20863152e-01 -1.83638595e-02 7.48854995e-01 4.35861856e-01
-6.72012508e-01 8.17602158e-01 2.17930391e-01 2.82625705e-01
7.36197293e-01 -1.35326290e+00 8.29003274e-01 9.00444806e-01
3.69024038e-01 6.82549894e-01 4.48442370e-01 -6.00079119e-01
-1.39158881e+00 -1.17529464e+00 1.48191750e+00 -1.85020879e-01
8.81745219e-01 -9.25904930e-01 -1.32934642e+00 7.53667235e-01
3.81319225e-01 1.21260047e-01 7.75296926e-01 3.38994056e-01
-5.80626607e-01 5.56907803e-03 -5.55197835e-01 6.05754316e-01
1.30275476e+00 -4.56157446e-01 -6.33817673e-01 4.55295861e-01
1.32298529e+00 -4.29955125e-01 -8.12710404e-01 3.55302721e-01
3.79980326e-01 -6.78708076e-01 6.81504846e-01 -8.78879428e-01
7.05357909e-01 5.31401811e-03 7.60536119e-02 -1.29680836e+00
-4.68916357e-01 -6.32245302e-01 -1.95537075e-01 1.26293385e+00
7.10354686e-01 -5.05921543e-01 8.56139123e-01 4.53342609e-02
-3.88502628e-01 -1.02520239e+00 -5.33976555e-01 -1.22094977e+00
4.18596298e-01 -6.18489206e-01 4.54553038e-01 9.22207892e-01
3.25638562e-01 8.89630318e-01 -4.77468342e-01 4.26270142e-02
4.29481477e-01 2.54621267e-01 5.11574149e-01 -7.66196907e-01
-4.55413342e-01 -5.28441787e-01 -1.69558764e-01 -1.62182164e+00
6.18237317e-01 -1.20919895e+00 2.75611103e-01 -1.60839784e+00
3.96152139e-01 -2.42956787e-01 -3.12931508e-01 4.29278076e-01
-4.41104800e-01 -7.75938258e-02 2.77272701e-01 3.99381667e-02
-9.56919849e-01 5.91279745e-01 1.04345083e+00 -1.51316971e-01
-1.41962260e-01 -3.21756095e-01 -4.39625114e-01 3.73394519e-01
1.05124819e+00 -4.12854761e-01 -3.73819977e-01 -8.88042808e-01
1.83392361e-01 -2.42315996e-02 -1.68564454e-01 -8.89994204e-01
3.91776174e-01 -9.64335799e-02 -1.94222368e-02 -7.49694288e-01
-2.22255439e-02 -3.54096323e-01 -5.54828644e-01 5.28149128e-01
-8.00556064e-01 2.66568035e-01 2.51173586e-01 7.01395631e-01
-4.44773704e-01 -2.97944069e-01 5.50642967e-01 -3.66599262e-02
-6.52077556e-01 5.06235659e-01 -3.66527349e-01 4.40790385e-01
8.27243090e-01 1.99610621e-01 -3.89945477e-01 -2.84856439e-01
-4.20997262e-01 6.59195304e-01 9.84747112e-02 5.82491159e-01
5.78316987e-01 -1.07550526e+00 -9.15135562e-01 -6.82827011e-02
1.83884472e-01 1.27906814e-01 2.86586821e-01 6.19081616e-01
-5.25273323e-01 7.18424857e-01 2.71497488e-01 -4.91162747e-01
-1.37024415e+00 8.10124457e-01 -1.60920590e-01 -9.15323794e-01
-7.31601238e-01 9.15116251e-01 1.35660291e-01 -3.00479561e-01
4.29619730e-01 -5.82341373e-01 -1.64737016e-01 4.05729674e-02
6.57930434e-01 -8.55707191e-03 1.59135893e-01 -4.46656674e-01
-3.85943115e-01 5.37593901e-01 -4.50783283e-01 1.58844203e-01
1.24265909e+00 -2.63437629e-02 -6.93106055e-02 4.47927207e-01
1.59278798e+00 6.19492121e-02 -9.02822673e-01 -5.21895468e-01
6.36900485e-01 -2.36979574e-01 -3.90543520e-01 -5.70111811e-01
-7.84697950e-01 1.16463995e+00 -1.34964556e-01 1.26411214e-01
8.47977102e-01 -1.22584216e-01 1.21796155e+00 5.14700890e-01
3.31039429e-01 -8.53712738e-01 2.76690632e-01 1.28541493e+00
8.78553629e-01 -1.20734990e+00 -7.42591172e-02 -4.82587308e-01
-5.86908162e-01 1.05969381e+00 6.16074860e-01 4.22429387e-03
3.84644508e-01 5.16781449e-01 -1.52778894e-01 2.75467336e-01
-1.26114178e+00 1.19037628e-02 3.50954294e-01 2.49981552e-01
9.61937964e-01 -4.30999286e-02 -6.41601264e-01 6.01673007e-01
-3.41043741e-01 -2.20251694e-01 4.14639592e-01 8.09687138e-01
-4.31246459e-01 -1.36307180e+00 1.15860492e-01 4.44138199e-01
-8.24516356e-01 -5.04743159e-01 -1.33185565e-01 6.62955940e-01
-2.65350014e-01 8.61258268e-01 9.17604342e-02 -2.83739209e-01
3.15499604e-01 1.36906892e-01 2.69141555e-01 -9.91038442e-01
-7.17281818e-01 -2.61956267e-03 4.26539510e-01 -5.33552587e-01
-1.66584507e-01 -6.31167233e-01 -1.30758870e+00 -2.52116591e-01
-9.80886221e-02 2.60597408e-01 3.12877625e-01 6.66038454e-01
4.52269435e-01 5.40195823e-01 3.76155168e-01 -4.07092690e-01
-8.14405918e-01 -1.12617123e+00 -5.20310938e-01 4.17103022e-01
2.67998040e-01 -4.77538556e-02 -3.07675213e-01 -3.40811908e-02] | [10.981687545776367, 8.614027976989746] |
8451b021-eed5-431d-b82f-758dc6baeca3 | codim-learning-with-noisy-labels-via | 2111.11652 | null | https://arxiv.org/abs/2111.11652v1 | https://arxiv.org/pdf/2111.11652v1.pdf | CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning | Labels are costly and sometimes unreliable. Noisy label learning, semi-supervised learning, and contrastive learning are three different strategies for designing learning processes requiring less annotation cost. Semi-supervised learning and contrastive learning have been recently demonstrated to improve learning strategies that address datasets with noisy labels. Still, the inner connections between these fields as well as the potential to combine their strengths together have only started to emerge. In this paper, we explore further ways and advantages to fuse them. Specifically, we propose CSSL, a unified Contrastive Semi-Supervised Learning algorithm, and CoDiM (Contrastive DivideMix), a novel algorithm for learning with noisy labels. CSSL leverages the power of classical semi-supervised learning and contrastive learning technologies and is further adapted to CoDiM, which learns robustly from multiple types and levels of label noise. We show that CoDiM brings consistent improvements and achieves state-of-the-art results on multiple benchmarks. | ['Xiao Han', 'Dimitris Samaras', 'Wei Yang', 'Junzhou Huang', 'Tian Shen', 'Kaiwen Xiao', 'Zixuan Liu', 'Xin Zhang'] | 2021-11-23 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 4.00321960e-01 1.49617836e-01 -3.39956760e-01 -6.28281593e-01
-1.33734667e+00 -7.52304375e-01 7.16598332e-01 2.98516899e-01
-4.73417610e-01 6.14588678e-01 1.40056051e-02 2.22567972e-02
-1.82212874e-01 -1.48812458e-01 -3.64739746e-01 -7.85521328e-01
3.78907546e-02 6.55648947e-01 9.08612534e-02 2.04326794e-01
-1.85526367e-02 1.46569014e-01 -1.75173867e+00 4.25634831e-01
6.76308334e-01 9.78513420e-01 -1.30037680e-01 5.87343037e-01
-4.12630171e-01 1.29633605e+00 -5.69865525e-01 -2.44939968e-01
2.11423948e-01 -5.66459000e-01 -1.37158620e+00 2.38254681e-01
7.59017408e-01 2.13725135e-01 1.16093725e-01 9.40373659e-01
6.58142149e-01 -1.12059735e-01 6.04329586e-01 -1.30443811e+00
-2.91993946e-01 9.92310107e-01 -5.48221648e-01 6.37724772e-02
9.92369354e-02 1.49683371e-01 1.09289181e+00 -9.65358675e-01
6.84675395e-01 1.28946590e+00 1.03373146e+00 8.28241110e-01
-1.47141302e+00 -6.44369483e-01 2.45798394e-01 -1.17652364e-01
-1.06383300e+00 -4.28974390e-01 7.20955074e-01 -4.87398475e-01
6.31099999e-01 2.85132051e-01 2.48993143e-01 1.18455255e+00
-4.62170839e-01 1.32809997e+00 1.97331715e+00 -8.32153857e-01
2.45694652e-01 1.51064262e-01 5.99943697e-01 9.05684292e-01
-4.74416465e-03 2.28212222e-01 -8.65637839e-01 -1.54444143e-01
-5.18827550e-02 -1.02601230e-01 -1.66043445e-01 -2.72988409e-01
-1.12696695e+00 5.74784458e-01 1.59568951e-01 1.93432853e-01
3.04315776e-01 1.96987107e-01 7.03183591e-01 4.77000564e-01
8.77694130e-01 6.89385593e-01 -7.24508286e-01 -2.38656878e-01
-1.19998825e+00 -8.27571228e-02 9.00724351e-01 1.12644517e+00
9.95335400e-01 -2.08177969e-01 -3.43083531e-01 8.91606033e-01
3.55522782e-01 3.34968090e-01 2.98718572e-01 -1.34127235e+00
1.11979663e-01 6.39559031e-01 -8.70176479e-02 -1.38329387e-01
-7.06908047e-01 -5.84993303e-01 -7.97281563e-01 3.04865569e-01
4.85240966e-01 -1.11439042e-01 -1.05570805e+00 1.59248555e+00
1.84127137e-01 2.00514317e-01 -8.59899446e-02 5.57321787e-01
9.54591751e-01 2.66114771e-01 3.13423306e-01 -4.75343466e-01
1.02573419e+00 -1.51124668e+00 -9.33164835e-01 -2.60146230e-01
1.12754214e+00 -7.91665852e-01 1.32906592e+00 7.60197639e-01
-1.05055809e+00 -5.21395743e-01 -9.64605749e-01 -1.70709983e-01
-5.10020196e-01 2.38556918e-02 5.75522065e-01 7.18748152e-01
-9.88880455e-01 8.34553957e-01 -8.03661585e-01 -9.83315408e-02
8.30580890e-01 1.82896137e-01 -1.28512815e-01 -3.03694993e-01
-9.17928040e-01 7.64014900e-01 3.87931913e-01 -1.13864101e-01
-1.07835972e+00 -5.60836494e-01 -7.18110502e-01 -2.25597158e-01
7.36085236e-01 -2.83781677e-01 1.61213398e+00 -9.45481777e-01
-1.31616545e+00 1.41338575e+00 -6.92815408e-02 -2.13404357e-01
6.31978333e-01 -3.75196069e-01 -2.13425711e-01 3.61629911e-02
1.12774603e-01 7.54697800e-01 6.68148339e-01 -1.72006702e+00
-8.75045121e-01 -3.39962870e-01 -4.06855494e-01 1.55934989e-01
-4.01235253e-01 1.76170483e-01 -2.06948161e-01 -5.62711179e-01
2.68635869e-01 -1.01272380e+00 -1.03522129e-01 4.28952724e-02
-5.44834256e-01 -5.87643862e-01 1.04486430e+00 -7.42253661e-03
1.01387012e+00 -2.22264504e+00 1.14659049e-01 2.89334003e-02
5.61127245e-01 4.78729010e-01 -8.25641677e-02 1.18457638e-01
-1.92821231e-02 4.73741621e-01 -3.93857598e-01 -8.57147992e-01
1.07544519e-01 4.80721712e-01 3.63744535e-02 3.57813448e-01
3.67663264e-01 9.84103203e-01 -1.55361748e+00 -7.20002353e-01
-2.89184954e-02 2.07636252e-01 -4.55471091e-02 2.08524421e-01
-2.46409699e-01 4.71720308e-01 -5.21507300e-02 9.12685812e-01
5.55282593e-01 -3.55049133e-01 2.51492947e-01 -1.36062667e-01
-2.46460941e-02 3.10180098e-01 -1.34580934e+00 1.72229755e+00
-3.39197308e-01 5.04130065e-01 1.10803165e-01 -9.05166626e-01
7.95141459e-01 3.19360882e-01 3.95874709e-01 -3.99315983e-01
3.63667943e-02 6.47781670e-01 -6.00588858e-01 -6.34492159e-01
1.44764557e-01 -1.94214463e-01 -1.43718287e-01 6.86614931e-01
6.33861959e-01 -3.07434142e-01 4.63044316e-01 1.93924963e-01
1.17005718e+00 5.41832209e-01 2.55990803e-01 -3.34766865e-01
3.30404609e-01 1.50258495e-02 6.17019951e-01 1.09768641e+00
-5.31027853e-01 7.25674629e-01 4.24657941e-01 -3.29812527e-01
-5.72051585e-01 -9.85520363e-01 -1.51130378e-01 1.53191948e+00
1.03176147e-01 -5.77971637e-01 -6.25026524e-01 -1.42972410e+00
-1.59351632e-01 3.10342461e-01 -6.25670552e-01 -3.85325141e-02
-3.95706803e-01 -7.89762318e-01 7.00482845e-01 5.76320946e-01
4.84516233e-01 -1.00104082e+00 -1.90839127e-01 3.80779095e-02
-1.97981223e-01 -9.97053444e-01 -2.71653622e-01 1.19626033e+00
-8.02942753e-01 -1.24262917e+00 -3.38286638e-01 -1.14497554e+00
6.35326445e-01 3.86330396e-01 1.64516068e+00 1.81792259e-01
-3.40694264e-02 4.80250448e-01 -6.29968107e-01 -4.37099904e-01
-6.81835234e-01 3.38455170e-01 2.38103736e-02 -1.83548525e-01
3.67608935e-01 -4.06376451e-01 -1.76117480e-01 3.85543555e-01
-1.00489438e+00 -1.78776383e-01 3.97853523e-01 1.15877342e+00
7.14845836e-01 7.69308433e-02 6.37369812e-01 -1.74653590e+00
3.56576234e-01 -3.76536936e-01 -2.67129302e-01 2.89628893e-01
-1.10576022e+00 2.04180792e-01 3.81580621e-01 -3.50617677e-01
-1.24813902e+00 3.40767145e-01 -5.19254915e-02 -5.55475391e-02
-4.65497971e-01 3.45713615e-01 -1.30699024e-01 -2.17850879e-03
1.14501226e+00 -2.08014309e-01 -1.78219259e-01 -7.74163246e-01
6.88057721e-01 9.82591331e-01 4.76751566e-01 -7.06113398e-01
6.26675069e-01 6.33142531e-01 -5.21574132e-02 -2.31149778e-01
-1.78902042e+00 -8.99225771e-01 -8.66924286e-01 -2.48241067e-01
4.93698329e-01 -1.05588078e+00 -3.34351301e-01 7.79205978e-01
-7.92682707e-01 -6.23268187e-01 -8.03209066e-01 3.40459347e-02
-6.04437768e-01 3.22325379e-01 -9.99424934e-01 -6.92804635e-01
-2.61686057e-01 -1.07253230e+00 1.18211424e+00 2.34185472e-01
-1.65455937e-01 -1.04322731e+00 7.84282461e-02 5.75492680e-01
1.80561975e-01 1.53819665e-01 3.92983675e-01 -9.59547937e-01
-1.56120434e-01 -3.02219540e-02 -1.21402167e-01 8.32106590e-01
1.42473772e-01 -1.28261670e-01 -1.51577902e+00 -2.97667742e-01
-1.22481406e-01 -1.25531006e+00 1.25728595e+00 3.11098117e-02
9.50426877e-01 -1.96073446e-02 -3.37903410e-01 3.72766167e-01
1.30245435e+00 -3.08135897e-01 2.25030795e-01 3.24408889e-01
9.71451342e-01 7.29618073e-01 7.71796882e-01 1.75195992e-01
1.93328857e-01 3.32018465e-01 2.67889977e-01 -3.39449257e-01
-5.59735596e-01 -1.69469908e-01 2.36963153e-01 1.07311225e+00
2.01228604e-01 4.49823551e-02 -1.03622389e+00 3.32582265e-01
-1.90717065e+00 -7.32298791e-01 -5.40400684e-01 1.77852201e+00
1.38960254e+00 3.73509049e-01 3.95009555e-02 5.24551511e-01
5.69333553e-01 8.46645012e-02 -4.39622849e-01 -5.93559258e-02
-4.96273905e-01 2.86744028e-01 3.13217968e-01 2.96878874e-01
-1.55184519e+00 9.39155579e-01 7.18248224e+00 1.11809337e+00
-7.33691514e-01 5.67876637e-01 8.42943013e-01 -5.98332435e-02
-8.21145624e-02 3.17324549e-02 -8.06996107e-01 2.97584951e-01
8.19402277e-01 6.24212086e-01 1.49686441e-01 8.97776842e-01
-1.35944530e-01 -2.63440281e-01 -1.32266724e+00 8.77545178e-01
8.07677060e-02 -1.13581145e+00 -4.57913667e-01 -3.07854205e-01
1.37788451e+00 2.34014913e-01 3.80096324e-02 5.73691905e-01
7.81263649e-01 -8.91203880e-01 8.39927733e-01 3.59426111e-01
6.74558401e-01 -5.71090281e-01 1.06608522e+00 5.38576126e-01
-9.17671204e-01 -2.25180089e-01 7.70869702e-02 -5.71276434e-02
-6.57076761e-02 9.85444248e-01 -3.77093077e-01 4.32736397e-01
9.36968625e-01 8.54653597e-01 -9.66198444e-01 9.89724934e-01
-6.29961371e-01 1.15794396e+00 -1.70832932e-01 3.03429306e-01
2.59480983e-01 2.32375219e-01 1.74030423e-01 1.67400014e+00
-2.21038163e-01 -3.70052099e-01 6.72270715e-01 5.38581431e-01
-2.86201566e-01 -6.83571026e-02 -3.58216047e-01 1.08115830e-01
5.78292429e-01 1.55414772e+00 -9.53585148e-01 -6.63031518e-01
-4.81269270e-01 7.38418579e-01 5.46515405e-01 1.59763992e-01
-3.77567202e-01 4.36529852e-02 8.56760368e-02 -4.03292477e-02
-2.41949335e-02 1.11330666e-01 -6.06851041e-01 -9.20294464e-01
-2.71052599e-01 -1.14673841e+00 6.05280936e-01 -5.36908090e-01
-1.63970280e+00 3.02521467e-01 -1.66983783e-01 -1.28781259e+00
7.72607625e-02 -4.97468561e-01 -4.05104309e-02 4.12510246e-01
-1.89902449e+00 -1.32872808e+00 -4.98736143e-01 2.88297653e-01
7.31796086e-01 -7.74119273e-02 9.08351123e-01 1.75048754e-01
-5.62642157e-01 4.67989326e-01 3.01640570e-01 4.65966128e-02
1.14838791e+00 -1.85035896e+00 9.80167910e-02 5.45418561e-01
4.56296414e-01 9.31884646e-02 4.40122247e-01 -3.22769880e-01
-7.28924096e-01 -1.22860193e+00 8.67163956e-01 -7.66002357e-01
4.38485533e-01 -4.09993857e-01 -8.92608941e-01 6.63110077e-01
1.66105136e-01 4.50112343e-01 8.97105038e-01 3.31591100e-01
-8.08303058e-01 -7.02671260e-02 -1.05703819e+00 1.90795466e-01
1.25573921e+00 -6.31797254e-01 -3.67318839e-01 6.79618359e-01
6.20127976e-01 -4.06793863e-01 -8.13144207e-01 4.65331554e-01
1.75485492e-01 -1.09778225e+00 6.27773762e-01 -3.72966796e-01
2.96574980e-01 -3.64506572e-01 -1.10277729e-02 -1.34478569e+00
-3.09089690e-01 -8.92781556e-01 -3.63170654e-01 1.68265724e+00
4.98063475e-01 -1.78245530e-01 7.99996555e-01 3.52399945e-02
-2.95604706e-01 -7.47626007e-01 -6.03130996e-01 -1.00985539e+00
1.67748645e-01 -4.89033788e-01 8.20043012e-02 1.35139740e+00
7.42746517e-02 4.81118321e-01 -4.28326845e-01 -2.55691022e-01
8.12796891e-01 -2.23302647e-01 4.89769310e-01 -1.66741645e+00
-1.84470654e-01 -3.67526054e-01 -2.32849400e-02 -1.00245166e+00
3.29901040e-01 -1.25467348e+00 5.56826770e-01 -1.28887904e+00
5.27495623e-01 -6.53969824e-01 -5.82648218e-01 9.17855918e-01
-5.49615622e-01 6.71433032e-01 1.24900937e-01 5.63894928e-01
-1.25520575e+00 1.29642650e-01 1.07101548e+00 -4.11843032e-01
-1.88179836e-01 -6.46249726e-02 -7.91309536e-01 8.79116774e-01
8.59552085e-01 -9.11684692e-01 -3.03485036e-01 -2.47871757e-01
4.03203845e-01 -7.37974465e-01 -1.55638403e-03 -9.36110675e-01
2.79602319e-01 1.03576869e-01 1.32109702e-01 -3.14325213e-01
-1.09517336e-01 -6.62964463e-01 -2.93936223e-01 3.01241964e-01
-8.81075919e-01 -3.44899744e-01 -2.27598846e-01 6.43620014e-01
-2.59852082e-01 -5.74287355e-01 9.82153475e-01 -3.66345853e-01
-7.80608654e-01 -1.44648507e-01 -3.06900203e-01 6.58194005e-01
7.44551599e-01 1.41713992e-01 -4.27468032e-01 -4.91573326e-02
-9.67618704e-01 4.61907715e-01 2.33310923e-01 1.68391168e-01
2.79407620e-01 -1.21629584e+00 -6.78646386e-01 -1.34147227e-01
4.10860866e-01 3.19581479e-01 -1.99337769e-02 9.04943943e-01
-1.71002582e-01 -5.21900691e-02 3.35582703e-01 -8.47380102e-01
-1.33530903e+00 5.36108851e-01 2.82687813e-01 -6.81481540e-01
-2.40718648e-01 1.09324086e+00 -5.09293497e-01 -9.61728990e-01
6.85780585e-01 -1.68604076e-01 -8.64730850e-02 4.93150145e-01
3.30253035e-01 6.69843197e-01 2.22697601e-01 -4.07352298e-01
-1.14799537e-01 4.49151397e-01 -2.42507026e-01 9.89022031e-02
1.20392144e+00 -2.39941105e-01 -1.38179556e-01 9.52532291e-01
1.13268244e+00 -1.21218354e-01 -1.47456551e+00 -9.04174328e-01
7.10777223e-01 6.54490292e-02 2.41248772e-01 -1.24939167e+00
-9.35991466e-01 7.26522684e-01 8.26615870e-01 4.14656907e-01
1.11032343e+00 4.36077118e-01 5.04065931e-01 4.33465481e-01
2.81832159e-01 -1.62125814e+00 5.73545754e-01 6.02075994e-01
1.74559161e-01 -1.83922851e+00 1.61831990e-01 -7.25970089e-01
-5.09933949e-01 9.34849560e-01 5.40353060e-01 3.02524239e-01
7.88397253e-01 7.42036521e-01 6.25341654e-01 -3.60881656e-01
-7.74230003e-01 -7.35229671e-01 1.03042610e-01 7.90082276e-01
6.02443039e-01 -8.69535729e-02 -2.13024095e-01 4.35484141e-01
2.77630091e-01 -2.85735335e-02 3.54713291e-01 1.45898390e+00
-5.32649696e-01 -1.25880337e+00 -3.15592140e-01 3.83149475e-01
-4.82827425e-01 2.60315668e-02 -5.76786160e-01 6.28033817e-01
7.53818512e-01 1.20446610e+00 -2.54279047e-01 -3.01884264e-01
3.64148766e-01 4.73358721e-01 3.95028859e-01 -9.75388288e-01
-8.33438694e-01 2.14589477e-01 1.66864023e-01 -4.18349504e-01
-1.10573781e+00 -5.63106656e-01 -1.20114470e+00 2.40421653e-01
-7.08565533e-01 1.71691719e-02 5.05035162e-01 1.12061608e+00
5.66137694e-02 5.12885988e-01 7.37260163e-01 -8.68270397e-01
-7.77322590e-01 -1.11503613e+00 -7.27677286e-01 6.62161648e-01
3.35002065e-01 -5.93521655e-01 -7.27420628e-01 2.99415350e-01] | [9.459659576416016, 3.9394540786743164] |
eb6a2827-e10a-44d7-8828-656fb976119d | clustering-by-maximizing-mutual-information | 2107.11635 | null | https://arxiv.org/abs/2107.11635v1 | https://arxiv.org/pdf/2107.11635v1.pdf | Clustering by Maximizing Mutual Information Across Views | We propose a novel framework for image clustering that incorporates joint representation learning and clustering. Our method consists of two heads that share the same backbone network - a "representation learning" head and a "clustering" head. The "representation learning" head captures fine-grained patterns of objects at the instance level which serve as clues for the "clustering" head to extract coarse-grain information that separates objects into clusters. The whole model is trained in an end-to-end manner by minimizing the weighted sum of two sample-oriented contrastive losses applied to the outputs of the two heads. To ensure that the contrastive loss corresponding to the "clustering" head is optimal, we introduce a novel critic function called "log-of-dot-product". Extensive experimental results demonstrate that our method significantly outperforms state-of-the-art single-stage clustering methods across a variety of image datasets, improving over the best baseline by about 5-7% in accuracy on CIFAR10/20, STL10, and ImageNet-Dogs. Further, the "two-stage" variant of our method also achieves better results than baselines on three challenging ImageNet subsets. | ['Svetha Venkatesh', 'Truyen Tran', 'Kien Do'] | 2021-07-24 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Do_Clustering_by_Maximizing_Mutual_Information_Across_Views_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Do_Clustering_by_Maximizing_Mutual_Information_Across_Views_ICCV_2021_paper.pdf | iccv-2021-1 | ['image-clustering'] | ['computer-vision'] | [-2.57876694e-01 -5.28299846e-02 -3.51382233e-02 -6.90228224e-01
-1.09022951e+00 -1.88224062e-01 3.42446506e-01 1.42398998e-01
-5.61759233e-01 1.61956564e-01 -7.23420316e-03 8.78135338e-02
1.38071701e-01 -3.95058125e-01 -9.43913162e-01 -1.02161300e+00
-2.18592778e-01 5.25295556e-01 2.69142121e-01 1.92882180e-01
1.50607631e-01 1.69016242e-01 -1.40489054e+00 5.93199790e-01
7.09533811e-01 1.14090598e+00 3.99512112e-01 2.60529190e-01
2.36616299e-01 1.00814199e+00 -4.06716198e-01 -1.55857652e-01
2.95647353e-01 -2.45304897e-01 -7.69882500e-01 4.65199023e-01
3.72013658e-01 -2.67868359e-02 -4.33833838e-01 1.16880131e+00
3.73499423e-01 3.20176840e-01 7.47108519e-01 -9.50178623e-01
-7.46064484e-01 7.52297938e-01 -1.08505297e+00 2.94597805e-01
-2.77412802e-01 2.63259739e-01 1.14285600e+00 -8.89361262e-01
3.38947535e-01 1.43856490e+00 5.25742412e-01 4.43793625e-01
-1.60144961e+00 -7.72410810e-01 7.22491145e-01 8.39096978e-02
-1.50509024e+00 -3.25681061e-01 6.61499143e-01 -6.30497694e-01
7.18715727e-01 -3.09984356e-01 1.50911063e-01 5.86621583e-01
-2.53377497e-01 1.06862640e+00 9.65143144e-01 -2.03508422e-01
2.69453466e-01 1.02142654e-01 4.93856728e-01 8.33909690e-01
3.25594470e-03 -2.39225507e-01 -1.14577729e-02 7.48747885e-02
6.18739128e-01 7.93615580e-02 -1.44033149e-01 -4.82320726e-01
-9.33816016e-01 1.06804752e+00 1.08745956e+00 6.59264699e-02
-5.19462109e-01 4.41019952e-01 3.14135253e-01 7.63389990e-02
6.13743484e-01 1.55851662e-01 -2.73092210e-01 5.32608151e-01
-9.42175448e-01 1.39270604e-01 4.73598152e-01 8.61570835e-01
9.37192738e-01 -1.53196096e-01 -3.19468796e-01 1.04854691e+00
5.12733400e-01 1.40224054e-01 4.62157875e-01 -1.00644040e+00
6.08743250e-01 7.18946874e-01 -1.04033522e-01 -6.89594626e-01
-2.42447183e-01 -7.69397080e-01 -9.09785569e-01 1.46122202e-01
2.25298226e-01 -8.74803588e-02 -1.15125108e+00 1.94059658e+00
1.79500163e-01 3.73367697e-01 -2.97509223e-01 1.11206150e+00
6.19789720e-01 7.96867847e-01 4.38228965e-01 1.47689417e-01
1.42963910e+00 -1.30166125e+00 -1.78230122e-01 -3.41598928e-01
4.56432223e-01 -4.51536119e-01 1.06083143e+00 2.50267506e-01
-1.15819824e+00 -7.25332975e-01 -8.82474363e-01 -7.68286660e-02
-2.49941498e-01 4.46448743e-01 3.85163724e-01 1.66351065e-01
-1.13354158e+00 5.85145473e-01 -9.02740121e-01 -2.58251894e-02
7.38895476e-01 4.72899944e-01 -8.21294934e-02 -1.37952849e-01
-5.88934362e-01 4.71133411e-01 2.24511743e-01 1.86297577e-02
-1.26379704e+00 -8.16708565e-01 -7.08496988e-01 2.06891865e-01
1.92115143e-01 -6.23199761e-01 9.90327239e-01 -1.00684190e+00
-1.22101378e+00 1.16824865e+00 -1.26473308e-01 -7.11311162e-01
3.64273876e-01 -3.66322964e-01 6.13288321e-02 2.79526234e-01
4.11253780e-01 1.14146280e+00 8.01274538e-01 -1.64881885e+00
-7.81493604e-01 -4.27006841e-01 -3.13647181e-01 2.19631582e-01
-2.72731353e-02 1.26379624e-01 -9.75147009e-01 -7.28113711e-01
-8.29182416e-02 -9.96259093e-01 -6.15738869e-01 -1.55484036e-01
-4.87967461e-01 -4.77732062e-01 6.64375186e-01 -4.50632215e-01
1.05463922e+00 -2.39981437e+00 1.71219528e-01 2.22922787e-01
2.35669032e-01 1.94659397e-01 -2.59923518e-01 -1.14873657e-02
-1.09673396e-01 -1.28984479e-02 -3.29268336e-01 -7.79919088e-01
3.84983309e-02 7.37754581e-03 -2.93540508e-01 5.88072717e-01
1.73818260e-01 7.40057588e-01 -7.25503147e-01 -2.82243818e-01
2.94732600e-01 5.01698673e-01 -6.66357100e-01 3.77736866e-01
-3.69678587e-01 3.54224920e-01 -3.02215397e-01 3.18330407e-01
5.82875073e-01 -8.11822534e-01 1.38822660e-01 -2.71376073e-01
2.74163783e-02 2.79202938e-01 -9.67046916e-01 1.72604525e+00
-4.66735177e-02 3.35342735e-01 1.42054334e-01 -1.21075082e+00
6.61941350e-01 2.15769410e-02 2.41211414e-01 -6.50413096e-01
2.27279797e-01 -2.65536398e-01 -4.36802991e-02 -1.91338032e-01
1.73011258e-01 2.56318510e-01 -1.47534266e-01 3.54384392e-01
1.15987279e-01 4.00127769e-01 3.55727434e-01 4.11903918e-01
8.98532271e-01 -6.08811304e-02 -1.49877846e-01 -5.09347737e-01
4.37295318e-01 -2.64213383e-01 7.03253627e-01 6.69798851e-01
-2.73891747e-01 7.25777626e-01 5.74210107e-01 -2.58249104e-01
-7.82037854e-01 -1.12757850e+00 1.62531048e-01 1.35926795e+00
1.16427504e-01 -1.77341804e-01 -9.54322517e-01 -7.52616763e-01
3.39017287e-02 4.98806149e-01 -9.04419482e-01 -1.07853338e-01
-5.95278203e-01 -1.01845646e+00 4.62329835e-01 7.69808650e-01
7.64514565e-01 -1.06724989e+00 -3.41225475e-01 2.11703181e-01
-2.82058835e-01 -1.25328052e+00 -6.25544429e-01 4.64193106e-01
-8.06165755e-01 -1.06999624e+00 -5.52214146e-01 -1.17639327e+00
9.54915762e-01 5.76374054e-01 1.20051968e+00 1.58403531e-01
-3.62408400e-01 -2.55236365e-02 -2.85757840e-01 -4.05747145e-02
9.47754756e-02 7.00542331e-02 -3.19973528e-01 3.10153514e-01
3.49033922e-01 -3.75192374e-01 -9.51821446e-01 3.17979276e-01
-5.59190750e-01 -4.41085756e-01 7.27121353e-01 5.89609325e-01
9.43895817e-01 9.68517736e-02 4.70568061e-01 -9.26381528e-01
1.97749510e-01 -7.62139559e-01 -7.33415544e-01 1.69714555e-01
-4.14357632e-01 2.00098768e-01 8.18403006e-01 -3.39611590e-01
-8.29944611e-01 3.01111013e-01 1.54137716e-01 -9.27076995e-01
-2.03591928e-01 1.80292040e-01 -2.23306119e-01 3.38507116e-01
4.71071899e-01 3.84556130e-02 -2.00776696e-01 -7.01243401e-01
6.16193831e-01 3.08449239e-01 9.72289741e-01 -6.08909667e-01
5.83260059e-01 6.78493917e-01 -5.31291187e-01 -5.37597597e-01
-1.00821221e+00 -8.39654863e-01 -5.51587045e-01 1.51039794e-01
1.20785403e+00 -1.47050643e+00 -6.29803896e-01 4.55772936e-01
-8.16852093e-01 -5.60879707e-01 -2.23323584e-01 3.94840986e-01
-4.46086168e-01 1.69053614e-01 -9.28197920e-01 -4.52489883e-01
-3.34740609e-01 -1.37131572e+00 1.06843126e+00 2.48302802e-01
1.57285437e-01 -7.69444704e-01 -9.72023234e-02 5.03235221e-01
5.26648425e-02 1.18345320e-01 1.11000800e+00 -5.52900612e-01
-7.32565701e-01 8.77161920e-02 -5.82646132e-01 5.39751768e-01
-2.98587978e-01 -6.88133612e-02 -1.02926457e+00 -4.12201703e-01
-1.79787874e-01 -5.50197542e-01 1.52778912e+00 6.04919910e-01
1.53441322e+00 -2.34723374e-01 -4.53362852e-01 8.01607013e-01
1.60395181e+00 9.94921401e-02 6.91485703e-01 3.28049250e-02
8.17985833e-01 5.05699754e-01 1.92789063e-01 2.69562453e-01
6.60676718e-01 5.56645393e-01 4.45010543e-01 -2.30484053e-01
-3.24408114e-01 -9.80129167e-02 3.21407408e-01 6.84815228e-01
2.60318130e-01 -9.12716910e-02 -5.86097062e-01 8.73936653e-01
-2.15364909e+00 -1.00057530e+00 1.42647550e-01 2.00473905e+00
7.42193818e-01 1.46050140e-01 4.84800458e-01 -1.03504978e-01
9.30150807e-01 1.94565445e-01 -6.52186096e-01 1.04350887e-01
1.83831781e-01 2.25783400e-02 3.77400279e-01 2.79993594e-01
-1.45207322e+00 1.18358457e+00 5.78592634e+00 7.58759856e-01
-9.58061755e-01 -6.03922689e-03 9.71192777e-01 8.91250968e-02
1.06849752e-01 -3.76174273e-03 -7.90011227e-01 6.87455058e-01
7.07463682e-01 4.32630271e-01 4.55010206e-01 9.99685526e-01
1.22508526e-01 -5.28066643e-02 -1.13786137e+00 8.86203408e-01
-1.12446748e-01 -1.28013778e+00 -1.26168234e-02 1.14872478e-01
8.25882673e-01 5.93520999e-01 3.28320861e-01 3.10349464e-01
8.80214930e-01 -9.21654403e-01 7.80906856e-01 2.44929835e-01
5.00111163e-01 -8.39944243e-01 3.90668511e-01 2.18551949e-01
-1.35345924e+00 -1.56441852e-01 -4.83256668e-01 2.34024942e-01
-2.22521499e-01 6.55781686e-01 -5.25799930e-01 1.59659371e-01
1.00179672e+00 7.26159096e-01 -6.08135819e-01 1.13018644e+00
-3.19620699e-01 8.18988323e-01 -2.26885706e-01 3.53925616e-01
5.89735091e-01 -3.22329164e-01 1.96459945e-02 1.48059392e+00
-3.03579599e-01 2.22604424e-01 5.02911568e-01 9.95184302e-01
-6.03457034e-01 -2.24881142e-01 -9.14138258e-02 2.07132101e-01
6.81928635e-01 1.41493022e+00 -1.13081992e+00 -5.06049275e-01
-2.73503184e-01 8.28369796e-01 7.80819178e-01 3.98194402e-01
-8.86714220e-01 -3.65265399e-01 4.54758644e-01 2.62257829e-02
8.74541461e-01 3.30804884e-02 -2.47737125e-01 -9.13723052e-01
-1.40887767e-01 -7.90309966e-01 5.87671876e-01 -5.21684885e-01
-1.53614223e+00 7.01998532e-01 -4.25556935e-02 -9.30508554e-01
-3.77142578e-02 -2.81033218e-01 -7.43987203e-01 5.88817239e-01
-1.46761072e+00 -1.15066111e+00 -2.54207343e-01 8.81081641e-01
6.17659032e-01 -1.32554665e-01 4.51716542e-01 4.11186934e-01
-8.23693573e-01 6.78665340e-01 1.57946214e-01 6.35714412e-01
5.05652726e-01 -1.33366346e+00 2.60174543e-01 8.72089744e-01
5.91566116e-02 5.44940889e-01 3.05593431e-01 -3.98519874e-01
-7.76165545e-01 -1.56110406e+00 4.67296332e-01 -2.58860648e-01
3.45435023e-01 -5.12534440e-01 -9.99806225e-01 8.60501409e-01
3.08836997e-01 2.36001387e-01 3.69107783e-01 1.17115434e-02
-1.00240016e+00 -2.16927975e-01 -1.11792064e+00 4.03175503e-01
7.30513155e-01 -3.99542481e-01 -5.19180059e-01 3.70164007e-01
8.80461991e-01 4.56227958e-02 -7.17313349e-01 3.07597876e-01
1.99818239e-01 -8.95024538e-01 1.01315129e+00 -3.61634344e-01
3.71106088e-01 -4.98558372e-01 -2.93588817e-01 -1.31131077e+00
-7.90283978e-01 -3.82103264e-01 1.68125987e-01 1.08716750e+00
2.93691039e-01 -1.85313910e-01 7.51882613e-01 1.53309733e-01
-2.97976345e-01 -7.74701893e-01 -4.73288625e-01 -6.71617210e-01
2.09538102e-01 -5.46880625e-02 3.21463853e-01 7.95996189e-01
-3.32244366e-01 6.42741501e-01 -4.63986658e-02 2.44186535e-01
1.08290327e+00 6.59624860e-02 6.23447955e-01 -1.17307174e+00
-3.43290180e-01 -6.84706986e-01 -1.65892616e-01 -1.26938510e+00
2.89269328e-01 -9.96551812e-01 2.69876808e-01 -1.37046599e+00
6.00083470e-01 -4.89000559e-01 -6.10725284e-01 5.28329074e-01
-3.37686151e-01 2.60107815e-01 3.08236659e-01 5.72901666e-01
-1.21205652e+00 4.49461520e-01 7.59037554e-01 -2.04053968e-01
-1.73159525e-01 -4.88528796e-02 -9.56472397e-01 9.08393562e-01
5.95465481e-01 -6.38966978e-01 -4.93240744e-01 -6.17797017e-01
-2.56853521e-01 -2.75807679e-01 4.06845599e-01 -1.15035975e+00
4.11941558e-01 2.57351577e-01 4.99770582e-01 -7.74218202e-01
2.18057752e-01 -5.19328356e-01 -2.92563558e-01 2.83394516e-01
-6.35376990e-01 2.55388133e-02 -1.88425362e-01 7.84429610e-01
-1.95335105e-01 2.62877852e-01 1.40179503e+00 -2.52203375e-01
-6.48493528e-01 4.33450460e-01 -1.36340722e-01 3.46016288e-01
1.07672465e+00 2.21551657e-01 -3.40698749e-01 -1.79270282e-01
-5.68118095e-01 6.51115596e-01 3.48675728e-01 2.40165874e-01
4.95293617e-01 -1.05227399e+00 -7.46728301e-01 1.65697992e-01
-1.39375161e-02 3.64635110e-01 1.35980174e-01 7.37721801e-01
-3.59492928e-01 3.16237330e-01 5.52984588e-02 -7.31786013e-01
-8.78075600e-01 7.22296715e-01 4.27873194e-01 -5.67768514e-01
-6.97077990e-01 1.25686955e+00 7.44310915e-01 -2.57380038e-01
7.44260252e-01 -2.12668702e-01 -2.95221759e-03 -1.40309092e-02
5.05145371e-01 3.00961405e-01 -5.91698401e-02 -6.79603338e-01
-5.02539635e-01 5.83025455e-01 -6.75193548e-01 -4.73358296e-02
1.39935231e+00 -1.98661551e-01 -2.95890328e-02 2.55477905e-01
1.58572066e+00 -5.84192574e-01 -1.69903922e+00 -4.01536822e-01
7.32647162e-03 -8.22692215e-02 -2.21832208e-02 -8.20019841e-01
-1.49108195e+00 9.52527523e-01 6.05182886e-01 -8.25345963e-02
1.14236021e+00 2.93078125e-01 7.98542380e-01 2.44347453e-01
1.42449558e-01 -1.15540481e+00 3.94341022e-01 3.13981861e-01
6.17971539e-01 -1.19315374e+00 -1.07457884e-01 -2.03154638e-01
-8.83001804e-01 3.66543233e-01 7.24044502e-01 -8.45595777e-01
8.18928003e-01 1.65697411e-01 1.12483740e-01 -3.83005977e-01
-9.83790398e-01 -4.38996643e-01 3.39039981e-01 4.70976204e-01
2.97762066e-01 1.54583991e-01 9.46362391e-02 4.41693127e-01
1.38149336e-01 -2.06976637e-01 -2.94988807e-02 5.85554719e-01
-6.61452174e-01 -7.17941582e-01 -1.87163785e-01 4.01735276e-01
-7.11279511e-01 2.14656368e-02 -1.84290200e-01 8.07378590e-01
1.54396594e-01 1.10263658e+00 4.24414754e-01 -3.74945015e-01
6.68447390e-02 -1.55291706e-01 3.34872127e-01 -6.17978871e-01
-7.52339423e-01 4.96843547e-01 -3.76856267e-01 -7.46174097e-01
-5.90685904e-01 -4.70529616e-01 -1.53392935e+00 -1.21547118e-01
-2.76796162e-01 2.68894404e-01 5.10064721e-01 8.60411823e-01
4.47353035e-01 4.43436831e-01 9.30189669e-01 -8.80137384e-01
-5.84235132e-01 -8.13839257e-01 -5.49672842e-01 7.84574568e-01
2.41461352e-01 -4.14606035e-01 -3.36134583e-01 1.82085186e-01] | [9.261744499206543, 3.137939453125] |
474399f8-2634-4acf-9a35-6579050dc4de | albmore-a-corpus-of-movie-reviews-for | 2306.08526 | null | https://arxiv.org/abs/2306.08526v1 | https://arxiv.org/pdf/2306.08526v1.pdf | AlbMoRe: A Corpus of Movie Reviews for Sentiment Analysis in Albanian | Lack of available resources such as text corpora for low-resource languages seriously hinders research on natural language processing and computational linguistics. This paper presents AlbMoRe, a corpus of 800 sentiment annotated movie reviews in Albanian. Each text is labeled as positive or negative and can be used for sentiment analysis research. Preliminary results based on traditional machine learning classifiers trained with the AlbMoRe samples are also reported. They can serve as comparison baselines for future research experiments. | ['Erion Çano'] | 2023-06-14 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-5.07563762e-02 8.64642486e-02 -6.94227815e-01 -9.77628946e-01
-6.48498893e-01 -6.16871476e-01 5.52459180e-01 6.90075696e-01
-8.80022109e-01 6.74643874e-01 5.01752853e-01 -4.56354886e-01
6.35046482e-01 -3.20372820e-01 -1.48626819e-01 -1.65049270e-01
2.21581217e-02 3.54791194e-01 -3.08502704e-01 -8.21480453e-01
6.63132787e-01 1.82044342e-01 -1.20257282e+00 7.49716222e-01
5.34698725e-01 4.64544505e-01 -5.21748997e-02 5.02568722e-01
-1.22424170e-01 9.74462748e-01 -9.53753948e-01 -9.62011337e-01
6.14801608e-02 -3.06912810e-01 -8.53100836e-01 1.78232342e-01
6.24706030e-01 -2.68436279e-02 4.90282238e-01 1.01199591e+00
3.67826879e-01 1.18230134e-01 6.64380074e-01 -9.18407083e-01
-9.15910780e-01 7.88079798e-01 -6.91974819e-01 5.80342889e-01
4.08720016e-01 -5.48902333e-01 1.54651737e+00 -1.28791416e+00
9.99381244e-01 1.35402727e+00 3.66793126e-01 5.98794699e-01
-4.48405564e-01 -5.83331764e-01 1.04372896e-01 2.31030565e-02
-7.08928168e-01 -4.02349085e-01 9.26680028e-01 -4.57224071e-01
1.25740254e+00 -4.18295451e-02 8.25925410e-01 7.24642813e-01
5.42907298e-01 9.08906102e-01 1.21375608e+00 -9.40577626e-01
1.08389877e-01 7.71711111e-01 4.84085530e-01 4.42392647e-01
5.21671534e-01 -8.14883411e-01 -7.90551364e-01 -1.32586449e-01
1.83350034e-02 -4.21063989e-01 2.42328748e-01 3.09277982e-01
-6.86668277e-01 1.37795103e+00 4.35340703e-02 4.97150391e-01
-3.05141360e-01 -4.17838603e-01 9.68817830e-01 7.60373473e-01
1.42120826e+00 8.33062708e-01 -9.94354010e-01 -2.11603925e-01
-6.03026569e-01 4.09109950e-01 7.50533581e-01 9.67067719e-01
4.94004548e-01 1.95872664e-01 4.38258797e-01 1.26042271e+00
4.77271616e-01 8.37075770e-01 6.80558503e-01 -4.53494102e-01
7.01042175e-01 8.38529825e-01 5.80543987e-02 -1.27083278e+00
-5.89287281e-01 2.44133830e-01 1.48139983e-01 -3.65773976e-01
1.06498480e-01 -6.61900163e-01 -5.37472844e-01 1.11582625e+00
1.87857673e-01 -9.08473909e-01 6.41619802e-01 8.86757255e-01
1.10333431e+00 7.69725382e-01 3.44465613e-01 -4.63021815e-01
1.70581567e+00 -1.10941887e+00 -9.39789295e-01 -1.01300144e+00
1.03763843e+00 -1.39245677e+00 1.59596884e+00 5.63628554e-01
-1.11090183e+00 4.01477935e-03 -9.60729837e-01 -3.75773042e-01
-8.01014423e-01 4.28099096e-01 1.16746831e+00 1.10067904e+00
-8.95913303e-01 -1.02186911e-01 -6.58019841e-01 -6.69007659e-01
3.47747624e-01 2.15764314e-01 -3.47209692e-01 5.01776524e-02
-1.15415835e+00 9.48735297e-01 -1.64722458e-01 1.05152920e-01
-5.63893095e-02 -1.36150792e-01 -1.39011824e+00 -7.39202261e-01
-1.89049467e-01 5.43063283e-01 1.73122418e+00 -1.45864689e+00
-9.25264299e-01 1.70037258e+00 -6.76582277e-01 -2.72629291e-01
-2.28499293e-01 -4.49432582e-01 -8.31948996e-01 2.23566800e-01
5.93396306e-01 3.44485968e-01 5.15577495e-01 -7.06864059e-01
-7.69065797e-01 -6.02450728e-01 1.00136362e-01 4.32988197e-01
-1.09913135e+00 1.20230031e+00 -5.24434522e-02 -8.99617791e-01
-3.08209546e-02 -8.39745224e-01 -2.43258253e-01 -7.31344461e-01
2.80110985e-02 -5.01079798e-01 5.05809009e-01 -6.52511120e-01
1.11239505e+00 -1.76304674e+00 -4.20002818e-01 -1.73500627e-01
-3.64550620e-01 4.84936908e-02 -7.10752830e-02 7.67255723e-01
-3.57690342e-02 2.82327980e-01 3.26531947e-01 -3.48506927e-01
-3.51224810e-01 1.07723251e-02 -3.56369346e-01 6.10869765e-01
4.69602346e-01 4.95471030e-01 -9.50033188e-01 -6.95796490e-01
-2.25903034e-01 1.34571403e-01 -3.58206213e-01 -2.70012826e-01
1.47879347e-01 -5.04559651e-02 -5.87045670e-01 8.39273393e-01
2.65648633e-01 1.81525320e-01 3.13850194e-01 1.32434800e-01
-2.55581737e-01 8.33505213e-01 -3.72270018e-01 1.03347278e+00
-6.80708528e-01 1.06113636e+00 8.39105397e-02 -8.44304264e-01
1.31995296e+00 1.92285538e-01 -7.41634797e-03 -6.70160890e-01
4.16855782e-01 3.04681957e-01 3.42854089e-03 -5.70824564e-01
1.18036580e+00 -5.18840969e-01 -5.16589701e-01 5.44775844e-01
-1.33045450e-01 -5.18906891e-01 8.22368860e-01 4.08259779e-01
4.13493782e-01 -3.11911821e-01 5.04131138e-01 -7.25278914e-01
7.99901545e-01 7.98740745e-01 5.49970329e-01 9.66951847e-02
-4.05560255e-01 2.16328532e-01 5.72002828e-01 -5.44526458e-01
-9.97272015e-01 -1.80254892e-01 -5.05803287e-01 1.75578821e+00
-2.83026427e-01 -8.37509811e-01 -6.09030366e-01 -7.18218088e-01
-3.62722605e-01 5.70206523e-01 -6.19744837e-01 3.31409335e-01
-6.44116819e-01 -8.99617076e-01 2.04048008e-01 5.67956686e-01
-1.43555313e-01 -1.35089672e+00 -3.74436498e-01 -3.33020128e-02
-9.58028659e-02 -1.28402877e+00 -1.28405184e-01 8.63504112e-02
-7.51018167e-01 -1.07255483e+00 -3.16541046e-01 -1.43915522e+00
7.58046091e-01 3.85221988e-01 1.30439794e+00 5.69168963e-02
1.44161984e-01 1.44227356e-01 -9.38865840e-01 -1.37342000e+00
-3.42465550e-01 1.06812663e-01 2.64945596e-01 -3.83374095e-01
1.37905955e+00 2.98109889e-01 -1.36490554e-01 -7.43325129e-02
-5.96856236e-01 -3.65824163e-01 3.57892290e-02 7.33406425e-01
1.49545416e-01 -1.52064815e-01 1.08065784e+00 -1.68131292e+00
1.00832522e+00 -5.67926168e-01 -2.16048822e-01 -1.42007545e-01
-4.83552277e-01 -8.02841961e-01 7.58790910e-01 -1.69435233e-01
-1.19555211e+00 -2.91985601e-01 -2.00093105e-01 6.50649428e-01
8.37706700e-02 9.19830918e-01 3.96277875e-01 2.80543983e-01
8.34619701e-01 -6.43548191e-01 -1.62183985e-01 -8.31884667e-02
-4.18070853e-02 1.40267277e+00 -1.15063585e-01 -2.91854411e-01
1.76558286e-01 3.94326746e-01 -6.20230496e-01 -1.33575487e+00
-1.37663114e+00 -7.07824707e-01 -5.80079973e-01 -3.66258115e-01
6.39594078e-01 -1.18121016e+00 8.31134766e-02 3.27007860e-01
-7.69868076e-01 -1.30121112e-01 -1.32157385e-01 6.57165527e-01
-4.06805538e-02 -8.48198384e-02 -9.93987203e-01 -1.02237189e+00
-6.19621217e-01 -7.48244107e-01 9.17935908e-01 3.04894835e-01
-7.19948947e-01 -1.52481031e+00 2.95675099e-01 8.82773876e-01
-8.31457451e-02 -2.13675767e-01 5.87713122e-01 -1.14634120e+00
9.01363552e-01 -8.77859533e-01 3.62687230e-01 6.23558402e-01
1.05357833e-01 2.96195745e-01 -9.02939022e-01 -2.63499290e-01
2.56757170e-01 -1.17399478e+00 4.18391258e-01 1.88554391e-01
4.04723912e-01 -5.20833969e-01 1.74617857e-01 -5.07783532e-01
1.08544672e+00 1.21250212e-01 1.69894069e-01 6.47092760e-01
3.05242985e-01 1.25146198e+00 1.39137053e+00 4.61716264e-01
7.72326350e-01 -1.36452511e-01 -3.80189687e-01 3.41118500e-02
4.99956995e-01 -2.40131766e-02 9.30222332e-01 1.38626027e+00
3.42175871e-01 -2.32108340e-01 -1.28125477e+00 1.08204031e+00
-1.42195892e+00 -6.65954411e-01 -3.72183710e-01 1.44686639e+00
9.79531407e-01 7.82569572e-02 -6.27183635e-03 2.47697249e-01
6.98389471e-01 5.29863596e-01 -6.73822537e-02 -1.21326971e+00
-3.99923980e-01 4.39002275e-01 1.50075719e-01 4.18379813e-01
-1.35039914e+00 1.51017368e+00 7.28310013e+00 4.07912791e-01
-1.12915468e+00 2.18453795e-01 8.13374877e-01 -7.97051117e-02
-1.76619217e-01 -9.84348431e-02 -1.03334451e+00 -1.86607316e-01
1.18614328e+00 -2.52453059e-01 -1.83138102e-01 1.21742785e+00
4.00964946e-01 -4.14734930e-01 -4.99321193e-01 4.21630353e-01
7.57643521e-01 -9.55708385e-01 -1.15690313e-01 -3.50446939e-01
1.05944240e+00 4.12115604e-01 -5.12489453e-02 4.51662809e-01
1.78002313e-01 -7.38481104e-01 7.28160858e-01 -3.59889954e-01
5.34165323e-01 -8.25352728e-01 1.40323424e+00 2.10462630e-01
-7.59349406e-01 1.30628467e-01 -7.88887858e-01 -9.17396903e-01
2.10426793e-01 2.70180047e-01 -8.49525869e-01 -2.94347405e-02
8.75723958e-01 1.25482595e+00 -9.26859796e-01 7.42780641e-02
-4.80463684e-01 1.21720505e+00 -7.06457794e-02 -9.02882755e-01
4.35140580e-01 -3.65212321e-01 1.53014600e-01 1.49395084e+00
-3.32962096e-01 1.45465553e-01 2.94777185e-01 -2.17250571e-01
-5.85118271e-02 1.24929500e+00 -7.57144511e-01 -6.84142590e-01
1.76110640e-01 1.77474463e+00 -1.30132473e+00 -3.52201849e-01
-1.11668599e+00 3.90945673e-01 4.20765311e-01 2.30363011e-01
1.92305539e-02 -5.81379890e-01 1.23459652e-01 1.38385504e-01
-1.86662927e-01 -5.06885275e-02 -5.81946313e-01 -1.42298758e+00
3.61707583e-02 -1.11218393e+00 4.48127955e-01 -7.23992288e-01
-1.70036912e+00 7.43014395e-01 -1.88515440e-01 -7.70682871e-01
-2.35795811e-01 -9.94456828e-01 -4.60463077e-01 6.39970839e-01
-1.37579668e+00 -1.19664741e+00 1.84289649e-01 2.98751265e-01
9.45767105e-01 -6.98415816e-01 7.55182445e-01 1.67697057e-01
-5.38427413e-01 3.89111638e-01 -3.13422620e-01 3.43359739e-01
1.02535892e+00 -1.18556893e+00 3.94305110e-01 7.30827928e-01
5.51931970e-02 7.57616103e-01 7.06771016e-01 -9.52738941e-01
-1.21585047e+00 -7.97412515e-01 1.71340442e+00 -5.76298833e-01
1.23525238e+00 -2.74198413e-01 -5.27032673e-01 8.22853148e-01
6.22732520e-01 -4.12472516e-01 1.36213005e+00 6.08587444e-01
-1.94138616e-01 3.56879123e-02 -1.10371029e+00 6.96000278e-01
2.95153201e-01 -5.52217126e-01 -8.15766394e-01 8.98651600e-01
3.83178294e-01 -1.36618435e-01 -7.69197702e-01 2.00885683e-01
3.10158402e-01 -4.42856669e-01 1.77127719e-01 -1.10549283e+00
8.65201890e-01 2.77568251e-01 -1.45286724e-01 -1.13482547e+00
2.53670692e-01 -3.53393257e-01 6.41379297e-01 1.31421959e+00
9.98312354e-01 -5.07730663e-01 7.51466393e-01 7.24328399e-01
-1.95913747e-01 -6.98482037e-01 -2.77678996e-01 -2.31308252e-01
6.22409701e-01 -8.65090668e-01 6.76872507e-02 1.28152144e+00
9.07344580e-01 1.15260375e+00 -6.16480671e-02 -5.00471175e-01
-1.44687638e-01 -7.73389637e-02 7.01449990e-01 -1.05780208e+00
5.46908319e-01 -1.51834205e-01 -3.56535405e-01 -4.44011569e-01
6.87108576e-01 -7.37466216e-01 -6.31142175e-03 -1.55344951e+00
9.06917080e-02 -1.47319317e-01 3.59466791e-01 5.68415046e-01
-2.47050542e-02 9.16797280e-01 -6.24917708e-02 4.01507504e-03
-3.54010791e-01 3.08966339e-01 1.02314055e+00 1.18813492e-01
-2.53147840e-01 -2.53113419e-01 -1.11906743e+00 1.30153394e+00
1.16674626e+00 -4.44028020e-01 -4.20411080e-01 -3.45562607e-01
1.02989507e+00 -5.40274501e-01 -6.08703017e-01 -4.00058836e-01
-1.86542556e-01 -4.38585371e-01 1.23321220e-01 -5.35168290e-01
1.37288466e-01 -4.23617393e-01 -1.07413888e+00 -9.61476490e-02
-5.51795900e-01 7.92504787e-01 3.32977384e-01 -1.40112564e-01
-5.57103217e-01 -8.70977759e-01 5.93354523e-01 -1.52364209e-01
-4.54420805e-01 -1.96226910e-01 -9.54293132e-01 4.70134288e-01
9.60606873e-01 7.93578401e-02 -4.26900625e-01 -5.25901735e-01
-3.17455947e-01 2.93106169e-01 6.19154334e-01 7.56901622e-01
5.72595417e-01 -8.61965537e-01 -8.93958211e-01 -8.29405244e-03
5.07464826e-01 -5.30831337e-01 -1.47857174e-01 4.75182503e-01
-9.82096314e-01 4.92691040e-01 -6.53935075e-02 6.66365027e-02
-1.31709802e+00 2.32979864e-01 -3.77407312e-01 -7.95724895e-03
-2.73354173e-01 7.31607199e-01 -2.85944253e-01 -6.20294750e-01
-1.51420370e-01 2.36447081e-01 -1.08625126e+00 6.56104386e-01
8.46457541e-01 1.04597032e-01 9.43231285e-02 -1.41262352e+00
-3.59463185e-01 2.33647674e-01 -4.12297964e-01 -6.49976254e-01
1.41561389e+00 -3.40840697e-01 -7.34678209e-01 1.19673705e+00
9.31528687e-01 5.68334103e-01 1.11487947e-01 9.37263891e-02
3.32295626e-01 -3.88998240e-01 1.17968515e-01 -4.39531296e-01
-9.37468231e-01 6.49588048e-01 1.62893988e-03 4.22286659e-01
8.99532795e-01 7.25793242e-02 4.94892776e-01 8.79030883e-01
1.26206353e-01 -1.66805470e+00 -4.04727086e-02 8.30467880e-01
9.38209176e-01 -1.60697818e+00 2.90420145e-01 -3.88037145e-01
-1.40452456e+00 1.11142278e+00 8.85945261e-01 -4.50050026e-01
1.01489115e+00 3.00180137e-01 8.92864823e-01 -6.17098451e-01
-8.68829966e-01 -3.00524980e-02 2.31291607e-01 3.94339681e-01
1.37362599e+00 -1.54369593e-01 -1.36330128e+00 8.58586192e-01
-8.90523553e-01 -5.00577390e-01 1.10569751e+00 1.13661754e+00
-4.16367829e-01 -1.08629644e+00 -2.49577478e-01 7.14429915e-01
-1.37666285e+00 -3.86321038e-01 -8.98198724e-01 7.99902976e-01
-3.92432243e-01 1.56934094e+00 3.56195450e-01 -6.48733135e-03
1.45463720e-01 -9.15452838e-02 -7.66134709e-02 -1.24586320e+00
-1.03524399e+00 3.34502041e-01 8.85672390e-01 1.03288433e-02
-1.11430097e+00 -1.11450756e+00 -1.31520772e+00 -1.89760432e-01
-4.99796748e-01 7.59233713e-01 8.01165223e-01 9.01218176e-01
-7.80508444e-02 -2.40344033e-01 5.38030028e-01 -5.00949681e-01
-6.61790296e-02 -1.47930622e+00 -7.83955097e-01 2.78052121e-01
5.29375970e-02 -2.12239176e-01 -3.91276032e-01 3.58057529e-01] | [11.206375122070312, 6.872526168823242] |
0c8a81bd-fd6e-4620-949f-0191183f77a5 | a-novel-blaschke-unwinding-adaptive-fourier | 1803.06441 | null | http://arxiv.org/abs/1803.06441v1 | http://arxiv.org/pdf/1803.06441v1.pdf | A Novel Blaschke Unwinding Adaptive Fourier Decomposition based Signal Compression Algorithm with Application on ECG Signals | This paper presents a novel signal compression algorithm based on the
Blaschke unwinding adaptive Fourier decomposition (AFD). The Blaschke unwinding
AFD is a newly developed signal decomposition theory. It utilizes the
Nevanlinna factorization and the maximal selection principle in each
decomposition step, and achieves a faster convergence rate with higher
fidelity. The proposed compression algorithm is applied to the
electrocardiogram signal. To assess the performance of the proposed compression
algorithm, in addition to the generic assessment criteria, we consider the less
discussed criteria related to the clinical needs -- for the heart rate
variability analysis purpose, how accurate the R peak information is preserved
is evaluated. The experiments are conducted on the MIT-BIH arrhythmia benchmark
database. The results show that the proposed algorithm performs better than
other state-of-the-art approaches. Meanwhile, it also well preserves the R peak
information. | ['Hau-Tieng Wu', 'Liming Zhang', 'Chunyu Tan'] | 2018-03-17 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [ 3.87436897e-01 -3.18511486e-01 -1.51495054e-01 2.51171850e-02
-5.07969379e-01 -2.05036610e-01 -3.40859443e-02 2.87745029e-01
-2.84416914e-01 8.15707624e-01 2.65023440e-01 -2.46518880e-01
-6.93752468e-01 -2.63903141e-01 -2.07646787e-02 -8.92435014e-01
-3.43441457e-01 6.62311018e-02 -2.43693292e-01 -4.26823609e-02
1.64580807e-01 4.61393625e-01 -1.15540063e+00 5.94970807e-02
1.03743088e+00 1.21940506e+00 -4.40278873e-02 6.52491748e-01
6.39977217e-01 4.78598118e-01 -4.02617991e-01 -6.51557371e-02
3.84121954e-01 -8.33109021e-01 -5.46578526e-01 -1.50499910e-01
-3.58059227e-01 -5.07368684e-01 -1.53662562e-01 8.88961852e-01
7.91231334e-01 -1.00253455e-01 6.24153376e-01 -7.05959558e-01
2.37408996e-01 5.05905867e-01 -6.24326468e-01 7.52763212e-01
3.51774007e-01 -2.77497858e-01 6.50895000e-01 -8.18744838e-01
5.98213434e-01 6.12622559e-01 1.00786972e+00 -5.52625433e-02
-1.26237345e+00 -4.56597060e-01 -6.18364751e-01 3.94456565e-01
-1.59077442e+00 -2.16513529e-01 1.11141396e+00 -3.23280424e-01
6.08956456e-01 5.54504514e-01 8.42094183e-01 5.45379817e-01
4.97078508e-01 2.86915570e-01 1.07354844e+00 -6.43396497e-01
3.23620364e-02 -3.96058261e-01 5.41674316e-01 4.62400973e-01
5.04029393e-01 1.83231980e-01 -5.37856758e-01 -5.49152553e-01
6.33578539e-01 -1.10319205e-01 -7.57543445e-01 -1.26780123e-01
-1.39346480e+00 5.34599841e-01 -1.51941955e-01 6.97013080e-01
-6.63318515e-01 1.37905190e-02 7.87130952e-01 2.86760271e-01
3.79143447e-01 1.23349801e-01 -1.86199509e-02 -3.35000604e-01
-1.48045754e+00 3.97479683e-01 5.37877560e-01 4.64417964e-01
1.81371287e-01 2.83395797e-01 -3.11741501e-01 4.86724436e-01
3.56295019e-01 4.07224745e-01 7.31908739e-01 -9.57935035e-01
3.19570243e-01 1.42100304e-01 -6.95988759e-02 -1.34139359e+00
-6.19448125e-01 -8.69972348e-01 -1.42461145e+00 -5.37071191e-02
2.07536966e-01 -1.18034340e-01 -2.99207270e-01 1.66055298e+00
2.51404047e-01 3.61587614e-01 2.65669107e-01 9.47221756e-01
7.00009942e-01 5.72071373e-01 -2.62047917e-01 -1.19962049e+00
1.32846177e+00 -3.40944529e-01 -1.34490979e+00 3.14562351e-01
1.59547120e-01 -8.63181055e-01 5.10281563e-01 8.12556624e-01
-1.29308164e+00 -8.26357663e-01 -1.39279556e+00 2.94328302e-01
4.64768350e-01 3.90138268e-01 3.85408312e-01 8.97293150e-01
-8.26713681e-01 8.24652970e-01 -7.59780467e-01 -2.04017296e-01
1.51722562e-02 7.94528723e-02 -2.91708797e-01 6.08374812e-02
-1.32007372e+00 7.38604546e-01 4.21268612e-01 3.51918340e-02
-4.85306859e-01 -7.60716319e-01 -4.80302185e-01 2.39804536e-01
-6.37041256e-02 -7.79384613e-01 8.90413105e-01 -4.91333693e-01
-1.19538212e+00 5.58692694e-01 -1.89103469e-01 -7.70417035e-01
6.12541139e-01 -1.60946578e-01 -5.23926556e-01 7.26087511e-01
-2.03090772e-01 -2.69280672e-01 1.04553759e+00 -8.45309675e-01
-6.40623197e-02 -5.02504408e-01 -6.63085878e-01 2.44642228e-01
-9.53276455e-02 -1.57545641e-01 -1.02038756e-01 -1.21250153e+00
6.69376552e-01 -6.29521906e-01 -5.71772642e-02 -3.37395608e-01
-1.48233280e-01 3.95685583e-01 6.01435661e-01 -9.91621375e-01
1.90240335e+00 -2.60702229e+00 2.90048480e-01 6.87288880e-01
3.87700558e-01 2.20373765e-01 4.42813933e-01 6.00781381e-01
-3.72139871e-01 -1.55981347e-01 -5.30916452e-01 4.45480905e-02
-3.93559098e-01 -5.50167263e-02 -3.08588505e-01 7.78133571e-01
-4.92187381e-01 3.30259025e-01 -5.08977413e-01 -4.50182974e-01
1.96235612e-01 5.76154649e-01 -4.16342914e-01 5.88210151e-02
7.66468525e-01 5.11098325e-01 -1.65376335e-01 2.86880761e-01
7.93028057e-01 8.13571438e-02 4.81077760e-01 -7.92637050e-01
4.11008857e-03 -4.37053218e-02 -1.35310876e+00 1.71211386e+00
2.46923149e-01 3.27390581e-01 2.75076795e-02 -1.15860486e+00
1.14618742e+00 8.06485355e-01 9.66958702e-01 -3.51111412e-01
1.72333792e-01 4.60833669e-01 3.10673594e-01 -6.04917347e-01
1.69191033e-01 -3.76311064e-01 2.72619784e-01 3.58999997e-01
-4.56205243e-03 -1.08665479e-02 4.22094822e-01 1.68807968e-01
9.51145828e-01 -4.62588929e-02 1.03171825e+00 -8.27659488e-01
7.56467044e-01 -3.36248159e-01 7.34478593e-01 4.85429406e-01
-3.78508955e-01 6.83660567e-01 4.61141288e-01 -4.23996955e-01
-8.49187911e-01 -9.29598868e-01 -4.81054366e-01 2.25867704e-01
2.36600135e-02 -7.92335272e-01 -8.45122576e-01 -7.12076724e-02
-1.45649955e-01 4.41593319e-01 -3.70244354e-01 -1.39180958e-01
-4.31675881e-01 -1.04657924e+00 7.18748808e-01 2.87856400e-01
5.57115972e-01 -5.80168724e-01 -1.16364384e+00 4.56581503e-01
-6.10255659e-01 -9.36197937e-01 -3.57678652e-01 -1.47918954e-01
-1.44989622e+00 -1.01141500e+00 -9.55858111e-01 -3.50350320e-01
3.45518112e-01 3.49288285e-02 9.07536626e-01 5.63336909e-02
-4.50534075e-01 3.49950671e-01 -6.34412706e-01 -2.47544110e-01
-4.75429386e-01 -2.31866226e-01 2.14925051e-01 2.66839385e-01
-7.06675882e-03 -1.03914487e+00 -8.11547995e-01 9.25301835e-02
-8.36563051e-01 -1.15300557e-02 5.35612404e-01 8.08680832e-01
8.23074102e-01 3.28498393e-01 6.90702379e-01 -8.02910268e-01
1.06890452e+00 -9.00669172e-02 -2.62079656e-01 -4.92444187e-02
-1.05497348e+00 -2.14810953e-01 5.52253842e-01 -1.16172142e-01
-7.02060878e-01 -2.57406011e-02 -2.05574632e-01 -4.18408364e-01
3.29564899e-01 6.25824690e-01 1.26257360e-01 -5.29590324e-02
7.10493028e-01 5.36228001e-01 2.25277573e-01 -4.47879672e-01
-7.83943236e-02 4.44620460e-01 7.61776686e-01 -5.12442052e-01
4.61967856e-01 3.86995733e-01 3.41278821e-01 -9.68256712e-01
-6.31522611e-02 -6.00621283e-01 -5.77615023e-01 -3.30214202e-01
7.31659293e-01 -5.87971985e-01 -7.28197455e-01 3.34955364e-01
-1.01384389e+00 3.93750995e-01 -3.75216454e-01 8.98552179e-01
-8.09438229e-01 9.78092015e-01 -5.87198019e-01 -9.35754538e-01
-9.95989978e-01 -8.73659492e-01 4.24956411e-01 -1.11380495e-01
-4.12491739e-01 -4.75165427e-01 3.66138190e-01 -3.05735003e-02
3.72377038e-01 7.39976585e-01 1.07639289e+00 -4.77981001e-01
4.23726216e-02 -3.12426388e-01 3.26435983e-01 4.74044204e-01
-3.66429836e-02 -1.73256293e-01 -6.18394852e-01 -5.03196776e-01
6.20189488e-01 3.29763383e-01 6.28468752e-01 6.28740728e-01
1.00856733e+00 -2.19627857e-01 -1.38953164e-01 8.05999517e-01
1.41991425e+00 3.54691535e-01 8.95843148e-01 4.04337049e-02
-9.64652449e-02 1.63304225e-01 5.49087584e-01 9.44110572e-01
-1.92286342e-01 4.84578490e-01 1.88447595e-01 -6.62227646e-02
-7.64606595e-02 9.24699232e-02 8.29816684e-02 1.44597566e+00
-6.75590277e-01 7.45790452e-02 -6.64129376e-01 3.68240267e-01
-1.85720932e+00 -1.17768610e+00 -3.38019818e-01 2.29409695e+00
5.63935935e-01 1.25600561e-01 1.87095627e-01 1.16205907e+00
5.63045681e-01 2.24135444e-01 -3.10871154e-01 -3.89346093e-01
-1.49227351e-01 4.32531059e-01 1.47912800e-01 3.71459603e-01
-8.09801757e-01 -1.65222719e-01 6.70214272e+00 7.04718232e-01
-9.65272725e-01 2.73578048e-01 3.85418385e-01 2.09845275e-01
9.37938169e-02 -1.72752634e-01 -2.52588749e-01 3.91226441e-01
9.59082544e-01 -6.07811809e-01 3.38084430e-01 4.99985218e-01
5.49008369e-01 -6.69105947e-02 -6.56347036e-01 1.44683409e+00
1.02857299e-01 -1.10725236e+00 -2.34395176e-01 -1.45516828e-01
1.90615937e-01 -5.09649158e-01 -1.88317373e-01 -2.13914469e-01
-1.02263236e+00 -6.89766586e-01 5.72846651e-01 7.19362438e-01
1.04285896e+00 -9.11348522e-01 8.40488315e-01 4.35060918e-01
-1.44453311e+00 -1.04457833e-01 -3.19516122e-01 1.01180650e-01
4.50003773e-01 1.19757974e+00 -5.78052044e-01 1.17551219e+00
4.11211520e-01 6.05583191e-01 -1.92921489e-01 1.10921741e+00
-1.20212048e-01 8.69633198e-01 -2.78363585e-01 4.58014190e-01
-3.49158257e-01 -4.27949458e-01 9.13853884e-01 1.18291748e+00
7.14732170e-01 6.12102330e-01 -9.30299461e-02 5.63952625e-01
3.36573839e-01 5.75596869e-01 -4.44164544e-01 6.40627891e-02
3.87262821e-01 1.01967132e+00 -5.32246053e-01 -4.59666371e-01
-1.69842303e-01 6.31413460e-01 -5.50326049e-01 1.75136179e-01
-6.57074213e-01 -6.29058838e-01 1.32488459e-01 3.48562777e-01
-8.68835002e-02 -2.19013274e-01 -4.45147842e-01 -1.04124415e+00
9.77371559e-02 -1.15245783e+00 6.35088325e-01 -6.14538908e-01
-8.33711982e-01 8.05783033e-01 2.35846564e-01 -1.60720980e+00
-4.49691772e-01 -1.37795702e-01 -4.45218742e-01 9.82712746e-01
-1.07471812e+00 -5.98519862e-01 -1.98976308e-01 7.16160417e-01
2.48013213e-01 -1.87154591e-01 1.05874729e+00 6.11739397e-01
-1.43168688e-01 5.73909581e-01 9.75231901e-02 -2.38698646e-01
4.39797789e-01 -7.15939879e-01 -2.58281499e-01 1.16634130e+00
-1.97706595e-01 6.70401096e-01 9.45831478e-01 -5.83103061e-01
-1.11828685e+00 -5.06134450e-01 9.19544041e-01 4.76733536e-01
2.92006675e-02 2.65639365e-01 -8.94317091e-01 2.18850300e-01
2.62535512e-01 -6.56960383e-02 8.40553761e-01 -2.83292443e-01
2.07485780e-02 -5.45027971e-01 -1.35196280e+00 -4.65175621e-02
5.34746349e-01 -3.88546526e-01 -9.70757782e-01 6.64559230e-02
1.80810764e-01 -4.41327423e-01 -1.32988715e+00 8.54983151e-01
8.04939508e-01 -1.22504508e+00 1.02142572e+00 -2.20208719e-01
2.55484253e-01 -4.72025931e-01 -2.41740167e-01 -9.41359878e-01
-6.04598165e-01 -1.10009027e+00 -3.13511431e-01 1.04444242e+00
6.88973367e-02 -5.91281772e-01 3.39839399e-01 1.67222135e-02
-5.29761836e-02 -7.21403837e-01 -1.19996202e+00 -7.04857111e-01
-5.02493799e-01 -2.55716622e-01 2.72583365e-01 8.80716622e-01
5.19283891e-01 2.68293798e-01 -7.15738952e-01 -8.85262489e-02
8.68368030e-01 1.61413744e-01 2.62544394e-01 -1.29336703e+00
-6.07651949e-01 -1.81673303e-01 -5.20203888e-01 -4.39117432e-01
-5.25207937e-01 -8.11206877e-01 -3.91613424e-01 -1.20293486e+00
2.50185728e-01 -3.53668667e-02 -6.63256347e-01 1.69191379e-02
-1.54803082e-01 4.09721792e-01 3.54744554e-01 4.12051558e-01
5.64913452e-02 2.63598353e-01 1.02304375e+00 1.85935453e-01
-2.85122454e-01 2.24975482e-01 -5.36620200e-01 5.04320979e-01
6.98262751e-01 -4.06632602e-01 -7.16280937e-01 2.11474404e-01
-7.03953626e-03 6.55059099e-01 9.22732463e-04 -1.41850221e+00
-1.08498007e-01 4.88639176e-01 2.33166263e-01 -8.48733664e-01
2.30361342e-01 -8.23958158e-01 7.44496942e-01 1.04439032e+00
-2.34384164e-01 5.13682544e-01 5.49784228e-02 4.71200377e-01
-3.75838459e-01 -2.97247708e-01 1.07619846e+00 1.14887841e-01
-3.37594636e-02 3.86872031e-02 -4.50459450e-01 -5.65157682e-02
9.57069159e-01 -1.92381084e-01 8.04460049e-02 -2.96169311e-01
-9.66627479e-01 -4.49847728e-01 -4.79862234e-03 -2.90124893e-01
8.72112393e-01 -1.36868703e+00 -1.02394056e+00 4.53321040e-01
-9.14934874e-02 -5.22034824e-01 6.43834591e-01 1.45233572e+00
-7.92183757e-01 3.59632045e-01 -4.69977349e-01 -6.82138920e-01
-1.42268145e+00 5.02422035e-01 2.06138089e-01 -5.60100138e-01
-1.06179070e+00 2.51903415e-01 -1.82977021e-01 4.61326391e-01
-6.63031638e-02 -3.25371623e-01 -3.85175973e-01 3.72747816e-02
7.12087691e-01 7.94593751e-01 3.64709198e-01 -5.83515167e-01
-4.37478185e-01 7.45842218e-01 4.67754722e-01 -2.88569272e-01
1.09867847e+00 -2.49909833e-01 -3.15568298e-01 3.04776818e-01
9.19590652e-01 2.27794439e-01 -4.59806919e-01 6.81833923e-02
-2.69634008e-01 -4.91995424e-01 1.78821519e-01 -7.04657555e-01
-9.75885570e-01 8.62702966e-01 1.02460420e+00 3.52939457e-01
1.73064327e+00 -9.18694437e-01 7.21158147e-01 -2.78551001e-02
3.88406992e-01 -8.03581119e-01 -2.90521473e-01 -5.22877313e-02
8.84674907e-01 -2.24165127e-01 5.68610251e-01 -4.23113793e-01
-4.92401212e-01 1.43580067e+00 -2.81365514e-01 -1.72086194e-01
9.09259677e-01 2.47635186e-01 -1.26205966e-01 1.34392187e-01
-3.34905326e-01 2.37168893e-01 2.72776514e-01 5.52329004e-01
5.52549362e-01 4.74184602e-02 -1.52812219e+00 9.60899472e-01
-1.18034460e-01 4.43299949e-01 4.55997229e-01 5.45937657e-01
-3.91784102e-01 -1.08327878e+00 -6.62294030e-01 4.73475218e-01
-8.08702528e-01 1.13578692e-01 1.30307540e-01 6.69261336e-01
6.88522682e-02 9.14686322e-01 -5.56299686e-01 -3.02773893e-01
3.21268171e-01 2.11189657e-01 4.79768842e-01 1.02803456e-02
-6.11066759e-01 5.58923781e-01 6.73116371e-02 -5.40869594e-01
-7.02557564e-01 -7.98503816e-01 -1.21080530e+00 -2.22347632e-01
-9.31924284e-02 4.93989766e-01 5.66298723e-01 7.21717894e-01
3.50812376e-01 8.32550883e-01 6.41780734e-01 -4.99208868e-01
-6.55224502e-01 -9.30581570e-01 -1.04802465e+00 3.99111569e-01
3.33273947e-01 -4.35274541e-01 -2.80506223e-01 3.04585636e-01] | [14.20455265045166, 3.225172996520996] |
6f73aacf-dab5-40fa-a0fb-f5adca378f1f | distilling-knowledge-from-deep-networks-with | 1512.03542 | null | http://arxiv.org/abs/1512.03542v1 | http://arxiv.org/pdf/1512.03542v1.pdf | Distilling Knowledge from Deep Networks with Applications to Healthcare Domain | Exponential growth in Electronic Healthcare Records (EHR) has resulted in new
opportunities and urgent needs for discovery of meaningful data-driven
representations and patterns of diseases in Computational Phenotyping research.
Deep Learning models have shown superior performance for robust prediction in
computational phenotyping tasks, but suffer from the issue of model
interpretability which is crucial for clinicians involved in decision-making.
In this paper, we introduce a novel knowledge-distillation approach called
Interpretable Mimic Learning, to learn interpretable phenotype features for
making robust prediction while mimicking the performance of deep learning
models. Our framework uses Gradient Boosting Trees to learn interpretable
features from deep learning models such as Stacked Denoising Autoencoder and
Long Short-Term Memory. Exhaustive experiments on a real-world clinical
time-series dataset show that our method obtains similar or better performance
than the deep learning models, and it provides interpretable phenotypes for
clinical decision making. | ['Sanjay Purushotham', 'Zhengping Che', 'Yan Liu', 'Robinder Khemani'] | 2015-12-11 | null | null | null | null | ['computational-phenotyping'] | ['medical'] | [ 2.50635684e-01 2.48222932e-01 2.80568246e-02 -8.50016832e-01
-5.64240873e-01 6.47310093e-02 -1.76905259e-01 3.37411761e-01
1.70869097e-01 7.90564358e-01 3.90127718e-01 -4.12123233e-01
-6.05899513e-01 -5.47103822e-01 -6.02014244e-01 -7.19419003e-01
-3.43827069e-01 8.67212057e-01 -7.71796882e-01 1.62184075e-01
-2.96557397e-01 2.80607313e-01 -1.18888032e+00 9.80969846e-01
9.34464455e-01 1.13997746e+00 2.27436516e-02 6.97026789e-01
7.05453306e-02 1.29224539e+00 -4.51604307e-01 -4.55391139e-01
-5.20787165e-02 -4.08539504e-01 -6.98031545e-01 -1.76715717e-01
-2.49354810e-01 -3.13963324e-01 -2.06519186e-01 7.08173037e-01
7.36784697e-01 -5.71554184e-01 5.79078853e-01 -1.03429770e+00
-1.02001178e+00 5.12149036e-01 -2.26442702e-03 1.24241427e-01
2.10069567e-01 2.00701922e-01 7.57418692e-01 -4.07080382e-01
4.11897629e-01 1.30986750e+00 1.22244489e+00 6.43240392e-01
-1.29800582e+00 -3.24999899e-01 -8.15096274e-02 4.10288304e-01
-1.02756536e+00 -1.79403022e-01 3.87665123e-01 -7.23210573e-01
9.69629288e-01 4.65047479e-01 6.08769655e-01 1.43086791e+00
7.59409785e-01 6.33660853e-01 8.84820342e-01 -6.43689781e-02
2.95203269e-01 -2.82044470e-01 4.69948441e-01 8.29589844e-01
3.47144872e-01 1.75010398e-01 -3.01654816e-01 -8.43383789e-01
5.91748357e-01 8.11769128e-01 -2.06878275e-01 1.55271992e-01
-1.21090293e+00 7.32502341e-01 3.67823839e-01 7.77721703e-02
-9.57431376e-01 6.63094893e-02 4.53290254e-01 4.57029372e-01
6.37624323e-01 5.34914076e-01 -9.44736660e-01 8.88298303e-02
-4.37862128e-01 1.56146407e-01 5.88789105e-01 4.07046348e-01
2.94932395e-01 3.39009166e-02 -3.66082400e-01 6.62048280e-01
1.86189920e-01 4.75536793e-01 8.74505043e-01 -7.18230665e-01
1.83720794e-02 1.04424942e+00 -5.09048700e-02 -9.59235013e-01
-7.80417085e-01 -5.00406086e-01 -1.36226165e+00 -2.63660073e-01
1.28738031e-01 -2.44631678e-01 -1.00854445e+00 1.46707749e+00
3.27170253e-01 4.36693281e-01 2.84666359e-01 6.78502917e-01
8.46972406e-01 4.33192939e-01 4.55518067e-01 4.40090932e-02
1.47671556e+00 -4.08697218e-01 -8.91330481e-01 2.22841740e-01
8.35024118e-01 -2.32829988e-01 7.01175272e-01 5.91282785e-01
-7.47122645e-01 -4.06409025e-01 -7.89511204e-01 -1.38094306e-01
-2.47999430e-01 2.30354309e-01 9.70244229e-01 4.06382799e-01
-6.80505335e-01 9.58558381e-01 -1.09563696e+00 -1.24317169e-01
8.13201725e-01 4.57499623e-01 -3.60306591e-01 -2.36844756e-02
-1.01010180e+00 5.09330392e-01 3.39686126e-01 2.23580629e-01
-7.45492518e-01 -1.10305583e+00 -6.67176664e-01 1.75590962e-01
-2.07434505e-01 -1.49272013e+00 1.11520314e+00 -1.07305670e+00
-1.27701116e+00 7.84384787e-01 -2.93105721e-01 -8.07694137e-01
3.28956306e-01 -2.84696579e-01 -6.01436079e-01 -6.01824000e-02
-1.63832188e-01 2.81308144e-01 7.38234341e-01 -6.23929501e-01
-3.53906810e-01 -9.48376298e-01 -5.81651449e-01 -3.77556562e-01
-2.72035986e-01 3.45790386e-02 2.63266832e-01 -7.75818825e-01
-3.77366804e-02 -7.98906028e-01 -4.86995578e-01 -1.14752613e-01
-6.65779591e-01 -2.97966212e-01 6.43355370e-01 -1.17431080e+00
1.14110732e+00 -1.94154406e+00 1.67012289e-02 -5.84713928e-02
7.83246934e-01 3.03081781e-01 -1.06408253e-01 1.75281063e-01
-4.60751832e-01 4.49526012e-02 -1.78738326e-01 -1.05727233e-01
-3.59236985e-01 4.42788333e-01 -2.73649216e-01 2.41409555e-01
6.59037411e-01 1.19433367e+00 -7.78603077e-01 -7.29231685e-02
1.42875314e-01 6.77572906e-01 -5.09144723e-01 6.06896996e-01
-2.77936459e-01 7.65725076e-01 -9.19461071e-01 7.52580464e-01
4.95875955e-01 -1.04514825e+00 5.01209617e-01 4.68301810e-02
6.64317548e-01 1.71079874e-01 -5.04435420e-01 1.30574048e+00
-3.65632802e-01 2.40116701e-01 -3.47220510e-01 -1.30107999e+00
8.74917328e-01 5.23410141e-01 5.86253226e-01 -4.71649200e-01
1.28539383e-01 -3.79487127e-03 2.83247918e-01 -1.02172565e+00
-5.32593131e-01 -2.33537987e-01 1.81461498e-01 1.96423039e-01
-2.87557513e-01 7.20669568e-01 -5.58447540e-01 -3.28783751e-01
1.56494415e+00 -6.58607408e-02 5.38606167e-01 -2.52254069e-01
4.20031250e-01 4.57501337e-02 1.07758904e+00 5.81518948e-01
-1.36246085e-01 5.77362418e-01 4.04863209e-01 -1.55637193e+00
-9.03027058e-01 -9.19736922e-01 -4.55722541e-01 8.59182417e-01
-5.26902020e-01 -1.34045005e-01 -6.28302097e-01 -7.00559378e-01
5.22940278e-01 3.89251202e-01 -9.60430145e-01 -2.70845711e-01
-4.51877028e-01 -1.31656957e+00 6.78271651e-01 6.57912552e-01
4.69739065e-02 -1.25309706e+00 -6.09713256e-01 4.59516227e-01
-1.10733606e-01 -7.85315037e-01 1.57539815e-01 2.73813635e-01
-1.24168229e+00 -1.41451502e+00 -4.65749204e-01 -5.57175100e-01
7.13896692e-01 -4.92715418e-01 1.36089003e+00 1.59247279e-01
-8.96940410e-01 -1.87198356e-01 -2.02351153e-01 -9.10508990e-01
-6.20400012e-01 -2.65958548e-01 2.04948470e-01 2.78975874e-01
9.09874082e-01 -5.07042885e-01 -7.95354724e-01 -6.39786422e-02
-7.30046809e-01 1.98651269e-01 7.28608370e-01 1.36146522e+00
8.20098341e-01 -1.46991387e-01 1.15002739e+00 -1.22912431e+00
7.28204310e-01 -9.50242102e-01 -1.96194381e-01 3.60679597e-01
-8.93409431e-01 4.39627707e-01 1.13726091e+00 -1.01568475e-01
-1.01282382e+00 -3.95417251e-02 -4.29756373e-01 -3.38138133e-01
-5.20214736e-01 5.86655915e-01 -2.72764601e-02 3.19005221e-01
6.42578483e-01 2.99408019e-01 1.66206792e-01 -1.00066507e+00
-1.76967114e-01 1.03370452e+00 2.23870546e-01 -2.31863305e-01
3.95793207e-02 5.19052744e-01 5.31860329e-02 -4.23996568e-01
-9.05178249e-01 -1.78071558e-01 -3.19009751e-01 4.00659114e-01
9.55633581e-01 -1.06333518e+00 -1.00072765e+00 4.43661153e-01
-1.06622064e+00 2.06803493e-02 -2.17503548e-01 3.94991159e-01
-6.32252216e-01 2.43196070e-01 -9.37567592e-01 -4.12478745e-01
-9.33692813e-01 -1.14166427e+00 1.24976826e+00 -8.59863535e-02
-6.35604501e-01 -1.24031734e+00 3.26930732e-01 5.47012389e-01
2.40663707e-01 6.73256636e-01 1.59099972e+00 -1.01976728e+00
-3.22103441e-01 -3.29687953e-01 -2.55213320e-01 3.51439387e-01
2.76487947e-01 -3.25185448e-01 -1.19867492e+00 -1.22256547e-01
1.22837134e-01 -2.81078964e-01 7.58996367e-01 6.92918241e-01
2.08151913e+00 -5.53608418e-01 -4.84707206e-01 1.11788702e+00
1.19477630e+00 4.07194525e-01 5.78874290e-01 -1.95567474e-01
6.53149068e-01 5.13154507e-01 7.30272755e-02 6.76600099e-01
4.79051679e-01 1.80393219e-01 5.59644103e-02 -4.47607636e-01
6.33658096e-02 -2.53500372e-01 -1.63882658e-01 8.85065317e-01
-1.40815198e-01 -9.06423777e-02 -1.04180002e+00 2.64941514e-01
-2.16916442e+00 -6.66525841e-01 -2.27738708e-01 1.82275224e+00
7.77290404e-01 -4.72523689e-01 -6.26254901e-02 4.17658165e-02
4.83149320e-01 -5.17337263e-01 -1.03098691e+00 -5.67010641e-01
-1.69955179e-01 3.34230334e-01 -5.79066463e-02 -1.95673108e-02
-9.07533109e-01 1.96001738e-01 6.97891378e+00 4.61778343e-01
-1.09154367e+00 1.88542500e-01 1.50143993e+00 7.37187490e-02
-1.84384823e-01 -5.70794344e-01 -2.26865038e-01 7.18547642e-01
1.30105603e+00 4.44254726e-02 1.14998922e-01 9.48013306e-01
5.48232257e-01 4.98320520e-01 -1.56097901e+00 1.25462091e+00
-4.01897907e-01 -1.64835501e+00 2.71242797e-01 1.39940932e-01
5.26749253e-01 1.29623115e-01 1.25312895e-01 4.46912944e-02
5.30449927e-01 -1.45674229e+00 -2.02935115e-01 9.38882947e-01
6.26138687e-01 -5.42460501e-01 1.05440104e+00 2.86382765e-01
-4.89957750e-01 -4.23894823e-01 -4.55870390e-01 -4.42793608e-01
-3.21275681e-01 1.07414711e+00 -1.21691608e+00 5.37909091e-01
7.70873368e-01 9.99064863e-01 -2.35955864e-01 7.33842790e-01
1.75329804e-01 8.43659580e-01 1.29386619e-01 6.33457527e-02
-1.90342531e-01 -9.08274949e-02 1.45244807e-01 1.09269774e+00
2.08105668e-01 1.46085009e-01 3.72696891e-02 9.56781983e-01
-1.28856048e-01 1.12793140e-01 -6.31361902e-01 -2.79693127e-01
3.26606221e-02 1.00125468e+00 -1.54713854e-01 -5.55246711e-01
-2.16715083e-01 8.70730698e-01 4.93858963e-01 2.53398776e-01
-6.55311763e-01 -6.33636937e-02 9.90049243e-01 1.04625359e-01
1.00381605e-01 4.92495447e-01 -6.51675105e-01 -1.24751163e+00
-5.96342348e-02 -1.20952189e+00 8.14055383e-01 -4.65643167e-01
-1.88468385e+00 6.87894762e-01 -6.36837602e-01 -8.91014338e-01
-3.61197501e-01 -8.03927481e-01 -5.37223577e-01 9.42701697e-01
-1.50744200e+00 -9.08372521e-01 -1.82587132e-01 5.00604212e-01
2.82680064e-01 -3.79883826e-01 1.39018774e+00 1.81766197e-01
-8.81596923e-01 5.28418183e-01 5.94145536e-01 1.57637358e-01
4.04819518e-01 -1.21506357e+00 4.33984846e-01 2.46013492e-01
-2.15555727e-01 7.48913646e-01 4.15311247e-01 -7.08200872e-01
-1.60390878e+00 -1.57732129e+00 1.04115069e+00 -6.70269430e-01
2.84008294e-01 7.41922669e-03 -1.00044119e+00 8.31628561e-01
-7.85259455e-02 7.39267468e-02 1.52232134e+00 3.75552654e-01
-2.79735744e-01 -3.24452490e-01 -1.40373397e+00 2.00236782e-01
8.59204412e-01 -2.98912048e-01 -5.99653721e-01 6.50194466e-01
5.03488660e-01 8.67676958e-02 -1.31406200e+00 7.19931304e-01
6.85941219e-01 -8.58242869e-01 1.03718460e+00 -1.48182452e+00
6.14387035e-01 -6.70857681e-03 2.43267715e-01 -1.33165944e+00
-7.44279742e-01 -5.01823068e-01 -4.86760139e-01 3.48784685e-01
3.78846377e-01 -8.83261204e-01 8.68442655e-01 7.79326022e-01
3.86713929e-02 -1.28966975e+00 -7.46496260e-01 -3.38282228e-01
-4.51415293e-02 -2.37753645e-01 1.06026304e+00 1.13706076e+00
2.71178409e-02 3.09613109e-01 -4.83410299e-01 2.72698760e-01
6.03128076e-01 3.50254685e-01 4.27152187e-01 -1.50084889e+00
-6.86919928e-01 -8.44047517e-02 -7.72076249e-01 -5.32497644e-01
4.13040854e-02 -9.31250572e-01 -4.84579831e-01 -1.36705470e+00
6.33272052e-01 -4.55243796e-01 -7.17739284e-01 7.32664287e-01
-6.72183514e-01 -6.17050529e-02 -3.27275187e-01 7.12924525e-02
-4.48167503e-01 3.60918909e-01 1.05362487e+00 -2.83544093e-01
-9.36957821e-02 1.66696161e-01 -9.24967349e-01 9.00252879e-01
8.55390429e-01 -5.16502738e-01 -2.95387030e-01 -6.65692866e-01
1.25871196e-01 3.68490756e-01 4.33547288e-01 -7.21808434e-01
-2.60431677e-01 2.22028838e-03 9.94797587e-01 -2.50409096e-01
7.96761438e-02 -5.81890523e-01 5.31883836e-01 9.86548245e-01
-4.01662916e-01 2.24777356e-01 3.08853621e-03 7.87311077e-01
-1.96530864e-01 3.07313412e-01 4.06459272e-01 -8.12742710e-02
-4.06227082e-01 4.12602246e-01 -4.65785950e-01 -1.35281488e-01
8.76224637e-01 -3.46773565e-02 -4.94787879e-02 -1.53088495e-01
-1.14419365e+00 2.16443658e-01 -5.67175597e-02 3.47741574e-01
8.19192410e-01 -1.16106188e+00 -1.07797658e+00 4.41060066e-01
1.96627736e-01 -1.80739418e-01 5.05330682e-01 6.76475763e-01
-7.39803910e-01 6.21639073e-01 -1.46259993e-01 -6.64365351e-01
-1.17482543e+00 7.08740592e-01 5.36551476e-01 -6.04905784e-01
-8.23591709e-01 7.88806498e-01 5.10033786e-01 -5.55598319e-01
9.02114138e-02 -7.52814531e-01 4.60319109e-02 -5.52477598e-01
8.99208784e-01 3.36821437e-01 3.11245620e-01 -5.85911497e-02
-3.99442077e-01 1.34402499e-01 -2.10182533e-01 1.05484378e+00
1.66534555e+00 3.34361553e-01 -1.24262273e-01 2.34476596e-01
1.13393307e+00 -7.66147375e-01 -8.76437843e-01 -6.14251709e-03
2.45614424e-01 -1.22118115e-01 1.96733885e-03 -1.24723482e+00
-9.31474566e-01 9.77042019e-01 8.79908741e-01 1.73596799e-01
1.31410515e+00 -2.17261031e-01 1.06550837e+00 7.95046747e-01
9.93133634e-02 -5.45655906e-01 -3.74822617e-01 4.08037864e-02
6.16341650e-01 -1.47014225e+00 -2.62330413e-01 -2.38338485e-01
-5.21920323e-01 1.17996466e+00 1.97437599e-01 -5.87664619e-02
7.96390533e-01 2.50871330e-01 3.91422719e-01 -3.09558392e-01
-1.14035189e+00 3.20880860e-01 1.71727449e-01 7.85700619e-01
6.20197177e-01 6.26936495e-01 -4.71735708e-02 1.21061540e+00
-7.05059431e-03 6.03377044e-01 -6.58088252e-02 5.13925314e-01
-7.14784637e-02 -1.10492074e+00 -2.98455387e-01 1.02813208e+00
-9.09778118e-01 -2.20644355e-01 -3.19504380e-01 1.67722732e-01
2.78917670e-01 7.84454346e-01 6.52173832e-02 -3.94079626e-01
7.69553781e-02 4.44008797e-01 2.47348502e-01 -4.26694900e-01
-6.81962311e-01 -2.75857091e-01 6.77141920e-02 -7.36681044e-01
-3.50788683e-02 -3.84739757e-01 -1.21270967e+00 -2.98602819e-01
2.42648393e-01 2.53001936e-02 2.51422107e-01 1.00335085e+00
1.43701899e+00 7.49670088e-01 5.17913640e-01 -1.64616540e-01
-8.72685671e-01 -8.74284267e-01 -6.15280747e-01 8.65158319e-01
5.83261907e-01 -2.85776645e-01 2.29539335e-01 4.28145945e-01] | [7.930600166320801, 6.307981014251709] |
b895801f-15c1-4745-bffa-953c40c5b50a | turing-at-semeval-2017-task-8-sequential | 1704.07221 | null | http://arxiv.org/abs/1704.07221v1 | http://arxiv.org/pdf/1704.07221v1.pdf | Turing at SemEval-2017 Task 8: Sequential Approach to Rumour Stance Classification with Branch-LSTM | This paper describes team Turing's submission to SemEval 2017 RumourEval:
Determining rumour veracity and support for rumours (SemEval 2017 Task 8,
Subtask A). Subtask A addresses the challenge of rumour stance classification,
which involves identifying the attitude of Twitter users towards the
truthfulness of the rumour they are discussing. Stance classification is
considered to be an important step towards rumour verification, therefore
performing well in this task is expected to be useful in debunking false
rumours. In this work we classify a set of Twitter posts discussing rumours
into either supporting, denying, questioning or commenting on the underlying
rumours. We propose a LSTM-based sequential model that, through modelling the
conversational structure of tweets, which achieves an accuracy of 0.784 on the
RumourEval test set outperforming all other systems in Subtask A. | ['Maria Liakata', 'Isabelle Augenstein', 'Elena Kochkina'] | 2017-04-24 | turing-at-semeval-2017-task-8-sequential-1 | https://aclanthology.org/S17-2083 | https://aclanthology.org/S17-2083.pdf | semeval-2017-8 | ['rumour-detection'] | ['natural-language-processing'] | [-2.47471675e-01 3.41291487e-01 -1.22026302e-01 -3.50927591e-01
-3.01248610e-01 -1.44236535e-01 1.27315795e+00 4.59551424e-01
-1.05106227e-01 8.47607374e-01 7.94469535e-01 -5.80049455e-01
4.15707737e-01 -6.25832856e-01 -4.66328084e-01 -3.80655676e-01
-7.35717043e-02 6.36252940e-01 7.35821277e-02 -8.18760276e-01
8.07253778e-01 -1.89212888e-01 -1.35874462e+00 1.06595051e+00
3.20966482e-01 1.03283989e+00 -4.70476210e-01 6.28098965e-01
-1.97796077e-01 2.17879653e+00 -9.60467935e-01 -4.52521771e-01
-4.28372443e-01 -6.12074673e-01 -1.70134056e+00 6.66251257e-02
5.06064117e-01 -3.64441574e-01 -1.33147493e-01 9.90776896e-01
2.02872843e-01 -8.00007731e-02 4.81643140e-01 -1.12247682e+00
-4.32158798e-01 1.22568679e+00 -3.01576048e-01 6.84138834e-01
6.16253257e-01 -1.12065442e-01 1.01025867e+00 -9.81956899e-01
4.30899501e-01 1.21143317e+00 9.97214139e-01 4.45714325e-01
-1.06463265e+00 -6.86397493e-01 -2.46408150e-01 6.31815016e-01
-6.53816819e-01 -6.40748799e-01 6.69439852e-01 -9.41598892e-01
6.33226871e-01 4.47344452e-01 4.29162085e-01 1.69159150e+00
4.27450210e-01 8.35185528e-01 1.86816633e+00 1.67637676e-01
1.45489156e-01 6.48117542e-01 7.11652994e-01 5.32074690e-01
-1.53561793e-02 -2.63766587e-01 -9.50422525e-01 -6.27843380e-01
6.33582324e-02 -3.89047354e-01 -4.36800182e-01 5.19302011e-01
-1.17509532e+00 1.23177814e+00 7.18158185e-01 1.93641543e-01
-9.22347844e-01 -3.79120439e-01 9.28651631e-01 1.11649561e+00
1.24484003e+00 6.19022131e-01 5.14655635e-02 -3.53363082e-02
-1.23820496e+00 6.13363624e-01 1.39399469e+00 2.53627419e-01
3.53876621e-01 2.37580448e-01 -1.24059372e-01 1.08666253e+00
-5.88949062e-02 5.02951853e-02 7.95558333e-01 -5.39164841e-01
3.18708032e-01 4.88915235e-01 5.44817209e-01 -1.13528526e+00
-8.70923102e-01 -4.82503951e-01 -8.87843132e-01 -3.21131982e-02
1.74682006e-01 -1.54208109e-01 -3.17531228e-01 1.17285776e+00
1.54124901e-01 -8.50380436e-02 6.97996691e-02 9.60733891e-01
1.35863543e+00 6.06405199e-01 -3.47948253e-01 -4.92154807e-01
1.48288691e+00 -8.38358164e-01 -8.02455425e-01 -2.53537625e-01
3.76613498e-01 -9.97136950e-01 5.91380239e-01 5.62901914e-01
-1.24528861e+00 -8.43357146e-02 -1.06295967e+00 1.73704535e-01
-8.22274834e-02 -5.83126128e-01 1.44621536e-01 2.45363325e-01
-9.11591589e-01 6.88880503e-01 -1.87643766e-01 -3.79977524e-02
4.03436720e-01 -1.66735068e-01 1.49988785e-01 1.41889557e-01
-1.80303299e+00 1.37694395e+00 2.44177178e-01 6.59914538e-02
-1.22551370e+00 -2.94518352e-01 -6.04210079e-01 -3.01058710e-01
1.40819669e-01 -3.79828215e-01 1.75881648e+00 -1.06193793e+00
-1.29445922e+00 1.33656454e+00 -1.20625816e-01 -1.26108205e+00
1.03827834e+00 -1.07948497e-01 -4.62317646e-01 -3.48973215e-01
2.06085086e-01 -2.19791695e-01 1.33058763e+00 -1.09071231e+00
-5.42771399e-01 -2.61080801e-01 -1.42495230e-01 -4.79226373e-02
4.37855348e-02 5.31328440e-01 9.80399489e-01 -9.61713195e-02
1.84815809e-01 -7.29722738e-01 3.70818526e-01 -1.16142976e+00
-8.27653766e-01 -5.11098862e-01 6.66638911e-01 -9.48740005e-01
1.08362806e+00 -1.42759395e+00 -2.53776163e-01 -1.82587653e-01
5.83672941e-01 3.11324865e-01 5.51775873e-01 6.49418354e-01
1.78098995e-02 -5.10842428e-02 1.69101283e-01 -4.58781660e-01
-9.57747921e-02 -1.50302216e-01 -1.02180910e+00 8.58989894e-01
-2.09231302e-01 6.68168604e-01 -9.69142318e-01 -8.82718489e-02
-3.37558955e-01 6.39300793e-02 -9.54406220e-04 4.96333390e-01
-2.61772513e-01 4.37896401e-01 9.23026726e-03 2.97886103e-01
5.76514542e-01 -4.06797767e-01 7.28873014e-02 2.82876134e-01
-4.15100455e-01 1.44270372e+00 -2.44434729e-01 2.72469968e-01
-3.54180932e-01 1.14357173e+00 3.10813040e-01 -6.07386172e-01
1.23003161e+00 7.25213170e-01 -1.10713176e-01 -4.57054794e-01
5.45718491e-01 4.53015029e-01 -6.50104359e-02 -5.60594320e-01
1.07343411e+00 -8.43917906e-01 -1.81969985e-01 1.12162352e+00
-4.81725216e-01 -1.37022868e-01 -3.77689809e-01 3.04187000e-01
7.82762766e-01 -5.63733160e-01 6.41027212e-01 -4.80796069e-01
6.80006623e-01 1.35284767e-01 9.56430882e-02 8.12646210e-01
-3.62841338e-01 3.27900082e-01 9.28215683e-01 -7.79029489e-01
-9.73317325e-01 -3.21069807e-01 -1.26895726e-01 1.29092598e+00
-2.74199009e-01 1.11417994e-01 -5.47179520e-01 -6.19387269e-01
1.62437372e-02 1.26614285e+00 -1.06683910e+00 1.53748319e-01
-5.35738826e-01 -8.59461606e-01 7.25526392e-01 -1.20201319e-01
5.56851029e-01 -1.28454602e+00 -3.62719506e-01 3.81141007e-01
-8.81874502e-01 -1.05495381e+00 -1.68562412e-01 1.33011927e-04
-5.36382794e-01 -1.18038082e+00 -3.67889404e-01 -5.39382577e-01
-2.49053184e-02 5.67748368e-01 1.35485601e+00 4.20734704e-01
6.09212816e-01 -3.59158546e-01 -3.92106742e-01 -5.54200411e-01
-1.23945522e+00 -6.57080039e-02 1.58841178e-01 1.12226032e-01
3.53994399e-01 -4.84055758e-01 -2.15063035e-01 5.63666999e-01
-4.86471683e-01 7.44005218e-02 -1.51869193e-01 9.28874850e-01
-4.36635017e-01 -4.58602548e-01 1.07741451e+00 -1.35504067e+00
1.35925519e+00 -1.28281415e+00 4.75326143e-02 -5.74770272e-01
-6.90736115e-01 -5.46633601e-01 6.45697474e-01 -8.40734988e-02
-8.54354680e-01 -1.24690342e+00 -3.39134127e-01 1.87584177e-01
3.98174077e-02 6.22409165e-01 8.34040344e-01 2.23213524e-01
1.09467602e+00 4.60550398e-01 2.20524386e-01 -4.21157658e-01
-1.34603322e-01 1.37477720e+00 3.63411605e-01 -2.76005603e-02
2.35646948e-01 5.75789094e-01 -4.06420887e-01 -9.36853230e-01
-1.88656688e+00 -7.15003610e-01 -1.73669159e-01 -3.31714660e-01
2.58357257e-01 -8.81872535e-01 -1.04022813e+00 7.11671054e-01
-1.47057843e+00 -2.21734658e-01 3.93283159e-01 -1.18510984e-01
-5.11519909e-01 2.87959039e-01 -1.22341847e+00 -9.97134984e-01
-8.40368807e-01 -6.33571565e-01 7.92277381e-02 -4.84020919e-01
-9.38760817e-01 -1.06966078e+00 3.40961576e-01 9.86793160e-01
7.11928189e-01 3.04479539e-01 5.86008072e-01 -1.39971459e+00
2.29608864e-01 -2.53230751e-01 -2.69889832e-01 4.50500816e-01
-2.04554483e-01 -5.66613376e-01 -1.13424754e+00 -3.63951981e-01
7.42130578e-01 -1.01315796e+00 1.18186283e+00 -8.85384083e-02
5.28743267e-01 -1.22346222e+00 3.72008532e-01 -1.16043791e-01
6.82965338e-01 -7.75990009e-01 4.98186678e-01 9.12665188e-01
3.57492357e-01 9.88128066e-01 4.81218636e-01 6.68664157e-01
8.01892519e-01 4.05345738e-01 9.52513158e-01 4.38770324e-01
1.84114166e-02 -2.04585686e-01 9.30935264e-01 9.26999331e-01
6.27645105e-02 -9.19701345e-03 -8.48201156e-01 7.53986478e-01
-1.72200549e+00 -1.25290668e+00 -1.15907240e+00 1.94072843e+00
1.02481484e+00 3.93838614e-01 5.62582195e-01 2.00133458e-01
8.04051399e-01 8.34596932e-01 -6.19802848e-02 -1.08177722e+00
-1.25572652e-01 -5.72102487e-01 5.96215837e-02 8.82892430e-01
-9.19814467e-01 7.12964773e-01 5.50228643e+00 -2.06352565e-02
-1.21514082e+00 6.09034479e-01 4.13211793e-01 -1.61211103e-01
-1.38967738e-01 -3.34865928e-01 -6.35414422e-01 6.57226384e-01
1.37351227e+00 -1.17548294e-01 3.58392119e-01 6.63277686e-01
5.96643984e-01 -4.64984067e-02 -1.04807305e+00 3.81911159e-01
5.80714703e-01 -1.40874362e+00 -1.74068715e-02 -9.61281285e-02
9.29515660e-01 6.94642127e-01 5.45198610e-03 5.65646291e-01
4.39609438e-01 -1.15452218e+00 1.14015484e+00 1.99575961e-01
-1.14867901e-02 -5.66898823e-01 1.31631625e+00 1.09492230e+00
2.41037846e-01 -3.39090861e-02 -3.93025309e-01 -7.16398001e-01
2.87087917e-01 7.09621489e-01 -1.90019047e+00 -1.32396324e-02
5.59359908e-01 7.24003315e-01 -2.11039141e-01 6.05120242e-01
-5.99146426e-01 1.22522712e+00 3.83095026e-01 -1.11676104e-01
5.29749453e-01 4.00309443e-01 1.04930997e+00 1.52084017e+00
-1.94000393e-01 -3.49991679e-01 9.81042907e-02 9.25286591e-01
-5.45790970e-01 -9.40988734e-02 -3.27891290e-01 1.02853984e-01
3.46302897e-01 1.03210795e+00 1.96527932e-02 -5.75679898e-01
8.47204477e-02 8.70907843e-01 4.40602869e-01 -2.49963209e-01
-5.86802900e-01 3.34151655e-01 4.51901913e-01 3.95621598e-01
-9.36437678e-03 3.34121317e-01 -4.14571136e-01 -1.00178397e+00
-4.48244601e-01 -1.25828218e+00 4.05144960e-01 -5.90317547e-01
-1.41715753e+00 9.39976692e-01 -6.22732699e-01 -8.09296668e-01
-4.70602959e-01 -1.65035412e-01 -1.11327553e+00 8.99413228e-01
-1.91427302e+00 -1.20981467e+00 -5.38295805e-01 2.61847913e-01
7.78092444e-01 6.07694201e-02 6.18488014e-01 -1.48040563e-01
-3.90045583e-01 4.36653234e-02 -2.29457736e-01 8.80423188e-02
8.90352786e-01 -1.18744230e+00 6.71140254e-01 4.93259802e-02
-4.18831617e-01 4.23463106e-01 1.70614314e+00 -5.65764785e-01
-6.49720252e-01 -1.02725482e+00 1.47187269e+00 -7.82501161e-01
1.35406637e+00 7.95804113e-02 -1.47723162e+00 9.69786942e-01
5.10426521e-01 -6.93047166e-01 6.22016609e-01 4.71491933e-01
-6.10226393e-01 6.22750580e-01 -1.07235360e+00 1.18025787e-01
2.50263751e-01 -6.83140814e-01 -1.22570479e+00 6.85410261e-01
3.14498276e-01 -3.96298617e-01 -7.21288681e-01 -2.07317293e-01
2.93272644e-01 -1.55420446e+00 5.08865952e-01 -1.10536838e+00
1.15069151e+00 2.11528689e-01 8.48824903e-02 -1.66416335e+00
-1.98672965e-01 -6.36026800e-01 -2.71847457e-01 8.24048102e-01
3.15553427e-01 -9.25284982e-01 5.61765790e-01 -1.98987976e-01
1.62052624e-02 -5.31581521e-01 -7.86175728e-01 -4.12830800e-01
3.55572730e-01 -3.47317368e-01 3.77963066e-01 1.64003527e+00
6.05195701e-01 8.43552232e-01 -1.14296758e+00 -7.25792572e-02
7.80909538e-01 4.22185868e-01 8.03701282e-01 -1.36252451e+00
9.15658176e-02 -6.21461868e-01 1.00107398e-02 -9.25702810e-01
3.68320018e-01 -1.08235633e+00 8.12129229e-02 -1.45474899e+00
4.39382941e-01 6.47173002e-02 -1.86030343e-01 1.18313812e-01
-6.23459406e-02 4.61723447e-01 -1.35923490e-01 8.42596292e-01
-5.51167667e-01 4.90429878e-01 9.37247157e-01 -1.52144596e-01
5.57660833e-02 6.46347106e-01 -8.59724164e-01 1.09742522e+00
9.44608927e-01 -5.38033426e-01 4.22883481e-01 1.47020981e-01
7.53427327e-01 4.29207355e-01 7.09741890e-01 -3.01538050e-01
2.27231592e-01 4.33377549e-02 -3.69918823e-01 -7.59403169e-01
4.99422789e-01 9.24621895e-02 -5.06169140e-01 5.35677075e-01
-7.95443654e-01 -1.42408624e-01 -1.87662959e-01 5.15760422e-01
-2.74686396e-01 -3.22797716e-01 9.84184444e-01 -3.71078253e-01
-2.52726227e-01 -2.65279144e-01 -9.20564055e-01 3.05279940e-01
3.04336041e-01 3.38788360e-01 -8.91642272e-01 -1.09616733e+00
-6.84387743e-01 1.90942377e-01 4.24354136e-01 3.96320641e-01
6.36207402e-01 -7.52095222e-01 -1.75266147e+00 -2.66939819e-01
2.09084842e-02 -6.80698752e-01 2.01797366e-01 1.59099424e+00
-1.38054237e-01 4.03075010e-01 -1.19358793e-01 -2.86291867e-01
-1.53925514e+00 2.85656571e-01 3.63604397e-01 -2.18651965e-01
-5.58986902e-01 9.50134099e-01 -3.49063575e-01 -3.84798139e-01
-2.29157984e-01 2.81913966e-01 -6.83595181e-01 5.19971728e-01
1.11868525e+00 8.73743296e-01 3.15841973e-01 -1.06232309e+00
1.13395823e-03 -7.33817458e-01 -6.11333609e-01 2.06270307e-01
1.24770582e+00 -3.42374951e-01 -9.83911157e-01 1.04554415e+00
9.21519756e-01 -2.73875035e-02 -4.86707449e-01 -6.58292770e-01
3.77891958e-01 -2.82203674e-01 3.74592245e-01 -1.00627327e+00
-1.23293132e-01 7.43502200e-01 -4.89557236e-01 1.17194664e+00
-3.02171689e-02 6.04178198e-02 1.09701443e+00 3.44401658e-01
2.11229980e-01 -8.36304367e-01 7.17043206e-02 1.26578832e+00
1.36351609e+00 -1.52574158e+00 2.76675016e-01 -2.63312738e-02
-9.76670623e-01 9.78873730e-01 1.77979663e-01 -1.82515919e-01
1.73519745e-01 -4.00388569e-01 2.18641266e-01 -7.21527576e-01
-1.23847091e+00 3.22940588e-01 1.49977878e-01 -1.71919838e-01
7.57022321e-01 4.25065160e-01 -4.39032882e-01 4.37054306e-01
-8.52669835e-01 -2.91146278e-01 1.36913264e+00 3.15582246e-01
-1.03005993e+00 -7.51125664e-02 -5.94399512e-01 5.35610914e-01
-9.28071618e-01 -2.31509715e-01 -1.00579810e+00 3.50791514e-01
-4.61976618e-01 1.32473266e+00 -1.80646688e-01 -6.14712536e-01
-1.42985824e-02 3.09502780e-01 -3.46887745e-02 -6.25439286e-01
-1.30251992e+00 -5.88514209e-01 1.15352130e+00 -3.46518487e-01
-3.00600886e-01 -7.85781920e-01 -7.94663846e-01 -1.16847551e+00
-2.92189598e-01 5.74273467e-01 6.38739645e-01 1.32123995e+00
-2.08659098e-01 3.01524013e-01 9.60649371e-01 -6.25710785e-01
-1.32833338e+00 -1.64284229e+00 -4.52513754e-01 4.55822021e-01
1.12058902e+00 -3.96638066e-01 -7.51929104e-01 -4.56646502e-01] | [8.219131469726562, 10.11571979522705] |
b58be912-63c0-42ec-96d0-c0843a196fc0 | discovering-dynamic-causal-space-for-dag | 2306.02822 | null | https://arxiv.org/abs/2306.02822v1 | https://arxiv.org/pdf/2306.02822v1.pdf | Discovering Dynamic Causal Space for DAG Structure Learning | Discovering causal structure from purely observational data (i.e., causal discovery), aiming to identify causal relationships among variables, is a fundamental task in machine learning. The recent invention of differentiable score-based DAG learners is a crucial enabler, which reframes the combinatorial optimization problem into a differentiable optimization with a DAG constraint over directed graph space. Despite their great success, these cutting-edge DAG learners incorporate DAG-ness independent score functions to evaluate the directed graph candidates, lacking in considering graph structure. As a result, measuring the data fitness alone regardless of DAG-ness inevitably leads to discovering suboptimal DAGs and model vulnerabilities. Towards this end, we propose a dynamic causal space for DAG structure learning, coined CASPER, that integrates the graph structure into the score function as a new measure in the causal space to faithfully reflect the causal distance between estimated and ground truth DAG. CASPER revises the learning process as well as enhances the DAG structure learning via adaptive attention to DAG-ness. Grounded by empirical visualization, CASPER, as a space, satisfies a series of desired properties, such as structure awareness and noise robustness. Extensive experiments on both synthetic and real-world datasets clearly validate the superiority of our CASPER over the state-of-the-art causal discovery methods in terms of accuracy and robustness. | ['Tat-Seng Chua', 'Yueqi Duan', 'Xiang Wang', 'An Zhang', 'Wenchang Ma', 'Fangfu Liu'] | 2023-06-05 | null | null | null | null | ['causal-discovery', 'combinatorial-optimization'] | ['knowledge-base', 'methodology'] | [ 7.26402923e-02 2.32044056e-01 -4.54371572e-01 -2.68591523e-01
-2.94372648e-01 -7.67619967e-01 8.06868196e-01 4.54993755e-01
1.60105169e-01 8.68039370e-01 3.51587474e-01 -7.08220065e-01
-9.65962648e-01 -9.83598292e-01 -8.65568042e-01 -6.85838044e-01
-6.37309849e-01 2.49280557e-01 1.44941643e-01 6.77846074e-02
2.79145718e-01 3.33398223e-01 -1.15013945e+00 -2.33766958e-01
1.36482990e+00 6.23991966e-01 -1.22742258e-01 2.72735268e-01
-3.62656116e-02 8.87444973e-01 -3.40215623e-01 -6.58764601e-01
-1.42997339e-01 -4.49872643e-01 -6.51915789e-01 -6.15226805e-01
2.97976404e-01 1.82209834e-01 -3.81737113e-01 1.13774049e+00
2.05739006e-01 -2.85885543e-01 5.74830294e-01 -1.48833001e+00
-7.08498478e-01 9.21580613e-01 -6.37838244e-01 2.36760631e-01
3.50564808e-01 2.12109461e-01 1.50675166e+00 -6.77970290e-01
5.01683056e-01 1.58981264e+00 4.45208728e-01 6.77710697e-02
-1.45005751e+00 -8.93503189e-01 5.35309374e-01 3.58588278e-01
-1.08748543e+00 7.22305551e-02 1.11280155e+00 -6.09789014e-01
3.28385919e-01 4.69482332e-01 5.53096175e-01 1.30777800e+00
4.18350935e-01 5.29033065e-01 1.01823354e+00 -5.95352128e-02
4.05587196e-01 -5.69033504e-01 2.24899009e-01 9.94190514e-01
7.23468125e-01 7.45823920e-01 -7.75435507e-01 -4.22195762e-01
7.25760758e-01 -1.88417986e-01 -1.90468863e-01 -6.91083610e-01
-1.34279668e+00 8.33689153e-01 5.90020597e-01 -1.84170797e-01
-1.76843271e-01 1.70398846e-01 3.40146720e-01 1.52052119e-01
3.33030760e-01 7.52301574e-01 -4.75832701e-01 1.25898138e-01
-3.52692097e-01 3.29966277e-01 5.54267347e-01 4.85899746e-01
4.48480666e-01 1.42719492e-01 -2.92285323e-01 1.90825522e-01
4.11618710e-01 3.18088830e-01 -7.52623081e-02 -4.75322694e-01
3.95608604e-01 1.27152717e+00 -3.30868624e-02 -1.32690334e+00
-5.98263741e-01 -7.42366970e-01 -1.12931144e+00 2.42028072e-01
5.81979096e-01 -1.71268761e-01 -6.30775034e-01 2.08488727e+00
6.70465052e-01 4.90853786e-01 -3.89101982e-01 1.01491332e+00
6.98942900e-01 2.98723608e-01 2.63413340e-01 -4.87626612e-01
1.01971626e+00 -2.41871163e-01 -6.94943786e-01 8.67546815e-03
3.18253845e-01 -1.17026515e-01 1.31846142e+00 4.14124221e-01
-5.36148965e-01 -2.59760857e-01 -1.25167191e+00 3.34167957e-01
-2.28433952e-01 -3.27904075e-01 1.11362350e+00 5.49283087e-01
-5.04694343e-01 7.45204091e-01 -8.12596083e-01 1.26176074e-01
4.41193134e-01 1.93009511e-01 -2.21855015e-01 2.63288647e-01
-1.50504887e+00 4.61330682e-01 3.41027737e-01 5.64081520e-02
-1.20013261e+00 -1.02055478e+00 -5.99464655e-01 2.68540621e-01
9.95746315e-01 -8.66977274e-01 6.60713017e-01 -4.14126843e-01
-1.15494931e+00 2.65128016e-01 -2.08553448e-02 -2.40989000e-01
6.47134066e-01 -1.16232395e-01 -5.94354987e-01 -2.86855698e-01
-5.72202429e-02 -1.73108965e-01 7.49667406e-01 -1.29996347e+00
-5.17810524e-01 -4.91306245e-01 1.35730401e-01 -3.51556480e-01
-2.80344903e-01 -3.31913412e-01 1.77061796e-01 -7.41712213e-01
1.62979662e-01 -4.78364468e-01 -1.48917675e-01 -2.55130082e-01
-7.80688584e-01 -5.42292893e-01 6.12621844e-01 -2.85141677e-01
1.71275675e+00 -1.74105680e+00 3.72850984e-01 5.09000957e-01
9.45217788e-01 -2.10255697e-01 -8.45673084e-02 4.33000982e-01
-4.56163824e-01 5.22042155e-01 -3.81841183e-01 2.81059146e-01
2.14068918e-03 1.55875152e-02 -4.17784452e-01 5.71663976e-01
5.44882357e-01 9.68394279e-01 -1.54403555e+00 -4.21819836e-01
-5.38533479e-02 4.26825844e-02 -4.74258751e-01 4.13562208e-01
-4.12524849e-01 6.98619783e-01 -7.09753811e-01 5.64439058e-01
4.56153661e-01 -3.73322248e-01 6.11214578e-01 -2.18638126e-02
-3.62664640e-01 4.62608457e-01 -1.23256326e+00 1.38041902e+00
2.71003675e-02 1.54861897e-01 -2.65596390e-01 -1.19881403e+00
1.12371659e+00 1.01377703e-01 3.68752211e-01 -7.75034070e-01
-2.53032923e-01 2.29343817e-01 2.99808502e-01 -3.62842709e-01
-1.40636683e-01 1.07942916e-01 -1.62049234e-01 2.62649924e-01
-6.52123317e-02 4.44972187e-01 2.74370402e-01 3.42506826e-01
1.25588036e+00 2.40686983e-01 4.62599456e-01 -4.96825039e-01
3.58644605e-01 -5.15062548e-02 9.89931107e-01 6.67339563e-01
1.95465431e-01 3.49312238e-02 1.33355594e+00 -5.21745145e-01
-6.27937376e-01 -1.40882528e+00 1.26596373e-02 8.46103728e-01
2.53468543e-01 -4.94257927e-01 -3.52382541e-01 -1.11907375e+00
4.15247291e-01 5.99775672e-01 -1.00127780e+00 -5.56681752e-01
-6.75900757e-01 -9.73765612e-01 3.76270354e-01 3.24623108e-01
6.30211160e-02 -5.94131589e-01 -3.56051445e-01 1.71670690e-02
5.65473177e-02 -4.20252949e-01 -4.23231155e-01 2.30850860e-01
-8.61776769e-01 -1.68364131e+00 3.63980210e-03 -2.43885189e-01
4.53605980e-01 -1.25480801e-01 1.23118663e+00 -1.91851147e-02
-2.72385329e-01 -1.33025840e-01 -1.88684091e-01 -3.00611049e-01
-3.98539960e-01 -6.32858649e-02 4.75536622e-02 1.35029182e-01
-5.50584048e-02 -1.10100305e+00 -5.73310733e-01 2.14131191e-01
-5.90849936e-01 -2.72953119e-02 6.70391440e-01 1.05094743e+00
6.12424612e-01 1.31280124e-01 8.82641256e-01 -9.53808308e-01
7.24102318e-01 -8.11746061e-01 -1.11286843e+00 5.05096197e-01
-1.12657297e+00 6.07804477e-01 8.01463008e-01 -3.40434283e-01
-9.11829412e-01 -2.20073119e-01 4.79509950e-01 -2.12279364e-01
1.98110804e-01 1.04790628e+00 -6.40554965e-01 2.58798927e-01
6.95940077e-01 -1.49639443e-01 -2.23364398e-01 -5.08048892e-01
5.06688356e-01 -2.10404560e-01 6.37443423e-01 -7.66712606e-01
9.79716063e-01 2.79031426e-01 6.61479294e-01 -1.62355870e-01
-7.54934669e-01 -4.22586538e-02 -5.44188201e-01 -2.59192228e-01
5.11809468e-01 -4.13202614e-01 -1.34263957e+00 1.95931438e-02
-1.18125916e+00 -1.48355052e-01 -6.85345232e-02 3.45782548e-01
-1.82326183e-01 2.30569988e-01 4.14546169e-02 -9.75002944e-01
-6.01965468e-03 -8.39686275e-01 7.38138199e-01 6.52783141e-02
-1.08231589e-01 -1.13933074e+00 3.40066731e-01 -2.50493735e-01
-2.13885978e-01 9.22476470e-01 1.60043669e+00 -4.11560893e-01
-6.80276215e-01 9.37993601e-02 -4.15531605e-01 -4.15593147e-01
2.93715298e-01 2.62309670e-01 -7.42666781e-01 -3.71840373e-02
-3.97815257e-01 1.32203057e-01 9.39115226e-01 3.16071868e-01
1.36551476e+00 -6.02621615e-01 -5.26340544e-01 6.34618640e-01
1.14929414e+00 1.85865149e-01 3.23619604e-01 2.58913133e-02
9.34034944e-01 6.87846422e-01 5.19754469e-01 4.75909114e-01
4.55043554e-01 4.96561766e-01 1.04921556e+00 -2.56907493e-01
-8.94316807e-02 -9.94886398e-01 1.55446440e-01 4.94507849e-01
-1.76537275e-01 -2.67462522e-01 -1.02977359e+00 1.89457357e-01
-2.11064267e+00 -9.07749772e-01 -6.34668529e-01 2.48847270e+00
9.85595942e-01 2.77347505e-01 3.55696708e-01 1.27466366e-01
6.54006124e-01 1.75268486e-01 -8.68739247e-01 6.27796128e-02
-2.29273468e-01 -2.33333260e-02 4.31127757e-01 4.59325761e-01
-9.97934818e-01 5.77721000e-01 5.33492184e+00 6.16474450e-01
-1.06604874e+00 -1.20406538e-01 5.89399099e-01 1.47962928e-01
-8.52204740e-01 2.54978180e-01 -2.02851132e-01 5.24583280e-01
6.62285030e-01 -5.04393637e-01 4.21542645e-01 4.77873504e-01
5.63816726e-01 1.72219113e-01 -1.30083048e+00 6.61197007e-01
-6.94021761e-01 -1.63979447e+00 1.58735335e-01 1.86724752e-01
6.60445273e-01 -4.27106529e-01 1.27175106e-02 1.81160308e-02
8.98067653e-01 -1.22233522e+00 8.60786736e-01 7.95834005e-01
8.13984454e-01 -8.69332552e-01 4.62504387e-01 1.05163656e-01
-1.18047726e+00 -3.00080419e-01 -1.50186699e-02 -1.26339465e-01
1.40235517e-02 1.02977085e+00 -7.45339334e-01 1.04325020e+00
5.29744685e-01 7.93422878e-01 -6.05061114e-01 9.55713272e-01
-6.46198988e-01 1.05786586e+00 1.19406119e-01 -1.86597601e-01
-1.42273484e-02 -3.46047461e-01 8.03502440e-01 9.68616784e-01
-4.87364223e-03 2.72402121e-05 1.34500772e-01 1.34522307e+00
-1.01015590e-01 -1.25046760e-01 -4.73800868e-01 -1.29481733e-01
7.43814409e-01 1.01411641e+00 -5.10368049e-01 8.48903283e-02
8.08454957e-03 2.73605287e-01 4.97937948e-01 5.06614923e-01
-8.61943722e-01 -2.36026615e-01 8.16003144e-01 1.16635375e-01
-2.12960631e-01 -2.76928127e-01 -7.46245861e-01 -8.75134468e-01
8.33406597e-02 -9.21866000e-01 6.91297233e-01 -1.63024142e-01
-1.41858292e+00 1.71325207e-01 3.06715686e-02 -9.97650266e-01
-9.06004533e-02 -4.72005486e-01 -8.86347115e-01 6.88645422e-01
-1.46729243e+00 -9.54160869e-01 -3.41883034e-01 3.76217186e-01
1.18601257e-02 8.31757560e-02 6.09811604e-01 1.02454320e-01
-9.41661000e-01 5.77777386e-01 -1.17545463e-01 -9.33060050e-02
5.73772192e-01 -1.65641487e+00 4.48668599e-01 9.02076364e-01
2.02858657e-01 8.22596014e-01 8.23301733e-01 -1.01683104e+00
-1.75930667e+00 -1.02820158e+00 6.08640552e-01 -4.99866307e-01
1.27330077e+00 -5.93358994e-01 -1.05264592e+00 2.81559676e-01
-2.13754773e-01 -1.78477969e-02 3.11867982e-01 7.09768534e-01
-7.06259787e-01 -3.07738245e-01 -6.84844077e-01 7.33209193e-01
1.74176002e+00 -2.13193700e-01 -4.11931008e-01 1.15592510e-01
1.01825011e+00 -7.20300246e-03 -8.24140310e-01 6.72755539e-01
4.81813282e-01 -9.07479823e-01 1.00914872e+00 -1.12804747e+00
5.59695661e-01 -4.87020075e-01 2.73243934e-01 -1.36105919e+00
-5.38800597e-01 -9.98300493e-01 -4.71479982e-01 1.30448020e+00
4.08098519e-01 -8.25530648e-01 5.44722438e-01 1.57074198e-01
-4.60070781e-02 -7.44708896e-01 -1.04255629e+00 -8.95224750e-01
1.62217822e-02 -3.45065325e-01 9.77524579e-01 1.34586751e+00
1.14493266e-01 4.91952688e-01 -1.90080613e-01 5.13510883e-01
1.05735981e+00 3.87851030e-01 6.04235053e-01 -1.97990894e+00
-2.02525675e-01 -8.48725259e-01 -1.93359032e-01 -4.96681333e-01
2.35616282e-01 -1.07912886e+00 -3.85918915e-01 -1.37408400e+00
1.62426606e-01 -6.14676535e-01 -3.97612363e-01 4.56428349e-01
-7.59381056e-01 -5.46710551e-01 -1.15967311e-01 1.73094980e-02
-2.28595942e-01 7.71725416e-01 1.24787283e+00 -2.13069588e-01
-1.99012414e-01 -7.69759938e-02 -9.06516016e-01 6.56498671e-01
5.93341887e-01 -6.66500747e-01 -5.88002324e-01 -8.04630145e-02
7.07079589e-01 2.46737123e-01 8.41623306e-01 -3.94769162e-01
7.07987547e-02 -7.22164989e-01 8.91598836e-02 -3.35102767e-01
-4.29555565e-01 -4.57475632e-01 3.44757050e-01 5.28445840e-01
-4.79243040e-01 9.44151357e-02 -2.22481675e-02 9.53540802e-01
-1.14485160e-01 3.75114322e-01 3.31470400e-01 2.86957800e-01
-6.35588586e-01 4.07579333e-01 3.55332881e-01 1.61037311e-01
6.72034562e-01 7.39493892e-02 -6.90783560e-01 -6.49132729e-02
-3.84858578e-01 6.25656843e-01 -3.71115953e-02 4.77582544e-01
5.15600562e-01 -1.23754907e+00 -9.97654021e-01 7.47443829e-03
1.75146937e-01 -5.05490378e-02 1.03912592e-01 1.00260544e+00
2.48146251e-01 3.16744059e-01 1.77900851e-01 -5.23416162e-01
-9.48881686e-01 7.27090657e-01 1.38192251e-01 -5.91036558e-01
-3.92638922e-01 7.02950776e-01 5.90931892e-01 -3.56065929e-01
1.75833210e-01 -4.18914557e-01 -1.21188775e-01 1.41422957e-01
2.29850695e-01 6.34138703e-01 -6.20704852e-02 3.86663795e-01
-5.38734019e-01 1.69799298e-01 4.57104921e-01 1.79765046e-01
1.33889759e+00 1.35778874e-01 -2.94307888e-01 4.94257808e-01
6.64909780e-01 1.69910744e-01 -1.52901721e+00 5.90698719e-02
6.45271540e-01 -5.04040837e-01 -1.72993485e-02 -1.23348975e+00
-8.55905354e-01 7.23257482e-01 2.79032618e-01 4.60302711e-01
9.66499448e-01 1.47845969e-01 5.09322062e-02 5.45541495e-02
3.32851350e-01 -5.85029006e-01 2.57873803e-01 2.73198217e-01
1.12220919e+00 -1.11878335e+00 3.90221104e-02 -4.62436736e-01
-2.14648411e-01 1.12010813e+00 5.53576291e-01 -7.05978647e-02
4.97895658e-01 1.24634489e-01 -4.50722426e-01 -4.58481103e-01
-9.48561907e-01 -3.50259021e-02 5.19699275e-01 6.02621078e-01
4.45698082e-01 3.73282522e-01 -4.90520418e-01 8.46405029e-01
-2.92733788e-01 -3.12198251e-01 2.23535001e-01 2.16738716e-01
-9.92470384e-02 -9.89117503e-01 -2.69896865e-01 3.66546452e-01
-1.18956961e-01 -1.12810552e-01 -7.30403721e-01 9.48506296e-01
4.58072759e-02 9.27864969e-01 -1.57471478e-01 -4.36935097e-01
4.76542801e-01 -2.29952544e-01 1.85221165e-01 -2.68462986e-01
-3.61624718e-01 -2.24639460e-01 7.65773878e-02 -7.27610230e-01
-5.82970381e-02 -6.75248921e-01 -1.24220228e+00 -3.81561160e-01
-2.28450537e-01 1.90777406e-01 3.57857198e-01 8.12810183e-01
5.17542124e-01 9.60355163e-01 9.33563828e-01 -9.78610143e-02
-6.66600823e-01 -6.44931674e-01 -4.55999345e-01 9.35724154e-02
4.98468190e-01 -1.08666193e+00 -3.17207366e-01 -2.26256251e-01] | [7.776580333709717, 5.3696513175964355] |
c630b647-8a6e-49ca-a18a-03fb6eb39791 | icnn-input-conditioned-feature-representation | null | null | https://openreview.net/forum?id=SJecKyrKPH | https://openreview.net/pdf?id=SJecKyrKPH | ICNN: INPUT-CONDITIONED FEATURE REPRESENTATION LEARNING FOR TRANSFORMATION-INVARIANT NEURAL NETWORK | We propose a novel framework, ICNN, which combines the input-conditioned filter generation module and a decoder based network to incorporate contextual information present in images into Convolutional Neural Networks (CNNs). In contrast to traditional CNNs, we do not employ the same set of learned convolution filters for all input image instances. And our proposed decoder network serves the purpose of reducing the transformation present in the input image by learning to construct a representative image of the input image class. Our proposed joint supervision of input-aware framework when combined with techniques inspired by Multi-instance learning and max-pooling, results in a transformation-invariant neural network. We investigated the performance of our proposed framework on three MNIST variations, which covers both rotation and scaling variance, and achieved 0.98% error on MNIST-rot-12k, 1.12% error on Half-rotated MNIST and 0.68% error on Scaling MNIST, which is significantly better than the state-of-the-art results. Our proposed model also showcased consistent improvement on the CIFAR dataset. We make use of visualization to further prove the effectiveness of our input-aware convolution filters. Our proposed convolution filter generation framework can also serve as a plugin for any CNN based architecture and enhance its modeling capacity. | ['Abhay Kumar', 'Chirag Singh', 'Suraj Tripathi'] | 2019-09-25 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [ 4.37276751e-01 -6.40166178e-03 2.59126097e-01 -4.35275018e-01
-4.04597282e-01 -5.27836382e-01 6.45855010e-01 -3.96461248e-01
-9.17986810e-01 5.76088309e-01 3.11879236e-02 -1.56724870e-01
-2.00711221e-01 -8.42117071e-01 -1.17974079e+00 -7.00364113e-01
3.49958688e-01 -1.54862143e-02 2.16641963e-01 -1.64772183e-01
4.03114893e-02 7.78145373e-01 -1.49100375e+00 7.69672811e-01
6.36004686e-01 1.02548397e+00 4.22750086e-01 6.53621972e-01
2.00441569e-01 7.64092684e-01 -6.14011765e-01 -4.20855165e-01
4.81647223e-01 -8.15965310e-02 -5.86469293e-01 2.09447574e-02
8.51785839e-01 -2.63535202e-01 -6.70988441e-01 9.72601652e-01
5.88795781e-01 2.57739037e-01 4.80177611e-01 -8.27087104e-01
-9.15170074e-01 7.15006053e-01 -2.62486160e-01 2.87568957e-01
-2.49581814e-01 3.11083764e-01 6.37068212e-01 -9.14719999e-01
6.56025708e-01 1.14256299e+00 4.08955038e-01 4.74665821e-01
-1.16689324e+00 -7.35151529e-01 2.70291090e-01 2.13753000e-01
-1.16485333e+00 -2.89360017e-01 5.75776458e-01 -3.19072962e-01
1.19120932e+00 1.92334652e-01 4.89419788e-01 1.01461649e+00
2.38633856e-01 6.18643701e-01 1.21650648e+00 -4.29579288e-01
-1.94641933e-01 2.26064891e-01 1.09443739e-01 6.33142650e-01
1.84050754e-01 1.89191774e-01 -3.18652928e-01 4.52053159e-01
1.04426026e+00 2.71385521e-01 -2.60780573e-01 9.41371620e-02
-1.37883997e+00 6.34154677e-01 9.46188450e-01 2.29838833e-01
-2.56300420e-01 4.73114371e-01 1.97085679e-01 4.24516618e-01
2.27605194e-01 3.24712783e-01 -6.01941943e-01 2.54998922e-01
-9.66727376e-01 1.68207586e-01 4.84178185e-01 9.15122509e-01
8.08821321e-01 2.37929821e-01 -4.39678669e-01 8.17682445e-01
2.08551083e-02 4.71521586e-01 6.09572530e-01 -6.75746620e-01
6.04674459e-01 6.03760242e-01 -2.13505819e-01 -8.11719000e-01
-5.11141241e-01 -1.00204802e+00 -9.22310352e-01 4.22551513e-01
5.04074872e-01 -2.93628097e-01 -1.30259824e+00 1.75102592e+00
-1.83779240e-01 2.91538566e-01 2.58855522e-01 9.69658613e-01
9.52594697e-01 5.13243318e-01 -9.57899168e-02 1.95829540e-01
1.27564979e+00 -1.26694477e+00 -5.38591683e-01 -2.11282417e-01
2.17471868e-01 -1.06685960e+00 1.08542764e+00 5.20900786e-01
-1.01778817e+00 -9.69474435e-01 -1.22469008e+00 -1.34392366e-01
-6.19250000e-01 8.76596510e-01 5.35008430e-01 4.51007545e-01
-1.11353493e+00 7.99165905e-01 -8.05024028e-01 -2.51342595e-01
5.73097408e-01 6.11485660e-01 -5.48337638e-01 7.32328594e-02
-8.60225439e-01 9.05924499e-01 5.56016624e-01 3.16464812e-01
-1.03159153e+00 -7.21867204e-01 -4.51908439e-01 2.79150307e-01
2.38172770e-01 -5.92142105e-01 1.02125156e+00 -1.12579107e+00
-1.45546949e+00 4.97404546e-01 -1.88728012e-02 -7.04534233e-01
6.02136016e-01 -5.27987838e-01 -6.22954249e-01 2.34247074e-02
-4.08189595e-01 8.57594132e-01 9.51269925e-01 -9.48446870e-01
-5.67387044e-01 -3.82239632e-02 2.02975944e-01 1.40114380e-02
-4.38670993e-01 -7.50535503e-02 -5.03770590e-01 -8.33748996e-01
9.19917598e-02 -9.33852971e-01 -2.00392514e-01 -9.68479179e-03
-4.95829195e-01 2.37758696e-01 8.55868399e-01 -5.26659667e-01
9.81713831e-01 -2.15146112e+00 3.63079160e-02 1.47523388e-01
-3.09680942e-02 6.07188880e-01 -3.59661043e-01 2.19659448e-01
-4.15714025e-01 -1.35151977e-02 -1.04981758e-01 -3.80610764e-01
-1.47454515e-01 7.04238787e-02 -3.78057212e-01 3.81905019e-01
6.08135283e-01 7.71120429e-01 -4.77320910e-01 1.09779008e-01
3.92631173e-01 8.72270763e-01 -7.27233648e-01 4.16257739e-04
-5.26175313e-02 2.87986219e-01 9.53677446e-02 1.84031680e-01
9.64937210e-01 -1.19754255e-01 2.07534526e-02 -7.15994656e-01
-2.28923753e-01 8.33298042e-02 -1.38557220e+00 1.88737726e+00
-6.06707871e-01 7.84419179e-01 -3.36478472e-01 -9.11370754e-01
1.03086257e+00 2.38340497e-01 3.97639647e-02 -7.37900555e-01
1.71761557e-01 5.78276254e-02 3.66490811e-01 -3.29675525e-01
3.17018360e-01 3.45170677e-01 3.60012829e-01 2.72369206e-01
6.39776468e-01 2.72092938e-01 1.82858720e-01 2.08578147e-02
9.09950614e-01 2.73790807e-01 -7.14166164e-02 -3.79534692e-01
6.46304786e-01 -3.91786963e-01 2.85524189e-01 8.43298912e-01
1.22292727e-01 8.85418236e-01 3.44750166e-01 -6.39382780e-01
-1.10096502e+00 -9.79379475e-01 -2.71749049e-01 9.63948548e-01
-6.25089630e-02 -2.34028280e-01 -7.80889630e-01 -5.64936817e-01
-1.79217756e-01 5.60610414e-01 -7.48423040e-01 -9.20348838e-02
-7.46605515e-01 -9.84012365e-01 5.84889114e-01 6.26703262e-01
9.49055254e-01 -1.13839650e+00 -4.77639198e-01 1.73842251e-01
2.56380051e-01 -1.31420612e+00 -2.98640132e-01 5.76103926e-01
-7.96039641e-01 -9.09107029e-01 -5.89129269e-01 -8.77855062e-01
7.37954021e-01 7.07004741e-02 7.66870618e-01 -1.78658292e-01
-2.34394193e-01 -3.96370590e-02 -1.20555148e-01 -3.34025115e-01
-8.82774219e-02 2.74556309e-01 -1.98085591e-01 2.69120723e-01
4.42228862e-04 -7.36154974e-01 -8.72433484e-01 3.40080202e-01
-9.55222249e-01 2.86666423e-01 8.77341270e-01 8.59464705e-01
5.92407823e-01 -1.97245162e-02 5.02603531e-01 -1.08609343e+00
3.48821491e-01 -1.44794285e-01 -6.34904623e-01 2.88448781e-01
-3.32742780e-01 3.49685162e-01 9.38290417e-01 -6.48142636e-01
-1.14059663e+00 4.19999003e-01 -1.64506063e-01 -4.25160110e-01
-2.76991546e-01 2.45092168e-01 -7.65389204e-02 -2.19507530e-01
8.28053176e-01 1.48970440e-01 -2.48462915e-01 -6.35985613e-01
5.83437860e-01 4.33431923e-01 8.78211379e-01 -3.55118752e-01
8.18396866e-01 5.73183835e-01 -1.33734187e-02 -5.37300527e-01
-6.00073099e-01 -1.01471990e-01 -7.79926002e-01 -3.02286516e-03
8.98874581e-01 -1.09643650e+00 -7.15548217e-01 5.83007097e-01
-1.28965425e+00 -3.15665603e-01 -1.71943128e-01 6.31523490e-01
-2.41980717e-01 -1.84085667e-01 -5.10635376e-01 -4.36040282e-01
-4.06108975e-01 -1.59012926e+00 6.28642976e-01 3.64053041e-01
3.03709626e-01 -7.13358581e-01 -3.34852606e-01 7.57429078e-02
7.92840958e-01 2.42043868e-01 6.60516977e-01 -7.21221805e-01
-7.78737783e-01 -4.34118472e-02 -5.65384924e-01 8.09892297e-01
6.01120777e-02 1.25229850e-01 -1.40964544e+00 -3.97224963e-01
-2.42031217e-01 -2.44318858e-01 1.30252075e+00 2.76929229e-01
1.37000334e+00 -2.12242529e-01 5.32770641e-02 9.42057073e-01
1.73505211e+00 1.50727838e-01 7.93028951e-01 3.25129092e-01
6.94844842e-01 1.23179860e-01 -3.16437031e-03 2.91694224e-01
-5.07435240e-02 6.57423854e-01 5.75070381e-01 -5.20815194e-01
-5.85621893e-01 2.20662039e-02 1.87572360e-01 5.55718422e-01
-3.41434479e-01 -2.37372920e-01 -5.82947969e-01 3.46621484e-01
-1.76645362e+00 -8.34258735e-01 -1.78546570e-02 1.93891418e+00
5.74905992e-01 3.74545991e-01 -2.12122843e-01 1.11822270e-01
6.00613594e-01 9.25700832e-03 -3.92836869e-01 -5.00237107e-01
-3.31396073e-01 8.19164097e-01 8.67067993e-01 4.16712701e-01
-1.28356719e+00 9.80431795e-01 5.67231798e+00 8.27745080e-01
-1.49713957e+00 1.06389523e-01 4.94583845e-01 -2.25097582e-01
7.24438950e-02 -1.63994104e-01 -8.91331851e-01 1.58950165e-01
8.78707051e-01 2.43209720e-01 5.94920158e-01 7.35312045e-01
2.42320746e-01 1.40969560e-01 -1.06516540e+00 9.48287725e-01
6.59962520e-02 -1.53946841e+00 2.34120876e-01 -1.40833884e-01
9.29041147e-01 3.10560822e-01 3.87882888e-01 2.49378994e-01
2.18677700e-01 -1.21373951e+00 8.88632655e-01 6.90425098e-01
8.68159354e-01 -9.04374480e-01 7.05836594e-01 2.19160970e-02
-1.05517769e+00 -1.68489546e-01 -6.34992599e-01 -2.01373130e-01
-2.92088598e-01 4.49555397e-01 -8.80058169e-01 5.34899592e-01
7.44100213e-01 6.41744614e-01 -7.73879409e-01 1.12785387e+00
-3.24058861e-01 4.99621570e-01 -3.45413536e-01 1.69864044e-01
4.72848296e-01 9.46523696e-02 1.05876960e-01 1.48959446e+00
2.52119511e-01 -2.03033522e-01 -4.85870391e-02 8.18325937e-01
-3.96921307e-01 2.84783524e-02 -2.83555686e-01 1.53379321e-01
2.24287093e-01 1.51270711e+00 -8.30020785e-01 -4.18618292e-01
-5.37682414e-01 1.07463241e+00 4.85809952e-01 6.66810393e-01
-9.78061497e-01 -6.59444630e-01 6.71428442e-01 -1.25043795e-01
8.49968135e-01 -1.85479045e-01 -3.15598816e-01 -1.23441315e+00
7.50861689e-02 -7.96229184e-01 3.20755169e-02 -6.54493690e-01
-7.61724710e-01 1.04106283e+00 -1.94039285e-01 -1.13773167e+00
5.20552695e-02 -9.80360031e-01 -6.03244841e-01 1.07119727e+00
-1.64339626e+00 -1.31520319e+00 -5.12671769e-01 7.64806449e-01
3.56808096e-01 -3.55713606e-01 8.27667058e-01 6.30669475e-01
-6.78930521e-01 8.60336423e-01 1.74687847e-01 3.10601741e-01
7.63575494e-01 -1.08012187e+00 4.65313256e-01 1.16344059e+00
4.30022031e-01 8.04893136e-01 3.99666160e-01 -2.30124727e-01
-1.18761015e+00 -1.48849106e+00 4.81138319e-01 -7.56055191e-02
3.87329519e-01 -5.22347093e-01 -7.23375797e-01 7.93779731e-01
4.27868247e-01 4.18301851e-01 2.00970247e-01 -2.33915061e-01
-5.20272732e-01 -4.66007560e-01 -1.01651049e+00 5.77494562e-01
1.06285846e+00 -4.68990147e-01 -3.50660235e-01 3.40795845e-01
8.59650612e-01 -6.04869246e-01 -6.69693828e-01 4.42101628e-01
4.95523304e-01 -1.04665840e+00 1.01427805e+00 -5.17316818e-01
5.32067180e-01 -4.67270553e-01 -1.77993923e-01 -1.31736851e+00
-5.70613682e-01 -3.54841202e-01 2.51532048e-01 9.68916059e-01
6.72817171e-01 -5.91229618e-01 5.71263850e-01 1.83046043e-01
-4.38803732e-01 -7.38482714e-01 -6.46688402e-01 -5.98488212e-01
1.54474182e-02 -7.09289074e-01 6.23383760e-01 5.09939313e-01
-7.65569866e-01 5.52938916e-02 -4.02420759e-01 3.98123354e-01
3.55243504e-01 -1.02879316e-01 7.15345085e-01 -7.26445615e-01
-4.82607633e-01 -3.79665613e-01 -5.05093873e-01 -9.26928580e-01
-9.10192654e-02 -1.03447878e+00 -3.53279710e-01 -1.35267258e+00
2.96868253e-02 -1.22415304e-01 -6.76446617e-01 7.71664679e-01
8.20776671e-02 6.87212527e-01 4.66619998e-01 -5.56737743e-02
-2.45631173e-01 2.05284774e-01 1.46313083e+00 -1.08676687e-01
-6.93034902e-02 -1.39740288e-01 -6.83527112e-01 6.37943268e-01
7.36194015e-01 -3.80666465e-01 -5.53158700e-01 -8.28083813e-01
-1.58110615e-02 -4.37244296e-01 5.83607852e-01 -1.49940467e+00
4.08085525e-01 1.21549152e-01 8.20858300e-01 -5.66963911e-01
1.95270926e-01 -8.84885728e-01 2.94889122e-01 5.26765168e-01
-4.84511524e-01 2.22512661e-03 5.42278349e-01 2.22148463e-01
-2.69248247e-01 -3.54605690e-02 8.69523883e-01 -6.35149106e-02
-7.29200125e-01 2.29215577e-01 -8.42977390e-02 -3.09170663e-01
7.07889855e-01 -7.96729475e-02 -5.44546247e-01 -1.14189973e-02
-9.52263832e-01 -2.67084271e-01 -2.69750748e-02 2.76599646e-01
6.45767808e-01 -1.21978951e+00 -7.58607090e-01 5.91815412e-01
-6.14834353e-02 -7.29887709e-02 2.77227879e-01 5.64897954e-01
-6.43092096e-01 4.05217797e-01 -5.41408241e-01 -4.54814970e-01
-9.29628432e-01 3.82466137e-01 4.27573681e-01 -1.55516475e-01
-6.34691477e-01 9.68293607e-01 2.41346791e-01 -4.05595869e-01
2.86558241e-01 -8.16985965e-01 -1.14251435e-01 -1.42222136e-01
5.86536705e-01 1.97623074e-01 5.36016226e-01 -3.41638625e-01
-2.76556551e-01 4.61751163e-01 -3.24450940e-01 -6.04644045e-02
1.51399446e+00 3.35175484e-01 1.08593650e-01 2.43846811e-02
1.33712506e+00 -2.71117538e-01 -1.42967606e+00 -4.05774474e-01
-3.30795169e-01 -2.14246273e-01 1.18747592e-01 -1.15243030e+00
-1.54547739e+00 9.89158034e-01 9.54366088e-01 -3.04277629e-01
1.36957920e+00 -5.02672315e-01 4.59738910e-01 6.81303144e-01
-1.37613356e-01 -9.32335019e-01 9.00427997e-02 6.19248271e-01
1.09812129e+00 -1.16228604e+00 -1.28313974e-01 -2.30183706e-01
-3.62547576e-01 1.50253808e+00 8.24654281e-01 -6.98072791e-01
7.04160988e-01 4.22346979e-01 9.78941247e-02 3.89832072e-02
-7.72855759e-01 -1.95772380e-01 5.59566438e-01 3.99347097e-01
5.47733486e-01 -1.16747245e-01 -1.55595779e-01 5.93537271e-01
-3.70141596e-01 1.07952558e-01 4.16863441e-01 6.89989626e-01
-3.30588937e-01 -9.59739566e-01 -2.68444240e-01 3.95999908e-01
-4.48131174e-01 -3.43945831e-01 4.83856499e-02 9.32061434e-01
5.31376302e-01 6.27275467e-01 1.21775292e-01 -5.47278821e-01
5.20132184e-01 7.86838401e-03 5.92949390e-01 -4.88683790e-01
-9.67735112e-01 1.38376027e-01 -1.97392091e-01 -5.82253933e-01
-5.20186305e-01 -2.46049643e-01 -1.06264651e+00 -1.33532181e-01
-2.57811993e-01 -3.87609929e-01 7.62254596e-01 8.73023152e-01
2.24702343e-01 1.07384884e+00 4.10872519e-01 -8.12512338e-01
-3.60699236e-01 -1.09677517e+00 -2.95011073e-01 5.28700531e-01
3.40753078e-01 -5.01340210e-01 -2.11880729e-01 2.79043078e-01] | [9.119053840637207, 2.1820895671844482] |
cb68bc2a-b1a6-4752-a0f8-87e514019928 | main-multi-attention-instance-network-for | 1904.05847 | null | http://arxiv.org/abs/1904.05847v1 | http://arxiv.org/pdf/1904.05847v1.pdf | MAIN: Multi-Attention Instance Network for Video Segmentation | Instance-level video segmentation requires a solid integration of spatial and
temporal information. However, current methods rely mostly on domain-specific
information (online learning) to produce accurate instance-level segmentations.
We propose a novel approach that relies exclusively on the integration of
generic spatio-temporal attention cues. Our strategy, named Multi-Attention
Instance Network (MAIN), overcomes challenging segmentation scenarios over
arbitrary videos without modelling sequence- or instance-specific knowledge. We
design MAIN to segment multiple instances in a single forward pass, and
optimize it with a novel loss function that favors class agnostic predictions
and assigns instance-specific penalties. We achieve state-of-the-art
performance on the challenging Youtube-VOS dataset and benchmark, improving the
unseen Jaccard and F-Metric by 6.8% and 12.7% respectively, while operating at
real-time (30.3 FPS). | ['Bernard Ghanem', 'Maria A. Bravo', 'Thomas Brox', 'Pablo Arbelaez', 'Juan Leon Alcazar', 'Guillaume Jeanneret', 'Ali K. Thabet'] | 2019-04-11 | null | null | null | null | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 4.34158325e-01 -1.78675145e-01 -3.85754138e-01 -4.55445647e-01
-1.09980810e+00 -6.27248168e-01 3.73465568e-01 1.14943244e-01
-6.67729437e-01 7.06232071e-01 -3.83441806e-01 -1.36195883e-01
-1.12684950e-01 -3.72988433e-01 -8.83102000e-01 -4.95677412e-01
-1.55286521e-01 4.79670763e-01 6.90280974e-01 1.13609672e-01
3.46856296e-01 2.89454430e-01 -1.65648782e+00 6.38096452e-01
1.08992672e+00 1.21549511e+00 2.97878951e-01 1.05547333e+00
-1.76233381e-01 6.61813855e-01 -6.26614988e-01 -4.10865188e-01
4.24719423e-01 -2.90394843e-01 -8.83724570e-01 3.24884772e-01
6.32604301e-01 -7.95140937e-02 -4.59617674e-02 7.00086653e-01
2.57526368e-01 2.32663542e-01 4.16040987e-01 -1.12555993e+00
-1.92288235e-01 1.40183643e-01 -7.03149378e-01 4.89461929e-01
1.05467029e-01 4.78184193e-01 1.06113160e+00 -5.41477144e-01
6.81539237e-01 7.12681055e-01 7.63525724e-01 5.10189295e-01
-1.21551013e+00 -3.23426813e-01 4.95374560e-01 4.06634152e-01
-1.23636699e+00 -3.22546124e-01 6.06809855e-01 -5.78470349e-01
9.67910111e-01 3.23405504e-01 6.74888551e-01 1.02260852e+00
-1.11858040e-01 9.94666398e-01 8.90591383e-01 -3.59302983e-02
1.87522724e-01 -9.27336216e-02 -3.98623981e-02 5.43870509e-01
-1.73395172e-01 -2.00685188e-01 -4.54224974e-01 3.18741113e-01
6.94634497e-01 -2.44284654e-03 -3.46596390e-01 -4.47978944e-01
-1.15062213e+00 4.86613721e-01 3.51544589e-01 1.71557114e-01
-4.50517684e-01 1.77100301e-01 5.99144459e-01 -3.80993262e-02
6.81414187e-01 3.70133132e-01 -8.46345186e-01 -5.58031023e-01
-1.50989509e+00 4.15123440e-02 6.72979414e-01 9.27471459e-01
6.62901402e-01 -1.33282796e-01 -4.89081532e-01 7.51952946e-01
-3.36367614e-03 2.41813615e-01 3.89593810e-01 -1.21627271e+00
5.44326365e-01 4.50168312e-01 2.31989563e-01 -7.55904734e-01
-2.42937490e-01 -7.12166429e-01 -2.75416195e-01 2.69347359e-03
5.44559002e-01 -1.77433640e-01 -1.33294952e+00 1.66961312e+00
4.42591876e-01 7.43017733e-01 -3.25587064e-01 1.05307996e+00
6.47552311e-01 6.39089763e-01 2.89627045e-01 -1.72633469e-01
1.00548029e+00 -1.40009868e+00 -3.67661744e-01 -1.71108067e-01
5.54453373e-01 -4.67880577e-01 1.24907839e+00 3.78605843e-01
-1.15049005e+00 -6.71851337e-01 -9.04131114e-01 6.27133325e-02
-3.60979944e-01 9.71127078e-02 3.04455101e-01 5.63951850e-01
-9.39151525e-01 9.05787110e-01 -9.58964229e-01 -2.13459268e-01
7.30395734e-01 6.41455233e-01 -2.21643969e-01 9.90240276e-02
-7.08173335e-01 3.40801984e-01 3.73501241e-01 3.64319561e-03
-9.39360738e-01 -1.17651653e+00 -6.33184254e-01 3.29957195e-02
7.86385655e-01 -6.41667724e-01 1.17326033e+00 -1.33765817e+00
-1.60374439e+00 7.70561278e-01 -1.38026223e-01 -6.28018439e-01
8.65429878e-01 -5.03951132e-01 -1.55399442e-01 4.31553304e-01
9.60694179e-02 9.92605865e-01 9.35073376e-01 -1.19649124e+00
-8.12083006e-01 -1.81292146e-01 3.37111801e-01 1.54214770e-01
-1.56213775e-01 -1.02039047e-01 -8.96642268e-01 -5.26380181e-01
-2.83773184e-01 -7.94327140e-01 -2.61405498e-01 -2.19013374e-02
-2.16558009e-01 2.32222900e-02 9.68024552e-01 -7.77775705e-01
1.16221130e+00 -2.10036254e+00 2.46117502e-01 -1.52964190e-01
8.55433568e-02 7.94218481e-01 -3.05794507e-01 -3.49645177e-03
2.39867121e-01 1.92325503e-01 -5.67172348e-01 -5.65329432e-01
-1.51956007e-01 1.84040487e-01 2.63805926e-01 2.94974238e-01
4.15584475e-01 9.48632479e-01 -1.12938631e+00 -6.54403985e-01
4.22848374e-01 6.54950738e-01 -8.82213533e-01 2.49965504e-01
-5.97089469e-01 5.45004427e-01 -1.70127258e-01 6.23250663e-01
4.46218431e-01 -3.89346361e-01 6.51104078e-02 -2.45926157e-01
-1.44894302e-01 9.43229496e-02 -9.79312181e-01 2.07851529e+00
-4.66474086e-01 6.93262517e-01 -5.51256500e-02 -1.20850646e+00
4.77456808e-01 2.26900980e-01 7.63145864e-01 -6.26121163e-01
-1.24510296e-01 1.84442729e-01 -2.26149380e-01 -6.91168010e-01
3.80517989e-01 2.31898397e-01 1.41467541e-01 -8.88894871e-02
2.50854045e-01 4.13690925e-01 4.71324801e-01 8.39923620e-02
1.09646451e+00 7.63439119e-01 -5.07936254e-02 -2.28076205e-01
5.54898620e-01 5.09895645e-02 8.24885964e-01 7.90732563e-01
-5.15498221e-01 9.88958836e-01 5.32360256e-01 -4.79338557e-01
-9.77971256e-01 -8.98539305e-01 5.91631196e-02 1.15782714e+00
2.29926229e-01 -4.66913819e-01 -9.61139023e-01 -1.15805018e+00
-1.17660932e-01 4.92754996e-01 -6.60867691e-01 2.46836424e-01
-6.26559794e-01 -6.37771964e-01 2.76777923e-01 5.98732173e-01
4.18515474e-01 -1.01475608e+00 -8.06643128e-01 4.74261791e-01
-2.34562755e-01 -1.46738946e+00 -5.68353593e-01 2.22386792e-02
-8.69699419e-01 -1.17807114e+00 -9.38161850e-01 -3.02457482e-01
3.98332953e-01 6.51364401e-02 1.34659624e+00 2.15243245e-03
-5.27996719e-01 3.46590191e-01 -4.00794804e-01 9.17997360e-02
1.24470703e-01 4.50575650e-01 -4.57951337e-01 3.77610981e-01
2.01117754e-01 -5.04799962e-01 -8.27242732e-01 2.65541315e-01
-7.85278797e-01 6.58778474e-02 5.57008684e-01 8.22994351e-01
8.30348730e-01 -4.01266992e-01 5.92730999e-01 -1.03351605e+00
9.87371728e-02 -4.47259247e-01 -6.47011459e-01 2.97376126e-01
-3.72942746e-01 -1.68723956e-01 7.17404068e-01 -4.17397380e-01
-8.68882477e-01 2.18719780e-01 -2.70150043e-02 -7.65233278e-01
-4.75437373e-01 2.34876439e-01 -1.11665756e-01 -2.10571453e-01
4.06414986e-01 2.27617785e-01 -2.14137316e-01 -5.19900560e-01
2.30946720e-01 4.48655307e-01 6.08692646e-01 -5.29941499e-01
3.21961492e-01 4.48481888e-01 -2.06852153e-01 -7.62983024e-01
-1.03183937e+00 -8.39118779e-01 -8.92172515e-01 -4.08842027e-01
1.13928854e+00 -8.21984589e-01 -5.54438412e-01 3.37757587e-01
-1.08467317e+00 -7.92264640e-01 -3.00834537e-01 2.37865195e-01
-7.58697927e-01 3.33637953e-01 -5.84387898e-01 -6.52002752e-01
-2.08230197e-01 -1.25110674e+00 1.23526835e+00 2.27542460e-01
2.37829667e-02 -8.90902102e-01 -1.28888115e-01 6.07627094e-01
2.71786690e-01 4.53074604e-01 2.88451016e-01 -6.30822122e-01
-1.06238580e+00 1.20661780e-02 -5.12820780e-01 3.48573685e-01
-1.14876948e-01 2.64338464e-01 -9.02441680e-01 -1.15370989e-01
-4.67605829e-01 -2.14133728e-02 8.78074586e-01 4.69470352e-01
1.54347348e+00 -2.90293187e-01 -1.95122376e-01 8.55944633e-01
1.69331932e+00 1.25771761e-01 5.38897812e-01 4.10823286e-01
8.06574821e-01 4.84233171e-01 7.82945812e-01 5.02719104e-01
4.16074187e-01 9.14410949e-01 5.17098129e-01 1.42568490e-02
-2.36303896e-01 5.17719984e-02 5.98678440e-02 3.63455057e-01
-3.26132774e-01 -3.58069032e-01 -8.04833233e-01 9.50382590e-01
-2.10622454e+00 -1.06341600e+00 -1.42796114e-01 2.23612857e+00
7.34727800e-01 2.53015012e-01 4.88824159e-01 7.66828880e-02
7.35523760e-01 2.31832460e-01 -6.12907887e-01 -3.64943564e-01
1.75793931e-01 1.40922382e-01 5.92654407e-01 4.52628762e-01
-1.44926035e+00 1.05628717e+00 5.88215256e+00 7.96991289e-01
-1.19049060e+00 1.54753104e-01 8.73935163e-01 -5.02763987e-01
-1.39415915e-05 -2.28075966e-01 -6.81312561e-01 8.26819599e-01
1.16547859e+00 2.80856907e-01 3.55020702e-01 7.27656126e-01
2.21037105e-01 5.17659485e-02 -9.88674939e-01 9.46118593e-01
-8.34442582e-03 -1.44626999e+00 -1.35237411e-01 -9.14226174e-02
8.12107444e-01 1.49034128e-01 -2.02626979e-04 3.21483999e-01
-1.53593063e-01 -8.48962307e-01 6.27172053e-01 5.06780744e-01
6.40077889e-01 -7.17866361e-01 6.61332071e-01 1.57339334e-01
-1.15969300e+00 -1.13968439e-02 4.17423807e-02 1.70185432e-01
4.43963528e-01 6.05223119e-01 -5.76679885e-01 5.75441957e-01
9.22864795e-01 9.03309405e-01 -4.04852778e-01 1.48159492e+00
3.76398414e-02 6.77831531e-01 -4.56157446e-01 2.40457729e-01
6.85467541e-01 -3.46002094e-02 5.52406192e-01 1.58273721e+00
3.70001830e-02 7.12785944e-02 3.47829193e-01 5.71165204e-01
-6.28804341e-02 1.36252716e-01 -2.07269564e-02 1.47679567e-01
1.13240220e-01 1.28105021e+00 -8.60092521e-01 -4.18630570e-01
-4.54384029e-01 1.26257932e+00 2.83684224e-01 3.94859403e-01
-1.25229716e+00 -2.80088961e-01 7.77330399e-01 1.83438182e-01
9.35204327e-01 -1.33323982e-01 -3.66036773e-01 -1.04060543e+00
2.50702560e-01 -6.69908464e-01 3.93324643e-01 -5.02684414e-01
-1.04260826e+00 5.68858147e-01 -6.51883259e-02 -1.08665001e+00
-1.39881238e-01 -5.24788678e-01 -4.78040457e-01 5.15968144e-01
-1.91772592e+00 -9.91467714e-01 -3.97771716e-01 4.49222594e-01
8.78143132e-01 2.00921863e-01 3.73938918e-01 6.45969152e-01
-8.16984594e-01 7.46089697e-01 -4.40375134e-02 5.40438443e-02
5.73585153e-01 -1.45264339e+00 3.14272821e-01 7.38203526e-01
2.39153534e-01 5.19357882e-02 6.63908422e-01 -4.00010496e-01
-9.56066370e-01 -1.33744264e+00 8.26631427e-01 -3.97495896e-01
4.75820422e-01 -2.95643151e-01 -1.01695490e+00 5.12939930e-01
9.03807208e-02 5.20703852e-01 6.38177276e-01 1.03815250e-01
-2.91668355e-01 -2.14543849e-01 -1.24544394e+00 5.65337896e-01
1.34705234e+00 -2.99407572e-01 -2.37069353e-01 3.96076232e-01
7.44574130e-01 -3.97701502e-01 -9.34095085e-01 6.04662716e-01
4.97062862e-01 -1.24822128e+00 1.01985085e+00 -8.17383051e-01
3.83909971e-01 -3.86978656e-01 4.70380373e-02 -9.08601940e-01
-2.06658140e-01 -7.73175895e-01 -3.45209926e-01 9.42004502e-01
5.36088407e-01 -2.55196512e-01 1.13475990e+00 6.15721166e-01
-3.54791611e-01 -1.22951555e+00 -8.53190720e-01 -8.85450423e-01
-2.99488395e-01 -5.24989307e-01 3.63773763e-01 6.77146614e-01
-2.77065873e-01 -2.05095112e-02 -3.81054312e-01 1.86653480e-01
6.17154300e-01 1.74725696e-01 5.98290682e-01 -1.00313687e+00
-4.41365004e-01 -5.07714212e-01 -6.48059487e-01 -1.21246970e+00
1.61796406e-01 -5.21858573e-01 1.57843098e-01 -1.45487034e+00
5.19662835e-02 -4.41857815e-01 -6.16577089e-01 3.93374264e-01
-4.23495114e-01 5.96447766e-01 4.00794029e-01 9.35911462e-02
-1.33510303e+00 3.91435951e-01 1.05626619e+00 7.24159107e-02
-2.23898187e-01 -1.70706399e-02 -2.25737646e-01 5.82388163e-01
8.49975944e-01 -4.54071641e-01 -4.00290102e-01 -5.44753194e-01
-3.24816853e-01 5.84381633e-04 3.50943774e-01 -1.31202877e+00
1.15903027e-01 -2.54951924e-01 2.31595904e-01 -5.42327762e-01
4.84373897e-01 -7.37050653e-01 -1.73685178e-02 1.53859124e-01
-3.89107674e-01 -1.59950063e-01 3.56701940e-01 8.05935383e-01
-1.77014574e-01 -1.42434448e-01 7.84940779e-01 -2.33611509e-01
-1.08107233e+00 6.19374990e-01 8.38492066e-02 4.20827478e-01
1.34577823e+00 -4.80218321e-01 -3.44077721e-02 -6.15983047e-02
-8.94023538e-01 2.57221818e-01 5.48835993e-01 3.88002872e-01
3.06480736e-01 -8.37821722e-01 -6.19605541e-01 -9.71448049e-03
-3.91091360e-03 -2.55406871e-02 5.07124841e-01 9.35168266e-01
-6.39161766e-01 4.57601875e-01 -6.65517300e-02 -1.02568424e+00
-1.14721382e+00 5.18356979e-01 4.21748906e-01 -3.78206611e-01
-5.53102374e-01 1.03038967e+00 5.30311465e-02 -1.77595913e-01
1.59451038e-01 1.02515295e-02 -1.09587006e-01 -6.53921664e-02
2.95002073e-01 3.54065686e-01 9.27876160e-02 -6.17285848e-01
-4.27685291e-01 6.42884672e-01 -7.01587126e-02 -4.63293605e-02
1.26153576e+00 -4.04641666e-02 4.52920109e-01 3.94529492e-01
1.31373298e+00 -3.47099423e-01 -2.12323499e+00 1.09111458e-01
1.30597770e-01 -7.42636025e-01 -3.50449756e-02 -9.44213033e-01
-1.36332548e+00 9.60262775e-01 4.08407807e-01 1.15444809e-01
1.12966847e+00 -1.47960976e-01 1.03507090e+00 -1.24843702e-01
1.61844552e-01 -1.07125366e+00 1.13200396e-01 2.77949125e-01
4.13898915e-01 -1.39450681e+00 -2.53356636e-01 -5.02269566e-01
-6.00681603e-01 8.38954806e-01 7.85044551e-01 -1.79959223e-01
3.83492857e-01 3.88784446e-02 3.52575406e-02 1.53980047e-01
-8.06951582e-01 -3.00551981e-01 5.94429076e-01 5.81159472e-01
4.16257322e-01 -1.16045900e-01 -2.64472932e-01 4.19493288e-01
4.34765249e-01 1.72367930e-01 1.35714591e-01 7.63535798e-01
-3.15879226e-01 -1.12564719e+00 1.06744476e-01 4.67542112e-01
-8.22360218e-01 -8.78014490e-02 1.06313996e-01 6.36168659e-01
2.12351114e-01 7.35603333e-01 2.19184309e-01 -2.65470266e-01
1.76022455e-01 7.15094209e-02 3.87618333e-01 -5.01541793e-01
-7.30298042e-01 2.05989718e-03 8.52102786e-02 -1.01168859e+00
-6.89678967e-01 -7.67169833e-01 -1.03955889e+00 -1.29758418e-01
-3.21026854e-02 3.26228179e-02 7.29182124e-01 9.69440222e-01
8.43641818e-01 7.09502995e-01 4.10824448e-01 -1.09812641e+00
-1.24174036e-01 -5.92223644e-01 -7.13269338e-02 6.65231526e-01
3.79177302e-01 -4.99290645e-01 -2.90700018e-01 2.15726867e-01] | [9.116019248962402, -0.03024616837501526] |
65bd9b46-7022-4bff-80c3-2b6384b70b48 | instance-smoothed-contrastive-learning-for | 2305.07424 | null | https://arxiv.org/abs/2305.07424v2 | https://arxiv.org/pdf/2305.07424v2.pdf | Instance Smoothed Contrastive Learning for Unsupervised Sentence Embedding | Contrastive learning-based methods, such as unsup-SimCSE, have achieved state-of-the-art (SOTA) performances in learning unsupervised sentence embeddings. However, in previous studies, each embedding used for contrastive learning only derived from one sentence instance, and we call these embeddings instance-level embeddings. In other words, each embedding is regarded as a unique class of its own, whichmay hurt the generalization performance. In this study, we propose IS-CSE (instance smoothing contrastive sentence embedding) to smooth the boundaries of embeddings in the feature space. Specifically, we retrieve embeddings from a dynamic memory buffer according to the semantic similarity to get a positive embedding group. Then embeddings in the group are aggregated by a self-attention operation to produce a smoothed instance embedding for further analysis. We evaluate our method on standard semantic text similarity (STS) tasks and achieve an average of 78.30%, 79.47%, 77.73%, and 79.42% Spearman's correlation on the base of BERT-base, BERT-large, RoBERTa-base, and RoBERTa-large respectively, a 2.05%, 1.06%, 1.16% and 0.52% improvement compared to unsup-SimCSE. | ['Yue Zhang', 'Zhenzhong Lan', 'Junlei Zhang', 'Hongliang He'] | 2023-05-12 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings', 'semantic-textual-similarity', 'semantic-similarity'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.06146231e-01 3.14100645e-02 1.01145342e-01 -4.29185569e-01
-6.73298717e-01 -3.19547802e-01 7.77069628e-01 7.37107754e-01
-8.33062530e-01 5.13858616e-01 3.89336854e-01 -5.49764521e-02
-1.32468045e-01 -6.41945422e-01 -5.58817506e-01 -5.59057713e-01
-7.77602717e-02 2.17529297e-01 4.54471767e-01 -4.08521354e-01
4.54561472e-01 6.95240349e-02 -1.49934638e+00 5.23085058e-01
1.11430693e+00 8.42255235e-01 1.90354526e-01 6.15789533e-01
-5.34895122e-01 4.01478648e-01 -6.45396769e-01 -3.84873509e-01
-1.24164641e-01 -4.22567070e-01 -6.37974441e-01 -4.13313895e-01
3.68088514e-01 1.04926407e-01 -3.70392531e-01 1.06895840e+00
5.10893703e-01 4.20224071e-01 9.66281652e-01 -1.16300273e+00
-1.08998120e+00 7.06712067e-01 -4.88924444e-01 4.95673954e-01
2.33794093e-01 -2.38171101e-01 1.15293050e+00 -1.30506778e+00
3.96991402e-01 1.23169196e+00 6.31586432e-01 5.72678506e-01
-1.00945783e+00 -6.30524278e-01 2.03199416e-01 3.74568909e-01
-1.16179836e+00 -4.09383237e-01 6.99057281e-01 -3.12990040e-01
1.30045557e+00 4.17904347e-01 3.95890653e-01 7.42573500e-01
2.24910855e-01 7.91835546e-01 1.01899803e+00 -5.38235962e-01
2.61710316e-01 3.49961996e-01 5.15912056e-01 4.22594726e-01
1.48288697e-01 -2.84738958e-01 -5.07993460e-01 6.90785870e-02
1.88021511e-01 2.22070560e-01 6.00564219e-02 3.43185253e-02
-1.02517962e+00 8.49883437e-01 8.20727289e-01 5.13257742e-01
-2.83962697e-01 -8.98694396e-02 7.34869897e-01 3.97616625e-01
7.09134161e-01 4.61266309e-01 -4.27470267e-01 -2.91726887e-01
-7.08365381e-01 1.89401343e-01 4.83301848e-01 8.31875622e-01
7.51622438e-01 -1.04363352e-01 -2.28385285e-01 1.12108219e+00
2.61920154e-01 4.55121577e-01 1.00345087e+00 -3.48098636e-01
4.68517393e-01 6.29661202e-01 -1.76225901e-01 -1.23321259e+00
-3.03901613e-01 -4.75170851e-01 -8.18412721e-01 -3.24124753e-01
-1.66959375e-01 1.76860198e-01 -7.07960665e-01 1.68574774e+00
1.03214331e-01 4.45105016e-01 2.59490788e-01 7.00463593e-01
9.46563125e-01 8.62779498e-01 2.00594202e-01 -1.44788653e-01
1.37913060e+00 -1.20788932e+00 -7.56558120e-01 -3.67846429e-01
9.43027198e-01 -6.84391856e-01 1.46902275e+00 1.04700774e-01
-9.31657374e-01 -7.10635364e-01 -1.14124894e+00 -2.84655727e-02
-6.74704254e-01 -1.30097508e-01 2.05310583e-01 3.48560184e-01
-9.28770900e-01 7.39113331e-01 -5.73862970e-01 -4.63573784e-01
2.27846995e-01 1.04614668e-01 -3.99968415e-01 -2.51787845e-02
-1.51541007e+00 9.69514191e-01 6.71566844e-01 -2.80100197e-01
-2.85182625e-01 -8.44854593e-01 -8.86703849e-01 2.23302796e-01
1.05954241e-02 -3.75490785e-01 8.22416306e-01 -6.51955545e-01
-1.17315185e+00 7.60117829e-01 -2.93072760e-01 -6.25165224e-01
8.05191044e-03 -4.67966884e-01 -6.40672386e-01 8.10064673e-02
1.78723827e-01 7.21509755e-01 5.57379425e-01 -9.93824482e-01
-4.40127343e-01 -3.74523550e-01 4.68917610e-03 3.69178921e-01
-1.09786808e+00 4.75430228e-02 -1.35950536e-01 -8.58807504e-01
1.10121883e-01 -7.44190037e-01 3.58958878e-02 -1.83618590e-01
-8.46307129e-02 -5.61192334e-01 7.70968258e-01 -7.09828496e-01
1.72410440e+00 -2.44036174e+00 1.34709984e-01 -1.48356557e-01
3.48479152e-02 5.34020305e-01 -3.36162329e-01 5.89625955e-01
-1.22779369e-01 2.41247371e-01 -3.28849196e-01 -4.64433491e-01
5.22827953e-02 1.19998343e-01 -1.21384919e-01 1.23690046e-01
3.73401284e-01 7.88072407e-01 -1.18227899e+00 -5.57604730e-01
1.76496089e-01 2.61613697e-01 -7.49799967e-01 2.67004699e-01
1.91423133e-01 -3.06990772e-01 -1.08178034e-01 2.77341008e-01
5.64364493e-01 5.05415797e-02 1.03980690e-01 -1.71594009e-01
3.38192098e-02 3.63367677e-01 -1.03020942e+00 1.67931402e+00
-7.39767075e-01 6.11051142e-01 -5.48862040e-01 -1.05619264e+00
1.26590896e+00 3.79108824e-02 2.03236148e-01 -7.12735236e-01
-6.20989539e-02 3.19712579e-01 1.25499651e-01 -4.15519476e-01
9.19182599e-01 -3.37183714e-01 -2.42543593e-01 4.04274613e-01
1.39057741e-01 1.04957238e-01 1.87137023e-01 4.30268675e-01
1.08324623e+00 -4.53479320e-01 2.14473203e-01 -4.56957191e-01
7.39221752e-01 -4.80587721e-01 2.65995413e-01 4.90889281e-01
-2.49884576e-01 6.12377048e-01 4.07470077e-01 -3.16456072e-02
-1.03775215e+00 -1.33704245e+00 -2.02015162e-01 1.18219554e+00
7.66779035e-02 -7.77175426e-01 -6.73205912e-01 -7.53515601e-01
1.79707021e-01 9.66285884e-01 -5.72857678e-01 -7.82259047e-01
-4.70229238e-01 -5.70163667e-01 2.77177542e-01 7.53440678e-01
4.41416472e-01 -1.06447637e+00 -7.16038197e-02 1.67234167e-01
9.19202566e-02 -7.98030615e-01 -7.00276911e-01 -1.01774978e-02
-8.86790872e-01 -6.46585166e-01 -5.07652462e-01 -9.76658821e-01
7.11866200e-01 3.33908826e-01 1.00298727e+00 3.64390798e-02
2.79999934e-02 4.20908444e-02 -7.18981087e-01 -1.26003429e-01
-2.54674882e-01 1.37385771e-01 3.30160141e-01 2.78340057e-02
6.80291653e-01 -4.77410108e-01 -3.79506886e-01 -2.21701562e-02
-1.00251818e+00 -3.79204422e-01 4.84408200e-01 1.18531954e+00
3.95475537e-01 -6.30491972e-02 9.26285446e-01 -7.03649282e-01
1.07674468e+00 -7.43755639e-01 5.63654292e-04 7.24611506e-02
-8.23519409e-01 4.13443744e-02 8.50024819e-01 -4.75381047e-01
-5.98332822e-01 -4.89301682e-01 3.77263241e-02 -3.77897680e-01
8.16210136e-02 6.49491787e-01 2.31493618e-02 4.76341546e-01
5.65210104e-01 3.98092896e-01 1.12240583e-01 -4.18867677e-01
4.94483650e-01 1.21951151e+00 2.07953706e-01 -2.28403211e-01
7.14325011e-01 1.58742424e-02 -5.76103389e-01 -9.97245967e-01
-8.40137780e-01 -6.19015872e-01 -5.97455204e-01 7.99239278e-02
7.74640143e-01 -6.12817585e-01 -2.15604335e-01 3.18349838e-01
-1.15753937e+00 1.63917422e-01 -2.67573237e-01 6.03340328e-01
-1.45818278e-01 5.64271629e-01 -4.90129083e-01 -7.92395413e-01
-5.57824671e-01 -8.44541013e-01 7.32389987e-01 4.07243609e-01
-4.72665727e-01 -1.16905820e+00 4.90846187e-02 8.20544437e-02
5.76762080e-01 -2.16190070e-01 9.44459260e-01 -1.11176026e+00
6.69262409e-02 -2.51317620e-01 -2.38713846e-01 7.86592543e-01
2.22936496e-01 -4.14112918e-02 -8.84434521e-01 -3.77861410e-01
-7.08674639e-02 -1.03176892e-01 7.86408484e-01 4.12670970e-02
1.18255711e+00 -2.28120640e-01 -1.48392960e-01 2.24812955e-01
1.32777643e+00 2.20883518e-01 5.65743744e-01 5.10942519e-01
4.92000729e-01 3.77122343e-01 7.70630717e-01 4.50636476e-01
3.75815928e-01 4.77939248e-01 1.17482059e-01 2.18879089e-01
-8.75190422e-02 -2.76938796e-01 6.11005008e-01 1.60501730e+00
4.32316929e-01 -4.56915013e-02 -8.17287862e-01 7.51473069e-01
-1.65766037e+00 -8.21851373e-01 -5.55380173e-02 2.14508986e+00
9.20213223e-01 5.10678232e-01 -4.70517315e-02 3.64027113e-01
8.52342308e-01 4.04692650e-01 -3.03802967e-01 -8.70469391e-01
1.21142520e-02 4.06222582e-01 1.08963989e-01 5.69551885e-01
-1.07618940e+00 9.25759614e-01 5.29941940e+00 9.32494760e-01
-8.97464454e-01 2.11580992e-01 4.05916125e-01 -4.80523221e-02
-4.94895250e-01 -6.38280511e-02 -7.08318174e-01 9.39192295e-01
1.24798834e+00 -5.60797393e-01 1.89960539e-01 7.69028068e-01
-2.57220268e-02 1.15033604e-01 -9.12151754e-01 9.57217515e-01
4.16363955e-01 -1.10475409e+00 2.39347592e-01 -3.64006460e-01
6.71278059e-01 -1.73104256e-01 1.55136108e-01 7.43740201e-01
-2.10668817e-01 -8.75071049e-01 3.59963298e-01 5.13686895e-01
4.66537416e-01 -9.91572261e-01 1.16548264e+00 2.97775686e-01
-1.22216356e+00 -6.55063018e-02 -6.34453833e-01 -5.94470836e-02
6.75545707e-02 7.14241326e-01 -7.43387878e-01 6.91994607e-01
5.84745109e-01 1.06466556e+00 -7.12246478e-01 6.94987714e-01
-2.09489271e-01 6.21517718e-01 -3.85460779e-02 -6.13208532e-01
4.18504447e-01 -1.91785082e-01 4.20561552e-01 1.41413438e+00
2.15726286e-01 -6.15549088e-02 1.26793040e-02 6.23983026e-01
-1.96299866e-01 3.18340391e-01 -4.03780520e-01 -1.89603180e-01
9.30640221e-01 1.06205571e+00 -2.85059452e-01 -6.73642457e-01
-3.94189835e-01 1.10751963e+00 4.96751517e-01 3.28747220e-02
-1.00096798e+00 -1.16590273e+00 7.37949789e-01 -5.81375249e-02
3.25147748e-01 -1.97408721e-01 -2.83477604e-01 -1.10371685e+00
2.50590950e-01 -5.11053085e-01 3.59570503e-01 -6.90317154e-01
-1.68174803e+00 6.65590763e-01 -3.06204371e-02 -1.24227071e+00
-9.67464894e-02 -4.85955983e-01 -8.03096056e-01 9.40265715e-01
-1.50701642e+00 -6.44913018e-01 -3.05335253e-01 1.02036074e-01
5.96784711e-01 -3.53880137e-01 9.22780395e-01 3.16974014e-01
-5.26911020e-01 9.60817516e-01 5.32074988e-01 9.53871310e-02
8.40900064e-01 -1.34036374e+00 5.01249492e-01 6.33611083e-01
1.16011731e-01 8.67887259e-01 5.29024541e-01 -3.60843629e-01
-1.24255228e+00 -1.13604593e+00 1.36510229e+00 -3.57897490e-01
9.66172814e-01 -4.49495316e-01 -1.32523000e+00 3.28768730e-01
3.52859139e-01 6.87224120e-02 8.15861225e-01 3.05160940e-01
-4.49998707e-01 -3.92200559e-01 -1.04230356e+00 7.41091788e-01
9.91078854e-01 -8.91312957e-01 -1.21390378e+00 1.15308285e-01
1.02615416e+00 1.21113859e-01 -1.09254777e+00 1.36635721e-01
2.07859978e-01 -8.11813772e-01 8.31116319e-01 -7.60578156e-01
6.47749662e-01 -1.70204371e-01 -2.29476020e-01 -1.58261204e+00
-4.76628304e-01 -1.90108314e-01 -2.32535750e-01 1.43349385e+00
3.66456360e-01 -8.64589334e-01 4.82824981e-01 2.15290263e-01
-4.78459865e-01 -9.68929350e-01 -8.73268425e-01 -1.20515215e+00
4.67688918e-01 -1.37963504e-01 5.48897803e-01 1.12304378e+00
3.94540250e-01 5.02808213e-01 7.53630400e-02 -7.26523995e-02
3.76072437e-01 -1.62508145e-01 5.15408218e-01 -9.82389987e-01
-7.65762776e-02 -4.66461658e-01 -6.74702704e-01 -1.00556278e+00
2.81847179e-01 -1.34650147e+00 -1.58495367e-01 -1.52170146e+00
3.21226746e-01 -3.48520577e-01 -8.95179868e-01 2.71971554e-01
-6.11118674e-01 -5.16375452e-02 1.94865227e-01 -2.60117073e-02
-5.41655898e-01 8.55469108e-01 9.54577625e-01 -2.07700372e-01
-2.14157477e-02 -5.99054933e-01 -6.70393050e-01 3.07336479e-01
1.06660914e+00 -5.37428856e-01 -5.00884771e-01 -4.62813765e-01
-1.45942628e-01 -5.19478619e-01 7.59788975e-02 -9.54823136e-01
2.62544692e-01 1.93451166e-01 1.24465741e-01 -5.33607006e-01
3.44500244e-01 -4.14887637e-01 -3.74157846e-01 6.61819160e-01
-5.67773104e-01 2.06653655e-01 1.85201481e-01 5.12613654e-01
-4.97115344e-01 -6.01381838e-01 6.50410652e-01 1.56503320e-01
-8.09548378e-01 4.42545786e-02 -7.45319501e-02 4.49406981e-01
8.71751249e-01 -2.38075897e-01 -3.46447825e-01 -2.27726880e-03
-5.34714997e-01 8.48860145e-02 2.01345161e-01 7.67386138e-01
8.90182436e-01 -1.75070405e+00 -8.23450506e-01 2.10761815e-01
3.67786735e-01 -2.06845939e-01 4.27931339e-01 6.76419377e-01
-1.11894146e-01 2.69927293e-01 -7.51194060e-02 -4.83016998e-01
-1.34915864e+00 5.97104073e-01 -1.65056631e-01 -4.71101217e-02
-4.76078331e-01 8.39753926e-01 -2.78191771e-02 -5.39736331e-01
-4.99804243e-02 -3.76174390e-01 -2.96609610e-01 3.28142822e-01
5.65410495e-01 4.11848217e-01 8.67379308e-02 -4.69203174e-01
-6.43779755e-01 5.43792248e-01 -4.03819710e-01 8.63977969e-02
1.30094969e+00 4.59819622e-02 -2.09654242e-01 6.22067690e-01
1.73462391e+00 -8.80983546e-02 -6.34389639e-01 -2.87076175e-01
2.84127593e-01 -5.44180870e-01 -9.97226760e-02 -5.42125762e-01
-6.96799457e-01 9.48867917e-01 6.03402078e-01 4.16645318e-01
8.54335666e-01 3.34201045e-02 9.44133818e-01 3.27814102e-01
1.91282090e-02 -1.26711893e+00 3.47403467e-01 7.84584820e-01
9.05293226e-01 -1.14418459e+00 -1.72668919e-01 -8.01590681e-02
-7.08407700e-01 1.13292694e+00 6.21391833e-01 -4.61608917e-01
6.87444925e-01 -2.82502770e-02 -2.44239524e-01 6.53090551e-02
-9.54760373e-01 2.24387217e-02 3.22353750e-01 3.75338942e-01
5.60043693e-01 1.33216679e-01 -7.12036848e-01 8.47362936e-01
-2.49329716e-01 -4.64837164e-01 3.23950112e-01 1.01453686e+00
-8.54532361e-01 -9.11334991e-01 9.40991342e-02 5.32340825e-01
1.34536820e-02 -4.19953734e-01 -1.96772113e-01 6.83014750e-01
-1.76498488e-01 9.09875214e-01 3.83609325e-01 -6.60457611e-01
6.07000113e-01 2.81314760e-01 1.46942809e-01 -8.85589242e-01
-6.13908947e-01 -3.22442919e-01 1.15329072e-01 -1.08646423e-01
-1.09676696e-01 -5.88988483e-01 -1.53593290e+00 -2.10128278e-01
-4.11284834e-01 3.92760813e-01 5.78810394e-01 7.55629539e-01
5.26712000e-01 7.26554215e-01 8.63560498e-01 -3.79492015e-01
-9.00740325e-01 -1.29475200e+00 -5.15282154e-01 7.20906913e-01
-3.61322761e-02 -6.05610430e-01 -8.15286160e-01 -3.04866225e-01] | [10.885794639587402, 8.646835327148438] |
6a31ad63-aad0-4d00-bccd-813189198da4 | fight-fire-with-fire-reversing-skin | 2208.10373 | null | https://arxiv.org/abs/2208.10373v2 | https://arxiv.org/pdf/2208.10373v2.pdf | Reversing Skin Cancer Adversarial Examples by Multiscale Diffusive and Denoising Aggregation Mechanism | Reliable skin cancer diagnosis models play an essential role in early screening and medical intervention. Prevailing computer-aided skin cancer classification systems employ deep learning approaches. However, recent studies reveal their extreme vulnerability to adversarial attacks -- often imperceptible perturbations to significantly reduce the performances of skin cancer diagnosis models. To mitigate these threats, this work presents a simple, effective, and resource-efficient defense framework by reverse engineering adversarial perturbations in skin cancer images. Specifically, a multiscale image pyramid is first established to better preserve discriminative structures in the medical imaging domain. To neutralize adversarial effects, skin images at different scales are then progressively diffused by injecting isotropic Gaussian noises to move the adversarial examples to the clean image manifold. Crucially, to further reverse adversarial noises and suppress redundant injected noises, a novel multiscale denoising mechanism is carefully designed that aggregates image information from neighboring scales. We evaluated the defensive effectiveness of our method on ISIC 2019, a largest skin cancer multiclass classification dataset. Experimental results demonstrate that the proposed method can successfully reverse adversarial perturbations from different attacks and significantly outperform some state-of-the-art methods in defending skin cancer diagnosis models. | ['Zhiqi Shen', 'Yuan Li', 'Yongwei Wang'] | 2022-08-22 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 7.53150523e-01 -1.46026582e-01 7.04735285e-03 -1.54547710e-02
-9.74027574e-01 -9.25089359e-01 4.85937983e-01 -9.06962603e-02
-3.64237726e-01 5.46149492e-01 4.44378033e-02 -2.73137212e-01
1.34024262e-01 -8.06671083e-01 -5.09349704e-01 -1.34259427e+00
1.34378031e-01 -4.77299541e-01 1.94663510e-01 -4.64986920e-01
2.52300408e-02 6.78337872e-01 -8.27652872e-01 4.88007039e-01
1.06472409e+00 7.60608852e-01 -6.03340864e-01 1.08495343e+00
1.68984249e-01 8.84378433e-01 -6.85019970e-01 -8.18589449e-01
2.88400054e-01 -4.92427409e-01 -5.68904281e-01 -6.30952492e-02
5.57859361e-01 -2.71672577e-01 -5.98213375e-01 1.69190466e+00
8.84187341e-01 -4.61949766e-01 6.99985981e-01 -1.03723741e+00
-9.76598263e-01 2.75379747e-01 -7.48489261e-01 1.87488914e-01
2.91549772e-01 4.42111015e-01 2.74477541e-01 -3.66860718e-01
4.80645388e-01 1.09517062e+00 7.46915519e-01 1.17765892e+00
-1.05232704e+00 -8.52317810e-01 -2.79926900e-02 4.09376323e-02
-1.16124475e+00 -2.07552105e-01 1.02579772e+00 -4.12612334e-02
1.25488024e-02 9.22156453e-01 4.50516790e-01 1.87454784e+00
7.93886483e-01 6.20881021e-01 1.45405602e+00 -1.85693815e-01
2.26036564e-01 1.29040971e-01 -2.83769280e-01 7.46846378e-01
2.52654612e-01 2.13216662e-01 -2.06466988e-01 -5.68887174e-01
5.80972910e-01 2.91313026e-02 -3.31324875e-01 1.39289483e-01
-6.33890629e-01 6.84553385e-01 7.84883618e-01 2.90831476e-01
-3.36674541e-01 7.56520852e-02 6.07126236e-01 2.56099224e-01
3.17458719e-01 3.17681015e-01 9.38049406e-02 5.32913864e-01
-3.33693355e-01 -1.72965124e-01 5.36361039e-01 2.69386023e-01
-4.10757633e-03 1.30471155e-01 -2.47906625e-01 6.17003322e-01
-1.25407562e-01 7.07017839e-01 4.99419183e-01 -6.04885519e-01
2.62332618e-01 3.19896728e-01 -2.59826273e-01 -1.30105877e+00
-1.04784705e-01 -5.81266642e-01 -1.53861964e+00 2.43045658e-01
2.09962949e-01 -4.33452539e-02 -1.03407109e+00 1.44138396e+00
7.09043860e-01 5.80841601e-01 3.88872117e-01 7.31820226e-01
5.68859756e-01 6.84519708e-02 5.25425315e-01 -5.00193126e-02
1.33465958e+00 -5.35470605e-01 -7.84062028e-01 -8.23476389e-02
3.11987102e-01 -8.08073759e-01 7.33257711e-01 4.76695269e-01
-8.87627363e-01 -4.01861399e-01 -1.13542831e+00 2.52177715e-01
-3.86841923e-01 -3.43953639e-01 5.95149517e-01 1.27609134e+00
-6.67775452e-01 4.18576628e-01 -1.02697980e+00 -1.96801126e-01
8.87375653e-01 1.90266222e-01 -6.03919446e-01 -3.39055002e-01
-1.45634317e+00 6.77244663e-01 -1.95036322e-01 2.01867849e-01
-1.13952374e+00 -8.73070896e-01 -7.70806491e-01 -4.44497228e-01
-4.87324037e-02 -6.59693241e-01 6.69285953e-01 -1.38258266e+00
-1.41338539e+00 9.81500685e-01 2.81843334e-01 -6.57238543e-01
8.30161929e-01 -1.77359536e-01 -7.51835525e-01 7.02863395e-01
-1.97546184e-01 7.65252039e-02 1.36023104e+00 -1.35979414e+00
-3.52959782e-01 -5.42875171e-01 -1.48744762e-01 -4.11366597e-02
-1.12094009e+00 1.59334257e-01 -1.69077545e-01 -1.06247556e+00
-2.32144341e-01 -9.44568455e-01 -7.28706598e-01 3.25937033e-01
-7.04289079e-01 5.38959682e-01 7.23669767e-01 -8.09934676e-01
1.16159916e+00 -2.15570688e+00 1.51449785e-01 4.13696676e-01
2.36432895e-01 6.26030862e-01 -4.50661004e-01 1.55930802e-01
-2.42955193e-01 4.22062963e-01 -4.24327224e-01 -5.63373603e-02
-4.57432419e-01 2.36536518e-01 -3.18314761e-01 8.78214002e-01
3.83573025e-01 8.64238143e-01 -9.83689785e-01 -6.63608015e-01
1.91605598e-01 8.01020265e-01 -3.28130603e-01 5.80380931e-02
3.12627673e-01 5.06380379e-01 -7.84528434e-01 1.07210672e+00
1.11269212e+00 2.28602707e-01 1.01345636e-01 -3.36484343e-01
6.55246556e-01 -7.26823568e-01 -7.80448914e-01 1.22902167e+00
-2.18220770e-01 9.58629027e-02 5.02331614e-01 -9.23325717e-01
5.25709808e-01 3.06967884e-01 5.27422130e-01 -4.14156735e-01
2.81057626e-01 2.75727548e-02 -1.37159243e-01 -6.34727299e-01
-1.50974914e-01 -2.54472017e-01 -2.54693568e-01 -2.62742370e-01
-3.43214869e-01 -2.20360249e-01 -4.55724895e-01 2.79287815e-01
1.39724171e+00 -5.03177285e-01 2.76262522e-01 -1.62437633e-01
9.88785028e-01 -1.28529221e-01 6.93394065e-01 7.43664682e-01
-7.25579679e-01 4.89181519e-01 3.08966339e-01 -3.39488357e-01
-5.42415202e-01 -1.47763944e+00 -1.80608541e-01 7.07124114e-01
2.81212211e-01 1.13325655e-01 -1.19153488e+00 -1.24531949e+00
6.63806051e-02 2.87586361e-01 -1.14770532e+00 -7.29577661e-01
-4.03400540e-01 -9.85478044e-01 1.39961302e+00 5.09373248e-01
7.49377608e-01 -6.20510161e-01 1.70834810e-01 -1.11706682e-01
8.64044651e-02 -8.63122642e-01 -5.83253920e-01 -1.58361316e-01
-5.48794329e-01 -1.44871092e+00 -7.77185023e-01 -8.02092731e-01
1.08924556e+00 1.40783846e-01 6.61748707e-01 2.97414213e-01
-9.87480879e-01 4.24277425e-01 -3.25834602e-01 -3.70614886e-01
-1.06236660e+00 -3.31245542e-01 1.96393937e-01 3.65342915e-01
2.03107968e-01 -3.54685605e-01 -7.23235011e-01 2.10938767e-01
-1.44069076e+00 -5.46110988e-01 5.89228094e-01 1.15935814e+00
6.32352114e-01 5.77814579e-01 3.32871467e-01 -1.08919287e+00
6.34700716e-01 -4.74352837e-01 -6.57355040e-02 3.74304652e-01
9.48240682e-02 -3.32770705e-01 1.12839425e+00 -8.52351487e-01
-1.08885181e+00 2.56120354e-01 -4.27441776e-01 -5.20227849e-01
-2.30326056e-01 1.35267809e-01 -3.40310395e-01 -7.95974612e-01
1.08668983e+00 4.44771588e-01 3.24188247e-02 -3.33314016e-02
2.46241271e-01 5.43442130e-01 8.38277698e-01 -3.00450206e-01
1.41466606e+00 9.28335488e-01 3.75872254e-01 -1.00639892e+00
-7.00467706e-01 -5.70519976e-02 -3.36255372e-01 -1.57751709e-01
8.80657732e-01 -8.86198640e-01 -4.90492880e-01 1.12255144e+00
-6.93337321e-01 4.34551351e-02 1.59648225e-01 -6.25142828e-02
-3.45348418e-02 7.93321967e-01 -1.09536064e+00 -6.16423666e-01
-6.55227065e-01 -8.53812277e-01 8.55234027e-01 4.32628989e-01
1.85869753e-01 -1.18876815e+00 6.84389891e-03 5.50314963e-01
5.04706740e-01 1.00989890e+00 6.88535333e-01 -3.88774663e-01
-6.84151426e-02 -7.53517210e-01 1.91375569e-01 7.76863813e-01
6.39407635e-01 2.33107671e-01 -1.02933240e+00 -6.57800198e-01
2.19938576e-01 -3.69140178e-01 8.52302551e-01 -1.31740104e-02
1.54957342e+00 -5.49165487e-01 -4.29255158e-01 9.24811244e-01
1.38250363e+00 -9.32296067e-02 8.27502906e-01 1.75604582e-01
7.52934456e-01 5.27120471e-01 4.44832742e-01 2.03016281e-01
-1.67866290e-01 -3.47030573e-02 7.83744812e-01 -5.89722633e-01
-3.30894031e-02 -1.34183124e-01 4.25333560e-01 3.45445961e-01
1.41681775e-01 -5.74114472e-02 -5.39050937e-01 3.67422163e-01
-1.22405696e+00 -1.04516077e+00 1.40151680e-01 1.99699104e+00
1.04745233e+00 1.58730343e-01 -2.71257192e-01 1.76180616e-01
7.76566327e-01 1.66850671e-01 -7.84323215e-01 -2.42209569e-01
-3.95518005e-01 4.93654698e-01 7.28907228e-01 3.71441275e-01
-1.66135192e+00 8.48238945e-01 6.07765532e+00 1.21473241e+00
-1.29756200e+00 1.15279080e-02 7.82105029e-01 5.14521711e-02
-2.37261131e-01 -6.61115706e-01 -2.49924883e-01 5.03458738e-01
5.46755552e-01 7.67591521e-02 1.10205948e-01 8.96998167e-01
-5.28963171e-02 3.82221192e-01 -5.60297012e-01 5.81994057e-01
9.43530723e-02 -1.27119684e+00 2.36521050e-01 -5.66870384e-02
8.24963927e-01 -5.48478484e-01 6.39642835e-01 -1.71867043e-01
4.78966683e-01 -1.16403246e+00 7.99005032e-02 5.88082492e-01
8.86640072e-01 -1.08946407e+00 7.86814630e-01 1.08895563e-01
-1.04672945e+00 -1.11938566e-01 -2.77319610e-01 5.31873822e-01
-3.75470519e-01 3.53412211e-01 -4.20107126e-01 6.60931945e-01
7.00973332e-01 4.72966552e-01 -1.01138890e+00 4.51560944e-01
-4.08338159e-01 7.52255380e-01 2.28710454e-02 5.87677099e-02
1.08848035e-01 2.00253084e-01 5.10725796e-01 1.31217396e+00
-3.65209430e-02 1.44012421e-01 4.54063974e-02 2.59668350e-01
-1.20145986e-02 -1.38069421e-01 -7.95910358e-01 6.64725015e-03
3.90160769e-01 1.39176953e+00 -5.41612983e-01 -7.63230841e-04
-8.20231289e-02 1.61039340e+00 -1.27790913e-01 3.28623444e-01
-1.03005552e+00 -4.64228243e-01 1.12864172e+00 -1.55371577e-01
-8.84127468e-02 2.71979392e-01 -2.46372163e-01 -1.09200370e+00
-1.58296645e-01 -1.41104543e+00 6.48187041e-01 -2.39662752e-01
-1.85809851e+00 6.70390546e-01 -6.46970034e-01 -1.27662957e+00
2.87917554e-01 -6.71804547e-01 -9.41646099e-01 6.45049393e-01
-1.41129196e+00 -1.58650994e+00 -3.64869297e-01 1.06607568e+00
1.44858971e-01 -2.44294256e-01 1.03989017e+00 3.43332961e-02
-8.07074666e-01 1.28667343e+00 2.20522523e-01 4.75449383e-01
8.08189094e-01 -1.23491263e+00 2.79765934e-01 1.24206924e+00
-2.94408232e-01 6.66502714e-01 5.60459733e-01 -7.47750044e-01
-1.65040410e+00 -1.36790562e+00 2.97435839e-03 -4.06846374e-01
9.26558971e-01 -2.04913184e-01 -9.57048476e-01 1.64006203e-01
2.31170848e-01 5.61406314e-01 1.11826968e+00 -7.41114914e-01
-6.27951860e-01 -3.64923865e-01 -1.78896201e+00 1.02694368e+00
8.02746236e-01 -8.28819454e-01 -2.69528419e-01 4.95465130e-01
5.95209956e-01 -2.78217167e-01 -1.06458855e+00 5.25803447e-01
4.81135249e-01 -7.40716398e-01 1.48370147e+00 -9.23604846e-01
5.27769864e-01 -1.66717663e-01 -9.89835560e-02 -1.28604960e+00
-2.82419235e-01 -9.52209294e-01 7.06058089e-03 1.05455208e+00
-1.94274247e-01 -7.46062458e-01 9.80075777e-01 2.58461267e-01
3.21702480e-01 -7.24233150e-01 -1.01440370e+00 -5.76931417e-01
4.86722440e-01 -6.32614174e-05 2.00036898e-01 1.12997353e+00
-3.97408247e-01 -5.36288381e-01 -3.67891103e-01 8.70309830e-01
1.19856679e+00 -6.02970898e-01 5.28009295e-01 -4.80740994e-01
-2.04044402e-01 -2.23254502e-01 -5.89325249e-01 -8.39590281e-02
1.30309120e-01 -5.36671340e-01 -1.65924296e-01 -7.02666938e-01
-1.83084644e-02 -1.13514006e-01 -8.39473665e-01 4.08826083e-01
-9.16189313e-01 8.30487430e-01 -3.58787589e-02 -1.48364618e-01
-1.75262436e-01 1.43097907e-01 1.44282746e+00 -6.12044871e-01
2.77098179e-01 9.85639989e-02 -9.42455947e-01 8.77171040e-01
8.42763662e-01 -4.97522414e-01 -1.66061565e-01 -3.48816849e-02
-3.73402178e-01 -1.83153078e-01 5.71257949e-01 -9.62297201e-01
2.63344496e-01 -4.74394590e-01 6.96139336e-01 4.05155215e-03
1.15855731e-01 -9.41863596e-01 4.12630513e-02 1.19970286e+00
-3.96903366e-01 -4.28229451e-01 2.42164120e-01 8.15181136e-01
-2.05452234e-01 1.08266652e-01 1.35277331e+00 -6.90210387e-02
-3.79237294e-01 3.39657634e-01 -4.42976922e-01 -3.34590554e-01
1.47517359e+00 -1.76555701e-02 -4.32123363e-01 3.23652215e-02
-6.88774765e-01 6.04942590e-02 5.15208542e-01 2.99914688e-01
9.32981789e-01 -1.28180635e+00 -9.40951228e-01 3.72536600e-01
2.76931450e-02 -4.28326488e-01 7.26887763e-01 5.45282602e-01
-6.88428402e-01 -3.70589197e-01 -3.26549441e-01 -3.25131297e-01
-1.81871641e+00 9.05546546e-01 6.98619723e-01 -4.50048506e-01
-1.96956038e-01 1.17524242e+00 2.01401830e-01 -1.46716058e-01
1.94435477e-01 1.64518028e-01 7.16234595e-02 -3.11843514e-01
8.35029125e-01 2.35816851e-01 -1.56182051e-01 -3.83172154e-01
-5.11399746e-01 6.48636281e-01 -4.24561471e-01 4.47455376e-01
1.00351357e+00 1.26511887e-01 -2.44267136e-01 -5.25091052e-01
1.22367966e+00 2.84353673e-01 -1.18319714e+00 -2.84772422e-02
-4.43686187e-01 -5.45189023e-01 -6.11810684e-02 -8.97825480e-01
-1.23289824e+00 7.27066934e-01 8.94471765e-01 4.01404172e-01
1.79460120e+00 -5.42623222e-01 8.97624075e-01 2.29031757e-01
1.53555259e-01 -8.65791798e-01 4.29574639e-01 -1.61053926e-01
8.80833089e-01 -1.18289840e+00 6.04754239e-02 -7.35358477e-01
-6.56679332e-01 9.46835637e-01 6.94513977e-01 -4.30307537e-01
6.16955876e-01 6.66028559e-01 5.59793472e-01 1.50204852e-01
-2.05167711e-01 3.33258092e-01 1.83988869e-01 9.23483014e-01
-1.52536020e-01 5.00059947e-02 -8.87044296e-02 6.31450534e-01
2.79113114e-01 -3.14954132e-01 3.34549278e-01 9.73026633e-01
-2.17262525e-02 -1.14376521e+00 -8.02872658e-01 1.66551128e-01
-1.03026533e+00 -4.43710163e-02 -7.72027969e-01 6.37307465e-01
1.47234201e-01 8.81940663e-01 -5.15381098e-01 -5.91000021e-01
1.37212411e-01 -4.18495536e-01 3.51849347e-01 -1.56508088e-01
-9.31690097e-01 -9.74405333e-02 -2.52485752e-01 -7.12601006e-01
-2.00852022e-01 -3.30590785e-01 -7.82797635e-01 -2.22482279e-01
-1.02735534e-01 -8.63915160e-02 4.46299702e-01 5.05239010e-01
1.61475092e-01 7.16329455e-01 1.23162889e+00 -5.23418009e-01
-1.09426844e+00 -5.31419873e-01 -4.64501858e-01 7.71548688e-01
5.24353385e-01 -7.82433525e-02 -6.26192927e-01 1.99429065e-01] | [5.505206108093262, 7.94072961807251] |
98dd6362-cd27-45c2-b31d-4b9bc91d19ce | conditional-support-alignment-for-domain | 2305.18458 | null | https://arxiv.org/abs/2305.18458v1 | https://arxiv.org/pdf/2305.18458v1.pdf | Conditional Support Alignment for Domain Adaptation with Label Shift | Unsupervised domain adaptation (UDA) refers to a domain adaptation framework in which a learning model is trained based on the labeled samples on the source domain and unlabelled ones in the target domain. The dominant existing methods in the field that rely on the classical covariate shift assumption to learn domain-invariant feature representation have yielded suboptimal performance under the label distribution shift between source and target domains. In this paper, we propose a novel conditional adversarial support alignment (CASA) whose aim is to minimize the conditional symmetric support divergence between the source's and target domain's feature representation distributions, aiming at a more helpful representation for the classification task. We also introduce a novel theoretical target risk bound, which justifies the merits of aligning the supports of conditional feature distributions compared to the existing marginal support alignment approach in the UDA settings. We then provide a complete training process for learning in which the objective optimization functions are precisely based on the proposed target risk bound. Our empirical results demonstrate that CASA outperforms other state-of-the-art methods on different UDA benchmark tasks under label shift conditions. | ['Toan Tran', 'Tuan-Duy H. Nguyen', 'Anh Tong', 'Lam Tran', 'Anh T Nguyen'] | 2023-05-29 | null | null | null | null | ['unsupervised-domain-adaptation'] | ['methodology'] | [ 6.47300601e-01 1.30491123e-01 -2.70553201e-01 -5.95281303e-01
-1.02918053e+00 -4.67606455e-01 7.62553990e-01 1.18402079e-01
-3.91372979e-01 9.57918823e-01 3.05063035e-02 -4.41873185e-02
-2.16394350e-01 -6.86719239e-01 -8.23543191e-01 -1.08546209e+00
3.43425751e-01 4.99805093e-01 1.22330382e-01 3.96129265e-02
1.20392598e-01 2.37388298e-01 -1.25883210e+00 -6.52390122e-02
1.13991916e+00 9.67244208e-01 1.61915999e-02 3.44676264e-02
1.89531855e-02 2.69275755e-01 -6.69079185e-01 -4.00811523e-01
4.95092630e-01 -6.08113229e-01 -7.23066211e-01 1.16884559e-01
2.51854628e-01 2.58695595e-02 1.50530729e-02 1.19664252e+00
5.15310466e-01 2.36463189e-01 1.39158130e+00 -1.37221611e+00
-6.83476269e-01 1.92271441e-01 -6.10961497e-01 1.26815766e-01
1.78151838e-02 -1.31174400e-01 7.64943779e-01 -8.26388299e-01
5.50575614e-01 1.00746143e+00 7.09595084e-01 7.13662565e-01
-1.55754757e+00 -8.87126863e-01 3.25777024e-01 1.14208080e-01
-1.35848057e+00 -1.91171825e-01 1.01739991e+00 -6.98613465e-01
3.31311256e-01 -1.74254641e-01 3.16264518e-02 1.42287612e+00
2.34476104e-01 5.59876382e-01 1.30906963e+00 -6.97139740e-01
6.74392164e-01 4.59493607e-01 1.25155807e-01 1.50263295e-01
3.52766395e-01 1.00217193e-01 -3.77011538e-01 -4.94453281e-01
4.21482861e-01 -1.07750699e-01 3.06952931e-02 -1.20715749e+00
-1.00304842e+00 1.13006866e+00 1.52176633e-01 9.95512009e-02
-5.25369644e-01 -4.46763396e-01 5.55471957e-01 2.96455085e-01
7.18959451e-01 2.10466906e-01 -4.65344489e-01 4.19517875e-01
-6.94814444e-01 4.13154542e-01 4.81558859e-01 1.01460350e+00
6.87155128e-01 -4.02179994e-02 -3.20531577e-01 8.91652346e-01
5.22953033e-01 5.62542737e-01 6.05093539e-01 -6.67018175e-01
4.64989096e-01 5.51733434e-01 2.87211925e-01 -5.60244262e-01
-7.86951855e-02 -5.41069329e-01 -7.45227456e-01 4.03742850e-01
8.39373231e-01 -2.15385437e-01 -8.24695408e-01 2.10687137e+00
6.58257306e-01 3.48758548e-01 5.51483512e-01 5.95271230e-01
2.19604835e-01 4.60249960e-01 3.07869613e-01 -2.58850366e-01
1.10628080e+00 -6.62843406e-01 -7.04347014e-01 -2.35841647e-01
5.81595421e-01 -7.26050913e-01 1.23440969e+00 2.05410883e-01
-6.78584337e-01 -5.38946033e-01 -1.31386638e+00 2.17509985e-01
-3.07458967e-01 5.29588610e-02 1.04110338e-01 8.12752068e-01
-5.84051788e-01 5.15255332e-01 -5.88367522e-01 -4.87599194e-01
6.13355756e-01 2.59660751e-01 -3.99382114e-01 -1.40080228e-01
-1.32341528e+00 9.85217869e-01 4.64265853e-01 -5.11744320e-01
-9.57391024e-01 -7.55996168e-01 -8.74812305e-01 -7.35192075e-02
3.34727883e-01 -6.72371745e-01 1.28198802e+00 -1.05186605e+00
-1.60298383e+00 1.10649252e+00 -1.93667151e-02 -7.37649262e-01
5.92777252e-01 -2.43356422e-01 -2.49415696e-01 -1.57675207e-01
2.54677206e-01 2.53438860e-01 1.25810349e+00 -1.34145808e+00
-6.08650744e-01 -4.86828059e-01 -3.85665894e-01 3.04540873e-01
-4.30361450e-01 -1.96558341e-01 1.53063565e-01 -1.00612450e+00
-1.87965393e-01 -7.58209229e-01 -9.46792886e-02 -1.22167416e-01
-1.93073690e-01 -4.55058962e-01 8.62543881e-01 -4.02353644e-01
1.08593690e+00 -2.32606363e+00 3.24560732e-01 2.88733780e-01
-2.27503762e-01 3.69957864e-01 1.06885843e-01 3.47320408e-01
-3.26009691e-01 -3.88407677e-01 -6.66956425e-01 -3.77531767e-01
2.57598519e-01 1.11174636e-01 -7.13158607e-01 7.47053027e-01
1.59923241e-01 2.28743866e-01 -8.49906445e-01 -4.87971842e-01
6.18816800e-02 3.29759002e-01 -5.27459562e-01 4.35871810e-01
-8.60420763e-02 7.00482249e-01 -5.15236378e-01 3.82166207e-01
9.48481083e-01 5.72255626e-03 1.88893095e-01 6.49579391e-02
4.07147437e-01 6.59800693e-02 -1.07581842e+00 1.68839264e+00
-2.75928289e-01 9.09983963e-02 -6.13562856e-03 -1.48729122e+00
1.31563497e+00 2.48315483e-01 4.22257781e-01 -3.13684911e-01
1.15360610e-01 3.68098199e-01 -1.66033998e-01 -4.62029278e-02
4.62681390e-02 -7.26569295e-01 -4.07957792e-01 2.68055618e-01
2.71342188e-01 -3.98955680e-02 -3.18378806e-01 -2.57357117e-02
7.95567453e-01 3.40413541e-01 6.83023751e-01 -5.50423145e-01
7.84456313e-01 -1.99761078e-01 8.23505461e-01 6.52068317e-01
-4.94231403e-01 5.13719916e-01 4.46714699e-01 -1.29255295e-01
-1.13885379e+00 -1.35133469e+00 -4.60836321e-01 1.11684585e+00
7.05714077e-02 2.20058724e-01 -8.99896562e-01 -1.26616132e+00
2.16250718e-01 1.15459633e+00 -8.47174764e-01 -5.17025471e-01
-2.55735278e-01 -5.92303932e-01 4.14049625e-01 4.50935215e-01
5.99075615e-01 -7.90767848e-01 -1.46169558e-01 6.60682172e-02
6.24190159e-02 -7.13937938e-01 -5.62395692e-01 3.78670990e-01
-8.61101866e-01 -1.00637162e+00 -1.15506506e+00 -9.47592974e-01
7.26755679e-01 -8.50521922e-02 7.69151807e-01 -8.68741333e-01
2.46652678e-01 3.02461028e-01 -4.29387033e-01 -5.97336650e-01
-5.62626898e-01 1.60807654e-01 3.80735099e-01 3.80412996e-01
6.44182503e-01 -6.69803441e-01 -3.77492607e-01 4.60681200e-01
-1.00767589e+00 -3.02227020e-01 4.73571569e-01 9.79552627e-01
7.32075095e-01 2.79177316e-02 1.23408604e+00 -1.29666781e+00
5.66100597e-01 -8.65757942e-01 -4.28257614e-01 3.90442878e-01
-8.41326356e-01 2.51287520e-01 7.12658703e-01 -6.40621543e-01
-1.34138322e+00 2.00640559e-01 1.16904914e-01 -4.05632526e-01
-3.70485723e-01 1.68963343e-01 -5.88203847e-01 1.96227223e-01
9.80989099e-01 3.47892582e-01 1.01593062e-01 -5.09779692e-01
2.90045351e-01 6.69537723e-01 6.28933907e-01 -8.72436404e-01
1.07983220e+00 5.41048527e-01 8.81026462e-02 -3.21245223e-01
-9.90686476e-01 -5.23326635e-01 -9.15413082e-01 2.34284729e-01
7.03432262e-01 -8.82485151e-01 7.08496617e-03 6.31802738e-01
-8.45668614e-01 -2.14776918e-01 -6.52779162e-01 6.18995726e-01
-9.11995530e-01 4.31010962e-01 9.26888585e-02 -5.77762902e-01
-2.64344692e-01 -9.31171894e-01 8.34818006e-01 2.11044237e-01
-1.95616767e-01 -1.18030298e+00 3.31835657e-01 1.21834099e-01
2.58466601e-01 4.38492805e-01 1.15801001e+00 -1.29383755e+00
2.76860714e-01 -2.27592230e-01 -2.41016820e-02 9.03123081e-01
4.80299413e-01 -6.85697377e-01 -8.93892348e-01 -6.69479489e-01
3.33974242e-01 -2.79099971e-01 6.79198444e-01 4.49811012e-01
8.55144322e-01 -1.97846428e-01 -4.03009385e-01 4.19890881e-01
1.22663987e+00 3.38244259e-01 3.66555929e-01 5.36711812e-01
2.55086809e-01 6.67485476e-01 1.18721533e+00 5.57823300e-01
-2.51945704e-02 6.73327088e-01 2.47099161e-01 7.35270604e-02
6.47572204e-02 -3.83181930e-01 4.32156622e-01 4.14023459e-01
4.77556080e-01 -1.28377616e-01 -7.03949094e-01 7.36337423e-01
-1.80831206e+00 -6.41561449e-01 4.30542052e-01 2.54305649e+00
1.05852842e+00 2.51468480e-01 3.39062303e-01 4.29228134e-02
1.07727075e+00 -5.01154810e-02 -9.60272610e-01 -3.20743293e-01
8.12798738e-02 2.43956894e-01 5.12454152e-01 4.25115287e-01
-1.41591668e+00 5.65878093e-01 5.94281769e+00 1.07213020e+00
-8.66716146e-01 2.59726763e-01 3.72954428e-01 2.11920232e-01
-1.93327546e-01 -8.87108371e-02 -7.70475566e-01 6.17718577e-01
8.43668520e-01 -6.67396724e-01 2.65350286e-02 1.27198291e+00
-1.38688862e-01 1.72903672e-01 -1.27769208e+00 6.55897081e-01
7.21806586e-02 -7.31731772e-01 1.96516067e-02 1.33030459e-01
8.89914215e-01 -4.31144685e-01 3.40160936e-01 5.95348716e-01
3.52322727e-01 -6.92551076e-01 5.39870441e-01 3.15584213e-01
1.02539694e+00 -9.28963423e-01 8.30041766e-01 4.86475915e-01
-7.24135637e-01 -6.98157772e-02 -3.75648469e-01 2.32518539e-01
-3.28639448e-02 4.41808075e-01 -9.88147259e-01 6.91513836e-01
4.27418262e-01 6.51366591e-01 -3.62068892e-01 8.11547041e-01
-1.37972414e-01 8.30912948e-01 -6.31138831e-02 3.38869959e-01
-2.54965145e-02 -3.14760298e-01 7.93244243e-01 9.89848495e-01
3.37426007e-01 -2.50606149e-01 6.03022575e-02 7.32388020e-01
-1.37923747e-01 3.16101760e-01 -8.81503999e-01 2.02981040e-01
6.66944146e-01 7.06280231e-01 -4.89977896e-01 -2.59852916e-01
-4.66719747e-01 1.03649306e+00 2.28958413e-01 3.68889987e-01
-8.73514235e-01 -5.11169910e-01 6.96375787e-01 2.94116974e-01
4.45031315e-01 1.85218036e-01 -3.76294464e-01 -1.00723529e+00
-3.19514535e-02 -8.34438443e-01 6.76982880e-01 -2.98831254e-01
-1.81455076e+00 3.83401453e-01 4.06684637e-01 -1.62185955e+00
-3.54934275e-01 -5.31580746e-01 -5.96750557e-01 1.14241076e+00
-1.49244356e+00 -1.15835822e+00 9.77499709e-02 9.03973520e-01
4.30353671e-01 -5.26365876e-01 1.00390530e+00 2.80902416e-01
-3.76938134e-01 9.34513986e-01 7.59924293e-01 -2.44295597e-02
1.32087743e+00 -1.31040311e+00 1.31705523e-01 8.01436603e-01
-4.12530005e-01 6.03223562e-01 8.03035378e-01 -6.34414673e-01
-6.21120334e-01 -1.21635032e+00 6.78277791e-01 -3.71035665e-01
4.43316579e-01 -3.97681355e-01 -1.14287603e+00 8.69998753e-01
7.13038594e-02 1.26169577e-01 8.58645499e-01 7.87397027e-02
-5.84955513e-01 -2.52216130e-01 -1.67274594e+00 2.13463530e-01
6.34236217e-01 -3.95993173e-01 -1.00771284e+00 2.25265503e-01
6.32247210e-01 -2.06451416e-01 -8.50111246e-01 5.72511852e-01
2.67623425e-01 -5.53780317e-01 9.89718795e-01 -8.88454735e-01
1.81760892e-01 -3.37682635e-01 -3.02047879e-01 -1.56526268e+00
-3.13446611e-01 -2.90960550e-01 6.22476172e-03 1.50128460e+00
2.52413712e-02 -9.27119672e-01 6.37203276e-01 3.04862976e-01
8.36184714e-03 -5.71055472e-01 -1.25393569e+00 -1.05490792e+00
6.07211053e-01 2.34355833e-02 5.07191896e-01 1.19271779e+00
-2.02594057e-01 3.52391750e-01 -2.48132601e-01 2.62732387e-01
9.00515556e-01 -1.36086019e-02 6.55607164e-01 -1.46789742e+00
-1.73388079e-01 -9.43862051e-02 -2.85178244e-01 -6.15150154e-01
4.82227176e-01 -1.05878878e+00 8.34408328e-02 -1.04474103e+00
2.20072702e-01 -5.02404451e-01 -6.68768466e-01 3.33768338e-01
-2.47580171e-01 -2.04548120e-01 -4.75707017e-02 2.88604826e-01
-3.42280507e-01 7.70469368e-01 1.00720870e+00 -3.91135849e-02
-2.36964375e-01 3.56864482e-01 -8.13107014e-01 8.03528666e-01
7.71444976e-01 -6.99311852e-01 -7.85767734e-01 6.53392151e-02
-3.86212438e-01 -3.45967174e-01 1.67997703e-01 -9.70102787e-01
-1.07109748e-01 -3.00166816e-01 3.16870660e-01 -3.31328511e-01
6.17500655e-02 -1.01062036e+00 -1.29544377e-01 3.75732690e-01
-6.14719510e-01 -5.01160860e-01 1.44897968e-01 9.71121848e-01
-1.64951220e-01 -4.31053251e-01 1.31742322e+00 3.20683211e-01
-4.99020427e-01 5.08375876e-02 -1.59546763e-01 2.51018018e-01
1.46026862e+00 -1.17048837e-01 -3.46557721e-02 -1.45469427e-01
-7.37313569e-01 -1.78576494e-03 4.76239771e-01 3.09193313e-01
3.07220489e-01 -1.66315663e+00 -8.82639945e-01 3.36029530e-01
3.47607702e-01 1.83980033e-01 2.46789888e-01 4.50042576e-01
-3.46980654e-02 2.96417147e-01 -3.92961323e-01 -4.75748003e-01
-1.09051478e+00 8.55289042e-01 2.40017742e-01 -5.28024554e-01
-4.42526489e-01 6.97308302e-01 7.08835125e-01 -7.38433123e-01
2.88944066e-01 3.88558544e-02 -1.62515968e-01 -2.86446437e-02
2.95562088e-01 3.37435365e-01 1.52798072e-02 -6.31398499e-01
-3.75506341e-01 4.27125007e-01 -3.50150913e-01 -6.03061058e-02
1.10274780e+00 -1.72461465e-01 4.94507253e-01 3.80174845e-01
1.17294538e+00 2.42516939e-02 -1.58027005e+00 -8.20889652e-01
5.53882234e-02 -5.75515687e-01 -3.03843409e-01 -8.19089413e-01
-5.07143855e-01 8.37605357e-01 9.71368670e-01 -4.46565114e-02
1.22007275e+00 -7.13989511e-03 4.77834791e-01 -2.77062710e-02
3.04790854e-01 -1.23160326e+00 1.48072228e-01 2.24760130e-01
9.24313903e-01 -1.11617696e+00 -6.40720427e-02 -4.10241693e-01
-9.32592571e-01 6.72351718e-01 5.65512419e-01 -3.77594471e-01
6.45428896e-01 -1.94487676e-01 4.11926173e-02 3.18116099e-01
-2.97195703e-01 5.62905222e-02 2.19948918e-01 1.08086991e+00
1.07426435e-01 -2.32490320e-02 -4.71054554e-01 1.11153746e+00
1.04580969e-01 -1.20054901e-01 6.75326586e-02 8.25823069e-01
-2.77971655e-01 -1.48696828e+00 -6.09767079e-01 1.76526517e-01
-4.47035968e-01 2.67876357e-01 -1.92332566e-01 7.69021094e-01
1.94837779e-01 6.75029397e-01 -2.35970363e-01 -1.88443303e-01
5.43524086e-01 6.11099362e-01 3.23547572e-01 -7.50598013e-01
-3.49229649e-02 -1.52169112e-02 -4.05088931e-01 -8.56335461e-02
-4.52812761e-01 -1.06366253e+00 -1.11057627e+00 9.51869264e-02
-2.88847208e-01 2.48032659e-01 4.49862331e-01 9.75745320e-01
2.94373125e-01 3.24088156e-01 8.67146969e-01 -6.44054770e-01
-1.25904000e+00 -1.15841186e+00 -9.08697486e-01 6.83186948e-01
2.60242552e-01 -1.14391530e+00 -5.18365145e-01 1.92172766e-01] | [10.367227554321289, 3.156066417694092] |
6e975359-76d3-43a7-bfd9-a7911c45782c | gae-isumm-unsupervised-graph-based | 2212.12937 | null | https://arxiv.org/abs/2212.12937v1 | https://arxiv.org/pdf/2212.12937v1.pdf | GAE-ISumm: Unsupervised Graph-Based Summarization of Indian Languages | Document summarization aims to create a precise and coherent summary of a text document. Many deep learning summarization models are developed mainly for English, often requiring a large training corpus and efficient pre-trained language models and tools. However, English summarization models for low-resource Indian languages are often limited by rich morphological variation, syntax, and semantic differences. In this paper, we propose GAE-ISumm, an unsupervised Indic summarization model that extracts summaries from text documents. In particular, our proposed model, GAE-ISumm uses Graph Autoencoder (GAE) to learn text representations and a document summary jointly. We also provide a manually-annotated Telugu summarization dataset TELSUM, to experiment with our model GAE-ISumm. Further, we experiment with the most publicly available Indian language summarization datasets to investigate the effectiveness of GAE-ISumm on other Indian languages. Our experiments of GAE-ISumm in seven languages make the following observations: (i) it is competitive or better than state-of-the-art results on all datasets, (ii) it reports benchmark results on TELSUM, and (iii) the inclusion of positional and cluster information in the proposed model improved the performance of summaries. | ['Radhika Mamidi', 'Subba Reddy Oota', 'Mounika Marreddy', 'Anudeep Ch', 'Lakshmi Sireesha Vakada'] | 2022-12-25 | null | null | null | null | ['document-summarization'] | ['natural-language-processing'] | [ 2.14290693e-01 2.90591598e-01 -2.06315249e-01 -2.00721145e-01
-9.37523484e-01 -4.49900031e-01 6.04285598e-01 5.11009753e-01
-1.86008528e-01 8.91462028e-01 1.09877670e+00 -3.12294215e-01
9.27771255e-02 -6.50862455e-01 -6.34967387e-01 -2.57248312e-01
4.49516512e-02 4.39140588e-01 4.28443588e-02 -3.57916206e-01
5.19151032e-01 1.04540521e-02 -9.21060205e-01 3.04203719e-01
1.63646996e+00 3.07439834e-01 3.47160220e-01 1.04877353e+00
-3.16400558e-01 1.01686907e+00 -1.19347858e+00 -4.91489559e-01
-2.50490755e-01 -1.08129609e+00 -9.90538239e-01 1.78304598e-01
8.41521263e-01 -4.31173056e-01 -7.19354391e-01 9.51598108e-01
8.76572073e-01 4.13196623e-01 9.83286023e-01 -6.94787085e-01
-1.07897937e+00 1.39453983e+00 -6.97447598e-01 2.07987502e-01
3.05254638e-01 -1.92593545e-01 1.18491805e+00 -5.27596772e-01
6.96283579e-01 1.01946974e+00 5.66911519e-01 5.96320331e-01
-8.18941593e-01 -3.39757293e-01 1.82519928e-01 7.03730583e-02
-9.31356728e-01 -5.99047363e-01 8.73507142e-01 1.80357337e-01
1.39030790e+00 4.92817312e-01 4.80899274e-01 8.81535411e-01
4.86408800e-01 1.25238287e+00 3.97520512e-01 -2.21148387e-01
3.02482218e-01 -3.93263102e-01 6.42266631e-01 5.93169093e-01
7.25516558e-01 -8.73326242e-01 -3.82157683e-01 -6.83633238e-02
2.09559917e-01 -3.15436214e-01 -3.35068405e-01 3.00452977e-01
-1.08813190e+00 1.06874585e+00 4.04293150e-01 3.12461942e-01
-6.17891848e-01 -5.29701598e-02 9.57949340e-01 1.35631189e-01
7.04506695e-01 4.98246610e-01 -9.69900712e-02 -4.63627744e-03
-1.34125686e+00 1.94207281e-01 9.55500782e-01 9.93519008e-01
3.15714329e-01 7.59910762e-01 -3.94888848e-01 1.12057972e+00
-2.10964590e-01 3.21184993e-01 9.28779900e-01 -7.54382432e-01
9.27843034e-01 5.59011579e-01 -4.39565986e-01 -1.00362444e+00
-4.62128580e-01 -4.17024851e-01 -1.58178961e+00 -7.14099705e-01
-4.50841337e-01 -5.15072703e-01 -9.23891723e-01 1.32856250e+00
-2.58445472e-01 -1.98777556e-01 7.37893045e-01 4.33401912e-01
1.91603124e+00 1.34068918e+00 -2.69992411e-01 -4.78392243e-01
8.85241807e-01 -1.37282121e+00 -8.72711480e-01 -3.56706232e-01
6.68404460e-01 -5.01179755e-01 8.56051922e-01 1.01280883e-01
-1.31125319e+00 -4.59694684e-01 -1.17223012e+00 -3.84899914e-01
-1.20097563e-01 5.00370443e-01 6.19926274e-01 2.29247361e-01
-1.29690337e+00 6.43284798e-01 -8.09621632e-01 -7.37288296e-01
3.08733255e-01 8.61355215e-02 -2.98666984e-01 1.18391730e-01
-8.95848811e-01 6.91962481e-01 9.59703624e-01 -1.40288966e-02
-3.09955746e-01 -3.30743462e-01 -1.28534567e+00 4.98112351e-01
1.64927885e-01 -1.02266467e+00 1.23846424e+00 -6.38161778e-01
-1.64670932e+00 4.69140261e-01 -2.66230583e-01 -8.02531242e-01
1.86515972e-01 -2.19345883e-01 -2.83338934e-01 4.05436337e-01
2.30125129e-01 5.21132350e-01 4.09258932e-01 -1.19442689e+00
-3.36228848e-01 -2.54020452e-01 -5.38841665e-01 3.74572575e-01
-3.33745897e-01 2.23065214e-03 -3.83464485e-01 -9.54486132e-01
-5.61860614e-02 -5.51072955e-01 -2.53718466e-01 -1.37113512e+00
-1.06829679e+00 -4.79526490e-01 8.39099407e-01 -1.36045313e+00
1.68495858e+00 -1.71489477e+00 3.12456250e-01 -3.73336613e-01
1.49670899e-01 6.05541706e-01 -4.27462071e-01 9.82509553e-01
2.04688355e-01 3.05628121e-01 -6.03876233e-01 -6.09164000e-01
9.88926739e-03 1.11742295e-01 -4.68326569e-01 1.01903915e-01
1.72529012e-01 1.13406229e+00 -7.88428903e-01 -8.48335147e-01
-3.70620489e-02 3.57829407e-02 -4.15516824e-01 1.71509579e-01
-2.21352905e-01 9.13123265e-02 -4.76064682e-01 4.37140852e-01
5.00436008e-01 9.59709063e-02 1.82249531e-01 4.83858846e-02
-5.16355457e-03 5.58662951e-01 -6.77943408e-01 1.82251930e+00
-1.51959121e-01 8.12114179e-01 -3.65563840e-01 -1.17349684e+00
9.95743930e-01 9.81070548e-02 9.02127549e-02 -5.21659136e-01
3.53562236e-01 2.48113826e-01 -1.15804054e-01 -3.25058341e-01
1.51823902e+00 4.29696560e-01 -3.80423754e-01 5.99245071e-01
3.45762044e-01 -6.43338859e-01 8.73642743e-01 8.55830908e-01
1.04686999e+00 -1.78162336e-01 7.13418424e-01 -2.00499117e-01
3.47902477e-01 -3.43810245e-02 5.12638450e-01 9.64389801e-01
9.84334424e-02 9.16435838e-01 7.11689413e-01 -4.53503430e-02
-9.86531496e-01 -7.12724686e-01 3.86852264e-01 8.23414266e-01
-7.66902640e-02 -8.03349555e-01 -1.02046204e+00 -7.41278052e-01
-4.87303734e-01 1.21607852e+00 -2.57853419e-01 -2.19955072e-01
-8.15567434e-01 -8.36956680e-01 8.05700898e-01 7.38043904e-01
9.01908100e-01 -1.34743512e+00 -2.04070479e-01 3.17546487e-01
-3.43923241e-01 -9.85232592e-01 -8.11847568e-01 -1.18615195e-01
-1.17300344e+00 -6.70237064e-01 -9.58418489e-01 -1.08358550e+00
3.91885579e-01 2.18062043e-01 1.19370353e+00 -5.95080331e-02
2.83290893e-01 1.81618974e-01 -5.38182020e-01 -5.96943557e-01
-7.60257363e-01 8.25340986e-01 -1.43012643e-01 -3.88132066e-01
-1.54119264e-02 -4.36357498e-01 -2.13803396e-01 -5.82893670e-01
-9.57769513e-01 1.66026279e-01 6.75494432e-01 6.95390046e-01
3.32176298e-01 9.00318697e-02 1.13679242e+00 -9.01624739e-01
1.43199849e+00 -3.12134802e-01 7.73284361e-02 3.94670784e-01
-2.28444248e-01 5.86941652e-02 9.17306185e-01 -1.67605594e-01
-1.04817986e+00 -6.14903390e-01 -2.56197214e-01 1.60450503e-01
2.12061107e-01 1.05220044e+00 -2.36349389e-01 7.61641920e-01
4.93548185e-01 7.19853878e-01 -2.86537558e-01 -4.05728161e-01
4.67018574e-01 1.07792687e+00 1.01737273e+00 -2.32451737e-01
2.38886788e-01 2.37135496e-02 -4.21377361e-01 -1.51662648e+00
-7.85298347e-01 -2.32590511e-01 -4.51294571e-01 2.99267232e-01
8.42273772e-01 -8.32969427e-01 -1.22622833e-01 6.46404266e-01
-1.29605460e+00 -2.77197272e-01 -2.56691009e-01 3.87337029e-01
-4.53524411e-01 9.92957413e-01 -9.01881754e-01 -5.77048361e-01
-1.40718031e+00 -6.21381879e-01 9.53018308e-01 6.59788370e-01
-4.40179735e-01 -1.14935100e+00 2.23414555e-01 3.98185402e-01
2.68437415e-01 3.09703737e-01 1.00403094e+00 -1.21538377e+00
6.94676712e-02 -2.76744485e-01 -1.84983268e-01 6.43937647e-01
2.69354999e-01 2.71211535e-01 -3.95446807e-01 -3.11425328e-01
-2.07093731e-01 -2.94076532e-01 1.52997708e+00 8.43888342e-01
1.06167924e+00 -8.22845161e-01 6.99311867e-02 2.84257233e-01
1.02390110e+00 -2.95990892e-02 6.07742548e-01 1.80448711e-01
9.07202542e-01 2.81369835e-01 2.06595019e-01 3.23843807e-01
7.95404077e-01 -6.67140707e-02 4.25348394e-02 -8.91695917e-03
-3.37951988e-01 -4.21845764e-01 5.86283386e-01 1.98604333e+00
-1.34082101e-02 -9.05168891e-01 -6.29559517e-01 7.92436123e-01
-2.14089251e+00 -9.95523810e-01 -1.91605628e-01 1.70089185e+00
9.39755321e-01 2.17588563e-02 1.09317034e-01 -7.63729662e-02
7.00033605e-01 6.89018846e-01 -5.08142352e-01 -1.03629935e+00
-5.00172794e-01 2.08530203e-01 3.63076717e-01 3.75235468e-01
-1.08878374e+00 1.21794200e+00 5.62032223e+00 9.10743058e-01
-1.02134788e+00 -2.55815029e-01 4.77145225e-01 1.30820617e-01
-3.05491716e-01 -2.55367428e-01 -4.98256415e-01 2.87195235e-01
8.64503682e-01 -5.74561119e-01 2.72323191e-02 5.85458338e-01
1.40905902e-01 -1.10842898e-01 -8.48837316e-01 7.28471041e-01
7.94137895e-01 -1.53878725e+00 7.31757760e-01 -3.71784687e-01
1.13098669e+00 1.55478433e-01 -3.37604284e-01 6.64327800e-01
5.87042749e-01 -8.87577891e-01 3.94281596e-01 3.14691067e-01
5.22367179e-01 -1.09815538e+00 9.97811019e-01 5.26401877e-01
-7.95463264e-01 2.33696833e-01 -6.89704061e-01 5.64294681e-03
1.65038764e-01 3.86054575e-01 -7.91617334e-01 1.20104599e+00
2.44125143e-01 1.11860394e+00 -9.29098308e-01 9.29334342e-01
-4.55922812e-01 1.09223509e+00 2.32230108e-02 -3.79245549e-01
5.65674543e-01 -4.47102100e-01 8.00732911e-01 1.72392774e+00
4.05877024e-01 -3.74222919e-02 2.95887649e-01 4.94569570e-01
-5.42468250e-01 4.00601625e-01 -4.49101269e-01 -5.50300241e-01
2.47785509e-01 1.04497492e+00 -7.10533261e-01 -8.12746108e-01
-7.57072642e-02 1.15744293e+00 2.36116573e-01 5.54171145e-01
-5.06577671e-01 -9.87577319e-01 3.05154640e-03 -4.42829072e-01
3.61684352e-01 -2.07766220e-01 -3.27362776e-01 -1.48561823e+00
-7.04579502e-02 -9.66508806e-01 5.29276431e-01 -8.22508156e-01
-1.04571271e+00 7.62175202e-01 -5.11922836e-02 -7.76253641e-01
-4.70902711e-01 8.12761337e-02 -1.31121647e+00 5.28020501e-01
-1.19985712e+00 -1.34433496e+00 1.84042882e-02 2.12231487e-01
1.23509729e+00 -6.32495284e-01 7.46116042e-01 -1.93589419e-01
-1.05539870e+00 4.76813465e-01 4.43455279e-01 3.81009459e-01
6.41423821e-01 -1.47334898e+00 8.85164976e-01 1.25104415e+00
2.08727941e-01 4.77905333e-01 8.34618390e-01 -9.18027401e-01
-1.17780721e+00 -1.38766861e+00 1.05048895e+00 -9.74629745e-02
5.31554282e-01 -2.92485785e-02 -1.07252312e+00 9.87893164e-01
1.07964659e+00 -8.36266637e-01 6.38348818e-01 -5.63625805e-02
1.14460498e-01 2.43335664e-01 -6.52217627e-01 8.36816430e-01
6.96543038e-01 -1.22672707e-01 -1.22252512e+00 2.80880362e-01
9.99621689e-01 -4.09522057e-01 -4.96006042e-01 2.44669050e-01
1.08818486e-01 -6.96325302e-01 4.34962094e-01 -6.85950637e-01
1.03415072e+00 -5.40808402e-02 -5.93944453e-02 -1.91862547e+00
-1.57089025e-01 -6.88380778e-01 -4.66505349e-01 1.73907137e+00
2.68399656e-01 -5.32555938e-01 4.36305553e-01 -1.13596037e-01
-6.91948056e-01 -6.36746407e-01 -5.22049129e-01 -6.02048278e-01
3.66769165e-01 3.54768410e-02 5.74879467e-01 8.12588751e-01
2.25455798e-02 9.67032552e-01 -3.81091177e-01 -1.83282658e-01
4.34953392e-01 3.12697172e-01 9.73926425e-01 -9.22244668e-01
2.32552029e-02 -7.57385612e-01 -9.60404947e-02 -1.04856455e+00
6.25262856e-01 -1.14199495e+00 -9.44699124e-02 -2.65480375e+00
5.75370610e-01 4.87556905e-01 1.16488107e-01 2.78909117e-01
-6.07794225e-01 -2.22982526e-01 1.77237689e-01 -3.65528725e-02
-9.13192630e-01 9.64521825e-01 1.02882290e+00 -5.22616386e-01
-5.28319895e-01 -3.17642242e-02 -1.16888869e+00 7.72053242e-01
1.03557348e+00 -1.73671052e-01 -3.56649280e-01 -5.82936168e-01
4.38119024e-02 1.32596925e-01 -2.56326228e-01 -7.44176328e-01
3.72873068e-01 5.58665320e-02 3.20458502e-01 -1.28658056e+00
-1.45073757e-01 2.45422900e-01 -6.02300823e-01 2.73639709e-01
-4.67409313e-01 7.77536780e-02 2.01360822e-01 3.67663175e-01
-4.19396579e-01 -5.08582950e-01 3.37567031e-01 -2.13128358e-01
-6.58756852e-01 9.07145217e-02 -6.05162203e-01 3.51146817e-01
3.55203986e-01 -1.43326089e-01 -5.12720883e-01 -6.98514819e-01
-2.69291289e-02 5.48902810e-01 3.11770052e-01 3.91489536e-01
6.70954227e-01 -9.92509902e-01 -1.46340835e+00 -2.83955365e-01
-9.07352269e-02 3.10878903e-01 3.39163512e-01 4.68958527e-01
-8.54130507e-01 5.21903694e-01 8.17787461e-03 -2.29452729e-01
-1.38415086e+00 3.44444484e-01 -1.72724485e-01 -5.03858984e-01
-9.41188931e-01 4.34591025e-01 4.70859446e-02 -6.25836492e-01
1.28562376e-01 -4.81641352e-01 -5.42191207e-01 2.50105143e-01
4.92245227e-01 6.45546019e-01 7.87178576e-02 -6.36814594e-01
-1.15712835e-02 2.46741325e-01 -3.32852781e-01 7.52257258e-02
1.49094021e+00 -1.35315716e-01 -4.49368328e-01 3.55845183e-01
9.48543072e-01 3.12633395e-01 -5.46289563e-01 -8.85711014e-02
9.49348211e-02 3.30911547e-01 5.85381836e-02 -5.82335353e-01
-9.02971804e-01 7.39215434e-01 -3.85669798e-01 2.61689693e-01
1.11980700e+00 -1.14925936e-01 1.26453781e+00 8.10171187e-01
-3.72961789e-01 -1.23461950e+00 -2.28263251e-02 8.73972237e-01
1.24180245e+00 -1.01447093e+00 4.65745032e-01 1.57578409e-01
-1.14636517e+00 1.07949150e+00 4.12995189e-01 -3.19661230e-01
-1.39185056e-01 -2.19534725e-01 -1.82424232e-01 -6.93691298e-02
-6.69416130e-01 -2.48952154e-02 3.71736854e-01 4.09502506e-01
5.55305123e-01 2.15380207e-01 -7.36304760e-01 9.84361351e-01
-8.53770792e-01 -3.42020810e-01 1.13649869e+00 8.08152139e-01
-6.22624695e-01 -5.72449088e-01 -9.22461003e-02 7.60811806e-01
-5.65807045e-01 -2.36408487e-01 -1.00545967e+00 6.96897388e-01
-7.73881912e-01 1.18827927e+00 6.59556240e-02 -6.34666458e-02
3.58232468e-01 -1.25343114e-01 2.83752978e-01 -8.59490991e-01
-6.49482846e-01 7.71922395e-02 3.15085322e-01 6.25728741e-02
-1.90397352e-01 -5.47100604e-01 -1.41979086e+00 -3.96813065e-01
-2.10071459e-01 4.08571541e-01 6.27262354e-01 7.08177388e-01
4.93016541e-01 9.59457457e-01 5.33213794e-01 -9.45427656e-01
-6.17187560e-01 -1.58252811e+00 -6.82673812e-01 1.04654886e-01
2.91878670e-01 2.88982302e-01 -1.17080972e-01 -8.20728838e-02] | [12.529449462890625, 9.516813278198242] |
d7d56e45-7efe-41ae-883b-49a36ec0ace2 | joint-learning-for-aspect-and-polarity | 2201.06313 | null | https://arxiv.org/abs/2201.06313v3 | https://arxiv.org/pdf/2201.06313v3.pdf | A Deep Convolutional Neural Networks Based Multi-Task Ensemble Model for Aspect and Polarity Classification in Persian Reviews | Aspect-based sentiment analysis is of great importance and application because of its ability to identify all aspects discussed in the text. However, aspect-based sentiment analysis will be most effective when, in addition to identifying all the aspects discussed in the text, it can also identify their polarity. Most previous methods use the pipeline approach, that is, they first identify the aspects and then identify the polarities. Such methods are unsuitable for practical applications since they can lead to model errors. Therefore, in this study, we propose a multi-task learning model based on Convolutional Neural Networks (CNNs), which can simultaneously detect aspect category and detect aspect category polarity. creating a model alone may not provide the best predictions and lead to errors such as bias and high variance. To reduce these errors and improve the efficiency of model predictions, combining several models known as ensemble learning may provide better results. Therefore, the main purpose of this article is to create a model based on an ensemble of multi-task deep convolutional neural networks to enhance sentiment analysis in Persian reviews. We evaluated the proposed method using a Persian language dataset in the movie domain. Jacquard index and Hamming loss measures were used to evaluate the performance of the developed models. The results indicate that this new approach increases the efficiency of the sentiment analysis model in the Persian language. | ['Sepideh Saeedi Majd', 'Fatemeh Sadat Masoumi', 'Milad Vazan'] | 2022-01-17 | null | null | null | null | ['aspect-based-sentiment-analysis', 'aspect-category-polarity', 'persian-sentiment-anlysis', 'aspect-category-detection'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.35482559e-02 -2.16560245e-01 3.28587629e-02 -4.80263174e-01
-4.18551534e-01 -5.67349911e-01 5.17667472e-01 6.08450472e-01
-5.40404618e-01 7.14350224e-01 1.23027347e-01 -1.62558645e-01
-3.61192636e-02 -9.44732487e-01 -2.49312133e-01 -5.74739456e-01
2.69200802e-01 3.44895154e-01 6.79396987e-02 -5.45150280e-01
7.52354443e-01 1.13429710e-01 -1.55911958e+00 5.24504900e-01
1.06077194e+00 1.06858826e+00 6.60282373e-02 2.97244340e-01
-5.17264366e-01 7.79230475e-01 -7.58476257e-01 -5.86904705e-01
5.87005205e-02 -2.34982148e-01 -5.07319272e-01 -2.64915109e-01
-1.12516761e-01 3.82137708e-02 5.24774075e-01 9.56562340e-01
5.88987172e-01 9.97164547e-02 6.91245675e-01 -1.09517372e+00
-3.73239040e-01 6.04400516e-01 -5.92216372e-01 -8.82635918e-03
1.87669963e-01 -4.28852379e-01 1.28102279e+00 -8.19528461e-01
2.12921113e-01 8.96282613e-01 8.32143724e-01 6.65320549e-03
-8.07921767e-01 -7.19409943e-01 1.95165321e-01 4.24904525e-01
-8.31965983e-01 7.62320608e-02 1.03868437e+00 -3.61053854e-01
9.78900909e-01 4.55088727e-02 8.00446749e-01 5.88100135e-01
7.24957407e-01 8.09008241e-01 1.39136100e+00 -2.83660382e-01
2.35007614e-01 5.29917300e-01 3.99689853e-01 3.12880337e-01
3.54635686e-01 -2.87044823e-01 -5.44926703e-01 -7.29954615e-02
-2.22631693e-01 2.32087150e-02 6.88592643e-02 -6.62654117e-02
-8.17878544e-01 1.06497467e+00 2.98182935e-01 3.79826039e-01
-6.06845379e-01 -4.15503651e-01 6.32988811e-01 7.48243108e-02
8.08130383e-01 6.68215573e-01 -7.56807029e-01 -1.74185634e-01
-1.02183962e+00 3.04195225e-01 9.06516314e-01 2.05687344e-01
7.47898757e-01 1.12242708e-02 -4.85300049e-02 8.47221792e-01
4.02682245e-01 4.36241001e-01 6.57508731e-01 -1.93673357e-01
3.40494782e-01 1.21994889e+00 -6.91749305e-02 -1.37011719e+00
-6.46097958e-01 -5.74424744e-01 -8.07162642e-01 4.24472481e-01
-9.82981268e-03 -5.10890067e-01 -8.34408402e-01 1.24511611e+00
1.27985492e-01 -5.15100241e-01 3.26366216e-01 7.49727428e-01
1.09855103e+00 8.15636694e-01 1.24598987e-01 -2.00075358e-02
1.71966207e+00 -7.75860965e-01 -9.41743910e-01 -3.90085906e-01
3.85507196e-01 -1.14936304e+00 7.52278090e-01 6.63955510e-01
-6.33509040e-01 -6.35526717e-01 -1.27183735e+00 2.82295138e-01
-7.42636263e-01 4.73973662e-01 6.54040992e-01 6.82606757e-01
-8.21056724e-01 5.49682856e-01 -5.63701272e-01 -1.55072093e-01
2.79115498e-01 3.78249139e-01 -3.62188756e-01 3.28058600e-01
-1.36234093e+00 1.05823898e+00 5.03493071e-01 2.83603340e-01
-3.29374552e-01 -2.26442114e-01 -8.46568167e-01 2.81789094e-01
3.08861043e-02 -2.10148871e-01 9.54665899e-01 -1.41272938e+00
-1.12802577e+00 4.02035773e-01 -2.04327151e-01 -4.28319842e-01
7.37890750e-02 -2.82804161e-01 -3.72052431e-01 -1.85429439e-01
-2.52828840e-02 4.21733528e-01 5.45964241e-01 -1.08586395e+00
-9.54135776e-01 -4.64417666e-01 1.39182195e-01 3.11196625e-01
-6.55843139e-01 8.85550082e-02 -2.38663152e-01 -5.15695155e-01
-1.67504903e-02 -1.11533546e+00 -2.06448361e-01 -7.43017733e-01
-2.70099282e-01 -1.68510631e-01 9.48002219e-01 -8.91289055e-01
1.17333925e+00 -1.88744915e+00 -2.50729650e-01 2.07124516e-01
6.51415884e-02 4.75616693e-01 8.16792324e-02 4.05920863e-01
-7.80724809e-02 6.93187341e-02 -2.79358059e-01 -2.93434054e-01
-2.84163743e-01 -1.53688595e-01 1.43402722e-02 7.35330060e-02
3.61169994e-01 6.22403622e-01 -2.25531071e-01 -2.89392352e-01
4.76172790e-02 7.08321929e-01 -3.47039163e-01 9.81563702e-02
-2.27654517e-01 2.23485366e-01 -1.87779382e-01 4.87015814e-01
8.26982856e-01 9.37032774e-02 9.24127847e-02 -4.96732384e-01
-2.18332529e-01 2.72964537e-01 -1.23050141e+00 9.47246432e-01
-7.22625196e-01 7.60740101e-01 -3.37126762e-01 -1.07609320e+00
1.35787642e+00 2.98444778e-01 4.81586933e-01 -6.06982410e-01
3.60394478e-01 2.10660815e-01 2.82286227e-01 -2.88273096e-01
8.28746557e-01 -1.92580879e-01 -1.05187893e-01 4.66718376e-01
-9.23501179e-02 -1.03918932e-01 3.78627777e-01 -5.66734597e-02
5.04872322e-01 3.39139090e-03 5.31819880e-01 -2.36080691e-01
8.89081657e-01 7.44648948e-02 7.51967371e-01 8.63567293e-02
-4.55129482e-02 3.96519184e-01 8.39335799e-01 -6.41552210e-01
-8.87057900e-01 -4.58983064e-01 -6.31025210e-02 8.80908191e-01
7.26336846e-03 -5.17370701e-01 -3.97932470e-01 -7.84957170e-01
-3.49427819e-01 8.16461325e-01 -6.57867432e-01 4.68555652e-02
-2.38160595e-01 -1.19058418e+00 8.37480053e-02 3.88210744e-01
7.73313165e-01 -1.28598201e+00 -6.68396235e-01 1.49181023e-01
-3.20097327e-01 -8.73511493e-01 9.07514319e-02 4.14007217e-01
-8.32895458e-01 -1.12626135e+00 -6.47204101e-01 -6.40172958e-01
5.63017845e-01 5.02732536e-03 9.42087173e-01 -2.37860471e-01
1.90331474e-01 -1.52799770e-01 -6.71481371e-01 -1.02891159e+00
-2.74450064e-01 3.06452274e-01 -3.13606173e-01 2.44760036e-01
8.77424121e-01 -2.03221604e-01 -6.09487593e-01 9.47981253e-02
-8.97644758e-01 -1.97112888e-01 7.30433345e-01 7.93374836e-01
5.15631080e-01 4.62962389e-01 6.24688268e-01 -1.07700384e+00
1.01504672e+00 -4.76832747e-01 -5.73899746e-01 4.72953804e-02
-1.07114363e+00 9.20770168e-02 8.99403691e-01 -4.75568920e-02
-1.23778844e+00 5.17956391e-02 -3.70974034e-01 2.73236096e-01
6.44146279e-02 1.00778806e+00 -2.94994954e-02 1.29831657e-01
4.29146230e-01 2.68625081e-01 1.62324291e-02 -2.64098078e-01
-2.34989554e-01 9.14059758e-01 -2.44333968e-01 4.71714772e-02
2.08626732e-01 2.39844441e-01 8.30962136e-02 -6.82524621e-01
-8.64719808e-01 -5.80497265e-01 -4.51587528e-01 -3.26775700e-01
8.13281476e-01 -9.74489570e-01 -5.91666341e-01 6.63329303e-01
-1.16973376e+00 3.78816932e-01 2.73723871e-01 5.19603789e-01
-1.02284580e-01 4.27597761e-01 -2.68214554e-01 -9.35197473e-01
-9.42493200e-01 -1.49236834e+00 8.48612607e-01 4.52363312e-01
-4.58275735e-01 -1.02533376e+00 1.73973963e-01 4.10614014e-01
7.04955697e-01 -1.04831043e-03 8.66248965e-01 -1.17832077e+00
-2.14961153e-02 -4.36382234e-01 -6.92552999e-02 7.17841208e-01
2.24168330e-01 1.12310745e-01 -9.98196363e-01 -1.89621672e-02
1.06620520e-01 -2.27010064e-02 8.64921093e-01 4.82824087e-01
8.38526666e-01 -1.79473877e-01 -1.40206544e-02 2.11848810e-01
1.51681256e+00 6.34446263e-01 6.61752462e-01 8.01554322e-01
4.97750968e-01 8.37859631e-01 9.04048502e-01 4.93609816e-01
5.54458499e-01 4.53700423e-01 4.31308240e-01 -6.17221445e-02
2.89993584e-01 2.08688408e-01 4.89230841e-01 7.95307040e-01
-3.08940038e-02 -2.43488044e-01 -1.02738452e+00 5.78767359e-01
-1.69274414e+00 -7.41352916e-01 -3.56471956e-01 1.91344798e+00
4.68289495e-01 3.36116850e-01 7.28889480e-02 5.32051802e-01
5.93939304e-01 2.01044142e-01 -2.43237332e-01 -1.03164542e+00
-2.14960501e-01 2.56702900e-01 2.91016281e-01 2.69950390e-01
-1.38232112e+00 7.06192195e-01 5.18819427e+00 6.24544322e-01
-1.35829437e+00 4.84198369e-02 7.92759717e-01 1.58999637e-01
-4.14501220e-01 -1.62841678e-02 -9.22730744e-01 4.65036273e-01
8.37112665e-01 -5.65756187e-02 -7.12473318e-02 1.08457041e+00
1.62022069e-01 -4.47697341e-01 -4.30464834e-01 6.84339106e-01
4.35247391e-01 -9.03033793e-01 2.07079314e-02 -1.61116928e-01
8.02467585e-01 -9.39726643e-03 1.95181593e-02 2.99679250e-01
6.16894476e-02 -8.94124568e-01 3.33506197e-01 2.64737815e-01
1.21101491e-01 -1.20937395e+00 1.69696093e+00 1.94178149e-01
-1.06296766e+00 -1.48839459e-01 -3.21350336e-01 -1.55856952e-01
6.40282631e-02 9.14744198e-01 -8.87004495e-01 5.78570843e-01
7.76681364e-01 7.73545027e-01 -6.50485396e-01 1.06804752e+00
-2.99718112e-01 5.46578705e-01 -5.76182716e-02 -6.86644614e-01
1.75620183e-01 -3.40173423e-01 4.25175875e-01 1.13902009e+00
5.07625937e-01 -3.95375222e-01 -1.16972886e-01 3.52177769e-01
2.88775951e-01 8.56219947e-01 -5.92798471e-01 -1.46637913e-02
-1.01610333e-01 1.59069777e+00 -1.02949965e+00 -2.28568107e-01
-5.18695593e-01 6.10986829e-01 1.92515031e-02 4.04381752e-03
-5.64687669e-01 -6.98530138e-01 4.95861322e-01 -1.56890526e-01
4.10894394e-01 7.37340376e-03 -8.46122742e-01 -9.80125606e-01
4.54700552e-02 -1.08582532e+00 3.09902787e-01 -5.08806825e-01
-9.48672593e-01 1.03432930e+00 -3.44391227e-01 -1.43763661e+00
-1.96270585e-01 -6.79831266e-01 -7.19852686e-01 9.09380496e-01
-1.59174216e+00 -1.13988626e+00 -2.67465413e-01 2.23733306e-01
7.28720725e-01 -5.17994642e-01 7.04531074e-01 2.23878041e-01
-3.72374117e-01 3.93183887e-01 1.63838249e-02 -1.05320541e-02
7.59101093e-01 -1.29583585e+00 1.87372882e-02 8.64382505e-01
1.00717433e-02 4.47579533e-01 8.60839128e-01 -6.97054148e-01
-8.81431997e-01 -9.29115653e-01 1.20084381e+00 -1.40741020e-01
4.11316961e-01 -8.14887360e-02 -6.86123490e-01 2.90311992e-01
4.06144589e-01 -7.78229356e-01 9.64134872e-01 4.28081095e-01
-3.18155885e-02 -5.88989437e-01 -1.02452314e+00 3.77879739e-01
-1.72241807e-01 -1.68673709e-01 -6.76153362e-01 -7.98124000e-02
2.49252573e-01 2.22195568e-03 -6.59226000e-01 5.75497210e-01
7.58899033e-01 -1.21157956e+00 4.28328931e-01 -1.22374691e-01
8.57907832e-01 -3.17437321e-01 -9.04338583e-02 -1.84785962e+00
-1.65475562e-01 3.31369758e-01 3.57663006e-01 1.43712759e+00
9.04521942e-01 -7.43644059e-01 6.90091789e-01 3.98487985e-01
1.59498692e-01 -8.86408985e-01 -5.46897531e-01 -1.04894541e-01
-2.70198286e-02 -6.25653863e-01 6.65176272e-01 5.88235438e-01
7.67982155e-02 6.93603456e-01 -3.59621584e-01 -1.13960549e-01
1.87364921e-01 2.78106511e-01 5.53988099e-01 -1.40703022e+00
-1.68483816e-02 -5.25957942e-01 -4.46928799e-01 -1.82375506e-01
2.11134180e-01 -6.30739868e-01 1.33127525e-01 -1.72996438e+00
2.44774804e-01 -2.60630369e-01 -5.25456131e-01 4.58275437e-01
-4.04735595e-01 2.95110226e-01 2.54696518e-01 -5.32473158e-03
-4.22635943e-01 6.09926522e-01 8.95733178e-01 -3.27304929e-01
-1.94484159e-01 4.53710675e-01 -1.04880905e+00 7.60239482e-01
1.18847692e+00 -4.48773682e-01 -4.76180941e-01 -1.19852670e-01
8.65955591e-01 -2.66587675e-01 -2.40640879e-01 -1.00015128e+00
1.27262101e-01 2.43698806e-01 5.39848924e-01 -9.24331367e-01
2.48963863e-01 -8.86816084e-01 -1.06134757e-01 4.91973430e-01
-2.18092903e-01 3.25195253e-01 4.20214802e-01 2.18827859e-01
-7.88415372e-01 -3.91961604e-01 5.93159497e-01 -1.38728335e-01
-7.17299759e-01 -2.65646651e-02 -5.94003379e-01 -5.21993160e-01
1.07202971e+00 8.87455698e-03 -2.15818271e-01 -4.96594846e-01
-3.39219004e-01 1.81442827e-01 1.10257693e-01 4.95543510e-01
6.19952798e-01 -9.10761118e-01 -7.18641818e-01 1.23191006e-01
2.40848124e-01 -2.79647410e-01 3.37447613e-01 8.37721229e-01
-6.00718558e-01 5.41197956e-01 -4.41274524e-01 -3.43094140e-01
-1.59318888e+00 2.98820764e-01 1.88776910e-01 -5.87665021e-01
8.26656595e-02 5.54684043e-01 -1.32187847e-02 -7.51599491e-01
-1.40482381e-01 -5.23311384e-02 -1.41966295e+00 7.69856811e-01
4.51511621e-01 2.00914443e-01 4.91991013e-01 -7.15386450e-01
-4.47498798e-01 7.16216683e-01 -2.68686593e-01 -6.56414852e-02
1.45746672e+00 5.79589419e-02 -5.21864176e-01 4.27675009e-01
9.59254801e-01 2.18014956e-01 -4.94157612e-01 2.36058116e-01
5.65736704e-02 -1.83694527e-01 3.48692089e-01 -9.89597261e-01
-1.09093988e+00 1.03714597e+00 6.98815167e-01 3.43667775e-01
1.44550860e+00 -5.20062923e-01 6.97517633e-01 3.53974372e-01
2.12842077e-02 -1.23375809e+00 -2.48976171e-01 7.13777661e-01
5.77107906e-01 -1.56206381e+00 7.37898573e-02 -9.10416469e-02
-1.11104190e+00 1.37904620e+00 7.13810682e-01 -2.91461330e-02
9.00746226e-01 1.61514997e-01 3.88673872e-01 -1.90890223e-01
-7.52529621e-01 -2.05620036e-01 4.51703876e-01 4.15344864e-01
9.25622165e-01 -9.19742092e-07 -1.00456607e+00 9.32388365e-01
-4.79705185e-01 -5.79756275e-02 6.50588930e-01 6.56431317e-01
-3.78642499e-01 -1.19089067e+00 -4.68171597e-01 7.50381112e-01
-1.07136619e+00 -2.15874568e-01 -3.66249889e-01 3.66839170e-01
1.56933889e-01 1.33741903e+00 -6.07183091e-02 -6.70106232e-01
3.95139962e-01 4.86682430e-02 -2.99281776e-01 -3.78105193e-01
-8.35464656e-01 -1.34266419e-02 3.61153781e-01 -1.88007683e-01
-6.37460589e-01 -5.97067952e-01 -8.83302271e-01 -1.19322658e-01
-3.37282330e-01 4.84272480e-01 1.15900421e+00 1.09744763e+00
3.47793281e-01 6.65088296e-01 8.41750622e-01 -5.08634508e-01
-2.50190049e-01 -1.23738611e+00 -3.88457119e-01 2.90009916e-01
5.39757647e-02 -4.85490203e-01 -2.76221693e-01 -2.51238078e-01] | [11.163753509521484, 6.740555286407471] |
b23b9ec3-bf3f-4990-b0d8-e1462f70dc2c | invaastcluster-on-applying-invariant-based | 2206.14175 | null | https://arxiv.org/abs/2206.14175v2 | https://arxiv.org/pdf/2206.14175v2.pdf | InvAASTCluster: On Applying Invariant-Based Program Clustering to Introductory Programming Assignments | Due to the vast number of students enrolled in Massive Open Online Courses (MOOCs), there has been an increasing number of automated program repair techniques focused on introductory programming assignments (IPAs). Such state-of-the-art techniques use program clustering to take advantage of previous correct student implementations to repair a given new incorrect submission. Usually, these repair techniques use clustering methods since analyzing all available correct student submissions to repair a program is not feasible. The clustering methods use program representations based on several features such as abstract syntax tree (AST), syntax, control flow, and data flow. However, these features are sometimes brittle when representing semantically similar programs. This paper proposes InvAASTCluster, a novel approach for program clustering that takes advantage of dynamically generated program invariants observed over several program executions to cluster semantically equivalent IPAs. Our main objective is to find a more suitable representation of programs using a combination of the program's semantics, through its invariants, and its structure, through its anonymized abstract syntax tree. The evaluation of InvAASTCluster shows that the proposed program representation outperforms syntax-based representations when clustering a set of different correct IPAs. Furthermore, we integrate InvAASTCluster into a state-of-the-art clustering-based program repair tool and evaluate it on a set of IPAs. Our results show that InvAASTCluster advances the current state-of-the-art when used by clustering-based program repair tools by repairing a larger number of students' programs in a shorter amount of time. | ['Vasco Manquinho', 'Mikoláš Janota', 'Pedro Orvalho'] | 2022-06-28 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-4.29372162e-01 -1.38208956e-01 2.49178782e-02 -3.16922605e-01
-3.95722598e-01 -9.04096723e-01 1.17998406e-01 1.03603852e+00
-2.81381235e-02 1.17877983e-01 -2.66992569e-01 -6.30681038e-01
-2.48613313e-01 -1.07778203e+00 -7.89641082e-01 -1.73416972e-01
1.69583693e-01 2.31937438e-01 7.25254178e-01 -2.23800465e-01
5.69029987e-01 6.51906729e-01 -2.22800875e+00 3.14983219e-01
1.41818595e+00 1.13498174e-01 1.80079907e-01 7.42926896e-01
-7.38554835e-01 6.68965042e-01 -7.73198843e-01 -4.54043359e-01
5.23120165e-03 -4.46510166e-01 -1.09641588e+00 -2.84122944e-01
6.58796787e-01 4.27272886e-01 -1.19985621e-02 1.41803563e+00
-1.22011915e-01 5.23943782e-01 1.83079839e-01 -1.36011446e+00
-4.22123641e-01 8.44167590e-01 -2.63008475e-01 -1.22667059e-01
6.65112317e-01 -1.72180399e-01 8.05904508e-01 -3.33665550e-01
7.27908432e-01 9.54754353e-01 6.08404636e-01 2.89774954e-01
-1.56608558e+00 -6.43378675e-01 -1.25512660e-01 1.50876015e-01
-1.42675912e+00 3.82823795e-01 6.68550372e-01 -9.09432828e-01
7.73547471e-01 7.27668345e-01 6.82989776e-01 2.71547318e-01
8.71564895e-02 5.21227539e-01 1.12114525e+00 -7.51351058e-01
3.81404668e-01 5.98004520e-01 1.00562871e+00 1.10508096e+00
2.70249516e-01 -2.84931988e-01 2.37536300e-02 -4.51863557e-01
1.63802765e-02 9.84431580e-02 -2.51249522e-01 -6.16560519e-01
-7.64745176e-01 7.22472429e-01 1.13302551e-01 7.94137776e-01
4.01746511e-01 -2.53172647e-02 4.85530019e-01 4.85963672e-01
-1.77933022e-01 7.52868950e-01 -1.75464094e-01 -4.36673880e-01
-1.26176822e+00 4.70097750e-01 1.05422235e+00 1.09528339e+00
1.09175110e+00 -1.10425569e-01 1.54496402e-01 5.42638898e-01
2.32319906e-01 -2.21477062e-01 7.36494541e-01 -5.66018343e-01
3.47429186e-01 1.51599646e+00 -6.17539942e-01 -1.30027533e+00
-4.89585102e-02 -1.54610336e-01 -2.70277500e-01 4.70287085e-01
4.70860660e-01 3.35546196e-01 -4.86708701e-01 1.46344876e+00
3.48823339e-01 2.42092282e-01 -1.43833369e-01 3.87555301e-01
8.38011324e-01 5.39797246e-01 -1.62209153e-01 2.75685072e-01
1.21349621e+00 -1.10912955e+00 -3.19160253e-01 4.22643334e-01
1.36207557e+00 -9.49925721e-01 1.25695443e+00 5.72817147e-01
-7.40677655e-01 -7.78138280e-01 -1.02031672e+00 1.56371742e-01
-8.13335121e-01 9.17951167e-02 3.39942515e-01 1.48049462e+00
-1.09925020e+00 9.98495221e-01 -7.28985310e-01 -4.03895706e-01
1.15099184e-01 5.15574396e-01 -5.27726173e-01 -1.30594701e-01
-3.65314037e-01 4.61908489e-01 4.65140224e-01 -6.41504407e-01
-4.79659140e-01 -1.02469945e+00 -8.24647784e-01 5.19983828e-01
1.68957472e-01 -4.18998182e-01 9.84491110e-01 -7.71538615e-01
-1.49143434e+00 8.17703664e-01 7.43355155e-02 -3.28538828e-02
1.98270008e-01 2.14761272e-01 -1.84663326e-01 2.12859106e-03
-3.17436665e-01 -5.58498837e-02 2.98314005e-01 -1.25077391e+00
-3.77940178e-01 -2.33982533e-01 -1.01072654e-01 -4.12261128e-01
-5.35347939e-01 2.69725591e-01 -1.00838736e-01 -3.94300848e-01
1.72712758e-01 -1.06478274e+00 -2.13193208e-01 -5.18448532e-01
-4.61312801e-01 -2.50442445e-01 9.31637108e-01 -4.07476693e-01
1.59801483e+00 -2.20643163e+00 3.84948999e-01 6.74700379e-01
1.99124485e-01 4.71826315e-01 -1.23703359e-02 3.96387100e-01
-5.17148614e-01 6.77582681e-01 -2.22371608e-01 -4.34759587e-01
2.21473664e-01 9.07564014e-02 -3.45685899e-01 2.53890455e-01
-5.80320835e-01 1.05879441e-01 -8.41917276e-01 -5.98954499e-01
4.75144506e-01 -8.53760764e-02 -9.50829744e-01 4.38584000e-01
-2.16718569e-01 2.27011457e-01 -3.80970269e-01 4.42223907e-01
7.60264277e-01 4.77504134e-01 1.77820593e-01 4.71819401e-01
-5.53330839e-01 5.83761148e-02 -1.48795831e+00 1.74477720e+00
-3.82131875e-01 5.46354294e-01 -8.99100304e-02 -1.09801447e+00
1.39923799e+00 2.29453132e-01 3.73807400e-01 1.20415956e-01
-1.67858191e-02 1.33636311e-01 -9.94429290e-02 -4.78422076e-01
7.55207658e-01 5.09247422e-01 -3.75312269e-01 7.45213151e-01
2.35401884e-01 -3.40180904e-01 4.30578768e-01 5.31585217e-01
1.48018563e+00 9.56360772e-02 6.12129048e-02 -4.54802006e-01
9.89224136e-01 1.96814641e-01 4.51144785e-01 7.45546758e-01
-9.40949097e-02 5.46008885e-01 8.28608811e-01 -4.29897785e-01
-7.73710907e-01 -8.56811941e-01 1.79091752e-01 9.50981557e-01
-3.91181201e-01 -1.23361158e+00 -1.22492087e+00 -8.58682513e-01
-1.02641605e-01 7.32962847e-01 -2.73624629e-01 -3.11026394e-01
-5.21038413e-01 -1.08797267e-01 7.91371882e-01 -5.65478876e-02
3.63759995e-01 -6.89702511e-01 -5.51672578e-01 2.19400204e-03
1.99084505e-01 -6.61159933e-01 -1.97115660e-01 -1.09238051e-01
-9.13947642e-01 -1.18434203e+00 2.73453910e-02 -9.87930179e-01
1.01063776e+00 2.70476907e-01 1.07931900e+00 8.59505177e-01
-5.39013028e-01 6.21651173e-01 -4.85348910e-01 -6.26646057e-02
-9.10247505e-01 2.91087985e-01 -5.65669797e-02 -2.28115126e-01
3.26103657e-01 -8.89071345e-01 1.36099130e-01 4.69610170e-02
-1.09068668e+00 -4.00214612e-01 -5.99884428e-02 4.09075111e-01
4.04464185e-01 6.74178481e-01 -7.62840435e-02 -1.42868757e+00
5.46869814e-01 -4.44692701e-01 -1.19886482e+00 5.55388153e-01
-5.79198480e-01 2.56010920e-01 1.04234540e+00 -3.16047430e-01
-8.60317230e-01 2.09531829e-01 -4.03128378e-02 -7.25971103e-01
-5.73622346e-01 6.80428922e-01 -2.37251282e-01 -5.62804103e-01
8.18796754e-01 1.37797117e-01 -2.78194875e-01 -6.26138031e-01
9.39305723e-02 3.59223932e-01 4.34773922e-01 -1.16896057e+00
9.63341951e-01 -2.92186499e-01 1.08343638e-01 -6.84096873e-01
-1.30150735e-01 -6.80544317e-01 -8.24715614e-01 -2.62433231e-01
6.39527619e-01 -4.67171788e-01 -1.01969516e+00 2.79950351e-01
-1.14049828e+00 -9.44720432e-02 -2.09081754e-01 2.49858558e-01
-1.01770110e-01 8.24200749e-01 -4.25955147e-01 -4.34672564e-01
2.32845787e-02 -1.76654363e+00 3.49627614e-01 3.83571833e-01
-3.69673669e-01 -8.89293849e-01 4.78271037e-01 5.75372517e-01
1.62766293e-01 4.07137752e-01 1.39416659e+00 -9.47166860e-01
-7.75913715e-01 -3.58009666e-01 4.06513304e-01 2.03935727e-01
-1.84950456e-01 8.60551894e-01 -7.50315666e-01 -9.69133079e-02
-1.50582999e-01 3.20087552e-01 3.58206183e-01 -3.01352888e-01
1.47804189e+00 -2.37159580e-01 -2.36795187e-01 6.62782013e-01
1.74034417e+00 -3.20366323e-02 7.05904365e-01 4.68220979e-01
9.66333449e-01 6.69213593e-01 3.24652314e-01 1.25492766e-01
2.14730144e-01 6.29153013e-01 3.52954268e-01 5.17016768e-01
8.58491510e-02 -2.97573775e-01 4.94650900e-01 1.04607451e+00
2.58140355e-01 2.71467209e-01 -1.22390020e+00 6.74077332e-01
-1.74699676e+00 -9.51209843e-01 -7.67219722e-01 2.43310285e+00
6.38168335e-01 -1.64808735e-01 2.23446310e-01 3.94267976e-01
7.30472386e-01 -3.24588597e-01 3.55014354e-01 -1.30625856e+00
4.60582942e-01 6.81312680e-01 2.31506541e-01 3.32375377e-01
-7.66906083e-01 7.74364948e-01 5.08668470e+00 7.99000978e-01
-8.72717023e-01 3.93039584e-02 -5.04117906e-02 3.62816781e-01
-5.24349689e-01 5.04792929e-01 -7.68631101e-01 6.17992282e-01
1.19961274e+00 -2.25877568e-01 4.97736514e-01 1.29141605e+00
-2.46678427e-01 -2.38210097e-01 -1.20894444e+00 5.86475074e-01
8.94006900e-03 -1.46725440e+00 -2.21970394e-01 -2.39243377e-02
1.05267358e+00 -5.09841442e-01 -1.33470088e-01 6.10866070e-01
5.68976104e-01 -9.57312644e-01 6.27860606e-01 3.16239148e-01
-3.09044924e-02 -1.05290151e+00 6.64355278e-01 3.65919352e-01
-1.33755612e+00 -8.89040455e-02 -3.04116458e-01 6.92139864e-02
-7.71652699e-01 6.39073074e-01 -7.70927548e-01 6.88465655e-01
8.47062826e-01 2.68150926e-01 -1.24648523e+00 1.40426517e+00
-2.82974422e-01 7.29148269e-01 -1.12519056e-01 -1.06149331e-01
-6.54033497e-02 -5.56052387e-01 6.17582560e-01 1.40674007e+00
6.23315930e-01 -9.36221890e-03 4.89191115e-01 1.46641588e+00
1.05043463e-01 4.41937089e-01 -7.38411069e-01 -5.45622921e-03
6.97263181e-01 1.42688107e+00 -8.01097274e-01 -2.62419939e-01
-2.69199759e-01 4.95691240e-01 4.26292658e-01 -6.87879026e-02
-7.04082072e-01 -8.95643890e-01 6.68145776e-01 7.03945979e-02
1.83180392e-01 -3.74526054e-01 -4.61723536e-01 -9.35689509e-01
-3.93589474e-02 -9.05726433e-01 4.38928068e-01 -5.36933124e-01
-4.85718757e-01 4.62005556e-01 -5.03640175e-02 -1.11627388e+00
3.99533734e-02 -2.39902273e-01 -1.24418557e+00 8.25981736e-01
-8.55087459e-01 -8.29390883e-01 -5.23828566e-01 6.84115469e-01
5.48454858e-02 -3.16062629e-01 1.08222747e+00 3.30422521e-01
-1.00489676e+00 9.92970586e-01 1.26848862e-01 9.49579924e-02
6.46795034e-01 -1.49467492e+00 -2.26611063e-01 1.23376763e+00
7.30957240e-02 9.20171082e-01 7.84381986e-01 -5.36844850e-01
-1.51577508e+00 -1.05225778e+00 9.06723022e-01 -3.19010615e-01
6.63098633e-01 -3.24523270e-01 -1.32516706e+00 7.10794091e-01
1.21523269e-01 4.77382205e-02 9.06664133e-01 2.70484030e-01
-5.88274658e-01 -1.53431401e-01 -1.36172235e+00 4.78039622e-01
3.59148264e-01 -5.19116044e-01 -7.45311618e-01 1.64405793e-01
7.14617014e-01 -4.93233323e-01 -1.35536575e+00 -2.72791058e-01
1.16760973e-02 -1.16501677e+00 6.45990610e-01 -3.56547624e-01
5.07726312e-01 -6.13678217e-01 1.57437902e-02 -1.30355346e+00
5.70792034e-02 -7.10107923e-01 2.17959940e-01 1.65537202e+00
1.57228690e-02 -3.70808989e-01 9.76353168e-01 1.02760506e+00
-5.93315125e-01 -2.47000962e-01 -6.14261210e-01 -7.64649928e-01
2.56986350e-01 -1.29536077e-01 8.91063571e-01 1.50164366e+00
5.87234557e-01 -1.37082338e-01 5.02897978e-01 2.44257286e-01
5.60996711e-01 5.25639296e-01 1.18819523e+00 -1.59161472e+00
-3.38581383e-01 -6.39064729e-01 -8.77638996e-01 4.83917408e-02
5.66748083e-01 -1.37340248e+00 -2.03733459e-01 -9.12054598e-01
-1.38945440e-02 -7.11942911e-01 1.56622499e-01 6.55401349e-01
4.50488590e-02 -3.85909766e-01 4.00852501e-01 -1.09293789e-01
-4.28498089e-01 7.55642483e-04 5.92980444e-01 4.27251123e-03
-4.94092405e-01 -8.38270336e-02 -2.86716014e-01 6.70176983e-01
7.76666224e-01 -8.47524226e-01 -1.94455311e-01 2.50308476e-02
9.06736851e-02 -1.12196676e-01 2.29760587e-01 -1.62931657e+00
6.45167410e-01 -2.70854205e-01 -2.32148632e-01 -3.75987023e-01
-4.26081032e-01 -1.04233706e+00 5.14609814e-01 6.18712723e-01
-3.23314220e-01 2.72421211e-01 4.39163953e-01 2.81516284e-01
-4.20774847e-01 -9.88188922e-01 7.76657224e-01 -2.09761843e-01
-6.91175878e-01 -2.49626502e-01 -8.30010056e-01 -3.01306099e-01
1.39453030e+00 -3.07351440e-01 -3.49532485e-01 2.88411021e-01
-5.83845258e-01 -1.57754451e-01 8.11272264e-01 2.18595311e-01
4.50276345e-01 -1.10242271e+00 -1.21487677e-01 3.73176426e-01
5.07574975e-02 8.65254328e-02 3.27315122e-01 5.82586348e-01
-1.01972985e+00 2.34988049e-01 -4.20428306e-01 -7.14732468e-01
-2.08072066e+00 8.05601060e-01 3.72733735e-02 -4.79556948e-01
-4.64723051e-01 5.24800301e-01 -3.85103881e-01 -1.17391264e+00
1.43722117e-01 -6.19840264e-01 -1.49656489e-01 -2.43899569e-01
3.59904647e-01 4.72973853e-01 4.72335726e-01 -2.03378737e-01
-1.70356669e-02 5.56455374e-01 2.33165659e-02 2.90244997e-01
1.20228112e+00 4.01541233e-01 -6.20803714e-01 1.15645751e-01
1.25403953e+00 5.32929063e-01 -3.31255287e-01 1.06240295e-01
3.12344193e-01 -7.46913016e-01 -1.87430587e-02 -3.87211591e-01
-1.17050362e+00 8.49477530e-01 5.03933728e-01 2.43446320e-01
8.89772356e-01 -4.91820306e-01 4.57182944e-01 4.74175662e-01
4.24584895e-01 -8.79765630e-01 2.66054198e-02 6.08200014e-01
1.54771864e-01 -7.86705136e-01 1.24898076e-01 -7.24689960e-01
-5.69942370e-02 1.45697784e+00 8.88482392e-01 -1.37064040e-01
5.63891947e-01 2.20434338e-01 -4.71487641e-01 -8.27026591e-02
-2.96922088e-01 3.17873478e-01 1.79919168e-01 4.04273748e-01
4.76456463e-01 2.26909906e-01 -7.18477964e-02 9.08015311e-01
-5.18694699e-01 -1.71412587e-01 1.21269596e+00 1.10223210e+00
-6.12788796e-01 -1.64828134e+00 -9.05715346e-01 1.08309552e-01
-2.35596552e-01 1.26089901e-01 -4.56388175e-01 8.29340935e-01
2.63287246e-01 8.63483608e-01 -1.37230158e-01 -7.42443085e-01
3.95929337e-01 3.48553807e-01 4.96546090e-01 -1.00218689e+00
-1.55707157e+00 -7.66242802e-01 -3.21167290e-01 -5.56183517e-01
-9.63170379e-02 -7.06457198e-01 -1.40961909e+00 -1.02547491e+00
-2.74014682e-01 4.84237343e-01 8.13421667e-01 7.15060771e-01
5.79741478e-01 6.20710313e-01 5.52054226e-01 -2.55509466e-01
-3.10658276e-01 -2.14014694e-01 -4.64100808e-01 4.82643723e-01
-1.42858952e-01 -4.20051366e-01 -1.63464740e-01 1.14003755e-02] | [7.861756324768066, 7.700655937194824] |
cf3f70c5-3f2c-4463-b27d-38a248dc246a | on-attention-modules-for-audio-visual | 1812.06071 | null | http://arxiv.org/abs/1812.06071v1 | http://arxiv.org/pdf/1812.06071v1.pdf | On Attention Modules for Audio-Visual Synchronization | With the development of media and networking technologies, multimedia
applications ranging from feature presentation in a cinema setting to video on
demand to interactive video conferencing are in great demand. Good
synchronization between audio and video modalities is a key factor towards
defining the quality of a multimedia presentation. The audio and visual signals
of a multimedia presentation are commonly managed by independent workflows -
they are often separately authored, processed, stored and even delivered to the
playback system. This opens up the possibility of temporal misalignment between
the two modalities - such a tendency is often more pronounced in the case of
produced content (such as movies).
To judge whether audio and video signals of a multimedia presentation are
synchronized, we as humans often pay close attention to discriminative
spatio-temporal blocks of the video (e.g. synchronizing the lip movement with
the utterance of words, or the sound of a bouncing ball at the moment it hits
the ground). At the same time, we ignore large portions of the video in which
no discriminative sounds exist (e.g. background music playing in a movie).
Inspired by this observation, we study leveraging attention modules for
automatically detecting audio-visual synchronization. We propose neural network
based attention modules, capable of weighting different portions
(spatio-temporal blocks) of the video based on their respective discriminative
power. Our experiments indicate that incorporating attention modules yields
state-of-the-art results for the audio-visual synchronization classification
problem. | ['Shervin Ardeshir', 'Naji Khosravan', 'Rohit Puri'] | 2018-12-14 | null | null | null | null | ['audio-visual-synchronization', 'audio-visual-synchronization'] | ['audio', 'computer-vision'] | [ 1.68535545e-01 -4.37533528e-01 4.31424826e-02 1.94859337e-02
-6.66096866e-01 -6.96317732e-01 5.85825503e-01 6.59677505e-01
-3.01067829e-01 9.84874219e-02 1.17305666e-01 8.53453670e-03
-1.15341626e-01 -3.29493642e-01 -5.72440028e-01 -7.08842158e-01
-2.99645245e-01 -1.19180650e-01 4.21213597e-01 -1.24603868e-01
3.52901936e-01 3.85531276e-01 -1.97819901e+00 5.99545658e-01
9.54345986e-02 1.19107878e+00 6.15977228e-01 9.98246014e-01
-1.66347891e-01 7.25121379e-01 -7.36315370e-01 -3.32937278e-02
-1.92402527e-01 -3.77697676e-01 -5.16444266e-01 1.52467906e-01
4.63071853e-01 -1.32512450e-01 -3.92369747e-01 9.97778475e-01
5.40334225e-01 2.59054393e-01 2.95411170e-01 -1.23463190e+00
-5.27795404e-02 5.91893554e-01 -4.27734643e-01 7.41224706e-01
7.52390325e-01 -2.37213206e-02 1.04442871e+00 -7.47843564e-01
7.17190146e-01 8.43709588e-01 2.96232373e-01 3.86165269e-02
-9.14423525e-01 -3.05887222e-01 2.95287192e-01 5.86275876e-01
-1.23644018e+00 -7.73553312e-01 9.89811420e-01 -6.42618299e-01
6.10829711e-01 5.15772581e-01 5.67447186e-01 1.14755571e+00
2.28281587e-01 5.71251631e-01 4.25590366e-01 -4.26682144e-01
1.60610035e-01 1.27402067e-01 -1.15812257e-01 7.94568136e-02
-5.06079376e-01 -2.71340281e-01 -8.41797829e-01 1.54706106e-01
4.04905826e-01 -1.06089271e-03 -4.64066178e-01 -8.85438174e-02
-1.12971556e+00 2.59926349e-01 -8.01327378e-02 7.94688523e-01
-5.18993855e-01 7.67521411e-02 5.67805886e-01 3.68378818e-01
1.45781845e-01 2.98900921e-02 -3.48380134e-02 -4.50395077e-01
-1.06864357e+00 2.16277838e-02 5.33596575e-01 6.67425573e-01
2.46717587e-01 -1.83066621e-01 -2.93733358e-01 5.47277093e-01
7.61027932e-02 2.18683407e-02 4.28622246e-01 -7.52700329e-01
4.15141016e-01 1.35865673e-01 1.63982332e-01 -1.33008599e+00
-2.32402325e-01 -5.38676642e-02 -4.98137027e-01 -3.07273469e-03
4.08608496e-01 1.49721786e-01 -3.47030222e-01 1.73864913e+00
2.84820646e-01 3.70508492e-01 -3.43892932e-01 1.08442116e+00
7.79048026e-01 8.15331697e-01 1.18873872e-01 -5.65991998e-01
1.53410757e+00 -4.62967485e-01 -9.50471103e-01 5.49756661e-02
6.88002631e-02 -1.02072358e+00 8.25210333e-01 4.66344327e-01
-1.20999765e+00 -9.94615674e-01 -8.40695977e-01 1.70519263e-01
-2.19335750e-01 -1.11727253e-01 -1.54266641e-01 1.28245220e-01
-7.78643847e-01 6.71284199e-01 -6.14018977e-01 -4.03855503e-01
-1.64942771e-01 7.34450072e-02 -5.12652695e-01 3.46922994e-01
-1.07104719e+00 3.91868830e-01 1.23201393e-01 1.69750720e-01
-9.27202880e-01 -4.70978916e-01 -6.28100574e-01 3.04542869e-01
4.02292639e-01 -9.94707737e-03 1.28897297e+00 -1.48855233e+00
-1.38631022e+00 8.34514081e-01 -8.70826319e-02 -1.92660317e-01
5.37033439e-01 -2.51721740e-01 -8.40658486e-01 5.05847216e-01
-2.75127292e-02 3.47748727e-01 1.37409890e+00 -1.07325637e+00
-8.15860510e-01 -1.15510099e-01 2.27703080e-01 1.49048001e-01
-4.22477514e-01 5.23039341e-01 -7.03473330e-01 -6.77020609e-01
-2.37663053e-02 -6.26900971e-01 3.90841961e-01 -9.60600674e-02
-2.55805731e-01 -5.86151481e-02 1.17122793e+00 -5.82229674e-01
1.41597891e+00 -2.45751023e+00 2.56122142e-01 1.53059661e-01
2.87417937e-02 1.77490100e-01 -1.30792081e-01 6.01110935e-01
-3.73281538e-01 -1.06321514e-01 2.90695339e-01 -3.33732933e-01
-2.53096551e-01 -1.66870117e-01 -3.97310436e-01 5.42174935e-01
5.95647581e-02 2.43807584e-01 -8.80472660e-01 -5.14749885e-01
2.01030642e-01 6.18115604e-01 -3.09550375e-01 5.05982339e-01
-8.30741748e-02 8.06383014e-01 -1.43401816e-01 4.27760154e-01
2.65357256e-01 -1.20160878e-01 2.27902725e-01 -3.45603973e-01
-4.50139791e-01 3.17459017e-01 -1.30110908e+00 1.67580211e+00
-4.52460080e-01 1.09706688e+00 1.85224563e-01 -9.68673646e-01
7.18121290e-01 7.97421098e-01 5.40747285e-01 -8.31300318e-01
2.93628186e-01 -1.30115241e-01 1.73856512e-01 -1.03052878e+00
7.26628423e-01 1.79044172e-01 2.54260246e-02 3.69506627e-01
6.91467524e-02 2.08529443e-01 3.68686676e-01 6.18216284e-02
9.75674272e-01 -5.68670146e-02 1.74182832e-01 1.02363691e-01
6.31537080e-01 -6.19665802e-01 2.24828437e-01 5.73729575e-01
-3.23788911e-01 8.56775165e-01 7.07772970e-01 -1.16762936e-01
-9.89849091e-01 -7.59990036e-01 1.11999929e-01 1.47923601e+00
2.15226367e-01 -6.71898246e-01 -6.30308628e-01 -1.65451512e-01
-2.73969114e-01 3.47753227e-01 -5.06339729e-01 -4.99412417e-02
-5.55901766e-01 8.74244869e-02 2.68937081e-01 2.92988479e-01
-7.49803483e-02 -1.20721543e+00 -1.00597358e+00 3.63428295e-01
-2.64702529e-01 -1.31162441e+00 -4.06878471e-01 1.32122338e-01
-4.75737393e-01 -9.98562872e-01 -6.66523039e-01 -6.44686043e-01
2.72936314e-01 3.82893234e-01 9.84319568e-01 -7.08354488e-02
-1.91065565e-01 5.75805664e-01 -4.65824902e-01 1.03777479e-02
-5.00820696e-01 -2.90876061e-01 1.76591665e-01 5.58012187e-01
-7.86271021e-02 -5.51832736e-01 -6.24278486e-01 3.55804920e-01
-1.04097974e+00 -1.11605413e-01 -7.41405413e-02 3.09482843e-01
4.19168949e-01 2.37193555e-01 2.38540724e-01 -1.90611884e-01
5.92112362e-01 -7.55693138e-01 -4.41623628e-01 1.39737666e-01
3.20902348e-01 -4.25322503e-01 7.57121921e-01 -8.78014565e-01
-6.17397606e-01 -1.99057102e-01 6.04989182e-05 -7.86554158e-01
-4.91095543e-01 4.97449428e-01 -1.13472894e-01 2.73553669e-01
3.51619363e-01 2.95716859e-02 -2.27040991e-01 -4.24428731e-01
1.82795059e-02 7.17044532e-01 7.63108075e-01 -2.86077708e-01
3.09392989e-01 4.19342846e-01 -1.99595049e-01 -1.15873373e+00
-3.25241804e-01 -6.87281370e-01 -4.36978549e-01 -8.64175379e-01
9.35207129e-01 -6.01991415e-01 -9.45768833e-01 3.82369697e-01
-1.33624494e+00 1.32455342e-02 3.00615337e-02 5.30688345e-01
-4.68716413e-01 4.82047856e-01 -4.18509305e-01 -9.01598513e-01
1.89288720e-01 -1.28307247e+00 9.86546516e-01 2.59425908e-01
-5.50672472e-01 -6.04714274e-01 1.36314943e-01 2.07659617e-01
1.17085919e-01 3.39250118e-02 6.74791455e-01 -5.54264069e-01
-2.88285226e-01 -3.53566825e-01 1.33165404e-01 1.82111427e-01
2.70740718e-01 5.04213512e-01 -1.21902335e+00 -1.01838060e-01
5.71663640e-02 1.92300454e-02 3.99388015e-01 5.07334352e-01
1.23031294e+00 -1.87801093e-01 -7.35337287e-02 1.36057556e-01
9.95854378e-01 4.32134748e-01 5.98964691e-01 5.74484169e-02
3.39032799e-01 9.28483665e-01 6.04147434e-01 6.78535938e-01
-3.55243497e-02 1.20616841e+00 5.98637938e-01 1.94312558e-01
3.23673673e-02 2.83391941e-02 6.82032049e-01 8.40999484e-01
-6.99297860e-02 -6.54902756e-01 -7.59468079e-01 5.51397026e-01
-1.70643592e+00 -1.23801684e+00 -2.35039487e-01 2.53961420e+00
4.92489189e-01 2.26380020e-01 2.93459833e-01 6.12471521e-01
1.04774117e+00 3.57128590e-01 -1.06665410e-01 -4.55920279e-01
3.86741087e-02 -2.44982988e-02 -2.61063099e-01 2.83359587e-01
-1.20132959e+00 3.52116555e-01 5.36888313e+00 7.89178014e-01
-1.53649282e+00 6.35819659e-02 3.75473171e-01 -4.38110769e-01
-4.32056822e-02 -1.84106603e-01 -3.73850226e-01 8.75514507e-01
1.05628109e+00 1.27683297e-01 2.17119500e-01 4.22339648e-01
6.13322377e-01 -4.10592943e-01 -1.38439643e+00 1.22421527e+00
9.52404886e-02 -1.08174777e+00 -2.40338057e-01 -2.53217638e-01
2.05343261e-01 -3.74755263e-01 2.28862360e-01 -1.91407949e-01
-6.79578424e-01 -9.08268511e-01 1.33930135e+00 4.22553092e-01
6.64952219e-01 -5.49830198e-01 5.33547461e-01 9.67940539e-02
-1.48343301e+00 -3.09274886e-02 2.17433050e-01 9.25744846e-02
4.37995791e-01 2.80503660e-01 -4.78720754e-01 3.82773429e-01
9.09115732e-01 4.77962047e-01 -1.80695981e-01 1.04793572e+00
-8.36327299e-02 4.08031166e-01 -2.68147618e-01 2.23799929e-01
3.12558077e-02 4.68408614e-02 9.68176305e-01 1.31121767e+00
5.09194314e-01 -5.71321025e-02 -1.35456055e-01 5.65119088e-01
8.33460689e-02 1.17576420e-01 -7.19281197e-01 -1.58854127e-01
3.40837330e-01 1.24216402e+00 -8.53601575e-01 -1.52639866e-01
-5.06448090e-01 9.47621644e-01 -1.54907957e-01 3.11049968e-01
-1.00566041e+00 -2.57939190e-01 6.61797643e-01 1.93939060e-01
5.55727482e-01 -1.54840156e-01 4.01499927e-01 -7.60822892e-01
2.35018700e-01 -8.30420196e-01 3.35023612e-01 -1.03360701e+00
-8.36365044e-01 6.62470400e-01 -2.19982713e-01 -1.58084130e+00
-2.98626602e-01 -2.07969517e-01 -7.49096155e-01 5.81693530e-01
-1.05540907e+00 -5.38621426e-01 -3.75773460e-01 6.22976124e-01
6.44606113e-01 7.30727240e-02 5.05257487e-01 7.47746348e-01
-4.95527864e-01 3.12019050e-01 -1.04047082e-01 -2.19842959e-02
8.97037864e-01 -7.95813560e-01 -9.01900902e-02 7.98802316e-01
5.25637567e-01 3.37490678e-01 9.31737065e-01 -3.87982935e-01
-1.30997133e+00 -5.00335157e-01 9.82517719e-01 -1.54204259e-03
8.18842053e-01 -3.46358210e-01 -9.63159740e-01 1.18092537e-01
3.42441589e-01 -9.81795564e-02 6.90940857e-01 -1.11249536e-01
-3.03308785e-01 -3.50684404e-01 -5.79422891e-01 4.83017385e-01
4.11440551e-01 -1.00973141e+00 -5.44540524e-01 2.43859738e-02
3.40206146e-01 -1.84880659e-01 -5.10537922e-01 1.08344525e-01
6.89011216e-01 -1.31873536e+00 7.81304896e-01 -5.87343991e-01
5.12982070e-01 -3.71874958e-01 -1.67763904e-01 -8.76652122e-01
-8.53589624e-02 -8.46014678e-01 -1.56450987e-01 1.40612328e+00
-1.62090778e-01 1.70162901e-01 2.07473904e-01 1.03861906e-01
9.61618051e-02 -8.65224525e-02 -1.10737610e+00 -5.28928816e-01
-8.00422370e-01 -7.26541877e-01 7.01085031e-02 8.86904657e-01
2.91458696e-01 1.66752085e-01 -3.89354289e-01 2.87323087e-01
1.77446306e-02 6.15951233e-02 5.56544185e-01 -1.11834872e+00
-3.67993683e-01 -6.47656560e-01 -7.10740983e-01 -7.80748010e-01
-4.17317962e-04 -4.29665297e-01 1.67349145e-01 -8.98392498e-01
-5.83883226e-02 -3.76646072e-02 -4.88551259e-01 -2.91691795e-02
1.18038319e-01 2.57553518e-01 3.77408326e-01 2.68805742e-01
-7.29358077e-01 1.53857961e-01 6.95137203e-01 -9.82516930e-02
-2.52031922e-01 1.86218899e-02 5.39795607e-02 6.58098817e-01
3.75291288e-01 -3.88354212e-01 -1.66830242e-01 -2.07575366e-01
3.15451056e-01 6.08777165e-01 4.92625058e-01 -1.19103515e+00
4.23388541e-01 7.71081224e-02 -9.67415888e-03 -4.61299866e-01
4.32907194e-01 -1.10335982e+00 3.89351338e-01 1.82208508e-01
-5.18499076e-01 2.06818074e-01 2.80678511e-01 5.59189796e-01
-6.14992440e-01 -2.35495239e-01 7.04692423e-01 1.43957898e-01
-6.68759167e-01 2.88890935e-02 -8.39433372e-01 -1.46653131e-01
1.00860393e+00 -2.80963838e-01 -7.11962283e-02 -7.40755022e-01
-8.94439161e-01 -2.80794203e-01 1.19871475e-01 6.98145866e-01
4.45163965e-01 -1.10005760e+00 -3.51583570e-01 1.79584354e-01
7.06228986e-02 -4.03315842e-01 6.50524437e-01 1.06287789e+00
-2.96675473e-01 2.93785483e-01 -2.62590677e-01 -7.59559393e-01
-1.62341261e+00 5.60298502e-01 1.55359581e-01 1.48359463e-01
-4.49468821e-01 7.44028866e-01 2.40226433e-01 6.92168593e-01
6.99410796e-01 -4.62050617e-01 -5.60625434e-01 7.78355956e-01
7.59329081e-01 3.33292067e-01 2.71191180e-01 -8.04986656e-01
-4.10540283e-01 4.83507603e-01 1.56042606e-01 -2.35644937e-01
1.16282570e+00 -3.30956429e-01 3.08859367e-02 1.02322948e+00
1.19432044e+00 2.98035979e-01 -1.15137720e+00 2.68804580e-02
-1.14321150e-01 -3.60644341e-01 7.56545272e-03 -2.06571847e-01
-1.03043079e+00 1.12077415e+00 6.27209246e-01 8.67282093e-01
1.19287479e+00 9.36234817e-02 5.11914313e-01 -1.00499019e-01
4.42878082e-02 -1.25426280e+00 3.70950222e-01 3.21117520e-01
8.82325470e-01 -9.76515651e-01 -1.98453724e-01 -1.41804833e-02
-6.13578320e-01 1.34895468e+00 1.40889600e-01 1.22731000e-01
4.92745250e-01 3.51552457e-01 1.23244468e-02 -7.26230890e-02
-8.08817804e-01 -1.52294561e-01 4.41828042e-01 3.19566131e-01
4.58817571e-01 -2.28422612e-01 9.04727206e-02 5.39268851e-01
1.19725525e-01 -2.57825673e-01 3.85497630e-01 8.17749619e-01
-3.09959203e-01 -7.51793325e-01 -5.92687786e-01 -1.92678720e-01
-6.96936488e-01 -2.66392995e-02 -2.57875204e-01 4.38461214e-01
1.15527362e-01 1.03244781e+00 5.79194188e-01 -4.24287379e-01
2.48851091e-01 1.00412734e-01 3.79120678e-01 -2.37119421e-01
-9.19458568e-01 5.26751339e-01 -1.27573431e-01 -6.44423783e-01
-6.55147016e-01 -7.21445680e-01 -8.98265421e-01 -4.02522162e-02
-6.71717897e-02 8.05004463e-02 6.89271867e-01 8.57389390e-01
2.49369189e-01 6.94943368e-01 6.55031025e-01 -1.12403882e+00
2.49334238e-02 -7.35545337e-01 -4.61385429e-01 6.85265958e-01
6.12406552e-01 -5.01169801e-01 -4.15995985e-01 4.38940167e-01] | [14.59400463104248, 5.027501106262207] |
4704f449-830f-4a2a-97cd-26478ef39827 | event-causality-identification-via-derivative | null | null | https://aclanthology.org/2022.coling-1.200 | https://aclanthology.org/2022.coling-1.200.pdf | Event Causality Identification via Derivative Prompt Joint Learning | This paper studies event causality identification, which aims at predicting the causality relation for a pair of events in a sentence. Regarding event causality identification as a supervised classification task, most existing methods suffer from the problem of insufficient annotated data. In this paper, we propose a new derivative prompt joint learning model for event causality identification, which leverages potential causal knowledge in the pre-trained language model to tackle the data scarcity problem. Specifically, rather than external data or knowledge augmentation, we derive two relevant prompt tasks from event causality identification to enhance the model’s ability to identify explicit and implicit causality. We evaluate our model on two benchmark datasets and the results show that our model has great advantages over previous methods. | ['Guilin Qi', 'Tongtong Wu', 'Heng Zhou', 'Shirong Shen'] | null | null | null | null | coling-2022-10 | ['event-causality-identification'] | ['natural-language-processing'] | [ 2.14523330e-01 1.08221039e-01 -5.83728433e-01 -3.86537552e-01
-7.49889672e-01 -4.22194093e-01 8.62934530e-01 5.96593142e-01
-2.81037331e-01 1.17021823e+00 7.81893909e-01 -3.97639722e-01
-2.36527190e-01 -6.32439554e-01 -6.20951116e-01 -3.03325772e-01
-3.18194449e-01 5.55090234e-02 2.94151604e-01 2.98192829e-01
1.27544522e-01 8.87019113e-02 -7.09235847e-01 3.10440958e-01
8.87597501e-01 7.31136322e-01 -1.05381191e-01 3.45341265e-01
5.32223321e-02 1.88430941e+00 -4.78471041e-01 -3.04446816e-01
-3.69641989e-01 -5.73360920e-01 -1.19137967e+00 -6.36500239e-01
-2.38246188e-01 -4.59117591e-01 -4.25931931e-01 6.02126122e-01
3.36416483e-01 -7.23715201e-02 7.71558285e-01 -1.48437023e+00
-5.87821245e-01 1.04773617e+00 -4.58279967e-01 7.93991327e-01
6.59289122e-01 -2.05280676e-01 1.38839459e+00 -9.30158913e-01
3.18405271e-01 1.34421682e+00 5.90296507e-01 2.78265804e-01
-9.50602770e-01 -8.46142530e-01 5.06757677e-01 7.39799678e-01
-1.14755571e+00 -2.18574330e-01 1.03742492e+00 -5.28650045e-01
8.96984339e-01 6.43893406e-02 1.02336489e-01 1.41824627e+00
1.05016708e-01 1.01063609e+00 7.50674963e-01 -3.42261106e-01
2.84043998e-01 -4.92716432e-01 3.42919856e-01 3.15882713e-01
7.69797526e-03 1.90710723e-01 -8.04664552e-01 -5.17764688e-01
5.75398743e-01 -2.39956096e-01 -2.70779669e-01 3.02970737e-01
-1.36123359e+00 7.42122233e-01 2.44573012e-01 3.60835530e-02
-6.56128287e-01 2.53906250e-01 5.40244043e-01 -3.77020538e-02
7.08463252e-01 2.30676502e-01 -7.42669463e-01 -9.99061391e-02
-4.12423790e-01 3.45370322e-01 7.67572880e-01 4.78144765e-01
3.15391980e-02 -2.25000948e-01 -6.79517925e-01 5.90348125e-01
3.59846115e-01 -6.58745179e-03 2.00116098e-01 -3.06004643e-01
5.21350086e-01 8.10310066e-01 2.76729137e-01 -1.02590764e+00
-6.06630206e-01 -1.54521242e-01 -6.48944497e-01 -5.01054347e-01
4.21043098e-01 -5.03110588e-01 -3.93123329e-01 1.97844338e+00
4.13320869e-01 1.08963990e+00 9.80262011e-02 9.30403352e-01
9.81560647e-01 7.64933586e-01 7.33746588e-01 -7.49750137e-01
1.14923358e+00 -7.00054646e-01 -1.18683791e+00 -1.61480248e-01
6.15307510e-01 -5.88869274e-01 7.73180366e-01 2.93267667e-01
-7.81266689e-01 -2.45926172e-01 -7.85721242e-01 1.17665477e-01
-9.05187149e-03 4.84187230e-02 9.11583841e-01 -2.52430458e-02
-3.66923183e-01 2.79343903e-01 -6.92333639e-01 -1.89840689e-01
3.30481470e-01 1.44343346e-01 4.19350481e-03 1.53148279e-01
-2.02433443e+00 7.09203541e-01 8.84083092e-01 1.47069932e-03
-8.99771810e-01 -1.02029395e+00 -6.85013771e-01 -2.18025707e-02
7.14407384e-01 -6.13472581e-01 1.38731980e+00 -5.52667558e-01
-1.07685387e+00 3.37723166e-01 -4.29838181e-01 -5.74326992e-01
3.29812706e-01 -6.63422346e-01 -6.90185249e-01 -6.44171908e-02
2.80407906e-01 7.74038360e-02 6.34600997e-01 -1.13016152e+00
-9.26079929e-01 1.61276497e-02 1.29953176e-01 4.74479701e-03
-4.87903237e-01 5.89881182e-01 -2.42831826e-01 -1.01666880e+00
-3.55263203e-01 -3.25022697e-01 -2.59297609e-01 -5.01853347e-01
-6.04135156e-01 -9.88224864e-01 7.79629350e-01 -5.88042676e-01
1.73875570e+00 -1.87808526e+00 -6.75338209e-02 -2.27208138e-01
2.42114395e-01 -9.33461636e-03 1.03249989e-01 4.89465475e-01
-6.53994083e-01 1.27720088e-01 -1.49549112e-01 -1.08422175e-01
-1.57981068e-01 2.37537622e-01 -9.30616856e-01 2.08984107e-01
6.92060411e-01 9.68170941e-01 -1.50280190e+00 -8.16397905e-01
-3.21970522e-01 2.16971844e-01 -2.53992528e-01 6.48293316e-01
-2.69406259e-01 6.03593647e-01 -6.91599607e-01 3.21310908e-01
2.53441393e-01 -4.50869113e-01 3.26743096e-01 -1.81138620e-01
-8.22303295e-02 7.73531079e-01 -1.18731737e+00 1.27870488e+00
-1.17106393e-01 -1.38105992e-02 -6.70888841e-01 -1.11051619e+00
5.94890594e-01 9.49234545e-01 6.66817427e-01 -6.97034836e-01
-9.19938534e-02 5.90505414e-02 -2.06992000e-01 -7.72385716e-01
3.00951879e-02 -1.05616100e-01 -1.49289742e-01 4.71635818e-01
-1.88538238e-01 5.91096520e-01 2.31160566e-01 3.28943700e-01
1.33350456e+00 1.07606888e-01 5.62692285e-01 3.25523615e-02
5.72494566e-01 -1.78423356e-02 1.11564696e+00 6.59114778e-01
-1.99319661e-01 2.46266082e-01 9.24434602e-01 -5.08366704e-01
-3.84594858e-01 -1.24622190e+00 -1.09499492e-01 7.33107448e-01
-2.79933512e-02 -6.23213887e-01 -1.38992935e-01 -1.33700383e+00
-8.87937769e-02 8.81147981e-01 -6.83349788e-01 -2.44618163e-01
-8.15706372e-01 -1.34813809e+00 6.12244070e-01 1.00004590e+00
3.45927447e-01 -9.70525980e-01 -4.02020544e-01 2.24930584e-01
-5.87856889e-01 -1.26584506e+00 -3.27489763e-01 1.26449481e-01
-6.67222440e-01 -1.33380413e+00 -1.15890756e-01 -6.28655851e-01
3.66389424e-01 -1.65038228e-01 1.15082657e+00 -2.68626571e-01
5.66331483e-02 1.58301935e-01 -4.15675074e-01 -5.75812578e-01
-1.62193269e-01 -2.99016476e-01 2.18596160e-01 1.67464331e-01
4.57848310e-01 -5.43058932e-01 -4.83418941e-01 -5.44808172e-02
-7.33614087e-01 -1.26567390e-02 4.64831203e-01 9.76069152e-01
4.68672395e-01 3.31863731e-01 1.36698937e+00 -9.27296519e-01
8.10477078e-01 -1.04075849e+00 -2.70174026e-01 1.68679044e-01
-7.91805267e-01 9.84664634e-03 6.30541146e-01 -5.31785727e-01
-1.45866513e+00 1.78525224e-02 2.42849037e-01 -7.77311921e-02
-1.62530780e-01 1.17961061e+00 -3.10637176e-01 7.64533937e-01
3.79013807e-01 -1.80809230e-01 -7.38629162e-01 -3.47699583e-01
3.29770774e-01 2.01607212e-01 6.00780189e-01 -6.39814317e-01
7.15122342e-01 3.22701186e-01 -8.29294547e-02 -1.09832212e-01
-1.37022972e+00 -4.44196969e-01 -6.05731606e-01 -3.51143368e-02
9.13278162e-01 -9.72667098e-01 -7.90491760e-01 1.50542855e-01
-1.69583130e+00 -2.46037722e-01 -4.29107919e-02 8.55516672e-01
-1.46316200e-01 2.86621422e-01 -6.80072963e-01 -1.04387355e+00
-1.79374859e-01 -4.04646486e-01 7.85584807e-01 3.95138636e-02
-6.56769454e-01 -1.30942297e+00 4.65774387e-01 -9.48702730e-03
-2.63845503e-01 3.47484082e-01 1.15074563e+00 -8.60398412e-01
-3.05147469e-01 -1.45885095e-01 -4.72366780e-01 -1.54905394e-01
4.54720467e-01 -1.87071890e-01 -8.42617214e-01 2.75580823e-01
-6.79868609e-02 -3.33348870e-01 8.94866943e-01 1.44273773e-01
1.17341435e+00 -6.26765668e-01 -6.15103185e-01 3.02829836e-02
9.02415037e-01 3.54259104e-01 3.04054528e-01 1.66726246e-01
8.97266150e-01 6.56638265e-01 7.16175914e-01 7.05400169e-01
7.91499197e-01 5.23485720e-01 3.58744472e-01 -1.83218583e-01
8.57254714e-02 -5.88107288e-01 2.35944539e-01 5.20516992e-01
1.76997352e-02 -3.69277537e-01 -1.02170336e+00 8.76570463e-01
-2.26561737e+00 -9.71742451e-01 -6.51579618e-01 1.82571602e+00
1.48235321e+00 1.88046455e-01 1.52876377e-01 2.72396863e-01
5.16765594e-01 9.55344141e-02 -3.64487082e-01 -5.14189377e-02
-1.30259618e-01 -2.76966672e-02 8.86606425e-02 4.21040446e-01
-1.52965379e+00 7.26260304e-01 6.56844425e+00 5.87891936e-01
-7.48650253e-01 2.03365654e-01 5.77599227e-01 -2.72843204e-02
-2.27187976e-01 2.16359258e-01 -5.66780984e-01 5.78186452e-01
8.82670283e-01 -3.68745536e-01 -2.99180388e-01 2.94228226e-01
6.67464018e-01 9.72977281e-03 -1.50374353e+00 7.39185929e-01
-2.80112475e-01 -1.22974801e+00 1.47148117e-01 -3.97222579e-01
5.98719001e-01 -4.89243358e-01 -2.98246711e-01 2.49323100e-01
5.30426085e-01 -1.08539021e+00 5.51920116e-01 6.80863738e-01
2.15049833e-01 -7.06068873e-01 6.76641822e-01 2.42475688e-01
-1.16177416e+00 -2.50365883e-01 3.09523910e-01 -5.91751754e-01
5.14844298e-01 1.04559851e+00 -9.80195165e-01 1.00409281e+00
5.65382957e-01 1.12028706e+00 -3.92871171e-01 8.22123528e-01
-7.73195207e-01 1.39857924e+00 -1.20849580e-01 1.43770590e-01
-1.85783193e-01 4.58514154e-01 4.69760388e-01 1.30157733e+00
-1.52900234e-01 4.50735569e-01 4.03584272e-01 1.06865239e+00
-1.29845008e-01 5.48782274e-02 -4.08070654e-01 -4.77065630e-02
5.89880228e-01 8.60650599e-01 -3.78703058e-01 -3.04814607e-01
-5.81943691e-01 7.98394799e-01 4.06573474e-01 4.33604300e-01
-1.15849853e+00 -3.55949774e-02 2.26492941e-01 -3.46845537e-01
-2.46546164e-01 -1.03696994e-01 -6.20013297e-01 -1.09319055e+00
3.28694284e-02 -3.38229150e-01 1.09519553e+00 -5.73249996e-01
-1.56948221e+00 8.59304592e-02 3.06910634e-01 -8.69287252e-01
-3.48636210e-01 -2.09681466e-01 -1.09091604e+00 8.33253980e-01
-1.82294130e+00 -1.28180301e+00 8.42511505e-02 5.05069137e-01
3.39678526e-01 3.10977191e-01 6.99364007e-01 5.78745723e-01
-9.96291578e-01 5.03695548e-01 -6.36703908e-01 5.02397835e-01
1.00031400e+00 -1.44620252e+00 2.27528319e-01 1.16988981e+00
-1.62019338e-02 5.87611854e-01 6.36403501e-01 -1.18447793e+00
-1.20045650e+00 -1.29171228e+00 1.60469699e+00 -7.43335366e-01
1.09811962e+00 -1.52434841e-01 -9.37913716e-01 8.24509561e-01
1.51235327e-01 1.37535352e-02 7.38481224e-01 5.39582193e-01
-4.03441221e-01 3.31854485e-02 -5.20695269e-01 5.34242570e-01
1.03256547e+00 -3.47835392e-01 -1.04130256e+00 3.39383125e-01
9.44538653e-01 -1.62119135e-01 -7.16297805e-01 7.41884172e-01
1.99331746e-01 -5.73927686e-02 1.02990949e+00 -1.09105623e+00
9.05646861e-01 -3.80712688e-01 1.49881408e-01 -1.08546901e+00
-3.75122428e-01 -6.06096387e-01 -6.85479462e-01 1.74073911e+00
5.49819350e-01 -4.54595268e-01 2.60806769e-01 6.29280984e-01
1.34705096e-01 -4.44195807e-01 -8.52617860e-01 -7.03118682e-01
-2.50336099e-02 -5.65059662e-01 4.68618989e-01 1.53179216e+00
4.68878567e-01 9.76968348e-01 -4.90206987e-01 7.11633444e-01
5.96412003e-01 1.76547274e-01 3.03393871e-01 -1.28313231e+00
-2.55925655e-01 -2.48405531e-01 1.81570396e-01 -6.68975234e-01
5.01824856e-01 -7.92231798e-01 -7.35123456e-02 -1.61542594e+00
4.21489716e-01 -3.37197125e-01 -7.49168515e-01 8.40150535e-01
-9.69758570e-01 -1.79119393e-01 -3.44701231e-01 7.88360089e-02
-7.81934977e-01 7.20558465e-01 9.21934068e-01 -2.62164976e-02
-2.61196345e-01 -1.31047577e-01 -5.41661322e-01 8.80460322e-01
5.62021255e-01 -6.37902915e-01 -6.75255775e-01 -4.51541156e-01
5.40257752e-01 2.63489902e-01 9.15207684e-01 -4.04035360e-01
3.71194750e-01 -5.18143475e-01 2.75132477e-01 -6.29425287e-01
-2.91568071e-01 -5.19464374e-01 -2.22925201e-01 3.39619279e-01
-6.53211474e-01 2.69721523e-02 1.45534590e-01 8.49700511e-01
-4.49612588e-01 1.63037837e-01 1.05647944e-01 2.16807947e-01
-9.17291224e-01 2.00500116e-01 -2.55873322e-01 1.81034490e-01
9.35336947e-01 6.09375298e-01 -3.48212719e-01 -1.64147377e-01
-5.31815529e-01 6.18211865e-01 -4.75707024e-01 6.72442853e-01
6.91922188e-01 -1.54246640e+00 -1.04981112e+00 -3.13329995e-01
2.15781003e-01 6.23380989e-02 1.11018931e-02 1.10948694e+00
3.83406103e-01 5.58875859e-01 5.43752909e-01 -2.01968089e-01
-9.07639086e-01 9.14128602e-01 3.32261622e-02 -5.92047751e-01
-5.32305062e-01 8.78933132e-01 5.12240887e-01 -1.47042135e-02
3.78858745e-01 -3.18537712e-01 -7.09847927e-01 1.76577732e-01
6.39198899e-01 2.20301241e-01 -2.27163821e-01 -9.58418660e-03
-6.38485670e-01 -6.17930144e-02 -1.29548073e-01 -6.00614361e-02
1.18532765e+00 1.89964343e-02 -3.94102931e-01 7.98956931e-01
7.92616308e-01 -8.02963898e-02 -1.11361051e+00 -4.61883217e-01
6.78775966e-01 -1.19075708e-01 9.29404050e-02 -1.21837783e+00
-5.92747390e-01 5.41948378e-01 1.81458622e-01 1.34176478e-01
1.24760044e+00 2.32879907e-01 7.74016023e-01 -2.80156434e-02
-4.98477817e-02 -8.51259351e-01 2.42512107e-01 7.27826774e-01
9.88146186e-01 -1.39221489e+00 -1.75109446e-01 -5.00056446e-01
-4.96735811e-01 9.53950703e-01 6.47636712e-01 2.78028380e-02
7.34829068e-01 3.82862270e-01 -1.39758676e-01 -1.89199403e-01
-1.18857503e+00 -2.16111884e-01 4.04083401e-01 2.21096545e-01
7.82895446e-01 -7.61083188e-03 -7.72688866e-01 1.34762383e+00
2.54594356e-01 3.86987090e-01 1.91897243e-01 7.83544421e-01
1.85465083e-01 -1.05982065e+00 -2.42023602e-01 1.48761228e-01
-7.48034239e-01 -3.54120344e-01 -5.69552183e-01 4.29973751e-01
6.83687329e-02 1.29663825e+00 -1.06146187e-01 -2.47664705e-01
4.00952339e-01 1.89387038e-01 1.69109926e-01 -4.57450479e-01
-2.94359505e-01 2.29406618e-02 4.55436319e-01 -5.26451468e-01
-7.48732507e-01 -7.73201585e-01 -1.64703345e+00 2.14514792e-01
-6.64149076e-02 2.16032669e-01 -1.59977451e-01 1.43894422e+00
1.27702668e-01 1.06475353e+00 7.12519109e-01 5.17693534e-02
-3.40754449e-01 -8.35090160e-01 -2.24569708e-01 4.49454874e-01
5.29929698e-01 -8.22696269e-01 -1.21472031e-01 4.91526932e-01] | [9.073692321777344, 9.11007308959961] |
d7d2d099-0fd0-4e25-be2d-17411ce14044 | deep-neural-network-for-blind-visual-quality | 2206.04363 | null | https://arxiv.org/abs/2206.04363v1 | https://arxiv.org/pdf/2206.04363v1.pdf | Deep Neural Network for Blind Visual Quality Assessment of 4K Content | The 4K content can deliver a more immersive visual experience to consumers due to the huge improvement of spatial resolution. However, existing blind image quality assessment (BIQA) methods are not suitable for the original and upscaled 4K contents due to the expanded resolution and specific distortions. In this paper, we propose a deep learning-based BIQA model for 4K content, which on one hand can recognize true and pseudo 4K content and on the other hand can evaluate their perceptual visual quality. Considering the characteristic that high spatial resolution can represent more abundant high-frequency information, we first propose a Grey-level Co-occurrence Matrix (GLCM) based texture complexity measure to select three representative image patches from a 4K image, which can reduce the computational complexity and is proven to be very effective for the overall quality prediction through experiments. Then we extract different kinds of visual features from the intermediate layers of the convolutional neural network (CNN) and integrate them into the quality-aware feature representation. Finally, two multilayer perception (MLP) networks are utilized to map the quality-aware features into the class probability and the quality score for each patch respectively. The overall quality index is obtained through the average pooling of patch results. The proposed model is trained through the multi-task learning manner and we introduce an uncertainty principle to balance the losses of the classification and regression tasks. The experimental results show that the proposed model outperforms all compared BIQA metrics on four 4K content quality assessment databases. | ['Guangtao Zhai', 'Tao Wang', 'ZiCheng Zhang', 'Qiyuan Wang', 'Jun He', 'Quan Zhou', 'Wenhan Zhu', 'Xiongkuo Min', 'Wei Sun', 'Wei Lu'] | 2022-06-09 | null | null | null | null | ['blind-image-quality-assessment'] | ['computer-vision'] | [-1.67882010e-01 -5.96043348e-01 1.77953586e-01 -2.93056101e-01
-1.07367921e+00 -7.63209388e-02 1.84708774e-01 9.86896008e-02
-2.51263976e-01 4.87504989e-01 3.72488379e-01 8.68192539e-02
-6.64517164e-01 -9.12220001e-01 -6.07629120e-01 -9.22864139e-01
-1.95745558e-01 -4.44041878e-01 2.00058937e-01 -4.29000370e-02
4.19568539e-01 2.15103611e-01 -1.89118254e+00 8.03835034e-01
1.27732146e+00 1.74685836e+00 3.08644414e-01 6.37918472e-01
1.17600381e-01 8.11541259e-01 -6.23685241e-01 -2.68146425e-01
2.55367994e-01 -3.78132969e-01 -5.79819977e-01 -6.13440685e-02
4.33933437e-01 -5.96159518e-01 -2.88330734e-01 1.31848204e+00
8.72629285e-01 4.74755317e-02 6.21976376e-01 -9.72410262e-01
-9.05515730e-01 5.66668026e-02 -4.11840677e-01 2.72788912e-01
2.37814218e-01 1.64188534e-01 7.57520974e-01 -8.23807776e-01
1.15931004e-01 1.13906634e+00 5.25469303e-01 2.03339569e-02
-8.29802155e-01 -4.65346009e-01 -9.16262195e-02 9.47308004e-01
-1.24761498e+00 -4.06905830e-01 9.92788613e-01 -3.11120301e-01
7.50716090e-01 3.55748236e-01 9.04314995e-01 6.18516564e-01
4.59972203e-01 7.94808507e-01 1.82922137e+00 -3.09077591e-01
1.70307547e-01 1.94065690e-01 -1.83502913e-01 7.57173359e-01
-4.55199815e-02 7.75431842e-02 -5.07346570e-01 1.31953344e-01
8.51256192e-01 -8.56736526e-02 -6.16005123e-01 -1.34105235e-01
-1.10784423e+00 5.67570329e-01 9.27526712e-01 2.40974098e-01
-4.10374343e-01 2.75775641e-02 3.10298681e-01 3.55358481e-01
1.83662906e-01 2.00734317e-01 -4.32483703e-01 6.31443635e-02
-7.04963923e-01 1.81848392e-01 1.65055767e-01 3.74459386e-01
8.17969024e-01 -1.43125355e-01 -4.71936226e-01 1.06149614e+00
4.81444091e-01 6.48484707e-01 7.17906237e-01 -9.64716852e-01
2.77824819e-01 5.83055854e-01 1.24049909e-01 -1.34147823e+00
-2.97390819e-01 -5.18549323e-01 -9.59731638e-01 7.40616739e-01
2.45432496e-01 2.18672395e-01 -9.36010182e-01 1.21012437e+00
4.75252932e-03 -7.84734040e-02 1.95230916e-01 1.30312788e+00
1.10385036e+00 9.23887849e-01 6.26218691e-02 -3.11223119e-01
1.43733823e+00 -8.02994430e-01 -7.90189624e-01 6.78947940e-02
3.86453979e-02 -7.98554182e-01 1.15654230e+00 6.22121811e-01
-9.64269936e-01 -8.76160502e-01 -1.26806819e+00 2.40928456e-02
-3.85620773e-01 4.95167375e-01 4.04383123e-01 6.56112134e-01
-1.04317319e+00 5.74841380e-01 -2.87145317e-01 3.66161019e-02
5.65135777e-01 1.55445203e-01 -3.69354784e-01 -3.14970195e-01
-1.27279770e+00 8.41664076e-01 3.01827282e-01 3.90119731e-01
-7.30918288e-01 -4.55430090e-01 -6.52538300e-01 7.47965500e-02
5.70756532e-02 -4.07559544e-01 8.01003993e-01 -1.14569461e+00
-1.68802249e+00 4.69526470e-01 1.23656638e-01 1.76134855e-02
1.46171212e-01 2.56346129e-02 -6.83731496e-01 4.45769548e-01
-6.63413033e-02 2.48611555e-01 1.03677511e+00 -1.37291193e+00
-8.27409446e-01 -3.14254135e-01 1.41553760e-01 4.37068671e-01
-4.05341804e-01 -8.52666721e-02 -5.79106569e-01 -6.24489784e-01
1.81500018e-01 -1.70196965e-01 5.06387092e-02 1.02760993e-01
-1.73991010e-01 -2.48575043e-02 3.75461847e-01 -1.18681812e+00
1.15539551e+00 -2.12647581e+00 2.02917576e-01 2.35912830e-01
1.63746238e-01 3.14166278e-01 -1.20704450e-01 4.19878364e-02
2.54697293e-01 -8.38684961e-02 -5.86460046e-02 9.36640948e-02
-2.11354524e-01 -2.73999751e-01 3.25147033e-01 3.57741952e-01
2.71406025e-01 7.01489747e-01 -6.28007829e-01 -5.42388022e-01
6.17089510e-01 5.48062444e-01 -4.00075018e-01 3.01108509e-01
3.20093073e-02 1.86376527e-01 -3.50960433e-01 6.34957492e-01
1.16856170e+00 -1.05909742e-01 -2.44002789e-01 -7.41010964e-01
-1.31688630e-02 -1.95239633e-01 -1.35359311e+00 1.64743066e+00
-7.39095569e-01 3.71815532e-01 -8.97966623e-02 -8.54094386e-01
1.00212204e+00 1.17543973e-01 8.35316777e-02 -1.41385019e+00
9.42659080e-02 3.37943941e-01 -1.19259536e-01 -8.25729907e-01
2.68273532e-01 -3.26884389e-02 7.98418820e-02 -8.59127194e-02
4.26513925e-02 2.97842652e-01 -4.25014585e-01 -2.18232334e-01
6.76025987e-01 -1.58639312e-01 1.71834067e-01 -1.26747355e-01
8.43563497e-01 -3.80312681e-01 6.27120554e-01 4.53370690e-01
-4.05001760e-01 5.19392490e-01 3.56252223e-01 -3.48522991e-01
-1.08772850e+00 -1.16052938e+00 -2.98306227e-01 5.89820325e-01
6.03971124e-01 -5.66298068e-02 -5.12098551e-01 -4.50206876e-01
2.73632333e-02 1.41705006e-01 -5.33513606e-01 -2.55216241e-01
-2.56416798e-01 -8.26693773e-01 5.85460402e-02 3.21716011e-01
1.05579519e+00 -1.15324795e+00 -3.73551607e-01 1.04890458e-01
-3.33060622e-01 -7.15466201e-01 -2.38538105e-02 -2.29579419e-01
-4.71058339e-01 -9.16216969e-01 -1.06398118e+00 -8.82831573e-01
2.64789343e-01 2.73121893e-01 6.83458388e-01 -1.27874747e-01
-2.97623843e-01 3.52118641e-01 -5.48598111e-01 -5.99960275e-02
-1.90442756e-01 -4.52379107e-01 1.83761842e-03 4.14138705e-01
2.65814930e-01 -4.77843434e-01 -1.13463163e+00 1.01462655e-01
-8.06450129e-01 -5.66801913e-02 9.48720634e-01 7.95349181e-01
7.96189487e-01 6.76525652e-01 8.56343508e-01 5.25780097e-02
6.67126775e-01 -1.22739099e-01 -4.20507640e-01 4.58371252e-01
-5.42413533e-01 -1.41152576e-01 6.63595855e-01 -2.47963130e-01
-1.19548869e+00 -4.97659057e-01 -4.70509939e-02 -3.16378236e-01
1.00139432e-01 4.27399307e-01 -5.82840741e-01 -5.24187148e-01
3.92171532e-01 5.14942646e-01 -3.68597321e-02 -6.23635709e-01
2.91309327e-01 9.87843275e-01 4.14949119e-01 -1.90197900e-01
5.20602345e-01 2.99554676e-01 -1.08117491e-01 -7.29932666e-01
-4.35527831e-01 -2.85714209e-01 -4.03361209e-02 -5.93933403e-01
9.82257724e-01 -1.12370145e+00 -9.37254131e-01 1.03721404e+00
-8.97761822e-01 -6.29572049e-02 -5.67120500e-02 7.19664931e-01
-4.11716551e-01 6.23372614e-01 -7.01376557e-01 -7.84211934e-01
-4.52568501e-01 -1.49924052e+00 9.03239310e-01 5.55530667e-01
7.79096484e-01 -4.68129605e-01 -3.00446898e-01 3.72841507e-01
4.92274970e-01 9.11704302e-02 1.08101451e+00 2.33820826e-01
-7.84447789e-01 -1.66816890e-01 -8.20290148e-01 9.14130926e-01
9.78206396e-02 -3.71397197e-01 -1.04334128e+00 -2.11357877e-01
2.19520591e-02 -3.91303748e-01 9.24599588e-01 7.35103130e-01
1.38780224e+00 -3.54133904e-01 1.86566651e-01 7.57228315e-01
2.04303527e+00 2.72897899e-01 9.78892922e-01 5.71841478e-01
6.51654363e-01 6.38310850e-01 6.24038815e-01 2.73042262e-01
2.90196002e-01 7.66096056e-01 5.23600817e-01 -1.20503508e-01
-3.21672887e-01 -9.88552496e-02 2.69965976e-01 9.84266460e-01
-1.29115492e-01 -9.29332972e-02 -4.43688810e-01 4.65511560e-01
-1.54065168e+00 -8.68257642e-01 -4.37260792e-03 2.12832832e+00
6.39186144e-01 5.95192499e-02 -1.32933870e-01 4.12526697e-01
6.49536669e-01 1.15318276e-01 -4.38621789e-01 -2.95430005e-01
-4.06631321e-01 2.59654969e-01 6.08299792e-01 3.49711150e-01
-1.20672071e+00 4.34303135e-01 5.53870726e+00 1.39361393e+00
-9.71120834e-01 5.95570393e-02 7.19037175e-01 -1.64906830e-01
-3.25803727e-01 -3.30747545e-01 -2.46174499e-01 6.49311483e-01
5.57487726e-01 8.84005427e-02 5.83368421e-01 6.48694873e-01
3.02427888e-01 -2.86314994e-01 -4.86700386e-01 1.49222434e+00
1.60979763e-01 -1.14188731e+00 3.09649289e-01 -7.59026268e-03
6.36080265e-01 -3.37932676e-01 5.16610920e-01 5.19398525e-02
-2.84110934e-01 -1.22630095e+00 7.04019845e-01 1.15835202e+00
9.78937745e-01 -8.26187849e-01 9.27127481e-01 -8.95425007e-02
-1.16829896e+00 -5.24761200e-01 -7.15142429e-01 8.39773566e-02
-6.17991053e-02 6.13515377e-01 5.42148277e-02 7.97046423e-01
1.32184529e+00 5.31579137e-01 -7.17135847e-01 1.53544319e+00
3.31229940e-02 2.86751390e-01 8.94504935e-02 -6.28638715e-02
8.49755853e-02 -2.15460509e-01 4.42420602e-01 8.70401144e-01
5.84606588e-01 9.40715522e-02 -1.76020265e-01 5.76856196e-01
2.48498291e-01 4.85795617e-01 -4.89957742e-02 4.11830515e-01
1.64713815e-01 1.27320373e+00 -4.80096281e-01 -3.06343228e-01
-5.39353371e-01 1.14086115e+00 7.22233057e-02 4.58648026e-01
-4.98551190e-01 -7.51543462e-01 6.25234962e-01 -1.33238971e-01
4.76690650e-01 1.31875336e-01 -2.56353378e-01 -1.10295165e+00
4.68097240e-01 -9.80382085e-01 7.39952475e-02 -1.12676716e+00
-1.38905334e+00 6.07793689e-01 -4.42351073e-01 -1.47291648e+00
4.30489153e-01 -8.08733821e-01 -3.40521097e-01 1.35564387e+00
-1.88162637e+00 -1.09109902e+00 -7.79959738e-01 5.74694693e-01
1.61559135e-01 -2.21683681e-01 7.00172603e-01 4.99244422e-01
-4.76965368e-01 6.25034928e-01 4.98843551e-01 6.38636872e-02
6.74319983e-01 -1.20152211e+00 -2.51177520e-01 7.45766103e-01
-4.69859898e-01 1.53838918e-01 3.80507320e-01 -3.17900389e-01
-1.27739584e+00 -9.04930174e-01 2.76666105e-01 -6.70131072e-02
1.23199664e-01 2.17494108e-02 -9.67291713e-01 -1.43698677e-01
-6.35349229e-02 4.52720597e-02 4.73965377e-01 -1.92503735e-01
-3.81462604e-01 -7.23743677e-01 -1.29669356e+00 2.40112156e-01
6.58954024e-01 -6.65205657e-01 -3.97732794e-01 6.84624724e-03
7.49190092e-01 -6.87799677e-02 -1.26715147e+00 5.72475433e-01
8.18136096e-01 -1.33861899e+00 9.98110354e-01 -2.46092528e-02
6.45576537e-01 -5.22348523e-01 -3.94070387e-01 -1.30812407e+00
-5.95527589e-01 2.52641112e-01 1.42797694e-01 1.23128760e+00
2.21765041e-01 -4.67831492e-01 4.05544400e-01 6.99123889e-02
-7.38924146e-02 -8.88747573e-01 -1.08613312e+00 -5.37423790e-01
-1.83958367e-01 -3.56423974e-01 7.54443944e-01 6.13510847e-01
-1.84107229e-01 -7.05804750e-02 -6.11212373e-01 4.13289934e-01
7.04266965e-01 7.93266445e-02 2.03333750e-01 -1.04479074e+00
-3.97807062e-01 -5.33383489e-01 -8.84626746e-01 -7.62428045e-01
-4.64598119e-01 -7.03635216e-01 -1.08932741e-01 -2.01076865e+00
2.01815873e-01 -5.67607880e-01 -7.19296515e-01 1.03366226e-02
-3.34817290e-01 5.09417593e-01 -6.44970089e-02 2.63525784e-01
-5.80540061e-01 7.81038821e-01 1.57212794e+00 -3.29907447e-01
-2.67573744e-01 -1.66254848e-01 -6.63595498e-01 3.01129997e-01
5.52259743e-01 1.96858391e-01 -3.20057452e-01 -4.64566469e-01
3.93838912e-01 2.39429295e-01 6.45666480e-01 -1.41213667e+00
-6.14791326e-02 5.12428470e-02 9.84297812e-01 -2.39853188e-01
3.59446734e-01 -8.22460115e-01 -1.29458115e-01 3.16985995e-01
-1.72032993e-02 -5.36678016e-01 1.34869978e-01 7.01865554e-01
-4.55617011e-01 5.65080084e-02 9.60238755e-01 -1.11017101e-01
-1.11221802e+00 4.23993886e-01 -1.33353651e-01 -4.92786646e-01
8.70987833e-01 -2.56062746e-01 -3.48044157e-01 -2.78038472e-01
-6.34182334e-01 -2.53560562e-02 3.73330444e-01 3.40711057e-01
1.01954067e+00 -1.72201502e+00 -7.57197618e-01 1.94201499e-01
4.16546911e-01 -3.74066263e-01 1.13506162e+00 5.07417262e-01
-7.38854706e-01 3.51946838e-02 -6.69079363e-01 -5.48280180e-01
-1.07584071e+00 6.88854694e-01 6.91227078e-01 -3.67368124e-02
-5.93227506e-01 7.88988948e-01 2.13092919e-02 7.36030331e-03
4.86243255e-02 -2.51725793e-01 -8.88971508e-01 1.07459351e-02
8.13994467e-01 4.94598389e-01 2.36589015e-01 -8.27073693e-01
-3.48576277e-01 1.01672769e+00 8.62426013e-02 5.33527993e-02
1.12315786e+00 -2.86254108e-01 -2.33922049e-01 1.80980682e-01
1.50588477e+00 -1.29455313e-01 -1.18892813e+00 -3.45980972e-01
-4.73161638e-01 -6.91408753e-01 4.69845563e-01 -1.09207702e+00
-1.13561976e+00 9.19497907e-01 1.44799459e+00 1.75365880e-01
1.77412820e+00 -1.61564097e-01 7.73265123e-01 -8.88128877e-02
2.87977278e-01 -1.04830146e+00 2.52703637e-01 -3.53082895e-01
9.66594398e-01 -1.22262335e+00 -6.66443333e-02 -2.19943896e-01
-5.05011380e-01 1.00531411e+00 5.21085680e-01 -1.80936068e-01
6.27129853e-01 -3.31091404e-01 2.33895242e-01 -1.40346348e-01
-1.97859764e-01 -2.07131982e-01 5.31469762e-01 8.72803986e-01
-1.27207279e-01 2.88240194e-01 -3.34370136e-01 7.98466444e-01
-1.00962438e-01 -1.63806658e-02 3.95588756e-01 2.29921028e-01
-6.43794656e-01 -6.16162479e-01 -4.01400805e-01 8.24727714e-01
-4.12532270e-01 -5.89471348e-02 3.02713752e-01 1.37002766e-01
6.08927310e-01 1.22021961e+00 -8.33248645e-02 -7.36112237e-01
4.64636743e-01 -1.83107898e-01 5.72256923e-01 3.96359060e-03
-1.73263147e-01 -3.72032784e-02 -2.96095520e-01 -7.54187465e-01
-4.73089218e-01 -2.90355057e-01 -5.97675085e-01 -1.75628051e-01
-1.87422484e-01 6.93520531e-02 6.32369399e-01 6.31961405e-01
1.43065184e-01 8.38333547e-01 8.34577203e-01 -6.99562311e-01
-2.25735068e-01 -1.08750725e+00 -9.77007687e-01 3.64508599e-01
5.99140882e-01 -6.99791789e-01 -5.19373894e-01 -2.94983894e-01] | [11.758156776428223, -1.9264256954193115] |
123e3cca-0b78-4bee-a3cd-ae3cad995ea3 | eye-movements-biometrics-a-bibliometric | 2006.01310 | null | https://arxiv.org/abs/2006.01310v1 | https://arxiv.org/pdf/2006.01310v1.pdf | Eye Movements Biometrics: A Bibliometric Analysis from 2004 to 2019 | Person identification based on eye movements is getting more and more attention, as it is anti-spoofing resistant and can be useful for continuous authentication. Therefore, it is noteworthy for researchers to know who and what is relevant in the field, including authors, journals, conferences, and institutions. This paper presents a comprehensive quantitative overview of the field of eye movement biometrics using a bibliometric approach. All data and analyses are based on documents written in English published between 2004 and 2019. Scopus was used to perform information retrieval. This research focused on temporal evolution, leading authors, most cited papers, leading journals, competitions and collaboration networks. | ['Karin Satie Komati', 'Jefferson Oliveira Andrade', 'Antonio Ricardo Alexandre Brasil'] | 2020-06-01 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [-3.55009586e-01 -4.98787612e-01 -6.46744132e-01 4.49682176e-01
2.32544124e-01 -4.31446731e-01 3.81360114e-01 4.83738750e-01
-8.23280334e-01 7.71477759e-01 1.69391721e-01 -4.34956104e-01
-2.58018792e-01 -5.24143219e-01 -1.35911509e-01 -3.19383681e-01
2.97780573e-01 -2.10032225e-01 2.18323022e-01 4.08715680e-02
1.04467404e+00 7.96992660e-01 -1.57082367e+00 -3.65828723e-01
9.18438494e-01 7.31038690e-01 -1.09644629e-01 1.18467435e-01
-2.96785414e-01 4.18285668e-01 -9.88014519e-01 -8.16966891e-01
-1.13628037e-01 -5.38664818e-01 -3.62124085e-01 -4.99937862e-01
1.16221812e-02 -6.57908618e-02 -6.31593913e-02 1.10949779e+00
5.97542226e-01 -4.31090266e-01 4.26497787e-01 -1.19102073e+00
-1.25188923e+00 9.29516461e-03 -8.77795160e-01 6.57351077e-01
9.62352276e-01 4.54611555e-02 3.09616774e-01 -6.95131719e-01
7.19498932e-01 1.04496384e+00 4.50454235e-01 4.27688211e-01
-6.55675054e-01 -1.07786131e+00 -1.74736530e-01 4.26427335e-01
-1.45995581e+00 -2.35026792e-01 7.02355802e-01 -7.63665974e-01
8.66304576e-01 2.78213084e-01 1.03263330e+00 1.05912375e+00
4.57380772e-01 -8.53921622e-02 1.49404943e+00 -8.53404224e-01
2.95912400e-02 6.55164301e-01 5.34205556e-01 2.25746617e-01
1.46527958e+00 1.05649762e-01 -8.22047293e-01 -1.23690769e-01
3.61155152e-01 2.36124113e-01 -2.68867314e-01 2.12742776e-01
-8.75727117e-01 4.13593769e-01 -1.39108803e-02 4.97669190e-01
-5.56944430e-01 -5.71776509e-01 1.65065214e-01 3.55839670e-01
2.96461254e-01 2.75369197e-01 -6.70738816e-02 -4.33013290e-01
-9.17299032e-01 -9.48218107e-02 9.59781945e-01 3.61349791e-01
1.70583501e-01 -1.25703096e-01 1.27751321e-01 3.33164126e-01
8.35774899e-01 6.46820188e-01 4.88385201e-01 -5.78930914e-01
2.00216155e-02 7.51875162e-01 3.49119827e-02 -1.19464874e+00
-1.81063354e-01 3.89980003e-02 -6.10001385e-01 2.56784856e-01
1.52985409e-01 -3.07530016e-01 -5.56174338e-01 1.12501776e+00
1.24991715e-01 -4.81450669e-02 -1.48145989e-01 6.68497324e-01
1.19047773e+00 2.15722069e-01 2.67967969e-01 -6.69706643e-01
1.67969906e+00 -4.40514535e-01 -1.11799443e+00 2.42064640e-01
7.23927468e-02 -9.67240572e-01 4.69778746e-01 4.50333506e-01
-1.02600789e+00 -2.01938659e-01 -6.41053081e-01 3.22302610e-01
-7.53653586e-01 -1.54829562e-01 3.40240955e-01 1.50403953e+00
-9.96958256e-01 2.20994696e-01 -5.74680030e-01 -1.02803493e+00
3.90839964e-01 3.26404482e-01 -3.80459547e-01 1.66023076e-01
-1.10962617e+00 1.21090555e+00 2.85512134e-02 -8.79127756e-02
3.57840151e-01 -5.66266954e-01 -4.29858178e-01 -6.05290771e-01
4.23973240e-02 -6.15333974e-01 4.97003794e-01 -5.15696168e-01
-1.51163650e+00 1.28671515e+00 -7.04064131e-01 -3.25653493e-01
1.21381618e-01 3.41355391e-02 -1.00044501e+00 3.10683459e-01
1.95420533e-01 -1.03440449e-01 3.99569452e-01 -7.56963611e-01
-9.18555379e-01 -9.89526153e-01 -4.19802547e-01 -3.09960872e-01
-8.00623059e-01 1.15058768e+00 -4.60269839e-01 -6.45169556e-01
-1.76550537e-01 -7.92530656e-01 1.92863077e-01 -3.59595656e-01
-1.01380274e-01 -5.97866118e-01 6.79093361e-01 -1.09499609e+00
1.78959835e+00 -1.99934638e+00 -2.84140050e-01 3.85799378e-01
3.62176090e-01 5.03112197e-01 7.17905164e-01 6.23852193e-01
2.03235656e-01 6.71312571e-01 3.62190366e-01 7.22235888e-02
-2.66447693e-01 -4.67729807e-01 1.25179172e-01 7.89628208e-01
-4.11849588e-01 7.44807124e-01 -4.68638748e-01 -5.36464274e-01
2.50524461e-01 6.07924044e-01 8.82439762e-02 -1.90873653e-01
8.59400868e-01 3.99139792e-01 -5.96217394e-01 1.01137757e+00
7.69182026e-01 -2.32568696e-01 -1.33876175e-01 -1.04448413e-02
-7.87256241e-01 8.25956613e-02 -9.63677347e-01 9.53297138e-01
-9.75706801e-02 1.12499833e+00 -2.17227295e-01 -4.42183971e-01
1.05661428e+00 5.66440761e-01 5.51211536e-01 -7.44243383e-01
2.97748148e-01 3.76869142e-01 -3.92424315e-02 -6.45158231e-01
2.69124746e-01 1.18793204e-01 4.58779514e-01 4.16287839e-01
-2.78209448e-01 5.74897230e-01 4.12229419e-01 1.17822096e-01
3.23121041e-01 -2.08383515e-01 5.44763267e-01 -1.26281261e-01
8.82647038e-01 -2.82755315e-01 3.75034600e-01 4.50165659e-01
-5.02490282e-01 -1.11772716e-01 3.38294983e-01 -1.39114574e-01
-6.15995467e-01 -5.45507431e-01 -5.38884282e-01 3.14177722e-01
2.45125338e-01 -3.56768847e-01 -8.85120988e-01 1.07066974e-01
-1.40735209e-02 1.83881491e-01 -5.01218617e-01 2.89629344e-02
-1.30783021e-01 -4.55830127e-01 4.37192380e-01 -1.44442581e-02
8.23215485e-01 -1.27867818e+00 -8.36541891e-01 -2.29620427e-01
2.71258444e-01 -1.01008701e+00 -1.92930419e-02 -8.83947372e-01
-8.57504666e-01 -1.41944325e+00 -1.20839107e+00 -7.47794390e-01
5.15871465e-01 1.95887357e-01 6.32009685e-01 9.32444185e-02
-4.32130367e-01 5.04341602e-01 -2.46509761e-01 -9.45465028e-01
1.49388108e-02 2.92600933e-02 3.60414565e-01 -5.49178496e-02
1.42277205e+00 4.00133468e-02 -5.08948803e-01 1.50841475e-01
-3.85188669e-01 -6.63498104e-01 3.74930054e-01 1.79298401e-01
1.19653948e-01 5.85451797e-02 5.00248075e-01 -2.72858709e-01
9.88730252e-01 -4.44781363e-01 -8.39066803e-01 2.32121006e-01
-1.27755737e+00 -5.38491607e-01 -2.22385377e-01 -1.04708299e-01
-8.94048810e-01 -8.37203801e-01 3.52585584e-01 1.38896480e-01
-3.78226519e-01 5.27486742e-01 -1.38502151e-01 -5.18476188e-01
5.96517205e-01 -5.86644895e-02 4.11634892e-01 -5.65282524e-01
-4.96314019e-01 1.24144506e+00 1.39554292e-01 -8.00332576e-02
5.54153025e-01 4.85397756e-01 4.32853736e-02 -1.20086551e+00
-3.37249525e-02 -7.30908811e-01 -3.36641043e-01 -5.36096215e-01
8.79301727e-01 -5.82449138e-01 -1.45386291e+00 9.41385865e-01
-1.17382538e+00 4.19077396e-01 3.49451452e-01 1.01471257e+00
5.31356037e-01 4.75759983e-01 -4.62009698e-01 -1.12919867e+00
-4.45122868e-01 -1.08199012e+00 2.16447026e-01 9.73541617e-01
-3.00784230e-01 -1.04548371e+00 1.90013707e-01 3.90303373e-01
4.78166014e-01 2.17180908e-01 9.04954672e-02 -3.86532456e-01
-3.08430910e-01 -4.61896062e-01 -2.74351478e-01 -4.59722243e-02
3.96684915e-01 4.71588165e-01 -6.43104374e-01 -2.21854463e-01
2.89255138e-02 4.39203471e-01 6.21819258e-01 7.43162811e-01
8.36862087e-01 -5.11722147e-01 -8.10751557e-01 3.40762496e-01
1.34142148e+00 7.78312147e-01 8.77151072e-01 1.09703839e+00
4.74171638e-01 9.98196423e-01 2.79782921e-01 4.24610943e-01
4.83926922e-01 6.64522111e-01 2.63851639e-02 2.60332584e-01
1.51589289e-01 3.90363485e-02 2.41949394e-01 4.09324110e-01
-7.72065818e-01 -8.46187547e-02 -1.21599317e+00 5.19132555e-01
-1.19332123e+00 -1.13954127e+00 -5.46799421e-01 2.54651904e+00
4.78978515e-01 1.57989580e-02 7.61197507e-01 1.95323348e-01
1.08672404e+00 -1.63089544e-01 -6.49997517e-02 -4.32036877e-01
-2.04584643e-01 2.38660783e-01 7.90435433e-01 1.13098808e-01
-6.94915771e-01 7.42248416e-01 6.74199343e+00 2.96987236e-01
-1.14503837e+00 8.74321163e-02 2.94341803e-01 -1.16220765e-01
1.39763975e-03 2.62753852e-02 -9.64134753e-01 8.92618895e-01
1.23393476e+00 -7.61833072e-01 2.34893635e-01 3.85815322e-01
4.71171945e-01 -3.31501305e-01 -9.15681645e-02 1.18759584e+00
3.20917845e-01 -1.58165693e+00 -1.81593105e-01 7.52224863e-01
7.40971208e-01 -4.28551614e-01 4.84675050e-01 -5.37281275e-01
-1.67202026e-01 -7.97653437e-01 3.77482116e-01 8.77654672e-01
8.09297740e-01 -6.92125916e-01 9.99802709e-01 -1.11326598e-01
-8.15576196e-01 -3.25020589e-02 -8.32150355e-02 -2.07851961e-01
3.01882356e-01 3.59626293e-01 -5.46955243e-02 5.60966432e-01
1.19885361e+00 8.00458133e-01 -5.83739758e-01 1.61278808e+00
-1.94564909e-01 6.36804581e-01 5.39533701e-03 -5.25752842e-01
-2.28816435e-01 -3.71118933e-01 5.07502317e-01 9.84776139e-01
3.48586500e-01 1.84504673e-01 -5.56849003e-01 4.79111820e-01
1.49495274e-01 3.78783673e-01 -8.02194476e-01 -4.40325081e-01
7.86144853e-01 8.32883179e-01 -8.75082910e-01 -7.52381757e-02
-8.45677018e-01 6.81846857e-01 -6.15872741e-01 4.93456781e-01
-1.35278970e-01 -6.96159184e-01 6.24885857e-01 4.87304628e-01
-2.82182842e-01 -7.73687363e-02 -4.48537856e-01 -7.36222744e-01
-7.01320097e-02 -7.52615333e-01 4.62936223e-01 -2.85778075e-01
-8.56168926e-01 3.38541418e-01 1.32922366e-01 -9.77039337e-01
5.26923910e-02 -7.89242446e-01 -2.82848954e-01 1.33559954e+00
-1.08154881e+00 -7.18941748e-01 -3.44190180e-01 3.34399790e-01
-2.94287056e-01 -8.51950049e-01 4.36413020e-01 2.07828701e-01
-9.39373195e-01 6.15851402e-01 -1.90038444e-03 1.37648657e-01
7.62949347e-01 -7.17648923e-01 6.13068752e-02 6.80762112e-01
-4.28176880e-01 1.19370115e+00 4.15170819e-01 -1.09159636e+00
-1.04086626e+00 -1.49575070e-01 1.52239931e+00 -5.90236127e-01
6.51427567e-01 4.62551653e-01 -5.36989987e-01 3.25102299e-01
5.72681248e-01 -3.69583786e-01 1.14660621e+00 -5.02810208e-03
-9.48146135e-02 -1.06438302e-01 -1.24809062e+00 6.91979766e-01
7.21293449e-01 -5.15225291e-01 -4.91364568e-01 -1.40788788e-02
1.84546620e-01 -6.73827752e-02 -8.61058235e-01 1.05054654e-01
9.67002213e-01 -8.74666989e-01 9.16688383e-01 -4.52368081e-01
-1.60355903e-02 -9.55895483e-02 4.38626707e-01 -5.43680072e-01
-1.99047506e-01 -9.07399476e-01 -1.12802394e-01 1.35417736e+00
1.03683304e-02 -1.33600926e+00 7.21633017e-01 7.56058514e-01
3.99887264e-01 -5.07173002e-01 -8.08422744e-01 -7.98743963e-01
-4.93277423e-02 1.64978221e-01 5.96961915e-01 9.58295524e-01
2.64955580e-01 -4.66442332e-02 -1.52283415e-01 -1.44443825e-01
9.48274732e-01 -3.92000198e-01 6.21412873e-01 -1.78412569e+00
4.53791767e-01 -1.04523408e+00 -6.74858809e-01 -1.36083350e-01
3.97354923e-02 -4.10422891e-01 -1.08600485e+00 -1.87870204e+00
6.87679723e-02 1.43990517e-01 -3.39261770e-01 6.71319142e-02
-1.83696732e-01 2.66940117e-01 4.38180566e-02 5.61126769e-01
7.22599104e-02 -2.58895069e-01 9.38705564e-01 1.47989765e-01
-4.22989070e-01 1.73017532e-01 -1.09964526e+00 6.59178674e-01
9.37349081e-01 -3.15535456e-01 2.32801288e-01 2.17537969e-01
4.10665035e-01 -3.56389195e-01 3.46714348e-01 -6.81567311e-01
5.71426988e-01 -4.58896548e-01 2.24102333e-01 -6.92930698e-01
-1.26349628e-01 -8.06171536e-01 3.11429322e-01 6.57076776e-01
7.57395327e-02 4.38037425e-01 3.99075866e-01 1.84868366e-01
-2.30247512e-01 -3.75856251e-01 3.16030651e-01 9.20015574e-02
-5.41015267e-01 2.80961275e-01 -5.01427710e-01 -2.72697717e-01
1.23811769e+00 -8.93739045e-01 -3.58730435e-01 -2.38237053e-01
-3.57213974e-01 -1.61100447e-01 8.13467681e-01 4.21908885e-01
6.16294563e-01 -1.24532092e+00 -4.74255264e-01 -7.24192634e-02
2.17706263e-01 -1.01272166e+00 1.75090611e-01 1.10928619e+00
-7.09074855e-01 8.19864988e-01 -5.77386856e-01 4.67646904e-02
-1.51122332e+00 6.96313381e-01 -1.20306171e-01 3.20698351e-01
-4.91957456e-01 4.55022097e-01 -5.01765728e-01 3.75073135e-01
5.02799988e-01 4.21999305e-01 -9.16490257e-01 3.83745492e-01
9.78537261e-01 1.01260734e+00 -2.41844788e-01 -9.92501736e-01
-9.13619399e-01 8.37973237e-01 1.36824057e-01 -5.13609767e-01
8.22087288e-01 -2.19398275e-01 -7.44061828e-01 5.84025681e-01
9.69478965e-01 3.77971441e-01 -2.26115867e-01 1.40357360e-01
2.82049507e-01 -7.00623274e-01 -1.13458775e-01 -6.10625684e-01
-9.77232933e-01 6.07601941e-01 9.03723001e-01 3.11367601e-01
9.08638418e-01 -2.07489371e-01 4.14600462e-01 -1.41738340e-01
4.17064071e-01 -1.23765910e+00 -6.55939579e-02 8.84991288e-02
5.59304655e-01 -9.55486715e-01 3.95016260e-02 -2.06926852e-01
-3.58230829e-01 1.04384780e+00 3.05437803e-01 2.93144375e-01
1.28455591e+00 -1.37850523e-01 2.16515347e-01 -2.08815813e-01
7.61942714e-02 -1.25348508e-01 4.23485219e-01 8.44867706e-01
8.33758652e-01 -1.15748145e-01 -1.39292574e+00 6.54730856e-01
-1.51441442e-02 4.57063913e-01 2.99998581e-01 8.06178868e-01
-2.30809808e-01 -1.39134061e+00 -8.32055807e-01 7.23873615e-01
-1.09759641e+00 -7.46295750e-02 -7.16294110e-01 7.56135881e-01
-8.83755460e-02 1.28224945e+00 -6.85811937e-02 -3.78691375e-01
2.00294465e-01 -3.29024047e-01 3.21006298e-01 -1.20110728e-01
-4.92084116e-01 -2.50796080e-01 -2.60370374e-01 -1.51026025e-01
-8.85657966e-01 -1.04366243e+00 -9.17244911e-01 -8.32471073e-01
-4.00019348e-01 4.62910742e-01 1.07352376e+00 8.48219514e-01
4.65157986e-01 1.84061676e-01 4.13449019e-01 -1.16211012e-01
3.07326317e-01 -8.94143939e-01 -6.14129543e-01 -6.83548898e-02
2.53796220e-01 -7.38229096e-01 -4.72579211e-01 -3.01009446e-01] | [13.34234619140625, 0.8238323926925659] |
31e16417-02ff-4e31-b95a-af8960ae63be | masked-contrastive-pre-training-for-efficient | 2212.00986 | null | https://arxiv.org/abs/2212.00986v2 | https://arxiv.org/pdf/2212.00986v2.pdf | Masked Contrastive Pre-Training for Efficient Video-Text Retrieval | We present a simple yet effective end-to-end Video-language Pre-training (VidLP) framework, Masked Contrastive Video-language Pretraining (MAC), for video-text retrieval tasks. Our MAC aims to reduce video representation's spatial and temporal redundancy in the VidLP model by a mask sampling mechanism to improve pre-training efficiency. Comparing conventional temporal sparse sampling, we propose to randomly mask a high ratio of spatial regions and only feed visible regions into the encoder as sparse spatial sampling. Similarly, we adopt the mask sampling technique for text inputs for consistency. Instead of blindly applying the mask-then-prediction paradigm from MAE, we propose a masked-then-alignment paradigm for efficient video-text alignment. The motivation is that video-text retrieval tasks rely on high-level alignment rather than low-level reconstruction, and multimodal alignment with masked modeling encourages the model to learn a robust and general multimodal representation from incomplete and unstable inputs. Coupling these designs enables efficient end-to-end pre-training: reduce FLOPs (60% off), accelerate pre-training (by 3x), and improve performance. Our MAC achieves state-of-the-art results on various video-text retrieval datasets, including MSR-VTT, DiDeMo, and ActivityNet. Our approach is omnivorous to input modalities. With minimal modifications, we achieve competitive results on image-text retrieval tasks. | ['Si Liu', 'Jinqiao Wang', 'Yousong Zhu', 'Xiaobo Li', 'Wenyu Sun', 'Shuwen Xiao', 'Yue Liao', 'Biaolong Chen', 'Fangxun Shu'] | 2022-12-02 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 3.22082102e-01 -4.39520925e-01 -6.29363179e-01 -3.08811069e-01
-1.16102695e+00 -3.65015715e-01 6.73261046e-01 -3.85880053e-01
-6.63475633e-01 1.50065646e-01 4.53349978e-01 -1.95459783e-01
1.81488827e-01 -1.42467812e-01 -9.05005276e-01 -4.81592119e-01
9.66494456e-02 2.80546516e-01 -2.46365368e-02 1.05540030e-01
1.01942830e-01 1.43972874e-01 -1.68940783e+00 1.03747702e+00
4.74518180e-01 9.22338665e-01 7.00301528e-01 9.89668429e-01
-1.09592527e-01 1.25379598e+00 -2.00825900e-01 -1.85192615e-01
2.94684350e-01 -3.84097964e-01 -7.80912161e-01 1.33446231e-01
1.00308549e+00 -7.09808111e-01 -1.05210376e+00 6.73937857e-01
5.93505085e-01 2.63613582e-01 6.72710478e-01 -1.01158381e+00
-5.13368309e-01 4.77721125e-01 -8.22096646e-01 1.58545732e-01
4.33183938e-01 1.11253396e-01 9.43924487e-01 -1.42027867e+00
7.10027456e-01 1.38340867e+00 3.98180962e-01 7.95809627e-01
-1.05800056e+00 -6.81590557e-01 3.74920994e-01 3.22726190e-01
-1.72048783e+00 -9.76231456e-01 4.18651938e-01 -2.51668692e-01
1.08317912e+00 3.43203157e-01 4.96721774e-01 1.20887196e+00
7.41715431e-02 1.35046077e+00 5.26317835e-01 -4.31823313e-01
-5.05840331e-02 -3.57389376e-02 -2.94699460e-01 8.56034696e-01
-2.87714213e-01 -2.55308840e-02 -1.08571088e+00 -1.75465252e-02
7.81933904e-01 2.66445339e-01 -3.35307568e-01 -3.16719264e-01
-1.38111174e+00 4.20711190e-01 2.85327524e-01 1.58465728e-01
-3.89344811e-01 5.69678783e-01 4.29138303e-01 4.66080457e-01
3.26464444e-01 2.37519736e-03 -2.18260720e-01 -3.30077350e-01
-1.80451620e+00 -3.42922360e-02 3.68402690e-01 1.07990551e+00
8.13894987e-01 2.63334930e-01 -4.53396171e-01 9.59292710e-01
4.41045731e-01 8.75143766e-01 6.60054803e-01 -1.07952416e+00
7.81719029e-01 1.15815118e-01 -2.62109935e-01 -9.35514569e-01
5.22593781e-02 -2.30939507e-01 -8.00950527e-01 -2.50187248e-01
-9.57128853e-02 2.60125726e-01 -1.43930316e+00 1.70762467e+00
-1.41967803e-01 3.12469155e-01 5.19081987e-02 1.19950581e+00
9.54004228e-01 9.17010725e-01 6.69493824e-02 -2.00799853e-01
1.08222640e+00 -1.33246899e+00 -7.50309110e-01 -3.85519356e-01
8.26340139e-01 -8.62040997e-01 1.21275342e+00 3.17958176e-01
-1.37939394e+00 -5.20693839e-01 -8.05047214e-01 -3.44312221e-01
-2.08443683e-02 4.01753247e-01 4.03868943e-01 2.67290056e-01
-1.25287008e+00 3.48289967e-01 -9.77812052e-01 -3.54122728e-01
3.68079871e-01 4.39189583e-01 -4.89836067e-01 -5.59852660e-01
-7.31415629e-01 5.37004709e-01 2.59625167e-01 2.74010766e-02
-1.30452096e+00 -6.12592995e-01 -8.96884024e-01 1.54814273e-01
3.94611299e-01 -8.07835340e-01 1.11830449e+00 -1.40725434e+00
-1.51447058e+00 9.40214157e-01 -6.66796088e-01 -4.61887062e-01
4.45568532e-01 -3.74118537e-01 -2.50239372e-01 8.31464589e-01
-2.38065720e-02 1.40591502e+00 1.41205418e+00 -1.22282398e+00
-3.05472314e-01 2.89682485e-02 -2.97175258e-01 5.81291318e-01
-5.49530864e-01 1.58650458e-01 -1.37069571e+00 -9.04773712e-01
1.03279769e-01 -8.72800171e-01 -1.36126488e-01 1.58960313e-01
-1.00337025e-02 2.30608553e-01 9.78324711e-01 -7.38172948e-01
1.43050909e+00 -2.38663816e+00 4.05164003e-01 9.96895432e-02
3.86952460e-01 2.47448146e-01 -6.80093884e-01 3.60110641e-01
-1.29104346e-01 -5.07277846e-02 -1.38855148e-02 -9.20399189e-01
-1.63918510e-01 1.89377591e-01 -5.17100811e-01 5.04756212e-01
6.61145449e-02 1.16056550e+00 -6.75298274e-01 -8.25948119e-01
4.99948531e-01 5.21991611e-01 -9.37121570e-01 3.91133577e-01
-3.26089203e-01 2.94116847e-02 -1.38742194e-01 9.35097218e-01
5.21659136e-01 -4.82405037e-01 1.09744549e-01 -3.87428492e-01
1.57887056e-01 1.23564862e-01 -9.12484825e-01 2.31979418e+00
-4.55333412e-01 9.62856293e-01 3.30379814e-01 -7.91340947e-01
3.00861686e-01 2.96053767e-01 6.11085534e-01 -1.04754829e+00
4.71676402e-02 5.01722023e-02 -4.81529772e-01 -4.70048994e-01
8.68045568e-01 3.72388393e-01 7.03171194e-02 4.01482314e-01
3.59100014e-01 2.57318497e-01 9.02211294e-02 7.18490124e-01
9.99776900e-01 2.30355546e-01 -1.64079458e-01 5.54678142e-02
3.64159703e-01 -5.95647879e-02 1.48957595e-01 8.91649783e-01
1.70051694e-01 1.04669988e+00 3.11205368e-02 -2.31666014e-01
-9.91053522e-01 -9.67520058e-01 2.00317994e-01 1.51225555e+00
2.28208974e-01 -9.65115070e-01 -4.56368715e-01 -4.61837441e-01
-2.68595904e-01 2.52298057e-01 -3.15512836e-01 -1.12705573e-01
-7.04863966e-01 -2.95989037e-01 5.99976122e-01 4.21218306e-01
4.79770690e-01 -8.76195490e-01 -2.30965629e-01 -1.16946111e-02
-5.22307515e-01 -1.45467949e+00 -1.05235064e+00 2.23642122e-02
-8.24234247e-01 -6.48235798e-01 -9.64626968e-01 -9.04252887e-01
7.75582373e-01 9.35198545e-01 1.08368683e+00 2.49537691e-01
-3.02984238e-01 8.50636959e-01 -4.34971035e-01 3.22603106e-01
-2.30349466e-01 -1.38530126e-02 1.27741262e-01 1.22554645e-01
5.63846789e-02 -4.10350591e-01 -7.97375619e-01 4.33186710e-01
-1.17624080e+00 3.31612319e-01 8.31208885e-01 9.69185650e-01
6.81778908e-01 -4.52281445e-01 -1.26257882e-01 -2.33368784e-01
1.31477460e-01 -4.00179476e-01 -2.89645553e-01 3.76852661e-01
-3.24487835e-01 -4.43818606e-02 4.49258804e-01 -8.48418593e-01
-6.11283004e-01 1.42394051e-01 7.09773824e-02 -1.45362437e+00
2.29308605e-01 5.68444610e-01 1.19415507e-01 -2.57504582e-01
4.56350863e-01 6.05434418e-01 1.12430125e-01 -3.56206447e-01
5.31276405e-01 6.42163932e-01 6.32372677e-01 -6.45903409e-01
7.25294709e-01 5.22081256e-01 -2.98753083e-01 -1.02594733e+00
-4.76769269e-01 -7.23268628e-01 -3.11615139e-01 -2.48668641e-01
7.94950962e-01 -1.67879868e+00 -5.49278915e-01 1.62214577e-01
-9.15052533e-01 -5.45736670e-01 3.21415402e-02 5.79586565e-01
-4.32066262e-01 6.45181835e-01 -7.21125841e-01 -6.76801741e-01
-5.84975958e-01 -1.21718371e+00 1.52395499e+00 -1.60741508e-01
-5.01670688e-02 -5.59236705e-01 -2.69276679e-01 4.34485137e-01
4.68358576e-01 -5.88650584e-01 4.06571567e-01 -2.69829631e-01
-9.72032666e-01 6.95906021e-03 -3.95075619e-01 2.47355551e-01
-3.16062063e-01 -1.90974608e-01 -9.90638554e-01 -7.93081224e-01
-3.89361054e-01 -6.65143371e-01 1.28261971e+00 4.31171983e-01
1.38967955e+00 -4.63911384e-01 -3.15155447e-01 9.57284629e-01
1.31484628e+00 -1.90607831e-01 8.35226536e-01 1.02628045e-01
1.00450909e+00 2.68742621e-01 5.89150906e-01 3.22777987e-01
3.72022182e-01 8.25833142e-01 2.65104562e-01 -3.26565444e-01
-4.62115854e-01 -4.65016335e-01 9.02514219e-01 9.62748110e-01
3.21972132e-01 -5.09929717e-01 -6.65891469e-01 6.21064067e-01
-2.04876161e+00 -1.14951754e+00 3.73126745e-01 2.12215948e+00
7.11835563e-01 -2.25280687e-01 1.48775920e-01 -3.06580752e-01
4.17915136e-01 4.20836240e-01 -2.81113535e-01 1.22043297e-01
-3.12879324e-01 1.60379633e-01 5.28645396e-01 6.46401286e-01
-9.71712947e-01 1.32110822e+00 6.44672298e+00 1.25830269e+00
-1.34420991e+00 1.81969449e-01 5.37100077e-01 -8.15871716e-01
-3.13859820e-01 -5.20077199e-02 -6.87880814e-01 4.37255740e-01
8.55399430e-01 2.44227067e-01 7.82215655e-01 6.66733444e-01
2.45214179e-01 -7.00186938e-02 -1.23501289e+00 1.60999382e+00
4.57377404e-01 -1.62403107e+00 4.84415144e-01 -7.01527297e-02
5.69346368e-01 3.09886813e-01 2.23848388e-01 3.80943626e-01
-1.82927787e-01 -1.10876024e+00 9.88215804e-01 3.68086904e-01
1.31084180e+00 -3.77856553e-01 3.21371615e-01 1.15759306e-01
-1.42671466e+00 -6.03153221e-02 -2.43237987e-01 3.24852854e-01
1.11554533e-01 2.81911522e-01 -5.37982821e-01 4.84247983e-01
8.03966343e-01 8.78866255e-01 -5.39669514e-01 7.59748936e-01
2.52283871e-01 5.64155161e-01 -4.71846372e-01 1.91672713e-01
4.05122072e-01 1.41802430e-01 4.75799352e-01 1.63220787e+00
2.09409922e-01 -4.19664234e-02 4.19511378e-01 5.21852791e-01
-2.70151824e-01 9.53769684e-02 -5.72407663e-01 -2.22977653e-01
4.38574731e-01 8.93552363e-01 -3.38617206e-01 -4.29725766e-01
-3.67360055e-01 1.42698443e+00 8.91652480e-02 7.13018298e-01
-9.20079112e-01 9.07772705e-02 5.47122419e-01 7.22611099e-02
5.38816273e-01 -2.99233168e-01 1.18623085e-01 -1.53318560e+00
2.24001974e-01 -1.24981058e+00 3.48138124e-01 -1.03877127e+00
-8.88313949e-01 5.09563088e-01 7.93278366e-02 -1.35630035e+00
-2.76753962e-01 -3.72833371e-01 -1.17238171e-01 5.98363757e-01
-1.59515798e+00 -1.35487497e+00 -3.54549080e-01 9.84863043e-01
9.30770576e-01 -3.73366684e-01 5.84382892e-01 6.58188522e-01
-4.60265338e-01 9.46504772e-01 4.33673859e-02 9.00028050e-02
9.26619112e-01 -5.94713211e-01 5.04085682e-02 9.07403171e-01
4.15833771e-01 6.68952703e-01 3.04038227e-01 -5.32136381e-01
-2.06480765e+00 -1.09837174e+00 6.00364923e-01 -2.66486108e-01
4.94710565e-01 -5.79984665e-01 -6.76948071e-01 6.32200897e-01
4.44773138e-01 7.98581690e-02 4.04148191e-01 -1.81059524e-01
-5.43013990e-01 -1.69844136e-01 -6.87325656e-01 8.85948539e-01
1.04231012e+00 -1.03182673e+00 -3.43026519e-01 3.72925401e-01
7.58014560e-01 -5.81422746e-01 -5.27491152e-01 3.63219291e-01
7.63972104e-01 -5.57513952e-01 1.11399245e+00 -3.93940479e-01
5.13328731e-01 -3.50291103e-01 -5.33837855e-01 -6.68308020e-01
-2.34184295e-01 -1.05125189e+00 -4.63809311e-01 8.91049743e-01
1.78472266e-01 8.82447138e-02 9.78239596e-01 2.21151546e-01
-1.26092449e-01 -6.87957704e-01 -9.49274123e-01 -5.16710520e-01
-3.16218495e-01 -6.43263400e-01 -5.80690987e-02 9.35685754e-01
-1.29651129e-01 2.65664726e-01 -8.97575617e-01 1.54979751e-01
5.21484494e-01 -1.96129128e-01 8.94540191e-01 -4.35499579e-01
-5.03691494e-01 -5.10693967e-01 -1.97659671e-01 -1.87105334e+00
2.72444367e-01 -9.30232167e-01 4.02879305e-02 -1.29101300e+00
4.16791558e-01 -1.30671874e-01 -1.77620396e-01 5.86029172e-01
-3.53872031e-02 5.08040428e-01 5.35874903e-01 5.74256122e-01
-1.12672579e+00 6.45648956e-01 1.03227007e+00 -3.19809169e-01
-1.46918148e-01 -5.84980309e-01 -2.93294132e-01 2.96743929e-01
2.66978145e-01 -4.69296783e-01 -5.25348127e-01 -9.61912155e-01
1.79610759e-01 3.39873523e-01 4.09536064e-01 -9.27383959e-01
4.78443176e-01 7.88166560e-03 4.74249899e-01 -9.51379001e-01
7.49644518e-01 -9.07549441e-01 2.54000947e-02 7.88058043e-02
-4.61274624e-01 2.71194130e-01 4.30167198e-01 4.65593219e-01
-3.54813844e-01 2.92687248e-02 5.20341098e-01 -6.83707967e-02
-7.91451693e-01 4.49974000e-01 -5.91395438e-01 -6.19264282e-02
5.40232480e-01 -2.70686775e-01 -2.54825026e-01 -7.52072573e-01
-5.98214686e-01 5.63356042e-01 5.69134831e-01 4.50158805e-01
1.05913901e+00 -1.33032095e+00 -6.65452778e-01 2.56363034e-01
1.42126054e-01 -1.56033069e-01 4.88532126e-01 9.79747713e-01
-5.29249728e-01 4.48633015e-01 1.74252063e-01 -1.10865748e+00
-1.60661912e+00 5.02574801e-01 2.93726951e-01 -1.58396691e-01
-6.52524650e-01 8.95900726e-01 3.64228547e-01 1.67167038e-02
7.94698358e-01 -4.56804670e-02 2.93326795e-01 -1.81853354e-01
7.88231611e-01 -1.11879416e-01 -9.56096798e-02 -4.84862924e-01
-3.59611839e-01 7.67747998e-01 -4.28852797e-01 -3.97504419e-01
1.06329334e+00 -3.03403437e-01 9.58997011e-02 1.20494291e-01
1.50961316e+00 -6.80402061e-03 -1.16845763e+00 -4.32012200e-01
-4.86527234e-01 -6.13177538e-01 3.91441315e-01 -6.05787039e-01
-1.06354582e+00 8.95405412e-01 6.32695913e-01 -4.85008031e-01
1.30055463e+00 -3.74810747e-03 7.66232967e-01 7.27165937e-01
6.29871488e-02 -1.04272008e+00 5.18283069e-01 5.54401696e-01
7.75441647e-01 -1.16407788e+00 1.98348179e-01 -1.05674885e-01
-7.66089737e-01 9.47700024e-01 5.72988451e-01 6.10565655e-02
3.42292905e-01 2.55309701e-01 -3.46286520e-02 -8.54408890e-02
-1.13788819e+00 -2.51199931e-01 6.10839725e-01 2.74092734e-01
3.25011194e-01 -3.88505965e-01 2.59651303e-01 4.93426174e-02
1.96727082e-01 -7.52715254e-03 5.84952980e-02 9.68394756e-01
-2.68287599e-01 -8.14883173e-01 -3.49271566e-01 2.69780695e-01
-4.39082325e-01 -6.10496342e-01 -2.01371640e-01 6.36744022e-01
-3.85539979e-01 7.92919755e-01 1.94669574e-01 -6.51844859e-01
-7.06824660e-02 -1.01225033e-01 7.04361260e-01 -3.85130674e-01
-4.28195357e-01 6.40383184e-01 5.65696992e-02 -1.00176239e+00
-5.52339852e-01 -3.27844143e-01 -1.04736781e+00 -4.65349227e-01
-1.85114712e-01 2.07975563e-02 5.89478195e-01 9.24699545e-01
6.73775136e-01 3.35355341e-01 6.61238074e-01 -1.19975793e+00
-1.28449008e-01 -8.44846368e-01 -3.94526571e-02 3.13677251e-01
5.18327832e-01 -3.50760788e-01 -3.24361354e-01 2.59272218e-01] | [10.300466537475586, 0.9696208834648132] |
15b30588-84db-4257-b0b9-b4d1d4c4e451 | spoof-face-detection-via-semi-supervised | 2005.10999 | null | https://arxiv.org/abs/2005.10999v1 | https://arxiv.org/pdf/2005.10999v1.pdf | Spoof Face Detection Via Semi-Supervised Adversarial Training | Face spoofing causes severe security threats in face recognition systems. Previous anti-spoofing works focused on supervised techniques, typically with either binary or auxiliary supervision. Most of them suffer from limited robustness and generalization, especially in the cross-dataset setting. In this paper, we propose a semi-supervised adversarial learning framework for spoof face detection, which largely relaxes the supervision condition. To capture the underlying structure of live faces data in latent representation space, we propose to train the live face data only, with a convolutional Encoder-Decoder network acting as a Generator. Meanwhile, we add a second convolutional network serving as a Discriminator. The generator and discriminator are trained by competing with each other while collaborating to understand the underlying concept in the normal class(live faces). Since the spoof face detection is video based (i.e., temporal information), we intuitively take the optical flow maps converted from consecutive video frames as input. Our approach is free of the spoof faces, thus being robust and general to different types of spoof, even unknown spoof. Extensive experiments on intra- and cross-dataset tests show that our semi-supervised method achieves better or comparable results to state-of-the-art supervised techniques. | ['Xuequan Lu', 'Wang Yuan', 'Chengwei Chen', 'Lizhuang Ma'] | 2020-05-22 | null | null | null | null | ['face-presentation-attack-detection', 'gan-image-forensics'] | ['computer-vision', 'computer-vision'] | [ 5.52106321e-01 -1.56313162e-02 -2.25260571e-01 -2.16920510e-01
-9.59891677e-02 -6.52043760e-01 6.73193872e-01 -4.67598081e-01
-3.94038297e-02 4.11760479e-01 -1.25910088e-01 -2.89598852e-01
2.27881387e-01 -8.26734424e-01 -7.68715024e-01 -1.02770948e+00
-2.15077907e-01 1.82205960e-01 3.04274447e-02 -1.10217698e-01
1.92272589e-02 5.89006066e-01 -1.33795381e+00 4.23876613e-01
5.78416705e-01 9.92200553e-01 -2.37696037e-01 6.75299346e-01
2.52018690e-01 9.94998991e-01 -8.30654621e-01 -6.82928622e-01
3.78342986e-01 -5.84074974e-01 -6.02343202e-01 3.01091611e-01
5.21257997e-01 -4.64464992e-01 -8.54540825e-01 1.27358508e+00
4.18888360e-01 -3.87965649e-01 4.62740690e-01 -1.78736246e+00
-6.08629525e-01 2.82216966e-01 -4.93394315e-01 1.38307467e-01
5.76226175e-01 2.32749850e-01 4.24874187e-01 -6.52658045e-01
4.81414407e-01 1.53730333e+00 4.02776212e-01 9.95055497e-01
-1.10205925e+00 -1.14189589e+00 -1.14888065e-01 1.01281472e-01
-1.19793606e+00 -1.04455590e+00 1.27159178e+00 -4.37178969e-01
2.56495059e-01 2.49821180e-03 2.46856272e-01 1.75643444e+00
3.97977196e-02 7.16551065e-01 9.73631918e-01 -2.26436019e-01
-1.54128656e-01 2.96215534e-01 -3.79425436e-01 8.78200233e-01
3.23389262e-01 6.23776793e-01 -4.86260653e-01 -2.58691072e-01
8.56289029e-01 1.65866271e-01 -5.35775006e-01 -5.09631872e-01
-9.26118076e-01 9.72660482e-01 3.66574943e-01 1.35211974e-01
-1.17712490e-01 -6.80591017e-02 5.46658754e-01 5.87113976e-01
2.77458847e-01 -2.98235975e-02 -1.35503694e-01 4.06404138e-01
-1.09780216e+00 -1.02839507e-01 8.82138014e-01 6.42107487e-01
5.77471018e-01 1.92044377e-01 1.10824168e-01 4.49020088e-01
5.57117224e-01 6.60834372e-01 4.81264353e-01 -6.14002407e-01
6.22813284e-01 1.42909139e-01 -8.69959295e-02 -1.54827476e+00
-1.95148122e-02 -2.75486887e-01 -9.56007719e-01 2.19421640e-01
3.67069870e-01 -1.24511451e-01 -9.01344597e-01 1.83440626e+00
2.47688070e-01 6.49467528e-01 1.41187236e-01 6.69476509e-01
4.91710365e-01 4.02184248e-01 -1.65218055e-01 -4.62538093e-01
1.04801261e+00 -1.11603701e+00 -6.94234192e-01 -1.65848359e-01
4.67617244e-01 -8.28273594e-01 4.04019803e-01 3.41159701e-01
-5.85217118e-01 -6.51347816e-01 -9.89916861e-01 4.49248314e-01
-3.55823845e-01 -2.53078877e-03 3.86969388e-01 1.26908267e+00
-8.85773480e-01 5.25097549e-01 -8.33773136e-01 -1.83361620e-02
9.18052614e-01 5.44628203e-01 -7.67114282e-01 -2.81540960e-01
-1.31573331e+00 4.05982196e-01 3.60481262e-01 1.01566389e-01
-1.69268024e+00 -2.64762282e-01 -1.06784022e+00 -1.43540859e-01
2.12624609e-01 -3.96031350e-01 5.43792248e-01 -1.28822815e+00
-1.61050487e+00 9.93972301e-01 -5.17547093e-02 -3.70134622e-01
7.51996279e-01 -2.10601296e-02 -8.19108009e-01 5.36777794e-01
6.67260066e-02 3.93166006e-01 1.71075606e+00 -1.48569942e+00
-2.23164439e-01 -5.07520795e-01 1.23618968e-01 -5.41721225e-01
-6.69783473e-01 2.47400865e-01 -1.87705353e-01 -7.80529916e-01
2.95043942e-02 -1.01090586e+00 3.35166931e-01 1.61894932e-01
-4.79172617e-01 1.08217590e-01 1.64616525e+00 -8.06951344e-01
9.41304147e-01 -2.28362942e+00 -6.40629381e-02 1.53012693e-01
2.01365232e-01 9.07214284e-01 -2.89029509e-01 2.05027759e-01
-4.44442272e-01 1.63345471e-01 -3.40844154e-01 -6.44058108e-01
-4.38501507e-01 2.08944455e-01 -4.46163028e-01 1.09360194e+00
4.27533507e-01 6.62137866e-01 -1.24422252e+00 -6.50549114e-01
2.19355419e-01 9.86861944e-01 -5.19313276e-01 4.21494991e-01
2.23361969e-01 9.53391731e-01 -4.11348850e-01 7.43644774e-01
1.12508380e+00 -7.23977759e-02 1.76910371e-01 -6.67280033e-02
6.00580156e-01 2.47383118e-01 -1.02279699e+00 1.46660686e+00
-5.13933063e-01 7.31696546e-01 4.92530942e-01 -1.25229549e+00
9.38161552e-01 7.00327635e-01 5.02341866e-01 -2.51836330e-01
2.02086926e-01 2.11498305e-01 -8.05918798e-02 -7.12044418e-01
-2.72369325e-01 1.48667479e-02 3.66062254e-01 4.75232154e-01
4.58895922e-01 2.86657929e-01 -1.35474443e-01 1.37466103e-01
1.07491612e+00 1.40545025e-01 -1.14457577e-01 -1.22281648e-02
1.09240329e+00 -7.77733564e-01 6.36600971e-01 5.71551621e-01
-6.08238280e-01 3.77415478e-01 5.19677937e-01 -3.63491654e-01
-7.57162154e-01 -8.70205224e-01 -2.06987724e-01 8.28727543e-01
3.61868143e-01 -4.97231483e-01 -7.92394578e-01 -1.39811742e+00
1.72900222e-02 -3.50052752e-02 -6.84262872e-01 -4.35424030e-01
-7.34749138e-01 -4.22941864e-01 1.01857722e+00 1.44382730e-01
7.35973060e-01 -1.03936124e+00 -4.00572300e-01 7.56541565e-02
-2.00254440e-01 -1.49514377e+00 -2.28726968e-01 -2.96795934e-01
-6.67818904e-01 -1.39196789e+00 -5.28457761e-01 -8.02507162e-01
8.09756279e-01 5.39317846e-01 6.54805958e-01 4.16587532e-01
-1.61473960e-01 2.83049401e-02 -2.55531460e-01 -5.46476319e-02
-7.23079801e-01 -4.39680815e-01 3.55357140e-01 7.24513412e-01
1.02748714e-01 -5.42071164e-01 -5.15325069e-01 6.10954404e-01
-9.41353858e-01 -3.42920333e-01 2.32508838e-01 1.06655312e+00
-8.85895938e-02 3.75892073e-01 2.12741882e-01 -9.40083802e-01
7.66806751e-02 -6.21025503e-01 -3.06580633e-01 1.59411207e-01
-3.83977681e-01 -1.23553008e-01 7.47511983e-01 -5.58797002e-01
-8.71446609e-01 5.93884401e-02 1.07241005e-01 -9.43297386e-01
-3.90272468e-01 -1.85021624e-01 -6.59946799e-01 -4.97570187e-01
4.59486306e-01 2.38889024e-01 1.96399286e-01 -2.84238249e-01
2.40787640e-02 8.00491691e-01 4.95167226e-01 -2.26811603e-01
1.31184876e+00 9.36356783e-01 8.79146233e-02 -8.68440211e-01
-4.84481663e-01 -3.23468536e-01 -6.94131076e-01 -1.75532997e-01
5.01901984e-01 -8.27990234e-01 -7.83337116e-01 9.51120138e-01
-1.36503875e+00 3.66034023e-02 2.10605904e-01 3.46656561e-01
-4.01811033e-01 6.93719387e-01 -7.63082862e-01 -8.54409158e-01
-1.06294952e-01 -1.35389721e+00 1.20356488e+00 6.99434280e-02
4.13061500e-01 -1.21679568e+00 -4.91760634e-02 4.76022005e-01
3.37407798e-01 4.74853694e-01 2.90609270e-01 -7.75039732e-01
-1.97715521e-01 -4.71732497e-01 -2.23966494e-01 7.05700636e-01
5.19823074e-01 1.73529494e-03 -1.41298759e+00 -8.04380298e-01
4.41028446e-01 -4.01024103e-01 1.09176910e+00 -1.84837997e-01
1.23868823e+00 -5.78532279e-01 -5.27614534e-01 9.25157547e-01
1.01070559e+00 3.53545584e-02 6.80003941e-01 -1.16631299e-01
1.01285279e+00 1.00463712e+00 1.88768968e-01 3.00593525e-01
-8.31395909e-02 6.50209963e-01 7.74267495e-01 1.00484550e-01
-8.43350440e-02 -3.83728683e-01 9.53189909e-01 3.51951838e-01
9.82848778e-02 -5.47564149e-01 -7.11987138e-01 2.41300717e-01
-1.47404802e+00 -1.19306457e+00 3.05083990e-01 2.14892030e+00
5.27628899e-01 1.17130235e-01 4.30649519e-02 5.78704476e-01
1.16845286e+00 5.86787522e-01 -3.15336764e-01 -3.28629129e-02
-1.32527173e-01 1.56384304e-01 3.60772938e-01 5.42025089e-01
-1.48811018e+00 1.07853484e+00 5.19674253e+00 7.06219912e-01
-1.48930848e+00 2.99113244e-01 4.57199186e-01 1.76266551e-01
1.50382549e-01 -6.51188865e-02 -6.84598744e-01 6.76683664e-01
8.80573332e-01 4.43622798e-01 5.51954865e-01 5.36723793e-01
-3.02973352e-02 6.51469111e-01 -1.02467358e+00 1.08964264e+00
5.25773346e-01 -9.32066381e-01 9.29203480e-02 3.30323815e-01
6.37836516e-01 -3.24559629e-01 1.70429260e-01 -6.36866838e-02
1.13633402e-01 -1.14874542e+00 3.89299482e-01 1.01692148e-01
9.72722352e-01 -5.52725971e-01 6.82796359e-01 2.78913468e-01
-1.16316926e+00 -1.66105300e-01 -2.20277473e-01 2.77854204e-01
1.47074595e-01 4.25755560e-01 -5.53665519e-01 5.92095733e-01
4.18342173e-01 9.99657512e-01 -4.27123368e-01 4.60869491e-01
-5.80262244e-01 7.79682994e-01 -2.19211847e-01 7.94044435e-01
2.78947991e-03 2.81393111e-01 7.52331078e-01 1.24764645e+00
2.32047737e-02 -3.55088472e-01 2.59661376e-01 7.02209294e-01
-2.52439886e-01 -3.21626365e-01 -9.67731774e-01 -2.70884961e-01
3.57386172e-01 1.06985104e+00 -5.42783499e-01 -2.35102743e-01
-3.06759864e-01 1.37794685e+00 4.18852363e-03 2.80280352e-01
-7.70704508e-01 -4.48404014e-01 8.15728307e-01 6.13051280e-02
3.63777161e-01 -1.19116098e-01 2.99574971e-01 -1.43823314e+00
-2.65508946e-02 -9.69692171e-01 4.58080441e-01 -1.65907130e-01
-1.22239614e+00 7.15784609e-01 -3.93816024e-01 -1.21369195e+00
-3.58796000e-01 -7.93131649e-01 -7.10677207e-01 5.94770551e-01
-1.87881291e+00 -1.32295334e+00 -3.84333044e-01 1.19465649e+00
2.44060948e-01 -5.51507592e-01 8.03477108e-01 4.76658344e-01
-7.61833966e-01 9.58911419e-01 -2.67809480e-01 7.85822153e-01
8.84139776e-01 -6.21327579e-01 2.99607962e-01 1.06760657e+00
3.44844937e-01 7.28906095e-01 3.83460313e-01 -6.90988779e-01
-1.46101069e+00 -1.09439552e+00 7.13471115e-01 -2.67373979e-01
6.34573281e-01 -4.21534121e-01 -8.92942965e-01 6.02234244e-01
2.78991479e-02 8.57853830e-01 6.74741864e-01 -6.24112666e-01
-9.57129121e-01 -6.93854541e-02 -1.47077644e+00 -1.69816669e-02
1.13504827e+00 -1.17898560e+00 -2.01718017e-01 4.61973339e-01
5.76377928e-01 -1.36915356e-01 -4.57287997e-01 4.20550734e-01
6.75276518e-01 -1.16260743e+00 1.16027725e+00 -8.16426575e-01
4.71284181e-01 -1.91358536e-01 -9.44519192e-02 -1.04459822e+00
2.51909979e-02 -1.02723372e+00 -6.24909222e-01 1.29439282e+00
-1.29530832e-01 -7.03633249e-01 9.25517023e-01 -2.27694318e-01
3.93390775e-01 -3.31992388e-01 -9.42686856e-01 -9.47003067e-01
-1.66552380e-01 -2.83604801e-01 6.61342919e-01 1.21695697e+00
-1.77498505e-01 -6.70360401e-02 -8.37698519e-01 6.48676097e-01
1.02506399e+00 -1.56527147e-01 7.30849385e-01 -1.16852355e+00
-3.91043365e-01 -1.40807405e-01 -7.49826372e-01 -1.07715881e+00
6.86489820e-01 -7.48767138e-01 -2.44070992e-01 -4.84101772e-01
-1.25381306e-01 -3.50495785e-01 -4.91398692e-01 3.21330160e-01
-4.24823537e-02 7.02637911e-01 1.75989687e-01 4.30591226e-01
-1.48997337e-01 3.93558472e-01 1.29810619e+00 -4.15882200e-01
1.71543472e-02 1.96257055e-01 -2.69873083e-01 7.01898515e-01
7.23290026e-01 -6.76697850e-01 -3.77852350e-01 -4.50117201e-01
-4.05321091e-01 2.03270257e-01 7.66708612e-01 -1.05650985e+00
2.49515027e-01 -3.79457399e-02 3.87194812e-01 1.70026541e-01
2.88281292e-01 -9.95858967e-01 -2.43581399e-01 8.24110627e-01
-2.42049381e-01 -3.11744958e-01 -1.33585766e-01 6.62944674e-01
-3.14519733e-01 -8.43724534e-02 1.04143810e+00 6.56016171e-02
-1.86116695e-01 6.94553792e-01 1.09821588e-01 -1.74383357e-01
1.01910782e+00 -1.76158339e-01 -4.78726953e-01 -4.12657678e-01
-7.55094409e-01 -1.31960198e-01 4.96180028e-01 6.61794364e-01
7.68863797e-01 -1.22188365e+00 -7.53027081e-01 8.90789747e-01
6.71342239e-02 -4.06185299e-01 -2.85075214e-02 5.37705421e-01
-3.42640907e-01 4.10249174e-01 -4.55998331e-01 -5.81855655e-01
-1.22212827e+00 1.00831485e+00 3.64728302e-01 -1.67231128e-01
-2.85915971e-01 9.40594435e-01 2.87155986e-01 -1.78054020e-01
2.39101201e-01 5.73093295e-01 -2.72404432e-01 6.62600473e-02
7.91285753e-01 3.18521887e-01 -2.37455547e-01 -1.33953667e+00
-5.62348068e-01 4.65953827e-01 1.01980148e-02 1.06097832e-01
9.89876330e-01 -7.15664104e-02 -1.80969238e-01 -1.01351902e-01
1.73529005e+00 1.27593443e-01 -1.37532818e+00 -3.71928424e-01
-2.99299538e-01 -9.02371228e-01 -1.32627059e-02 -1.69719398e-01
-1.65040469e+00 1.22032130e+00 6.86917901e-01 3.05379540e-01
1.06184483e+00 -1.80099055e-01 7.74630547e-01 5.98844700e-02
5.21048725e-01 -3.69393229e-01 3.59355509e-01 1.25639305e-01
6.15272939e-01 -1.49921477e+00 -3.44116151e-01 -7.30233252e-01
-2.29955927e-01 1.19095659e+00 5.63282073e-01 -1.49801567e-01
9.37017441e-01 1.62085757e-01 3.21874209e-02 -9.65760276e-02
-3.03239107e-01 1.99863881e-01 -5.74856400e-02 9.58205700e-01
1.36389419e-01 -6.92778081e-02 3.26131910e-01 -3.41880769e-02
-1.46685898e-01 -1.01302065e-01 1.65598273e-01 9.72572267e-01
7.88905695e-02 -1.37446105e+00 -5.35287797e-01 -2.83344891e-02
-6.61289096e-01 2.46263742e-01 -3.38604420e-01 3.31999779e-01
4.52264398e-01 1.35603368e+00 -4.13652323e-02 -7.25109875e-01
-1.99748501e-01 -9.68226567e-02 4.72679079e-01 -4.51405615e-01
-3.23478222e-01 -1.78273574e-01 -3.80712301e-01 -7.72750258e-01
-6.85212374e-01 -5.22277474e-01 -6.88155293e-01 -5.32766223e-01
-5.88307619e-01 1.45569116e-01 5.86088538e-01 9.05978203e-01
1.83020890e-01 1.90428138e-01 1.43469989e+00 -9.49553609e-01
-4.87506539e-01 -8.22772026e-01 -4.74816561e-01 6.55170321e-01
1.12668347e+00 -8.32216740e-01 -7.25673795e-01 2.97724843e-01] | [13.039645195007324, 1.1734243631362915] |
1f3fbe2d-fb73-4466-9374-6b3bd03cc421 | neural-face-editing-with-intrinsic-image | 1704.04131 | null | http://arxiv.org/abs/1704.04131v1 | http://arxiv.org/pdf/1704.04131v1.pdf | Neural Face Editing with Intrinsic Image Disentangling | Traditional face editing methods often require a number of sophisticated and
task specific algorithms to be applied one after the other --- a process that
is tedious, fragile, and computationally intensive. In this paper, we propose
an end-to-end generative adversarial network that infers a face-specific
disentangled representation of intrinsic face properties, including shape (i.e.
normals), albedo, and lighting, and an alpha matte. We show that this network
can be trained on "in-the-wild" images by incorporating an in-network
physically-based image formation module and appropriate loss functions. Our
disentangling latent representation allows for semantically relevant edits,
where one aspect of facial appearance can be manipulated while keeping
orthogonal properties fixed, and we demonstrate its use for a number of facial
editing applications. | ['Eli Shechtman', 'Sunil Hadap', 'Ersin Yumer', 'Kalyan Sunkavalli', 'Zhixin Shu', 'Dimitris Samaras'] | 2017-04-13 | neural-face-editing-with-intrinsic-image-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Shu_Neural_Face_Editing_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Shu_Neural_Face_Editing_CVPR_2017_paper.pdf | cvpr-2017-7 | ['facial-editing'] | ['computer-vision'] | [ 5.84918082e-01 4.12596226e-01 4.12285239e-01 -5.94012678e-01
-5.19583642e-01 -6.94788694e-01 6.00100398e-01 -6.33021593e-01
-6.97583109e-02 6.69776857e-01 -4.18100208e-02 2.63712313e-02
5.53464442e-02 -7.04764009e-01 -8.99309456e-01 -8.43704998e-01
7.98254982e-02 4.02174085e-01 -3.89401793e-01 -2.74516940e-01
-3.27153224e-03 9.43555355e-01 -1.37786472e+00 -1.14283644e-01
5.01497567e-01 6.33982897e-01 -4.06014770e-01 9.26488519e-01
4.40382630e-01 5.79137146e-01 -4.88363385e-01 -7.87473083e-01
6.40010118e-01 -5.69090307e-01 -3.46918046e-01 2.30554730e-01
1.00659728e+00 -5.31481981e-01 -4.70205486e-01 1.00014639e+00
5.14583707e-01 2.17473194e-01 8.17909300e-01 -1.32607341e+00
-8.59509945e-01 -1.29247248e-01 -6.46206081e-01 -4.68249142e-01
3.40899944e-01 3.44887137e-01 6.38595760e-01 -9.43937898e-01
7.65457153e-01 1.35543311e+00 7.59310007e-01 8.09086680e-01
-1.80870140e+00 -9.21358407e-01 -2.73691684e-01 -5.32570124e-01
-1.34580207e+00 -8.39637995e-01 1.11348939e+00 -5.25653422e-01
4.71664041e-01 3.96804899e-01 5.84096909e-01 1.16405189e+00
2.40735143e-01 -4.48160851e-03 1.11226714e+00 -5.06648362e-01
-4.07274552e-02 -2.64986306e-02 -6.88840568e-01 1.25006866e+00
1.53741539e-01 2.47165784e-01 -4.21934694e-01 -2.12282956e-01
1.28611588e+00 -1.09539129e-01 -2.76113242e-01 -7.78425097e-01
-8.07601333e-01 8.30105245e-01 1.38958842e-01 -4.00206327e-01
-1.16958752e-01 5.08685172e-01 -8.29377323e-02 3.74710828e-01
7.53734052e-01 5.32371342e-01 -2.18983918e-01 2.25265563e-01
-1.08745492e+00 4.46100205e-01 8.44220459e-01 7.13468611e-01
1.03713274e+00 2.13759080e-01 -4.21494506e-02 8.92348468e-01
3.27998221e-01 7.19596386e-01 -1.66278660e-01 -1.39308393e+00
6.64086714e-02 7.65603632e-02 9.57203433e-02 -1.12661362e+00
-1.47798760e-02 1.38171706e-02 -7.27234662e-01 1.01825917e+00
3.67496789e-01 -3.58312190e-01 -1.28637052e+00 2.20273256e+00
3.82887632e-01 2.87141889e-01 -3.29335183e-01 6.20445549e-01
5.03100574e-01 2.88315475e-01 -1.66563764e-01 -3.64822149e-02
1.17913520e+00 -7.28442371e-01 -6.66672707e-01 -1.98253632e-01
-2.43472993e-01 -1.12738287e+00 9.41459596e-01 1.76190197e-01
-1.59438217e+00 -2.50042975e-01 -1.10059643e+00 -5.47950327e-01
-1.48926198e-01 1.61090493e-01 5.38335145e-01 7.70785332e-01
-1.28215611e+00 8.42556953e-01 -8.84467244e-01 -1.94296136e-01
4.96648014e-01 6.30537629e-01 -7.44649708e-01 4.36382182e-02
-7.31579363e-01 8.35100114e-01 -2.32936054e-01 1.84532017e-01
-1.07841885e+00 -9.33106899e-01 -1.02443671e+00 3.77390422e-02
2.65824020e-01 -1.20108569e+00 1.00911820e+00 -1.12789011e+00
-2.06329942e+00 1.07534683e+00 -1.06691331e-01 3.26349199e-01
7.47771263e-01 -1.29997477e-01 -1.11622572e-01 2.98893213e-01
-8.21853802e-02 7.18497396e-01 1.58625090e+00 -1.45682859e+00
2.30712071e-01 -4.77871716e-01 9.73473266e-02 1.99871004e-01
-1.93234578e-01 1.28695354e-01 -5.22902369e-01 -1.04331481e+00
4.16355021e-03 -1.10831249e+00 2.39623562e-02 9.86772954e-01
-4.21380103e-01 6.39173627e-01 9.83406842e-01 -9.21956480e-01
4.87270474e-01 -2.11004782e+00 3.71563405e-01 3.42202395e-01
3.94078314e-01 1.12155847e-01 -3.68128330e-01 1.94863170e-01
-4.10142779e-01 1.85582414e-01 -3.05445939e-01 -7.45080352e-01
-8.35578740e-02 2.08185181e-01 -2.11228222e-01 6.02420092e-01
4.87000078e-01 8.48703325e-01 -7.69844472e-01 -3.10829043e-01
1.37972906e-01 1.04486609e+00 -7.97603428e-01 3.74370635e-01
-2.45120764e-01 7.73344874e-01 -9.95333716e-02 4.01188642e-01
8.30547869e-01 1.87627137e-01 1.06800236e-01 -3.42072278e-01
2.59194076e-01 1.44671891e-02 -8.40293407e-01 1.77761829e+00
-6.47995234e-01 7.99992323e-01 5.46844065e-01 -4.30802464e-01
7.65686929e-01 3.88297737e-01 3.80600601e-01 -1.36394635e-01
2.29096085e-01 -8.59589130e-02 -2.09053844e-01 -3.34087104e-01
1.71104804e-01 -4.30262566e-01 3.18955719e-01 5.91764390e-01
1.42783746e-01 -6.53503299e-01 -2.37091839e-01 6.71300068e-02
9.14003491e-01 4.30153161e-01 -2.99817584e-02 -2.35661909e-01
1.27493694e-01 -6.17551386e-01 4.10346180e-01 2.74308592e-01
2.56491989e-01 9.85821009e-01 6.36852503e-01 -2.52109915e-01
-1.31815088e+00 -1.35562932e+00 5.11024147e-02 8.45390379e-01
-2.72698551e-01 -3.17535065e-02 -9.75732803e-01 -4.24907386e-01
2.09302574e-01 5.29583037e-01 -9.16347384e-01 -3.91527861e-01
-5.56828558e-01 -3.46129447e-01 7.47210741e-01 3.48358363e-01
2.63106853e-01 -7.03481317e-01 -1.65823981e-01 -1.36316672e-01
1.66456908e-01 -1.06088769e+00 -8.82320583e-01 -2.00554699e-01
-6.21526122e-01 -1.01573503e+00 -5.61876416e-01 -5.45463622e-01
1.20072389e+00 2.94557959e-02 1.09271610e+00 9.65325683e-02
-5.09164631e-01 3.68491739e-01 2.59656280e-01 -4.60518271e-01
-4.82954293e-01 -3.78278404e-01 2.89179683e-02 3.18347573e-01
-4.20779526e-01 -1.14847207e+00 -6.90299571e-01 2.67782182e-01
-8.67483020e-01 1.57362223e-01 8.21768641e-02 7.73217261e-01
4.56296951e-01 -2.81105489e-01 1.83235601e-01 -1.07845843e+00
5.10692239e-01 9.58338752e-02 -6.79719985e-01 1.87191293e-01
-2.74108142e-01 9.63682011e-02 4.51262265e-01 -4.37519908e-01
-1.30059659e+00 2.64201730e-01 -8.18075016e-02 -7.14425266e-01
1.03573337e-01 -6.99363947e-02 -5.88702202e-01 -6.85437679e-01
6.94965363e-01 -1.32118776e-01 5.82986236e-01 -2.33885601e-01
6.92768097e-01 1.34958578e-02 7.57156610e-01 -8.28003645e-01
1.36745715e+00 7.45305300e-01 3.42468143e-01 -7.78293431e-01
-7.48327136e-01 3.72095644e-01 -7.32622325e-01 -5.55259921e-02
8.27541947e-01 -8.42757940e-01 -6.52779579e-01 6.76051736e-01
-1.19539070e+00 -4.35884088e-01 -4.47874010e-01 1.23275369e-01
-6.47056460e-01 2.62273997e-01 -5.33638120e-01 -4.94808793e-01
-2.61893749e-01 -1.10092020e+00 1.21203065e+00 1.87521800e-01
-1.80289939e-01 -9.68135834e-01 2.98191868e-02 3.56764376e-01
4.06360209e-01 6.94557190e-01 8.11742067e-01 1.51765481e-01
-7.32851446e-01 -1.46458805e-01 -2.62455136e-01 5.17333031e-01
3.46334547e-01 4.71374243e-01 -1.19344807e+00 -4.32299525e-01
-6.41244948e-02 -2.60616273e-01 6.28090978e-01 2.06352606e-01
1.24154747e+00 -6.32601976e-01 -1.64502747e-02 1.25277019e+00
1.17588639e+00 -1.61287606e-01 8.20875585e-01 -4.04394418e-01
1.00861764e+00 4.26566333e-01 -5.15089072e-02 3.38527560e-01
-2.85863765e-02 6.96146905e-01 3.74717921e-01 -2.93969721e-01
-3.85361880e-01 -2.93826729e-01 3.20302725e-01 3.66216451e-01
-3.19282651e-01 -1.25031814e-01 -2.72906333e-01 1.30061418e-01
-1.43852007e+00 -1.03410017e+00 3.78224164e-01 2.11395264e+00
1.04748762e+00 -1.47657186e-01 -1.73114359e-01 -3.40882182e-01
4.85578656e-01 2.96684384e-01 -6.49094880e-01 -4.43849534e-01
8.49448293e-02 9.92166877e-01 3.22519481e-01 8.59127641e-01
-8.95490229e-01 9.57611799e-01 6.85087585e+00 4.49196219e-01
-1.28274238e+00 1.26679316e-02 4.72498745e-01 -3.56667161e-01
-7.38354623e-01 1.53091833e-01 -2.85782486e-01 1.53102472e-01
3.93075943e-01 1.81558952e-02 8.90728176e-01 4.95065510e-01
1.99056596e-01 1.24536611e-01 -1.20130622e+00 8.50305378e-01
4.16569084e-01 -1.29059327e+00 1.71779960e-01 1.69230998e-01
6.66886210e-01 -4.63746756e-01 4.12679315e-01 -2.42112920e-01
4.83596981e-01 -1.32377970e+00 7.16190875e-01 7.85899758e-01
1.46193504e+00 -6.94263935e-01 -1.03160545e-01 -1.19737372e-01
-7.61639535e-01 4.04434115e-01 -4.04202379e-02 2.40531147e-01
7.62357339e-02 5.11396825e-01 -4.98154134e-01 1.98944271e-01
3.21563005e-01 2.83606857e-01 -3.00520301e-01 3.78668010e-01
-6.30067170e-01 2.83713192e-01 -4.24896598e-01 5.53278804e-01
-3.84708881e-01 -5.24942160e-01 7.26721585e-01 8.35436344e-01
3.60781640e-01 2.09183186e-01 -2.21156612e-01 1.26250291e+00
-5.51112771e-01 -3.49152178e-01 -7.56740093e-01 -2.00163387e-02
2.73931623e-01 1.49428356e+00 -4.92878258e-01 6.47969618e-02
-1.02089606e-02 1.38645291e+00 3.46802324e-01 6.49837077e-01
-8.55225503e-01 -4.92840827e-01 1.11976087e+00 2.89510041e-01
2.42660433e-01 -3.89635444e-01 -2.13155672e-01 -1.29221904e+00
8.92859697e-02 -8.99528384e-01 -2.76888430e-01 -1.07379401e+00
-1.18860877e+00 3.88606519e-01 -8.87045860e-02 -6.45288885e-01
-3.10346872e-01 -6.54298902e-01 -7.98814178e-01 1.15174925e+00
-1.31797206e+00 -1.56306219e+00 -4.16204900e-01 7.36059248e-01
3.22729424e-02 -9.57797319e-02 8.69469345e-01 2.94258595e-01
-6.20507360e-01 9.96827960e-01 -4.45479378e-02 2.91877002e-01
8.94817233e-01 -1.02672184e+00 4.88507718e-01 8.15275133e-01
1.01022087e-01 8.87620986e-01 8.40801060e-01 -4.38284218e-01
-1.44688857e+00 -1.06570375e+00 5.86771131e-01 -5.39033115e-01
3.29339355e-01 -7.34542251e-01 -6.06043875e-01 9.67501104e-01
2.41737634e-01 3.43430847e-01 6.23581707e-01 -1.44234434e-01
-7.93610394e-01 -1.87551662e-01 -1.45886922e+00 9.41695213e-01
1.14660180e+00 -9.29052532e-01 -5.60697801e-02 4.25589949e-01
6.10227346e-01 -6.36695027e-01 -7.60918498e-01 1.23092160e-01
9.97952640e-01 -9.04480159e-01 1.20948160e+00 -6.72702789e-01
3.18229049e-01 -2.76920289e-01 1.32946046e-02 -1.45260191e+00
-1.98224425e-01 -1.22772551e+00 4.82268520e-02 1.15799129e+00
1.73676625e-01 -6.82793260e-01 8.19505274e-01 8.81182373e-01
1.78957749e-02 -5.76289117e-01 -7.08762884e-01 -4.46616203e-01
-8.86261649e-03 -6.80866688e-02 6.37589037e-01 1.02470016e+00
-6.88808799e-01 2.25062504e-01 -5.41961074e-01 2.94511765e-01
7.24239528e-01 -1.07757956e-01 1.02129483e+00 -1.15376329e+00
-3.09746623e-01 -1.85894102e-01 -3.18009704e-01 -6.26867771e-01
4.07140046e-01 -6.74647391e-01 -1.01648375e-01 -8.77552569e-01
8.07389691e-02 -2.41433978e-01 3.24695021e-01 7.01381505e-01
-2.13288739e-02 5.53506136e-01 9.49044004e-02 -5.94264641e-02
2.88485080e-01 6.53508067e-01 1.48380744e+00 1.92686766e-02
2.08766814e-02 -2.11989895e-01 -7.32074738e-01 9.54034388e-01
5.87657034e-01 -4.36230212e-01 -6.40568674e-01 -5.45102060e-01
2.05632836e-01 2.83394586e-02 7.21347094e-01 -7.92143345e-01
-6.36443868e-02 -2.99810618e-01 6.48864686e-01 1.16900653e-01
9.42551255e-01 -7.06652820e-01 4.17183131e-01 -7.79657736e-02
-2.43218243e-01 7.70933321e-03 2.56197959e-01 3.86376262e-01
1.80326328e-01 -3.76891047e-02 1.13936114e+00 -4.44080532e-02
1.60666957e-01 7.09470809e-01 1.17714226e-01 1.14183351e-01
9.53337371e-01 -2.72636294e-01 -1.35816425e-01 -6.44001544e-01
-7.93301761e-01 -3.65702480e-01 9.02938426e-01 1.20238163e-01
6.72082663e-01 -1.34275973e+00 -7.89611995e-01 7.65158117e-01
-2.96618730e-01 6.18625544e-02 8.97399783e-02 3.37609887e-01
-9.80726123e-01 -5.43210924e-01 -4.29856181e-01 -1.99103504e-01
-1.43868256e+00 2.53388077e-01 7.14662433e-01 -3.48909833e-02
-4.40503120e-01 9.08086598e-01 2.84446716e-01 -5.37906766e-01
-1.77164301e-01 2.21958309e-01 5.22772014e-01 -1.92003772e-01
1.89719975e-01 2.55247280e-02 -4.58784141e-02 -6.56793416e-01
-3.73107046e-02 8.41288090e-01 7.11471364e-02 -4.50113922e-01
1.33983004e+00 -4.11388204e-02 -4.14644748e-01 -1.28016815e-01
1.37996638e+00 3.81705880e-01 -1.74924028e+00 1.84870228e-01
-8.38581800e-01 -7.93963253e-01 5.30602075e-02 -6.68791533e-01
-1.52959776e+00 9.73402262e-01 2.99631029e-01 -2.16774926e-01
1.17894900e+00 -3.99344116e-01 6.52546704e-01 1.48076385e-01
1.46929353e-01 -7.59962142e-01 2.22805575e-01 2.39192605e-01
1.30227506e+00 -8.02554846e-01 2.97265828e-01 -7.42128015e-01
-2.04279274e-01 9.92736280e-01 3.82282376e-01 -2.68465668e-01
9.82547522e-01 4.49333936e-01 1.00499921e-01 -2.47895449e-01
-3.89686108e-01 3.25955719e-01 4.82696265e-01 7.28058994e-01
3.48043084e-01 -1.56750288e-02 2.70954102e-01 -1.81548879e-01
-3.66290987e-01 -1.45617500e-01 3.61215651e-01 8.06093991e-01
1.57683417e-01 -1.32777500e+00 -2.42965966e-01 2.69288272e-01
-3.78858149e-01 -5.00177406e-02 -4.22579855e-01 7.75595307e-01
2.52792388e-01 5.20502806e-01 3.03785726e-02 -1.58635825e-01
2.01730892e-01 1.27390280e-01 1.13367343e+00 -7.14224637e-01
-3.93174469e-01 -4.20049690e-02 2.14499468e-03 -7.06403553e-01
-3.00444037e-01 -5.99000514e-01 -9.59223449e-01 -6.33997977e-01
-2.32131183e-01 -5.02336442e-01 6.64817035e-01 6.85536623e-01
6.08079314e-01 3.31794173e-01 6.98349893e-01 -1.16148591e+00
-3.96979094e-01 -6.36002541e-01 -5.96392930e-01 5.80401838e-01
5.29040575e-01 -8.39281917e-01 -3.34002316e-01 4.18543130e-01] | [12.617537498474121, -0.3037738800048828] |
ca8d7db6-de26-4798-a330-1f1d76e0acda | improving-video-text-retrieval-by-multi | 2109.04290 | null | https://arxiv.org/abs/2109.04290v3 | https://arxiv.org/pdf/2109.04290v3.pdf | Improving Video-Text Retrieval by Multi-Stream Corpus Alignment and Dual Softmax Loss | Employing large-scale pre-trained model CLIP to conduct video-text retrieval task (VTR) has become a new trend, which exceeds previous VTR methods. Though, due to the heterogeneity of structures and contents between video and text, previous CLIP-based models are prone to overfitting in the training phase, resulting in relatively poor retrieval performance. In this paper, we propose a multi-stream Corpus Alignment network with single gate Mixture-of-Experts (CAMoE) and a novel Dual Softmax Loss (DSL) to solve the two heterogeneity. The CAMoE employs Mixture-of-Experts (MoE) to extract multi-perspective video representations, including action, entity, scene, etc., then align them with the corresponding part of the text. In this stage, we conduct massive explorations towards the feature extraction module and feature alignment module. DSL is proposed to avoid the one-way optimum-match which occurs in previous contrastive methods. Introducing the intrinsic prior of each pair in a batch, DSL serves as a reviser to correct the similarity matrix and achieves the dual optimal match. DSL is easy to implement with only one-line code but improves significantly. The results show that the proposed CAMoE and DSL are of strong efficiency, and each of them is capable of achieving State-of-The-Art (SOTA) individually on various benchmarks such as MSR-VTT, MSVD, and LSMDC. Further, with both of them, the performance is advanced to a big extend, surpassing the previous SOTA methods for around 4.6\% R@1 in MSR-VTT. | ['Dong Shen', 'Fan Yang', 'Xiangyu Wu', 'Hezheng Lin', 'Xing Cheng'] | 2021-09-09 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 1.25461131e-01 -6.22507572e-01 -1.79200828e-01 -1.05276167e-01
-1.12496352e+00 -3.32293868e-01 7.97524333e-01 -8.40762928e-02
-4.71054614e-01 2.30366990e-01 2.36601070e-01 6.85337465e-03
-1.49852872e-01 -2.26461962e-01 -5.65832436e-01 -6.50600970e-01
2.93095708e-01 4.61194605e-01 2.44533852e-01 -2.70245701e-01
3.70597899e-01 1.32187292e-01 -1.53854859e+00 5.33617079e-01
7.24785805e-01 1.19041610e+00 5.28000653e-01 5.29760778e-01
-3.41048747e-01 7.96958089e-01 -3.40863645e-01 -6.08371198e-01
4.70583409e-01 -1.38316095e-01 -5.57854712e-01 9.73497797e-03
5.53550601e-01 -4.90257770e-01 -8.07252228e-01 9.02891695e-01
7.54671097e-01 1.86226293e-01 7.28119373e-01 -1.11501396e+00
-4.46044475e-01 5.05069792e-01 -1.11824608e+00 1.84559479e-01
4.64738578e-01 4.08343971e-02 1.25517154e+00 -1.33683383e+00
6.26879930e-01 1.25065351e+00 4.24755603e-01 3.04987431e-01
-6.23905957e-01 -6.56323016e-01 3.56701732e-01 3.59319210e-01
-1.70353889e+00 -6.37156606e-01 5.98857701e-01 -2.87833720e-01
1.11160743e+00 3.23936790e-01 4.53327507e-01 1.08032954e+00
1.23845182e-01 1.41027474e+00 5.71213961e-01 -2.84291565e-01
-8.73257145e-02 1.37057394e-01 -1.03690356e-01 6.58933520e-01
-1.92204192e-02 -4.77821887e-01 -7.26699173e-01 -4.73806262e-02
6.75646782e-01 1.96086675e-01 -3.13056260e-01 -1.67077750e-01
-1.30290699e+00 4.97350395e-01 3.75749357e-02 4.27311361e-01
-2.32787043e-01 5.95621914e-02 5.84885836e-01 3.49002033e-01
4.41529393e-01 2.44422749e-01 -3.28719139e-01 -2.88134277e-01
-1.34828901e+00 1.50746927e-01 5.17239332e-01 1.20373547e+00
4.11933899e-01 -3.87916751e-02 -4.67812002e-01 1.14969432e+00
4.06553000e-01 6.40487373e-01 7.59188294e-01 -6.26346648e-01
1.01610005e+00 6.11075044e-01 -5.09695187e-02 -1.18333709e+00
-9.32456404e-02 -5.54212689e-01 -8.76792252e-01 -4.49907124e-01
2.05393918e-02 1.96041733e-01 -1.03558481e+00 1.37031472e+00
1.42601833e-01 3.17768097e-01 3.68456542e-02 9.80785131e-01
9.54369009e-01 1.00918746e+00 -2.04357490e-01 -3.49321365e-01
1.27749145e+00 -1.28678954e+00 -7.62232661e-01 -3.95161659e-01
6.93966806e-01 -1.12688541e+00 1.01021135e+00 4.18802649e-01
-1.22685635e+00 -5.94619989e-01 -1.00965011e+00 -2.09270209e-01
-1.47488564e-01 4.89044249e-01 4.08901453e-01 1.66942343e-01
-9.90457416e-01 3.94502312e-01 -6.87297285e-01 -2.55964756e-01
3.76313269e-01 2.26175934e-01 -4.31369096e-01 -3.16781312e-01
-1.16176558e+00 6.52085006e-01 2.57722497e-01 2.07873300e-01
-7.84295440e-01 -6.01915061e-01 -6.50587440e-01 1.90584943e-01
5.23662567e-01 -7.07585156e-01 1.01618195e+00 -8.04390371e-01
-1.43292499e+00 7.52382696e-01 -3.11134666e-01 -8.27489197e-02
6.33770943e-01 -4.77286607e-01 -5.68047762e-01 3.50118697e-01
1.13625318e-01 6.63061500e-01 1.20609462e+00 -1.08621562e+00
-7.79976010e-01 -2.09406391e-01 -9.79064032e-03 5.29511154e-01
-6.51690781e-01 3.15104693e-01 -1.32887387e+00 -9.02103841e-01
1.64919659e-01 -8.01787078e-01 -3.03938258e-02 -3.68836783e-02
-2.99416572e-01 -2.55345196e-01 6.15941167e-01 -7.76113391e-01
1.89421701e+00 -2.38219523e+00 4.06099945e-01 5.24336137e-02
2.85276592e-01 5.69161296e-01 -3.53226066e-01 5.37403166e-01
6.19671047e-02 -5.68701997e-02 -3.60355377e-02 -7.16674805e-01
1.33048803e-01 -9.05230120e-02 -4.03992504e-01 4.15002137e-01
4.43164408e-02 8.58299136e-01 -7.86338747e-01 -9.21251833e-01
1.88931137e-01 2.74286360e-01 -5.05842090e-01 2.86834002e-01
-2.08370522e-01 -5.91651350e-02 -6.43436313e-01 7.25729525e-01
6.77746177e-01 -4.80705768e-01 1.88570246e-02 -4.21541959e-01
-7.95881078e-02 8.10668170e-02 -1.41705394e+00 1.94758964e+00
-3.22113723e-01 5.91429353e-01 -2.65614521e-02 -9.72657979e-01
7.82552421e-01 4.54006672e-01 6.95528686e-01 -8.03695798e-01
2.57605165e-01 3.05670708e-01 -3.02151620e-01 -7.36074328e-01
7.57566750e-01 3.35824519e-01 -6.42063469e-02 4.88508791e-01
1.87552258e-01 2.50815719e-01 3.71928453e-01 4.71256286e-01
1.03418493e+00 1.52328357e-01 1.70878664e-01 7.05194846e-02
7.08394051e-01 -3.03504437e-01 6.23834014e-01 7.58381486e-01
-1.13746291e-02 8.74876678e-01 1.94179669e-01 -2.15927705e-01
-1.01453364e+00 -7.92493284e-01 1.03238121e-01 1.01953661e+00
4.04622644e-01 -7.53393590e-01 -5.83346963e-01 -5.59740663e-01
-2.03449383e-01 3.69426370e-01 -2.70388067e-01 -2.30887234e-01
-7.71624684e-01 -7.11645067e-01 5.07907331e-01 4.24540669e-01
6.56808794e-01 -7.62352228e-01 -1.79203361e-01 1.14770740e-01
-5.30477464e-01 -1.44474697e+00 -8.60300124e-01 -2.31880754e-01
-6.26016974e-01 -7.29337275e-01 -9.85949099e-01 -7.95192003e-01
4.36717004e-01 6.80418551e-01 9.41687047e-01 1.81296378e-01
-2.07269475e-01 3.36788774e-01 -7.19512463e-01 -1.76664554e-02
-2.24437546e-02 1.14268512e-01 3.47518064e-02 2.89768547e-01
4.35673833e-01 -3.73867542e-01 -7.40382731e-01 4.47841346e-01
-1.10678148e+00 8.12641382e-02 6.73013926e-01 7.65517294e-01
6.93779409e-01 -1.29122660e-01 2.19706923e-01 -4.98708695e-01
4.09035981e-01 -6.18705273e-01 -3.53070468e-01 5.59287786e-01
-5.89268148e-01 -1.56901866e-01 6.72298312e-01 -6.18357718e-01
-9.51710820e-01 -2.05357954e-01 -2.83097941e-02 -8.60160351e-01
2.19560698e-01 6.71567678e-01 -2.04109728e-01 1.24485850e-01
1.38970435e-01 4.75908726e-01 -2.81407237e-01 -5.25711834e-01
1.39266342e-01 9.11405027e-01 2.59835094e-01 -4.78169560e-01
8.74233007e-01 3.15175921e-01 -2.55664647e-01 -6.96981788e-01
-8.37680459e-01 -8.27012897e-01 -4.49480325e-01 -1.72082141e-01
6.07455432e-01 -1.26433003e+00 -3.10922384e-01 6.28030241e-01
-9.41523552e-01 -9.83765721e-02 1.22075506e-01 5.76492608e-01
-3.73977125e-01 6.97323680e-01 -6.71298325e-01 -5.09239435e-01
-5.46985865e-01 -1.33712339e+00 1.29578435e+00 1.60269916e-01
1.00751795e-01 -6.89458728e-01 -6.79804534e-02 5.21630347e-01
2.78788418e-01 -3.53203207e-01 6.57117009e-01 -8.19392800e-01
-7.03632593e-01 -4.34112608e-01 -3.87532800e-01 2.33442158e-01
-1.38817921e-01 1.59149557e-01 -7.76643932e-01 -6.55160189e-01
-9.52324942e-02 -1.99613690e-01 8.57383251e-01 2.83617944e-01
1.14708018e+00 -1.39300391e-01 -4.98155028e-01 7.63261259e-01
1.41779590e+00 3.05652022e-01 7.69329190e-01 3.73897314e-01
7.10552931e-01 2.19279587e-01 9.64718878e-01 6.82220161e-01
4.17026848e-01 1.02185798e+00 2.15367511e-01 -7.94271678e-02
-1.49222270e-01 -1.88938648e-01 5.96571863e-01 1.33221114e+00
-2.83012781e-02 -6.54309988e-01 -6.34017229e-01 4.78464127e-01
-1.99155509e+00 -8.83141577e-01 3.36395786e-03 2.08220649e+00
6.75780654e-01 5.44122644e-02 -6.66545630e-02 -6.41030446e-02
6.68863177e-01 4.65605706e-01 -3.45043987e-01 -2.96126995e-02
-2.28669181e-01 1.22687437e-01 2.21991956e-01 2.10997403e-01
-1.10867333e+00 1.08038366e+00 5.64323282e+00 1.43855727e+00
-1.14347613e+00 6.51418194e-02 2.77270377e-01 -4.85125989e-01
-1.37261018e-01 -7.15140104e-02 -1.02646041e+00 5.68785429e-01
6.61750078e-01 -1.17700463e-02 5.72241545e-01 4.45968181e-01
1.05305009e-01 -1.39660398e-02 -1.08712971e+00 1.37852383e+00
4.05178905e-01 -1.13882172e+00 3.48971277e-01 -1.42224178e-01
6.95570350e-01 4.71780188e-02 9.80391428e-02 4.58892018e-01
-1.50085360e-01 -6.92981362e-01 6.90631747e-01 5.43580711e-01
8.55669260e-01 -5.22471666e-01 7.85951316e-01 3.03135425e-01
-1.56682789e+00 -1.29929349e-01 -4.03507382e-01 4.24455941e-01
2.80115545e-01 4.38687742e-01 -3.21338356e-01 9.17586625e-01
7.47085512e-01 8.56480360e-01 -6.15624368e-01 1.02863014e+00
9.42228734e-02 4.24308628e-01 -4.08583701e-01 -6.15670392e-03
4.17539597e-01 -1.33430138e-01 7.35409498e-01 1.36802876e+00
4.76841033e-01 -1.34373307e-02 3.46428692e-01 4.14825439e-01
-2.20090151e-01 4.16206956e-01 -1.88406900e-01 -1.24268845e-01
4.76020306e-01 1.30307245e+00 -5.67788124e-01 -4.70743924e-01
-5.93831241e-01 1.11944807e+00 2.19415694e-01 4.17857230e-01
-1.01847255e+00 -4.20355350e-01 4.34062630e-01 -2.24851761e-02
5.95625818e-01 -3.13586220e-02 1.14845142e-01 -1.49649191e+00
4.80851859e-01 -1.18911624e+00 3.88496250e-01 -7.43447542e-01
-1.12139273e+00 7.03436613e-01 1.14168497e-02 -1.71874583e+00
-1.44193932e-01 -3.60230535e-01 -3.06962818e-01 6.15452409e-01
-1.57431424e+00 -1.19953334e+00 -2.16373190e-01 6.51518583e-01
9.24701750e-01 -4.32677448e-01 4.49591696e-01 9.48551178e-01
-9.06682491e-01 1.13175797e+00 3.45139623e-01 1.29152760e-01
1.16491926e+00 -8.06570172e-01 2.44606331e-01 9.19581831e-01
3.11440408e-01 6.07725322e-01 4.28022385e-01 -5.64188480e-01
-1.69361091e+00 -1.00259268e+00 7.54383087e-01 -1.75264701e-01
6.16585433e-01 -2.66209662e-01 -9.43130672e-01 5.15988231e-01
1.13489166e-01 -1.41480268e-04 4.47401136e-01 -2.41981015e-01
-4.39098746e-01 -3.18093985e-01 -7.73233891e-01 7.13932276e-01
1.04777265e+00 -6.41617119e-01 -4.36865538e-01 2.78959543e-01
8.04403365e-01 -4.87403244e-01 -9.35216248e-01 4.78238344e-01
7.17798471e-01 -7.29400873e-01 9.58414197e-01 -3.92048568e-01
6.94131672e-01 -4.13467407e-01 -3.98971021e-01 -9.01778281e-01
-2.96591282e-01 -7.20645428e-01 -3.41660947e-01 1.44920576e+00
2.81169116e-01 -3.00277442e-01 5.10063231e-01 3.11607569e-01
-3.14245194e-01 -8.97029877e-01 -8.89338553e-01 -6.77182198e-01
-2.61270672e-01 -3.84967685e-01 4.50275898e-01 9.67271864e-01
-7.29870647e-02 5.19690692e-01 -6.91263795e-01 2.71889027e-02
3.89563560e-01 2.20336661e-01 7.68345594e-01 -8.50655258e-01
-4.60553139e-01 -5.70381582e-01 -1.21027514e-01 -1.64484453e+00
2.07056012e-02 -9.30898964e-01 -3.44594521e-03 -1.48263717e+00
4.57603395e-01 -3.59442413e-01 -4.12019819e-01 2.26720050e-01
-4.16780144e-01 1.32858843e-01 4.17034715e-01 4.11234617e-01
-1.08862603e+00 6.96865618e-01 1.26100922e+00 -2.14138508e-01
-1.88835397e-01 -1.25405192e-01 -4.83985364e-01 6.14252210e-01
4.93435830e-01 -4.23886508e-01 -3.55538577e-01 -7.21109986e-01
2.65894294e-01 2.15270281e-01 3.93033847e-02 -8.87023032e-01
5.23359776e-01 3.93713973e-02 2.61338592e-01 -9.60127056e-01
5.23531020e-01 -8.44696939e-01 1.81058973e-01 -2.84888931e-02
-2.50049353e-01 9.29507390e-02 7.47218654e-02 4.96154875e-01
-5.34505069e-01 -4.03852701e-01 4.87973064e-01 -2.02868506e-02
-7.30346680e-01 5.50542831e-01 -2.69123912e-01 1.14164613e-01
6.68658257e-01 -2.74267614e-01 -1.92384958e-01 -3.91287327e-01
-2.59215057e-01 6.10940814e-01 2.75512546e-01 5.55832744e-01
7.92961240e-01 -1.23028398e+00 -7.31750429e-01 4.46521603e-02
2.17790008e-01 5.70385605e-02 5.13303041e-01 9.32367206e-01
-2.68353015e-01 3.98047090e-01 9.81614366e-02 -6.39057398e-01
-1.33717585e+00 5.93584538e-01 -2.89655440e-02 -6.83167994e-01
-6.36426151e-01 8.06286156e-01 2.67529249e-01 -3.46674360e-02
4.62959439e-01 6.76574856e-02 -3.47262144e-01 2.50966519e-01
6.26486361e-01 2.94647068e-01 1.88459307e-02 -7.36794531e-01
-1.97795391e-01 8.91120791e-01 -6.07195497e-01 1.64203241e-01
1.25293541e+00 -4.00113612e-01 -8.08802173e-02 1.94521204e-01
1.40497220e+00 -2.88335420e-02 -9.98716891e-01 -4.55715656e-01
-2.28393719e-01 -6.07812285e-01 1.38164312e-01 -4.52839285e-01
-1.17012358e+00 8.45033109e-01 4.72223580e-01 -1.87567025e-01
1.25143611e+00 -1.67577267e-01 1.20305657e+00 5.25808632e-01
2.44729131e-01 -1.27692223e+00 3.66985470e-01 5.20010769e-01
7.86400437e-01 -1.14343464e+00 2.84005105e-01 -3.43727946e-01
-6.24549210e-01 9.20046449e-01 6.78877354e-01 -1.31885409e-01
5.23604512e-01 1.59522966e-01 -1.52379394e-01 -6.46519661e-02
-9.35752213e-01 -1.15798227e-01 5.18167794e-01 3.11825741e-02
1.92101657e-01 -3.37517291e-01 -2.32838154e-01 4.47231591e-01
2.14235470e-01 -4.80344892e-02 1.50426999e-01 9.89164293e-01
-2.70797223e-01 -1.04032242e+00 -9.66185182e-02 5.36329031e-01
-6.25694394e-01 -3.11109096e-01 -1.24493390e-01 7.06431329e-01
-2.15565696e-01 7.93182909e-01 -3.17422189e-02 -6.39584780e-01
3.40367705e-01 -1.07245855e-01 4.25649166e-01 -2.52880991e-01
-6.07288957e-01 6.02477908e-01 -5.95573373e-02 -6.63014174e-01
-4.56397474e-01 -5.89897871e-01 -9.01407719e-01 -1.79041535e-01
-6.41027272e-01 1.93615463e-02 3.78308654e-01 1.02588296e+00
4.88328785e-01 3.75358582e-01 8.70274484e-01 -7.80137658e-01
-6.14276707e-01 -1.04229951e+00 -4.06686634e-01 3.86239856e-01
1.66656941e-01 -6.46825433e-01 -3.28781545e-01 1.51688922e-02] | [10.324411392211914, 0.9434652328491211] |
229aab39-6405-4b69-b18b-96d3ccf97d17 | detecting-and-tracking-small-and-dense-moving | 2111.12960 | null | https://arxiv.org/abs/2111.12960v1 | https://arxiv.org/pdf/2111.12960v1.pdf | Detecting and Tracking Small and Dense Moving Objects in Satellite Videos: A Benchmark | Satellite video cameras can provide continuous observation for a large-scale area, which is important for many remote sensing applications. However, achieving moving object detection and tracking in satellite videos remains challenging due to the insufficient appearance information of objects and lack of high-quality datasets. In this paper, we first build a large-scale satellite video dataset with rich annotations for the task of moving object detection and tracking. This dataset is collected by the Jilin-1 satellite constellation and composed of 47 high-quality videos with 1,646,038 instances of interest for object detection and 3,711 trajectories for object tracking. We then introduce a motion modeling baseline to improve the detection rate and reduce false alarms based on accumulative multi-frame differencing and robust matrix completion. Finally, we establish the first public benchmark for moving object detection and tracking in satellite videos, and extensively evaluate the performance of several representative approaches on our dataset. Comprehensive experimental analyses and insightful conclusions are also provided. The dataset is available at https://github.com/QingyongHu/VISO. | ['Yulan Guo', 'Wei An', 'Zaiping Lin', 'Yingqian Wang', 'Feng Zhang', 'Hao liu', 'Qingyong Hu', 'Qian Yin'] | 2021-11-25 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 1.35702584e-02 -9.38183188e-01 -2.25328520e-01 -1.53877407e-01
-9.29822028e-01 -7.00603485e-01 3.87734026e-01 -4.17484283e-01
-4.60890651e-01 5.62232375e-01 8.62262100e-02 -8.15451145e-04
-3.73793952e-02 -7.11323977e-01 -7.53691673e-01 -1.04943168e+00
-6.21370077e-01 -9.07658637e-02 5.91529548e-01 7.74673522e-02
-2.36958280e-01 4.50237215e-01 -1.51374090e+00 1.82981845e-02
8.12223136e-01 9.71875906e-01 3.69942933e-01 7.70401299e-01
6.77921236e-01 6.55739963e-01 -2.10378513e-01 1.36084557e-01
4.56366509e-01 -6.14338964e-02 -4.83169198e-01 4.82630223e-01
9.02076364e-01 -9.64366376e-01 -7.66957760e-01 1.23830128e+00
3.63723069e-01 4.04653579e-01 2.85853863e-01 -1.28574371e+00
-4.44185853e-01 2.01185748e-01 -6.56976819e-01 9.21663344e-01
-8.40911525e-04 5.47888339e-01 1.01785684e+00 -9.61935520e-01
5.16033828e-01 1.00651777e+00 5.31684577e-01 2.38323972e-01
-8.78099501e-01 -7.30189025e-01 2.44603932e-01 3.73554498e-01
-1.71509290e+00 -5.71917951e-01 6.45757690e-02 -5.78745365e-01
4.25751239e-01 4.71247017e-01 8.96458983e-01 6.06481731e-01
-3.81355345e-01 8.40220273e-01 5.92434704e-01 1.77705675e-01
-1.26100183e-01 -4.44479227e-01 3.17162089e-02 5.16598701e-01
3.97426963e-01 1.63730383e-01 -1.79478467e-01 -8.50522146e-02
7.28710711e-01 3.86561960e-01 -7.95715630e-01 8.92802477e-02
-1.44217122e+00 5.94866514e-01 6.99690640e-01 1.64108902e-01
-3.49167287e-01 9.03356373e-02 1.20085880e-01 5.75075373e-02
4.83578920e-01 -4.26281691e-01 -2.21699417e-01 1.32767260e-01
-1.08017564e+00 1.43472269e-01 1.57909155e-01 8.97751570e-01
6.39264345e-01 3.22528005e-01 -2.81555742e-01 5.17840922e-01
3.78886610e-01 1.30316734e+00 8.24483335e-02 -8.95191789e-01
5.44954240e-01 3.62145245e-01 4.12235528e-01 -1.18111098e+00
-4.57095861e-01 -2.96518087e-01 -8.34130943e-01 -1.94566801e-01
2.72947252e-01 -1.66909307e-01 -6.95890367e-01 1.12071002e+00
7.03716159e-01 7.37286210e-01 -7.97505230e-02 1.40089476e+00
1.04464126e+00 1.18366754e+00 -9.79441702e-02 -2.20654771e-01
1.42044544e+00 -9.85299826e-01 -5.36959648e-01 -2.31946081e-01
5.40428698e-01 -3.97734255e-01 7.07515657e-01 7.99015388e-02
-4.72966045e-01 -4.93606508e-01 -5.79749286e-01 1.51723236e-01
1.76410645e-01 7.10916996e-01 5.28709173e-01 2.64264762e-01
-7.84777284e-01 5.55115715e-02 -1.17047143e+00 -3.87387127e-01
6.93920732e-01 3.89842801e-02 -3.40363890e-01 -3.72059494e-01
-9.50568199e-01 4.14179444e-01 3.65107596e-01 6.62948191e-01
-1.34054589e+00 -7.01455355e-01 -7.84036756e-01 -9.07246321e-02
4.50456917e-01 -2.32616082e-01 1.03659701e+00 -8.77016008e-01
-7.37514615e-01 6.51138842e-01 -2.08569523e-02 -4.21174794e-01
4.75331396e-01 -3.00363034e-01 -7.43469179e-01 3.58983636e-01
2.08179608e-01 6.43562675e-01 7.17425525e-01 -8.34320188e-01
-1.21332383e+00 -2.88653135e-01 -6.04240559e-02 8.25650617e-02
-3.83018434e-01 2.83909500e-01 -9.13224816e-01 -8.61263454e-01
1.38561979e-01 -1.17631853e+00 -1.99559882e-01 3.84017557e-01
-4.79826741e-02 1.60506651e-01 1.12878036e+00 -9.52565134e-01
1.28150141e+00 -2.30768466e+00 3.72534506e-02 -1.95484206e-01
1.09836943e-01 5.06441474e-01 -3.06682616e-01 2.24938706e-01
1.30479977e-01 -7.38807991e-02 -3.87307465e-01 8.83427858e-02
-5.06074786e-01 -4.42777425e-02 -3.28448176e-01 9.52288389e-01
1.30380422e-01 6.91105604e-01 -9.57533777e-01 -3.78179461e-01
2.89777011e-01 3.82538468e-01 -2.41601527e-01 -4.25691344e-03
-8.78284276e-02 6.70971274e-01 -6.54917359e-01 1.27121723e+00
9.29466128e-01 -3.78180355e-01 4.74680029e-02 -4.17649716e-01
-2.47053042e-01 -1.69802547e-01 -1.24040782e+00 1.28608012e+00
7.70521909e-02 1.11429477e+00 2.74210721e-01 -5.37743330e-01
5.04915714e-01 4.17536870e-02 7.41461277e-01 -3.25565964e-01
-4.70602773e-02 1.23196214e-01 -2.58478463e-01 -7.13765323e-01
6.95865273e-01 1.14170499e-01 1.28977954e-01 3.25788371e-03
-4.25230622e-01 2.72123784e-01 4.91814971e-01 2.62043267e-01
9.19236898e-01 -2.09218245e-02 9.84796882e-02 -1.10487580e-01
5.55836082e-01 4.06686068e-01 9.30289268e-01 5.36614776e-01
-3.51075262e-01 4.64865953e-01 -2.46255428e-01 -4.89846826e-01
-6.44447744e-01 -7.83657849e-01 -4.04093921e-01 8.20481777e-01
4.55539286e-01 -2.20219746e-01 -2.24479228e-01 -3.68373573e-01
1.72854885e-02 1.66746885e-01 -4.33344573e-01 2.59551585e-01
-4.08602417e-01 -1.23155046e+00 4.76399779e-01 3.93107116e-01
6.41267002e-01 -7.56991148e-01 -3.28268647e-01 1.41456321e-01
-7.79201806e-01 -1.31100714e+00 -5.81653118e-01 -6.41389608e-01
-1.12095714e+00 -1.17229056e+00 -7.47184396e-01 -5.79342902e-01
5.59748709e-01 1.35079598e+00 7.88363636e-01 3.99257243e-01
-4.02917475e-01 3.03480238e-01 -6.21964872e-01 1.65404782e-01
1.39782190e-01 -2.76912034e-01 2.28763372e-01 2.30147824e-01
2.03020543e-01 -1.78138286e-01 -7.55365133e-01 7.48764873e-01
-1.18360448e+00 -2.16792122e-01 4.35825258e-01 4.66383696e-01
4.71621901e-01 2.26080745e-01 -1.25350147e-01 -2.47640107e-02
-4.98854995e-01 -7.48586595e-01 -1.13236141e+00 6.35212138e-02
1.71863273e-01 -6.98194563e-01 1.20253107e-02 -4.04754072e-01
-7.36652017e-01 3.92641455e-01 1.51126623e-01 -4.60320383e-01
6.61777332e-02 5.54336071e-01 1.93591937e-02 -2.72912681e-01
3.40959728e-01 5.20381987e-01 -1.56674370e-01 -4.44128513e-01
1.08366512e-01 7.56134868e-01 5.77567697e-01 -1.09030940e-01
1.10749328e+00 1.01163507e+00 -3.50157976e-01 -1.36274529e+00
-9.00185645e-01 -8.12112927e-01 -4.78966951e-01 -5.16898990e-01
7.52493620e-01 -1.85564172e+00 -5.21116853e-01 7.09400058e-01
-6.70387745e-01 -3.74047726e-01 2.16697469e-01 8.52922976e-01
3.94380949e-02 6.30231977e-01 -5.66967964e-01 -7.18163431e-01
-3.26956302e-01 -9.61947143e-01 1.12628710e+00 4.39760208e-01
6.39681458e-01 -4.96634662e-01 -9.00671855e-02 4.55227107e-01
3.39456320e-01 1.45341173e-01 -2.48245269e-01 -4.29337770e-02
-1.26256573e+00 -2.28132457e-01 -3.12470287e-01 2.31791347e-01
2.09592562e-02 3.62464130e-01 -4.98481542e-01 -7.56105125e-01
-3.70147765e-01 -1.15080290e-01 1.26980841e+00 6.59247220e-01
8.56792808e-01 -1.79706365e-01 -4.97264713e-01 9.24049973e-01
1.58152044e+00 -8.05202406e-03 6.76066875e-01 5.03952026e-01
9.71222341e-01 2.71983057e-01 1.05869699e+00 6.53190076e-01
5.16287804e-01 8.52094710e-01 6.17760837e-01 1.59955457e-01
7.02041686e-02 2.09140852e-01 8.16309154e-01 5.08492768e-01
-4.05273288e-01 -2.42091462e-01 -9.14233804e-01 8.37882817e-01
-1.76497138e+00 -1.37921381e+00 -7.96984613e-01 2.16788673e+00
2.53822565e-01 -4.51253176e-01 2.66174108e-01 -3.37216467e-01
1.01639378e+00 4.32862103e-01 -4.68711555e-01 1.39872634e+00
-4.54852283e-01 -7.29498267e-01 9.23220694e-01 3.13137442e-01
-1.65460825e+00 8.96961153e-01 5.26987314e+00 7.79905140e-01
-1.08749115e+00 1.18262552e-01 1.78442597e-01 -4.24843103e-01
2.84892917e-01 1.38533209e-02 -1.26778293e+00 7.33458936e-01
5.65483391e-01 2.82327216e-02 1.75237343e-01 6.33626044e-01
6.43959939e-01 -1.44175813e-01 -3.22647274e-01 9.84080672e-01
-2.92721558e-02 -1.49854636e+00 9.51690308e-04 1.93180606e-01
8.79334867e-01 7.27939904e-01 -1.92205921e-01 7.94554502e-02
1.95196327e-02 -4.50149596e-01 7.74877191e-01 5.29329240e-01
5.01658678e-01 -3.93061042e-01 5.07956922e-01 3.49547952e-01
-1.73738289e+00 -2.99462825e-01 -5.02055764e-01 5.51046245e-02
4.71639603e-01 5.06398857e-01 -7.32708499e-02 6.11422181e-01
1.21652150e+00 1.46433318e+00 -6.49269521e-01 1.62909293e+00
-9.88129005e-02 8.39492679e-01 -5.26866317e-01 2.57656604e-01
3.63634467e-01 -4.80544835e-01 9.40917969e-01 8.90084267e-01
6.66113675e-01 6.28343463e-01 4.97656941e-01 2.79485077e-01
-2.43897177e-02 -1.12021148e-01 -3.89952362e-01 -7.02431053e-02
7.08941400e-01 1.52806771e+00 -5.50987184e-01 -4.19878393e-01
-6.03553236e-01 6.38599455e-01 -2.51481980e-01 2.57374674e-01
-1.27127600e+00 2.05222383e-01 9.87344801e-01 -8.10874477e-02
7.66420484e-01 -6.52361691e-01 8.44475627e-01 -1.76765549e+00
1.37028202e-01 -9.73884284e-01 5.83460212e-01 -8.14917386e-01
-1.01190984e+00 3.75439882e-01 -6.02974333e-02 -1.99248028e+00
4.63611454e-01 -4.73955393e-01 -6.26958251e-01 3.02753091e-01
-1.55702984e+00 -1.28647494e+00 -9.68732774e-01 4.24117267e-01
6.82627201e-01 7.64197633e-02 8.65044817e-02 8.23592961e-01
-9.57184136e-01 -1.36107951e-01 6.25102222e-01 4.57176358e-01
4.76153970e-01 -4.92721766e-01 4.09985304e-01 1.39570355e+00
2.16950446e-01 8.24502856e-02 4.26344514e-01 -6.53901935e-01
-1.81552958e+00 -1.62287986e+00 2.74854481e-01 -4.10101801e-01
8.48892093e-01 4.83397394e-02 -1.12777555e+00 9.03027117e-01
-3.03535521e-01 6.11240864e-01 3.28657836e-01 -6.48292005e-01
1.81731120e-01 -1.69755012e-01 -6.37491167e-01 2.07428396e-01
1.00776672e+00 -1.18579917e-01 -5.33881485e-02 7.21275330e-01
5.61019301e-01 -4.92805511e-01 -8.60871553e-01 2.87935495e-01
4.89103913e-01 -6.58300996e-01 1.23407674e+00 -4.75960076e-01
1.67279527e-01 -1.12779331e+00 -4.81459171e-01 -8.62175822e-01
-4.43417460e-01 -2.12802604e-01 -7.80449957e-02 1.23886430e+00
1.35458922e-02 -1.77299216e-01 4.81356114e-01 2.45622858e-01
1.19252585e-01 -4.57086338e-04 -7.29803383e-01 -1.11546445e+00
-5.56527555e-01 -3.71745646e-01 3.44024181e-01 9.21470582e-01
-8.42598736e-01 -2.72344053e-02 -1.03283238e+00 9.91411805e-01
9.48066652e-01 5.10548115e-01 9.83888865e-01 -9.90866184e-01
-1.74484238e-01 -1.39335990e-02 -6.38648510e-01 -1.43401778e+00
-1.03426278e-01 -6.76768661e-01 1.74984112e-01 -1.32936418e+00
4.52498496e-01 -1.43776849e-01 1.81317583e-01 3.70715946e-01
-3.56465459e-01 7.30605543e-01 2.48056084e-01 8.03116381e-01
-8.77620518e-01 8.66156816e-01 1.03066599e+00 -2.99743801e-01
-9.45397094e-03 1.34619057e-01 -1.41875446e-01 6.86334252e-01
5.59202254e-01 -7.41115212e-01 3.06459814e-01 -8.29010367e-01
-9.97127220e-02 1.68214485e-01 8.54881644e-01 -1.15840578e+00
2.15136349e-01 -3.58912855e-01 3.84385973e-01 -9.35968578e-01
2.67242432e-01 -9.62797821e-01 4.48600352e-01 7.48088300e-01
1.81486219e-01 -8.00657123e-02 2.72105485e-01 1.01896822e+00
-3.09954733e-01 -5.41609228e-02 7.84321427e-01 1.40316218e-01
-1.55465364e+00 7.92059720e-01 -4.39910620e-01 2.15894431e-02
1.20574367e+00 1.45078292e-02 -5.06473839e-01 -2.75664389e-01
-6.17741823e-01 7.48976827e-01 5.38802743e-01 5.25523663e-01
4.27003473e-01 -1.43269289e+00 -1.19308913e+00 -9.30816010e-02
3.09879869e-01 -9.89261717e-02 7.22742677e-01 1.24255180e+00
-7.35557675e-01 1.49645627e-01 -3.13956380e-01 -9.85678375e-01
-1.56049335e+00 2.08260849e-01 4.00292963e-01 4.90958363e-01
-8.81588161e-01 6.25636578e-01 7.03438837e-03 -3.31806503e-02
-2.20733471e-02 -3.34575504e-01 -1.22336239e-01 1.62643731e-01
1.15018070e+00 5.62946260e-01 -2.39119917e-01 -1.15208614e+00
-6.58834279e-01 6.91857755e-01 2.57475823e-01 1.99541509e-01
1.43111134e+00 -5.03379643e-01 -6.14912398e-02 8.69090110e-02
9.13136363e-01 -1.32895112e-01 -1.75513339e+00 -5.14277399e-01
-2.33685702e-01 -1.09362793e+00 3.65507364e-01 -5.70049249e-02
-1.46038258e+00 6.32292330e-01 8.19279611e-01 -9.11696777e-02
1.11929941e+00 -3.19333039e-02 7.08332419e-01 5.75183272e-01
3.21188658e-01 -6.89157128e-01 -1.83828980e-01 4.34678227e-01
7.27522016e-01 -1.59136987e+00 3.43928099e-01 -4.03038085e-01
-6.44303262e-01 9.18490469e-01 5.64599156e-01 8.45783055e-02
3.30021590e-01 -1.29235595e-01 9.38261822e-02 -1.00001380e-01
-5.56895733e-01 -7.02964246e-01 3.29351515e-01 3.93102139e-01
-7.63216475e-03 4.99380864e-02 -2.78513264e-02 1.37979969e-01
5.44817626e-01 -1.08346939e-02 5.76979816e-01 9.55219746e-01
-7.74524987e-01 -4.86710846e-01 -7.95506835e-01 4.14642364e-01
-3.68993908e-01 -1.96893513e-01 1.63779825e-01 6.94705069e-01
-1.92553744e-01 9.81266081e-01 9.98967364e-02 -1.73931658e-01
1.04790032e-01 -6.76902890e-01 9.77096558e-02 -2.45198369e-01
4.22257278e-03 1.39592320e-01 4.62603644e-02 -4.93420750e-01
-9.18771207e-01 -1.01144588e+00 -1.07369006e+00 -2.78183401e-01
-4.31290776e-01 -8.65315087e-03 4.61741924e-01 6.91246152e-01
3.35729122e-01 6.12620339e-02 6.31844282e-01 -1.22366452e+00
-4.74787980e-01 -8.35822463e-01 -7.59798765e-01 1.91971108e-01
5.77588379e-01 -7.10561514e-01 -4.67537433e-01 5.10160744e-01] | [8.912240982055664, -0.7323412895202637] |
f307281c-f376-4aab-966a-7ec815df8c12 | large-scale-mixed-bandwidth-deep-neural | 1907.04887 | null | https://arxiv.org/abs/1907.04887v1 | https://arxiv.org/pdf/1907.04887v1.pdf | Large-Scale Mixed-Bandwidth Deep Neural Network Acoustic Modeling for Automatic Speech Recognition | In automatic speech recognition (ASR), wideband (WB) and narrowband (NB) speech signals with different sampling rates typically use separate acoustic models. Therefore mixed-bandwidth (MB) acoustic modeling has important practical values for ASR system deployment. In this paper, we extensively investigate large-scale MB deep neural network acoustic modeling for ASR using 1,150 hours of WB data and 2,300 hours of NB data. We study various MB strategies including downsampling, upsampling and bandwidth extension for MB acoustic modeling and evaluate their performance on 8 diverse WB and NB test sets from various application domains. To deal with the large amounts of training data, distributed training is carried out on multiple GPUs using synchronous data parallelism. | ['Wei zhang', 'Khoi-Nguyen C. Mac', 'Xiaodong Cui', 'Michael Picheny'] | 2019-07-10 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension'] | ['audio', 'speech'] | [-2.62297913e-02 -4.75398481e-01 1.33843452e-01 -4.93023664e-01
-1.10718811e+00 -6.02168776e-03 2.67396271e-01 -1.63590789e-01
-6.36168838e-01 3.36452663e-01 7.79913142e-02 -9.33310688e-01
3.31313372e-01 -4.72430944e-01 -5.46145380e-01 -6.54457688e-01
-1.38262674e-01 4.28378940e-01 2.69048095e-01 -1.21383041e-01
-2.94343024e-01 6.33208215e-01 -1.47092962e+00 4.53466505e-01
2.91319042e-01 1.06485176e+00 7.05284894e-01 1.49311173e+00
-2.26255268e-01 6.00028574e-01 -1.29974759e+00 -3.47300209e-02
1.68632288e-02 7.93770105e-02 -3.75098556e-01 -2.58171350e-01
1.32785648e-01 -6.04293227e-01 -6.84705377e-01 7.48647094e-01
1.20080197e+00 6.40023470e-01 1.86444506e-01 -8.05050671e-01
-1.19217567e-01 8.15717518e-01 -6.99046850e-01 7.07082748e-01
-9.29279551e-02 -7.34733418e-02 5.76069295e-01 -9.92800772e-01
-3.24345648e-01 1.41200757e+00 7.11625934e-01 6.75452828e-01
-9.49453831e-01 -7.80637920e-01 -1.44955620e-01 5.29673517e-01
-1.40579987e+00 -1.21531355e+00 7.97942221e-01 2.34036013e-01
1.70399654e+00 3.72138053e-01 6.30206287e-01 1.32619762e+00
-1.91388443e-01 8.17917883e-01 6.32903337e-01 -4.97743100e-01
4.48083311e-01 -2.20358372e-01 5.41905880e-01 4.36324655e-04
-2.06603721e-01 1.90734506e-01 -9.72963929e-01 -4.65153277e-01
5.31030357e-01 -5.16065955e-01 -9.18435454e-02 9.16013956e-01
-6.91057682e-01 7.64374375e-01 -2.70765007e-01 5.54165170e-02
-3.51583600e-01 5.28102875e-01 8.49511683e-01 4.41904038e-01
6.17473900e-01 -3.08160275e-01 -8.32351863e-01 -7.33238816e-01
-8.80493343e-01 -1.12714224e-01 7.71556556e-01 1.03061616e+00
3.74164283e-01 1.06370735e+00 3.61345708e-01 1.90449202e+00
5.06702781e-01 8.27162266e-01 1.24058795e+00 -4.55682456e-01
5.14562905e-01 -6.06451511e-01 -3.13964009e-01 -3.11298072e-01
-3.17945838e-01 -3.52326035e-01 -1.07442963e+00 -4.50603783e-01
3.65177309e-03 -3.85087937e-01 -1.26436973e+00 1.23729432e+00
4.33282405e-01 6.54360533e-01 2.45882720e-01 7.03651130e-01
1.11514819e+00 1.36948478e+00 2.34436709e-02 -1.22872956e-01
1.29871166e+00 -1.16045761e+00 -7.63867021e-01 -4.44935083e-01
7.13440180e-01 -1.14530790e+00 1.19833875e+00 6.43331826e-01
-1.02089787e+00 -8.62737536e-01 -1.07487130e+00 -2.24556535e-01
-8.22825506e-02 3.42905939e-01 3.39181840e-01 1.41440082e+00
-1.23091733e+00 3.89053635e-02 -1.22674143e+00 2.49605794e-02
9.66894776e-02 4.43046629e-01 4.11941223e-02 -2.95636454e-03
-1.24181151e+00 3.75087440e-01 7.73331337e-03 2.17444360e-01
-1.15748501e+00 -9.46745396e-01 -7.15602875e-01 4.43670973e-02
-5.47566116e-02 -6.60517216e-02 1.68788159e+00 -5.82035899e-01
-2.10927224e+00 4.02099907e-01 -3.64403874e-01 -7.53830910e-01
-3.35601605e-02 -3.92272353e-01 -8.88037145e-01 -1.50914595e-01
-9.06234205e-01 2.58558661e-01 1.06282365e+00 -7.79293776e-01
-5.44690073e-01 -3.83174151e-01 -5.29274762e-01 3.07745878e-02
-5.88084877e-01 3.70477438e-01 -2.93155760e-01 -7.46970534e-01
8.26274231e-02 -6.92805052e-01 -1.56005517e-01 -8.94790053e-01
-1.96844235e-01 -1.61787406e-01 1.01368022e+00 -1.10645247e+00
1.20493782e+00 -2.42695928e+00 -3.26479971e-01 1.67480156e-01
-1.24090716e-01 5.31971216e-01 -2.85301030e-01 8.54729861e-02
3.95020097e-02 -3.51248294e-01 2.72498727e-01 -7.26998270e-01
-5.70922568e-02 3.51101637e-01 -6.39034271e-01 3.85110080e-01
-4.06560600e-01 2.92669117e-01 -2.57212698e-01 -1.41399026e-01
3.43080997e-01 8.72608423e-01 -3.48623961e-01 4.43677425e-01
2.72929400e-01 8.45124759e-03 -2.07965121e-01 5.81228018e-01
7.25171566e-01 3.54027927e-01 6.00383915e-02 -2.61738539e-01
2.79138535e-01 7.09336340e-01 -9.93817925e-01 1.42821455e+00
-1.12189138e+00 1.02426708e+00 6.29715264e-01 -1.13902426e+00
1.26824915e+00 7.20191538e-01 1.31517768e-01 -6.62186801e-01
1.90660097e-02 5.10812104e-01 2.75244057e-01 -1.26700103e-01
6.85230255e-01 -1.37649387e-01 4.73848939e-01 3.00786108e-01
1.13170601e-01 -4.09455895e-01 -4.47528869e-01 -2.89030463e-01
1.06558979e+00 -7.32882321e-01 -4.39620428e-02 -1.14490390e-01
4.02407199e-01 -4.29998010e-01 2.61790782e-01 9.17445362e-01
-3.55381548e-01 4.65605021e-01 -2.01646641e-01 -4.07995939e-01
-1.23251927e+00 -8.32499206e-01 -2.58659601e-01 1.51824808e+00
-4.71063316e-01 -2.63198644e-01 -7.07619011e-01 1.03235431e-01
-3.08834106e-01 7.92478263e-01 1.92768663e-01 -1.29782204e-02
-1.01677382e+00 -1.22326887e+00 9.70738828e-01 6.08423650e-01
3.07278991e-01 -9.81270432e-01 -5.25404453e-01 4.10989910e-01
1.04050398e-01 -1.40132022e+00 -4.20133889e-01 5.49444377e-01
-7.59771109e-01 -2.04970837e-01 -7.50019610e-01 -7.53248692e-01
-1.54538944e-01 4.37484115e-01 1.00697172e+00 -1.19271830e-01
-2.68580496e-01 3.58785987e-01 -4.50420141e-01 -5.42764902e-01
-8.37218702e-01 1.63339823e-01 4.44127411e-01 -1.58566356e-01
4.56351936e-01 -6.75636768e-01 -3.36955577e-01 3.54488552e-01
-8.03219736e-01 -1.60622105e-01 2.76356190e-01 9.86457884e-01
4.23461705e-01 1.64669063e-02 8.74368787e-01 -2.91975111e-01
8.63533199e-01 -3.25760245e-01 -6.33284926e-01 5.86834066e-02
1.62825943e-03 -4.67272967e-01 7.03266561e-01 -1.03059006e+00
-1.12955904e+00 -4.36999828e-01 -1.02441847e+00 -5.94470799e-01
-3.39203179e-01 4.62130487e-01 -1.43247202e-01 1.38884768e-01
5.74140370e-01 4.72573251e-01 -2.18694806e-01 -6.97785020e-01
2.52524644e-01 1.51541626e+00 1.53255805e-01 -8.28948557e-01
9.99525562e-02 -5.92735223e-03 -6.80310488e-01 -1.93934643e+00
-1.28001096e-02 -5.45848131e-01 1.30623924e-02 -5.64034916e-02
3.49948466e-01 -1.02201521e+00 -5.96955776e-01 9.15609062e-01
-1.20819151e+00 -9.16931033e-01 -6.88431114e-02 9.92771089e-01
-3.93637419e-01 3.55648547e-01 -1.16084242e+00 -1.29140508e+00
-8.47189426e-01 -1.48882127e+00 1.17721391e+00 -1.36748642e-01
-6.23946153e-02 -5.55611491e-01 -4.53080498e-02 4.08874720e-01
7.89001644e-01 -1.04678607e+00 6.68537974e-01 -9.29577351e-01
5.83689958e-02 -3.10051888e-01 -6.97328430e-03 7.16705680e-01
1.11733325e-01 -5.59427179e-02 -1.41039681e+00 -5.61939776e-01
4.63691980e-01 -4.22778577e-01 5.25012374e-01 8.35770726e-01
1.55162895e+00 -4.75408435e-02 -1.15207955e-02 6.27494335e-01
7.65122473e-01 8.35339010e-01 4.30614591e-01 -1.74157638e-02
5.75244129e-01 3.54989201e-01 2.45985582e-01 5.45987189e-01
-1.76621437e-01 7.69167006e-01 -2.41378129e-01 1.07607603e-01
-2.99033910e-01 1.95185810e-01 6.88877046e-01 1.81982756e+00
2.14326411e-01 -4.93151903e-01 -1.07786536e+00 5.34710765e-01
-1.14956975e+00 -4.62217718e-01 -3.48943099e-03 2.15805197e+00
9.70870137e-01 1.76726412e-02 1.26676843e-01 3.07484895e-01
7.30504990e-01 4.77829814e-01 -4.80447173e-01 -7.63406873e-01
-7.20369220e-02 6.30117178e-01 5.10129869e-01 5.44635057e-01
-8.00993204e-01 9.18614209e-01 6.84803057e+00 1.81845450e+00
-1.26414859e+00 6.34901106e-01 1.24716687e+00 -5.32050610e-01
3.26234177e-02 -6.81480289e-01 -1.11549497e+00 3.64304572e-01
1.90178609e+00 2.41243526e-01 5.75068653e-01 1.01945865e+00
3.20536792e-01 1.92098141e-01 -8.21887195e-01 1.35596907e+00
-2.70553589e-01 -9.45126474e-01 -2.34888136e-01 -7.85653740e-02
2.61410862e-01 7.88763046e-01 2.43671551e-01 4.85383779e-01
4.00921583e-01 -1.07859421e+00 6.18864655e-01 -3.62396508e-01
9.53220785e-01 -9.85034347e-01 7.25172937e-01 3.16325814e-01
-1.33328521e+00 1.15323044e-01 -8.21173131e-01 2.37373114e-01
3.06583792e-01 6.97485328e-01 -8.08487833e-01 1.24465890e-01
8.41915309e-01 9.78418440e-02 5.60864694e-02 5.93423963e-01
4.14943665e-01 1.44234562e+00 -8.94228280e-01 -3.19128543e-01
1.09367214e-01 -5.93448319e-02 5.16341567e-01 1.50134051e+00
6.59173906e-01 1.72845617e-01 -2.43598133e-01 2.16377318e-01
8.79869089e-02 8.89231265e-02 -7.32301250e-02 1.21037669e-01
7.85752356e-01 8.86267424e-01 -4.04593438e-01 -5.02307415e-01
-4.68580812e-01 5.22984087e-01 1.06019586e-01 6.10339224e-01
-9.60293412e-01 -2.77282327e-01 1.01751018e+00 -3.98611307e-01
6.69371188e-02 -8.15991402e-01 -2.24805564e-01 -7.96962738e-01
-3.39404821e-01 -1.15071630e+00 3.90817039e-03 -6.71890795e-01
-1.03976309e+00 9.06347990e-01 -2.01246381e-01 -5.18941760e-01
-9.32852104e-02 -7.21828520e-01 -5.10910690e-01 9.99507129e-01
-1.39718437e+00 -7.61174142e-01 -8.10058729e-04 3.77562225e-01
1.40266871e+00 -6.85695887e-01 7.65581310e-01 7.15767920e-01
-7.14489400e-01 8.76045704e-01 3.91952366e-01 -4.65109907e-02
2.94312626e-01 -7.96872497e-01 1.06052661e+00 5.77902138e-01
3.48540634e-01 4.39369112e-01 6.85972154e-01 -2.54022568e-01
-1.64124858e+00 -9.36651826e-01 2.53304780e-01 1.87193185e-01
8.70965362e-01 -6.54023051e-01 -1.22018373e+00 4.56894070e-01
7.62217641e-02 1.65304095e-02 8.34504187e-01 1.18414305e-01
-9.30984914e-02 -4.29299682e-01 -7.52092242e-01 4.75376874e-01
6.56850100e-01 -6.79529905e-01 -1.65786788e-01 3.27631205e-01
1.27006495e+00 -5.72274089e-01 -9.34716702e-01 2.04619288e-01
4.40378308e-01 -5.41429996e-01 1.33337426e+00 -4.02126878e-01
-1.57030165e-01 1.94803298e-01 -8.11234117e-01 -1.32139838e+00
3.61545861e-01 -7.28733361e-01 -3.87071520e-01 1.08063102e+00
3.14122230e-01 -7.84306884e-01 8.46001863e-01 3.54055196e-01
-5.01138449e-01 -6.14604473e-01 -1.46803701e+00 -7.78328419e-01
-1.37091041e-01 -1.23596096e+00 8.49842012e-01 4.63287145e-01
-3.81173849e-01 1.26038685e-01 -5.71609676e-01 3.80286932e-01
3.12223434e-01 -5.03573895e-01 7.44999051e-01 -3.46262991e-01
-6.45317197e-01 -3.10373217e-01 -2.19545275e-01 -1.62757111e+00
1.17123455e-01 -3.04000437e-01 9.23298821e-02 -7.26907849e-01
-4.76715922e-01 -6.38081014e-01 -2.88609028e-01 1.76560134e-01
1.74451113e-01 1.86547320e-02 -2.66013801e-01 -9.78719443e-02
1.25140786e-01 8.49884212e-01 5.52652657e-01 -2.26388678e-01
-2.52084076e-01 2.31311142e-01 3.71659219e-01 4.95735109e-01
8.31067026e-01 -2.50929505e-01 -5.86829662e-01 -8.31975877e-01
-4.84597951e-01 3.82496744e-01 -1.61173239e-01 -9.87284541e-01
2.94765443e-01 7.05492571e-02 4.34772037e-02 -6.68385744e-01
9.93838072e-01 -5.06393671e-01 7.99995735e-02 2.26809829e-01
-3.34644705e-01 -4.80304174e-02 6.47503853e-01 4.43206221e-01
-3.76141667e-01 -4.07366425e-01 8.67854714e-01 2.77706891e-01
-5.29031157e-01 1.71240419e-01 -8.80493760e-01 -2.04917133e-01
3.17665219e-01 -4.11177278e-02 -2.60483593e-01 -4.38157946e-01
-7.61904716e-01 -3.49951357e-01 -4.01891172e-01 3.72018158e-01
1.05292988e+00 -1.02576995e+00 -7.80282140e-01 6.04868591e-01
-3.20371956e-01 1.62194312e-01 6.66529894e-01 5.20866632e-01
-7.58674860e-01 3.98458868e-01 4.12232697e-01 -7.27212548e-01
-1.60326791e+00 5.85258491e-02 5.34107268e-01 2.01386407e-01
-6.56677425e-01 1.27875614e+00 1.73868507e-01 -3.94621640e-01
5.27967691e-01 -3.15447211e-01 3.53361503e-03 -3.38341445e-01
6.00517452e-01 6.61798477e-01 5.94856083e-01 -5.49027562e-01
-1.63001984e-01 1.13474555e-01 -1.05752930e-01 -4.92534548e-01
1.28549063e+00 -6.03580400e-02 8.63630995e-02 6.21327221e-01
1.37823570e+00 -1.08020484e-01 -7.31406569e-01 -1.39475867e-01
-2.13212386e-01 -3.89968097e-01 6.03412330e-01 -2.01021329e-01
-1.01800001e+00 1.19001055e+00 8.06498230e-01 3.32704395e-01
1.10329223e+00 -1.36187345e-01 1.27828693e+00 5.51853180e-01
3.37560475e-01 -1.32014239e+00 -8.76020193e-02 6.37209952e-01
7.47752130e-01 -9.52928424e-01 -4.81423974e-01 5.27706323e-03
-3.20494145e-01 1.14381504e+00 5.27615011e-01 1.31552920e-01
1.02859390e+00 8.96031618e-01 2.70688802e-01 2.90546089e-01
-9.26974773e-01 3.44395638e-01 3.88328582e-02 6.21740818e-01
4.04720068e-01 2.75723547e-01 2.13257685e-01 7.54521608e-01
-6.01735055e-01 -5.37665427e-01 3.71226966e-01 7.08842397e-01
-5.16243875e-01 -8.45120013e-01 -7.01802433e-01 5.70815027e-01
-5.17780542e-01 -3.54186833e-01 3.54822665e-01 2.39664793e-01
-6.93108201e-01 1.32238483e+00 3.14000696e-01 -4.69586432e-01
-3.19722877e-03 5.91106042e-02 1.87487826e-01 -4.61016297e-01
-4.49578255e-01 7.90317059e-01 4.16503876e-01 -2.06936970e-01
1.59836277e-01 -1.91036373e-01 -1.14642203e+00 -4.89301920e-01
-6.07865393e-01 9.10834000e-02 1.41554987e+00 8.04277301e-01
2.68783212e-01 8.89452040e-01 5.46762764e-01 -8.94229412e-01
-8.84042144e-01 -1.48995936e+00 -8.73991549e-01 -1.66870132e-01
4.85433877e-01 -1.43686935e-01 -5.49376607e-01 -1.28150523e-01] | [14.513089179992676, 6.438774108886719] |
61f98861-7854-462f-9a50-30c580efcda6 | a-robust-kernel-machine-regression-towards | 2201.05060 | null | https://arxiv.org/abs/2201.05060v1 | https://arxiv.org/pdf/2201.05060v1.pdf | A robust kernel machine regression towards biomarker selection in multi-omics datasets of osteoporosis for drug discovery | Many statistical machine approaches could ultimately highlight novel features of the etiology of complex diseases by analyzing multi-omics data. However, they are sensitive to some deviations in distribution when the observed samples are potentially contaminated with adversarial corrupted outliers (e.g., a fictional data distribution). Likewise, statistical advances lag in supporting comprehensive data-driven analyses of complex multi-omics data integration. We propose a novel non-linear M-estimator-based approach, "robust kernel machine regression (RobKMR)," to improve the robustness of statistical machine regression and the diversity of fictional data to examine the higher-order composite effect of multi-omics datasets. We address a robust kernel-centered Gram matrix to estimate the model parameters accurately. We also propose a robust score test to assess the marginal and joint Hadamard product of features from multi-omics data. We apply our proposed approach to a multi-omics dataset of osteoporosis (OP) from Caucasian females. Experiments demonstrate that the proposed approach effectively identifies the inter-related risk factors of OP. With solid evidence (p-value = 0.00001), biological validations, network-based analysis, causal inference, and drug repurposing, the selected three triplets ((DKK1, SMTN, DRGX), (MTND5, FASTKD2, CSMD3), (MTND5, COG3, CSMD3)) are significant biomarkers and directly relate to BMD. Overall, the top three selected genes (DKK1, MTND5, FASTKD2) and one gene (SIDT1 at p-value= 0.001) significantly bond with four drugs- Tacrolimus, Ibandronate, Alendronate, and Bazedoxifene out of 30 candidates for drug repurposing in OP. Further, the proposed approach can be applied to any disease model where multi-omics datasets are available. | ['Hong-Wen Deng', 'Hui Shen', 'Md ashad Alam'] | 2022-01-13 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 2.73739785e-01 -1.53202206e-01 -2.75329143e-01 -1.89552516e-01
-7.16351151e-01 -2.31772915e-01 4.01208460e-01 5.40822208e-01
-1.92222372e-02 1.01696670e+00 1.06118575e-01 -3.96030694e-01
-8.15496564e-01 -6.67572260e-01 -1.06895292e+00 -8.02658737e-01
-4.08133477e-01 3.77313465e-01 -3.05530697e-01 1.94180831e-02
2.21376926e-01 5.50740063e-01 -1.38026321e+00 2.74138629e-01
1.35523725e+00 5.35433352e-01 -1.93709671e-01 5.51128983e-01
2.40756422e-01 1.89210668e-01 -2.71560818e-01 -4.54381227e-01
-1.01226635e-01 -4.98059183e-01 -1.72264233e-01 -5.49465716e-01
1.13443635e-01 6.20615706e-02 -8.91420618e-03 1.06524694e+00
6.38858914e-01 -4.10111964e-01 1.11255980e+00 -1.46787798e+00
-5.45265973e-01 2.87564903e-01 -7.86564052e-01 -1.43422022e-01
8.29474702e-02 1.86594307e-01 7.27965713e-01 -1.01794291e+00
5.72681904e-01 1.54479218e+00 5.29808939e-01 6.14081882e-02
-1.29185045e+00 -7.82101810e-01 -3.41905236e-01 3.02245580e-02
-1.40352297e+00 -2.63168037e-01 1.55675590e-01 -6.46389186e-01
3.70572418e-01 6.11727297e-01 4.33762938e-01 9.63430882e-01
8.92915666e-01 3.39620531e-01 1.33189130e+00 -1.11087844e-01
4.11348730e-01 -2.90643871e-01 9.78307873e-02 6.39866471e-01
8.57274055e-01 3.23614866e-01 -6.03953838e-01 -8.93780351e-01
4.55043375e-01 3.15453947e-01 2.07846224e-01 -8.87887366e-03
-1.20091486e+00 6.29626155e-01 -1.69819623e-01 5.11842035e-02
-4.12921399e-01 2.33620510e-01 2.01099232e-01 1.29996732e-01
4.11844641e-01 1.73668236e-01 -7.13122785e-01 1.74798340e-01
-5.47842085e-01 3.27529103e-01 4.17778164e-01 5.68968713e-01
5.15629470e-01 -1.38343289e-01 -1.30822286e-01 8.95314634e-01
5.27114391e-01 9.76218939e-01 4.03742552e-01 -5.52179277e-01
2.18848959e-01 6.78682745e-01 4.87449616e-02 -1.20747268e+00
-7.07896054e-01 -1.81524277e-01 -1.10407197e+00 -2.65002280e-01
3.74106765e-01 -9.32657197e-02 -8.30391467e-01 1.66394269e+00
5.96663058e-01 4.66648370e-01 8.27397779e-03 6.47842050e-01
6.71014845e-01 1.91680133e-01 2.24442393e-01 -2.91019350e-01
1.53606176e+00 -1.47455364e-01 -5.82366884e-01 2.48307467e-01
5.05253851e-01 -5.70587993e-01 7.47131348e-01 6.29295528e-01
-5.39111137e-01 -2.56340235e-01 -9.07182574e-01 3.00882459e-01
-4.92700070e-01 1.18885607e-01 7.16373384e-01 5.62282383e-01
-2.20369756e-01 7.40232825e-01 -8.42061639e-01 -3.11773509e-01
4.40899700e-01 5.64417660e-01 -6.16823018e-01 -3.49191278e-01
-1.22000098e+00 7.07803190e-01 1.24820001e-01 2.10325912e-01
-8.89352262e-01 -9.71737087e-01 -4.48638678e-01 -2.35615417e-01
-1.20727926e-01 -8.21476936e-01 2.27717608e-01 -2.62694955e-01
-1.20584869e+00 4.91226017e-01 -1.27400219e-01 -4.73997667e-02
3.10720205e-01 -3.76124471e-01 -8.68418872e-01 1.44946083e-01
2.05354393e-01 -1.49105070e-02 5.96658945e-01 -1.00943851e+00
-2.89858967e-01 -9.78622735e-01 -7.22616673e-01 -2.72075266e-01
-1.53713807e-01 1.77688479e-01 1.78133816e-01 -8.10099483e-01
3.89882267e-01 -1.08498549e+00 -4.21054244e-01 -1.76844403e-01
-9.29408789e-01 1.72835618e-01 1.46900937e-01 -6.77031755e-01
1.04933584e+00 -2.15503478e+00 7.70893767e-02 5.22630215e-01
9.19631049e-02 -3.76345329e-02 6.11528456e-02 5.97298265e-01
-3.01139921e-01 5.20319045e-01 -3.19907576e-01 4.21997160e-01
-9.44008231e-02 -1.22412078e-01 1.23473391e-01 1.07800841e+00
5.14164329e-01 8.78158152e-01 -8.89596462e-01 -2.26789653e-01
7.64711499e-02 2.84539044e-01 -4.61749494e-01 -5.33454008e-02
-7.68785626e-02 4.45532531e-01 -6.45115137e-01 1.15539372e+00
9.21566546e-01 -8.95176083e-02 3.01442444e-01 -1.94202140e-01
8.43383521e-02 -4.96020526e-01 -1.29567218e+00 1.25734448e+00
9.97890383e-02 -7.47038946e-02 -2.30867818e-01 -7.69904673e-01
9.34671640e-01 3.05359531e-02 6.22649133e-01 -5.26515603e-01
7.42188320e-02 5.23155928e-01 4.47830617e-01 -7.45700955e-01
-1.43658400e-01 -2.63353646e-01 -6.73574284e-02 7.05268532e-02
-9.90177765e-02 3.78334016e-01 -9.00889561e-02 -5.28169014e-02
1.35847187e+00 -9.08876210e-02 4.45487350e-01 -5.15952110e-01
4.32490557e-01 -1.50007352e-01 6.79688573e-01 6.47911847e-01
4.29375693e-02 5.47229469e-01 8.30437124e-01 -7.69603997e-03
-9.50340927e-01 -1.33599496e+00 -7.66281009e-01 5.63064933e-01
-1.67889684e-01 -5.09832650e-02 -6.77373931e-02 -1.44758269e-01
8.98566961e-01 5.36311984e-01 -6.69181705e-01 -3.62666816e-01
6.52643899e-03 -1.45804274e+00 9.61484671e-01 -4.28321287e-02
4.74613681e-02 -2.27298573e-01 1.60181791e-01 2.21980527e-01
1.56770468e-01 -5.73355854e-01 1.40526369e-01 2.08336890e-01
-1.07189119e+00 -1.47810233e+00 -5.86344719e-01 8.91235918e-02
6.81159079e-01 -2.69108802e-01 5.37211597e-01 -7.43037462e-02
-6.00412548e-01 -1.66437253e-01 -1.58210054e-01 -6.17719769e-01
-3.30598682e-01 -6.18301630e-01 5.67206323e-01 2.28387028e-01
3.62772495e-01 -6.41586840e-01 -6.68309987e-01 4.62043345e-01
-8.07358921e-01 -3.63495797e-01 8.81380498e-01 8.52352023e-01
6.26953542e-01 -1.27598926e-01 1.00869823e+00 -9.78630722e-01
4.34467256e-01 -1.29802883e+00 -2.67710507e-01 2.19191372e-01
-8.04112494e-01 -5.98202758e-02 3.95627588e-01 -5.42076766e-01
-5.88440716e-01 -1.68080926e-01 2.65516698e-01 -1.99504510e-01
-1.77358359e-01 1.01530802e+00 -4.10064250e-01 2.63027936e-01
5.86378038e-01 -3.38010192e-02 3.09394866e-01 -5.87910712e-01
1.73641950e-01 6.72323287e-01 3.57895404e-01 -7.43446231e-01
6.33819699e-01 5.68539500e-01 5.59541762e-01 -8.54382277e-01
-1.96030483e-01 -2.80880719e-01 -3.37537140e-01 1.44519120e-01
6.21009707e-01 -1.22203445e+00 -9.99046862e-01 5.94492912e-01
-6.14931285e-01 2.74990290e-01 4.60313171e-01 1.01205778e+00
-2.11443350e-01 2.45476767e-01 -5.77800214e-01 -8.75901818e-01
-2.72513270e-01 -9.07000244e-01 9.59390581e-01 -3.82535271e-02
-3.28218639e-01 -5.61010599e-01 2.56070048e-01 4.48132545e-01
-8.69062021e-02 9.53310728e-01 1.49974048e+00 -1.00734174e+00
-1.37731120e-01 -2.37594724e-01 -2.96824098e-01 1.11147650e-01
3.27638805e-01 4.87260103e-01 -8.04159582e-01 -3.72941326e-03
-4.12668645e-01 -2.78857276e-02 3.96794975e-01 6.51771247e-01
9.66321886e-01 -1.94956630e-01 -5.36868393e-01 5.54444790e-01
1.32323360e+00 2.16472939e-01 7.70432711e-01 2.21178398e-01
6.36283755e-01 4.89380777e-01 8.60745907e-01 6.46248460e-01
1.39807582e-01 4.69443500e-01 4.18166220e-01 -1.11669756e-01
2.74264097e-01 -2.24479079e-01 4.24756885e-01 3.98679793e-01
-5.23542315e-02 -1.72204167e-01 -9.41072464e-01 4.35941994e-01
-1.76875257e+00 -5.46023965e-01 -9.85623240e-01 2.28699708e+00
8.04402530e-01 -1.68374896e-01 -1.75793339e-02 -2.56637663e-01
8.15259218e-01 -6.30658805e-01 -9.91948605e-01 -2.84672648e-01
-4.87542242e-01 4.32462126e-01 5.93596697e-01 3.88498828e-02
-7.15944707e-01 5.17952502e-01 5.97518396e+00 6.28518999e-01
-6.36674643e-01 -1.55552968e-01 6.15408778e-01 -1.62923262e-01
-3.67385209e-01 2.63149925e-02 -5.82556367e-01 7.94840991e-01
1.28075516e+00 -2.83632845e-01 -2.47134566e-02 3.95476758e-01
9.04415846e-01 -4.06736523e-01 -1.01615191e+00 6.86348021e-01
-2.03058705e-01 -1.09418714e+00 -6.93772137e-02 5.01319051e-01
7.17467666e-01 -2.56984844e-03 -3.01307514e-02 -9.36052501e-02
3.96372288e-01 -1.24480951e+00 1.56754032e-01 9.70164180e-01
8.13991547e-01 -9.25026059e-01 7.91235805e-01 5.62990606e-02
-4.87985849e-01 -7.30163679e-02 -4.98291463e-01 9.92419049e-02
-2.78419405e-01 1.39226162e+00 -8.07471454e-01 9.79732871e-01
5.34106910e-01 5.08804917e-01 -6.26081586e-01 8.67453158e-01
-9.16519538e-02 7.13043392e-01 -3.41942340e-01 -3.43921520e-02
-4.40847009e-01 -2.27764145e-01 4.45414037e-01 7.78714061e-01
7.32115924e-01 2.55288273e-01 -2.28245586e-01 8.00680220e-01
2.55049974e-01 4.63341743e-01 -4.83267516e-01 -3.67621809e-01
5.82583666e-01 1.06113899e+00 -4.30310845e-01 -4.58856672e-02
-2.40266889e-01 5.00726402e-01 -1.71968743e-01 3.43389213e-01
-6.67829752e-01 -1.89338967e-01 1.14744031e+00 9.88775864e-02
-1.38299420e-01 9.34660435e-02 -3.83708894e-01 -9.56023455e-01
-2.71984071e-01 -1.29018879e+00 5.59418201e-01 -5.94230473e-01
-1.59741032e+00 -3.22248995e-01 -4.93055284e-02 -1.21221387e+00
-8.82329606e-03 -6.32651627e-01 -4.18732911e-01 9.26331580e-01
-1.15967321e+00 -1.14778090e+00 1.25698894e-01 4.39832777e-01
-3.36105347e-01 -2.24817678e-01 8.61483991e-01 3.71888131e-01
-8.18324447e-01 3.71931493e-01 7.60525584e-01 -2.24812716e-01
1.04097521e+00 -8.99796903e-01 -4.07516919e-02 5.30547738e-01
-6.23519182e-01 1.14825428e+00 7.50775933e-01 -1.33629882e+00
-1.86089289e+00 -1.15887511e+00 5.12097180e-01 -2.14957908e-01
8.75080109e-01 -2.15151906e-01 -6.34310961e-01 4.60975409e-01
-5.98394156e-01 1.31472554e-02 1.38073504e+00 2.19069228e-01
-3.59417200e-01 -5.16587459e-02 -1.36168385e+00 5.84245384e-01
6.47907317e-01 -1.48781344e-01 -2.29728714e-01 4.53397930e-01
4.82724249e-01 -2.33014133e-02 -1.62892127e+00 6.51447237e-01
9.39839303e-01 -6.30406618e-01 1.12079179e+00 -1.22731733e+00
7.32681394e-01 -6.18271530e-01 -4.85117257e-01 -1.01697850e+00
-2.06481650e-01 -3.10873628e-01 -1.72934741e-01 1.19217229e+00
4.04737115e-01 -6.33943439e-01 4.21846837e-01 5.52178681e-01
-9.32262465e-03 -6.11320198e-01 -1.18780851e+00 -8.91706586e-01
2.28628278e-01 -3.31866622e-01 5.57155430e-01 1.16911161e+00
1.16618350e-01 -9.94444564e-02 -5.61528623e-01 6.36418521e-01
7.38088608e-01 -1.23439915e-01 9.51372266e-01 -1.55120993e+00
-2.09265649e-01 -2.77170707e-02 -6.37314200e-01 1.32265776e-01
-2.82997906e-01 -9.35460925e-01 -4.78122115e-01 -1.31921053e+00
4.19576943e-01 -3.16824049e-01 -6.57710612e-01 3.47936928e-01
-2.94284582e-01 -1.67763576e-01 -2.29881927e-01 -3.25089060e-02
1.64511397e-01 3.89978260e-01 9.77974296e-01 -2.39086986e-01
-1.40282214e-01 -4.36222404e-02 -8.32275152e-01 5.86786926e-01
5.93504012e-01 -7.38365173e-01 -2.15683326e-01 2.49814361e-01
3.15526068e-01 7.95102678e-03 6.83268487e-01 -5.40859282e-01
-8.90038610e-02 -6.25731528e-01 8.75981569e-01 -6.31488502e-01
-2.22187191e-02 -5.41781604e-01 8.33883762e-01 9.94785666e-01
-6.78449636e-03 -7.51346797e-02 1.46324039e-01 7.20950961e-01
1.61956444e-01 -3.33944298e-02 5.36796451e-01 5.61988987e-02
-7.65188411e-02 2.59786874e-01 -3.56268197e-01 -1.07780434e-01
8.83170128e-01 -4.52173091e-02 -5.75196326e-01 1.43382266e-01
-7.38644361e-01 2.01508343e-01 1.29707128e-01 4.60008442e-01
6.67209089e-01 -1.31801963e+00 -1.16533244e+00 4.60769124e-02
4.01234418e-01 -2.65881598e-01 4.39555526e-01 1.10065424e+00
-4.23734426e-01 7.41738677e-02 -1.89167053e-01 -6.63437724e-01
-1.03806007e+00 4.16421294e-01 -3.61199453e-02 9.42694489e-03
-1.42034173e-01 5.43684661e-01 1.58942536e-01 -4.37458485e-01
-2.90549129e-01 -2.17496306e-01 7.27904961e-02 2.45089501e-01
3.76483530e-01 8.20506513e-01 1.31594136e-01 -3.13778341e-01
-6.78984344e-01 3.68699700e-01 1.49845764e-01 2.19929621e-01
1.62786794e+00 2.11249662e-05 -6.30957127e-01 6.86683774e-01
1.19409001e+00 1.76991314e-01 -3.75239462e-01 1.97301030e-01
2.74988055e-01 -4.33177888e-01 -4.68394756e-01 -1.03902113e+00
-5.11874199e-01 3.90840471e-01 9.73090410e-01 -3.60171229e-01
7.26129711e-01 -7.96195567e-02 1.45969808e-01 1.10926650e-01
1.71861380e-01 -9.97238755e-01 -2.84159303e-01 -3.19888107e-02
8.09951723e-01 -9.78600740e-01 4.25348192e-01 -4.36734319e-01
-1.65168107e-01 1.14420843e+00 4.19822544e-01 1.22052111e-01
7.45853007e-01 6.94885179e-02 -2.19430953e-01 -3.62594724e-01
-6.74373269e-01 1.79301172e-01 3.51318806e-01 5.42264104e-01
4.51764613e-01 3.78873736e-01 -7.41392434e-01 1.01567483e+00
-1.20701924e-01 6.55944720e-02 4.42144901e-01 5.11297643e-01
-3.43581617e-01 -9.82710838e-01 -6.06886327e-01 9.13633227e-01
-6.39391363e-01 -2.29052156e-01 -3.30473632e-01 1.07820511e+00
2.49295950e-01 1.00680017e+00 -2.10025996e-01 -3.63143563e-01
3.48738045e-01 2.03772187e-01 1.32704685e-02 -3.77724975e-01
-3.74822348e-01 4.93563861e-01 1.93132550e-01 -5.51916420e-01
-2.58062243e-01 -8.77724290e-01 -1.33169091e+00 -3.60604942e-01
-4.17268962e-01 -5.08857444e-02 8.55289042e-01 9.52532709e-01
6.80841506e-01 4.79728103e-01 6.46189630e-01 -1.31323889e-01
-4.48207706e-01 -1.04313087e+00 -1.03141880e+00 3.23546976e-01
1.12116121e-01 -7.93535769e-01 -4.30761695e-01 -1.50545329e-01] | [6.429522514343262, 5.523967742919922] |
3965592a-cfa4-400e-aaac-4f459ca66e82 | improving-adversarial-robustness-via-mutual | 2207.12203 | null | https://arxiv.org/abs/2207.12203v1 | https://arxiv.org/pdf/2207.12203v1.pdf | Improving Adversarial Robustness via Mutual Information Estimation | Deep neural networks (DNNs) are found to be vulnerable to adversarial noise. They are typically misled by adversarial samples to make wrong predictions. To alleviate this negative effect, in this paper, we investigate the dependence between outputs of the target model and input adversarial samples from the perspective of information theory, and propose an adversarial defense method. Specifically, we first measure the dependence by estimating the mutual information (MI) between outputs and the natural patterns of inputs (called natural MI) and MI between outputs and the adversarial patterns of inputs (called adversarial MI), respectively. We find that adversarial samples usually have larger adversarial MI and smaller natural MI compared with those w.r.t. natural samples. Motivated by this observation, we propose to enhance the adversarial robustness by maximizing the natural MI and minimizing the adversarial MI during the training process. In this way, the target model is expected to pay more attention to the natural pattern that contains objective semantics. Empirical evaluations demonstrate that our method could effectively improve the adversarial accuracy against multiple attacks. | ['Tongliang Liu', 'Yibing Zhan', 'Xiaoyu Wang', 'Bo Han', 'Xinbo Gao', 'Nannan Wang', 'Dawei Zhou'] | 2022-07-25 | null | null | null | null | ['adversarial-defense', 'mutual-information-estimation'] | ['adversarial', 'methodology'] | [ 3.62629890e-01 4.42063510e-01 2.87380248e-01 -4.30817157e-01
-3.32305968e-01 -9.84540582e-01 6.72855139e-01 -2.52058804e-01
-4.02887553e-01 4.91334766e-01 8.36892724e-02 -2.82720868e-02
8.84554163e-02 -1.11751354e+00 -1.08663511e+00 -5.65896392e-01
1.26169622e-01 2.04886228e-01 7.63576999e-02 -2.11501583e-01
1.62219942e-01 6.06407523e-01 -1.04374528e+00 2.51173288e-01
9.35140371e-01 1.04272020e+00 -2.96628743e-01 4.93593246e-01
1.61781743e-01 1.10105729e+00 -1.03912807e+00 -8.70728016e-01
4.71350610e-01 -5.76636791e-01 -6.22005463e-01 -1.19236693e-01
3.81722718e-01 -2.93697536e-01 -8.91106784e-01 1.80906093e+00
1.74861044e-01 2.22683325e-01 8.07496369e-01 -1.48473597e+00
-1.02393377e+00 8.93144310e-01 -7.12238550e-02 9.03706774e-02
9.92797762e-02 5.61762810e-01 7.63948917e-01 -6.75822735e-01
4.18643981e-01 1.59859848e+00 2.52418667e-01 1.04775059e+00
-1.10677028e+00 -9.61029053e-01 4.17532504e-01 -5.43561354e-02
-1.16526186e+00 -4.63742018e-01 9.83238459e-01 -2.53414214e-01
2.43274286e-01 4.46070880e-01 2.65849032e-03 1.59311616e+00
2.49977246e-01 8.29303980e-01 8.72204006e-01 -1.41991094e-01
3.83427143e-01 5.03551126e-01 -2.23095641e-01 4.08172876e-01
1.11579694e-01 5.35077989e-01 -8.68626311e-02 -9.17013213e-02
4.44381982e-01 2.41220415e-01 -4.00390625e-01 -1.15531690e-01
-7.66838372e-01 7.83251226e-01 8.98869038e-01 2.17839390e-01
-1.77871183e-01 2.59295218e-02 4.10161942e-01 7.04250157e-01
1.66777819e-01 7.29993820e-01 -4.33903188e-01 3.83328706e-01
-2.75811195e-01 2.07395598e-01 7.43350804e-01 8.42952311e-01
5.41405082e-01 2.62976915e-01 -1.17381349e-01 6.84783936e-01
3.83039057e-01 8.67251635e-01 5.77493608e-01 -6.56764984e-01
6.73705816e-01 5.05840838e-01 -3.40593420e-02 -1.35776997e+00
4.27292325e-02 -4.53307331e-01 -1.19028962e+00 4.13734823e-01
5.98335624e-01 -2.91478992e-01 -8.54391336e-01 2.16391373e+00
-3.55658829e-02 -2.70973463e-02 5.56184888e-01 7.62014866e-01
4.14246261e-01 4.51463938e-01 1.26330614e-01 -6.64283708e-03
7.57032812e-01 -5.88937938e-01 -5.51768661e-01 -6.77793145e-01
2.73539931e-01 -5.33937633e-01 1.24395549e+00 1.98245525e-01
-9.46429253e-01 -4.17759836e-01 -1.11680233e+00 6.24114454e-01
-3.64599496e-01 -4.57119882e-01 -4.70360406e-02 7.97337115e-01
-4.67587739e-01 1.05275130e+00 -7.00924337e-01 1.85955346e-01
7.42456734e-01 2.46358067e-01 -4.39145565e-01 -1.26574963e-01
-1.60547638e+00 9.12274361e-01 5.93349993e-01 5.99931553e-03
-1.28597105e+00 -4.37687248e-01 -8.23937953e-01 7.12746195e-03
3.29457015e-01 -3.23571742e-01 1.02917862e+00 -1.63962138e+00
-1.13205516e+00 6.30698264e-01 3.43195975e-01 -6.80136204e-01
8.99355471e-01 -3.29674393e-01 -4.86166894e-01 -3.20697241e-02
-2.74817735e-01 3.81096900e-01 1.29995513e+00 -1.44398689e+00
-1.19367979e-01 -4.72244322e-01 2.30696052e-01 3.17715257e-02
-8.40711594e-01 -1.91027597e-01 -4.57007065e-02 -1.12573671e+00
5.57379760e-02 -8.37785840e-01 -3.87288749e-01 -9.31132864e-03
-8.49722981e-01 1.32912070e-01 7.60689020e-01 -2.23362252e-01
9.41031277e-01 -2.40230012e+00 7.01879710e-02 4.80178684e-01
3.62096965e-01 5.11969745e-01 -1.25819668e-01 5.44638485e-02
-1.75237492e-01 5.01548648e-01 -3.14363986e-01 1.39647156e-01
1.20523758e-01 4.06693131e-01 -7.63121724e-01 4.49165523e-01
4.50982571e-01 8.66619468e-01 -9.64978039e-01 2.44028606e-02
-1.56484433e-02 9.37245265e-02 -5.34696519e-01 6.08494759e-01
-2.43488863e-01 3.81066412e-01 -5.80824554e-01 3.83513063e-01
7.80786216e-01 5.10775298e-02 -6.95652068e-02 -1.07119724e-01
8.41713786e-01 -8.85316357e-02 -9.47449088e-01 8.62112105e-01
-2.64944732e-01 5.18864989e-01 -2.21481323e-01 -9.15971160e-01
1.10690582e+00 2.37455547e-01 -1.74724460e-01 -4.75752592e-01
3.56862068e-01 4.13418328e-03 3.08988661e-01 -2.73148358e-01
-2.31718421e-02 -4.66274828e-01 -1.96924627e-01 4.37937140e-01
-9.86334309e-02 8.33519325e-02 -4.45571244e-01 3.17377746e-01
1.17525959e+00 -4.25720274e-01 1.88067570e-01 -7.81906620e-02
5.90839446e-01 -4.23590064e-01 6.87234461e-01 1.06403649e+00
-4.15107906e-01 3.38117391e-01 6.93973780e-01 -3.69555295e-01
-1.02873635e+00 -1.56389284e+00 1.11846425e-01 7.63823628e-01
3.35093915e-01 1.19828574e-01 -9.99554276e-01 -1.29216588e+00
7.41714761e-02 9.68408346e-01 -8.52041245e-01 -1.06175876e+00
-4.42849070e-01 -3.27343613e-01 1.00290465e+00 5.95282495e-01
6.32841408e-01 -1.27375567e+00 -8.31773728e-02 7.86538608e-03
1.65678427e-01 -8.81062567e-01 -6.09819531e-01 1.15721285e-01
-6.67956591e-01 -1.03093600e+00 -3.34629893e-01 -4.33049887e-01
8.66389275e-01 -1.80122286e-01 1.08345604e+00 5.93588389e-02
1.38639435e-01 -2.41021477e-02 -4.27929074e-01 -5.52562356e-01
-1.04224503e+00 -2.93548197e-01 4.78399605e-01 1.28619060e-01
2.65766323e-01 -5.90040624e-01 -3.76114666e-01 5.15409410e-01
-1.46942675e+00 -3.11222136e-01 5.68117738e-01 9.00686085e-01
2.74429679e-01 3.34899485e-01 7.21309245e-01 -1.17605412e+00
6.77075267e-01 -7.06442833e-01 -4.07537848e-01 2.37232432e-01
-5.29312074e-01 2.44966820e-01 1.55691350e+00 -1.01643503e+00
-7.81840682e-01 -2.89363146e-01 -1.97963819e-01 -9.71887469e-01
-3.02226663e-01 6.32634759e-02 -1.01178181e+00 -4.02336642e-02
9.46233511e-01 1.41278684e-01 -2.02736005e-01 -1.76684782e-01
3.71506363e-01 4.54194039e-01 5.18027365e-01 -5.02107322e-01
1.36317945e+00 1.30677819e-01 -1.51581854e-01 -3.22596669e-01
-1.03974068e+00 4.61570293e-01 -3.35131496e-01 -2.27399662e-01
4.77243662e-01 -3.48527640e-01 -4.20104027e-01 6.50718629e-01
-9.58755374e-01 -8.37590098e-02 -3.06432694e-01 3.17038625e-01
-4.47605669e-01 3.02093118e-01 -5.25333881e-01 -7.76183069e-01
-2.02772424e-01 -1.19144058e+00 2.95705974e-01 2.04438120e-01
-1.60529986e-01 -1.12617230e+00 -2.72705406e-01 2.33003739e-02
3.22160542e-01 5.35150290e-01 9.72860336e-01 -1.36510694e+00
-3.48016769e-01 -6.36190951e-01 -1.38046771e-01 9.99839425e-01
1.88180313e-01 -1.35663047e-01 -1.20316541e+00 -1.89257145e-01
5.38372099e-01 -4.12204862e-01 7.10025132e-01 -8.67654309e-02
1.29617763e+00 -1.03421366e+00 2.19760444e-02 7.58238733e-01
1.31240165e+00 3.01302224e-01 7.46591151e-01 2.34583437e-01
8.59279513e-01 5.23842454e-01 6.04914069e-01 2.90791214e-01
-3.73720229e-01 2.51412272e-01 9.67507720e-01 2.72198677e-01
3.81793112e-01 -6.03093863e-01 6.28092527e-01 4.25456941e-01
5.46573997e-01 -7.12434351e-01 -8.46721530e-01 2.57556021e-01
-1.52036691e+00 -9.84688163e-01 3.29092711e-01 2.22246051e+00
8.90021622e-01 7.48501003e-01 -2.73287922e-01 2.25693747e-01
7.72404432e-01 2.79094785e-01 -1.06777847e+00 -3.95713478e-01
-2.77233392e-01 -2.58221501e-03 5.21326602e-01 3.62239569e-01
-1.20510232e+00 9.87974524e-01 5.68945646e+00 1.00935638e+00
-8.62998664e-01 -2.10519269e-01 9.66605544e-01 -6.22988902e-02
-6.04696095e-01 -3.07171673e-01 -3.68574023e-01 6.60807252e-01
9.50394034e-01 -3.85037392e-01 6.43058956e-01 1.04837584e+00
-2.28948027e-01 6.48140788e-01 -1.34274268e+00 4.12559628e-01
-2.39564613e-01 -9.46438968e-01 4.59001124e-01 1.81836337e-02
6.70625389e-01 -2.06159189e-01 4.52085972e-01 3.35375249e-01
7.47251987e-01 -1.21498585e+00 7.08389521e-01 7.12060630e-01
4.87102330e-01 -9.96697068e-01 8.49175215e-01 7.07770288e-01
-6.86039507e-01 -1.32333964e-01 -5.01696348e-01 1.70847207e-01
-4.70229477e-01 5.08650184e-01 -5.69719017e-01 1.47744030e-01
3.60420495e-01 5.45117438e-01 -4.68749791e-01 2.69956231e-01
-4.60901439e-01 5.51599801e-01 -1.10927708e-01 -1.16585232e-02
2.10561872e-01 3.36690769e-02 7.23785460e-01 7.02852964e-01
-1.69388339e-01 -1.40178308e-01 9.14830565e-02 1.12435818e+00
-6.75628781e-01 -1.68531567e-01 -1.16910958e+00 -3.05413693e-01
6.11185968e-01 7.20828474e-01 -2.69055158e-01 -3.18463832e-01
-3.81173119e-02 1.16624558e+00 3.66246670e-01 4.43430066e-01
-8.12395692e-01 -4.34963286e-01 9.71295774e-01 -3.27288836e-01
-1.82913363e-01 3.41845751e-01 -4.86488104e-01 -1.08114398e+00
2.36650780e-01 -1.18493116e+00 3.03617448e-01 -4.33576316e-01
-1.91291356e+00 8.94747138e-01 -4.19139117e-01 -1.36047137e+00
-1.83957219e-01 -6.39063060e-01 -8.37858021e-01 9.61958766e-01
-8.60545576e-01 -6.71681762e-01 1.29034564e-01 7.66164064e-01
3.50230724e-01 -4.40834045e-01 8.94293427e-01 -2.87789963e-02
-6.84862077e-01 1.28933275e+00 2.41718993e-01 6.51902497e-01
4.40679729e-01 -1.13321686e+00 6.82783544e-01 1.06340849e+00
2.98447963e-02 7.10133553e-01 8.38676512e-01 -5.70524156e-01
-1.04360998e+00 -1.51276398e+00 3.90484184e-01 -5.20118892e-01
8.97654474e-01 -8.38626400e-02 -1.09295630e+00 7.43484199e-01
-1.75847232e-01 4.74589095e-02 6.82984829e-01 -4.28970605e-01
-7.46079683e-01 -1.96915530e-02 -1.76265728e+00 9.29481864e-01
9.63845968e-01 -7.52231061e-01 -7.00316250e-01 3.11138242e-01
1.06391823e+00 -1.76014662e-01 -8.17909300e-01 4.35279191e-01
4.13047791e-01 -8.71349573e-01 1.04814422e+00 -1.09128928e+00
6.47614360e-01 1.73325557e-02 -3.78249526e-01 -1.55117583e+00
-2.03677088e-01 -3.28144819e-01 -2.66529292e-01 1.13728261e+00
4.01138842e-01 -7.36622453e-01 7.26331651e-01 8.28255713e-01
9.63227600e-02 -5.98395288e-01 -8.10979545e-01 -8.92892063e-01
4.11358416e-01 -4.51334864e-01 6.82134509e-01 1.02712953e+00
-3.20279121e-01 -4.15339619e-02 -5.09256840e-01 5.75280368e-01
7.37306774e-01 -2.25124270e-01 5.83325922e-01 -8.96298528e-01
-3.83126140e-01 -4.97717500e-01 -6.21269107e-01 -8.12457621e-01
4.00411785e-01 -7.63994217e-01 1.60448104e-01 -6.72478616e-01
-1.59487590e-01 -4.10455227e-01 -6.37622058e-01 3.79564852e-01
-5.80929935e-01 2.40464389e-01 3.73738647e-01 2.64668941e-01
-2.62910128e-01 6.48479223e-01 1.26214027e+00 -3.95634711e-01
9.96513292e-02 3.72230411e-01 -9.28668618e-01 1.13684535e+00
9.70524192e-01 -8.35689008e-01 -5.16398907e-01 -3.42280239e-01
3.98040973e-02 -1.77316874e-01 2.94382393e-01 -9.55418050e-01
-9.63106453e-02 -4.33535397e-01 5.26374578e-01 -6.41063452e-02
8.97625089e-02 -1.08434737e+00 -1.41310364e-01 7.00047255e-01
-9.06769812e-01 -1.69255689e-01 -2.04496551e-03 8.56350660e-01
-1.06387585e-01 -4.21684921e-01 1.11138916e+00 -2.51208901e-01
-2.07795039e-01 5.23781478e-01 -2.32401624e-01 3.55839610e-01
1.08348453e+00 2.34901413e-01 -3.42998534e-01 -4.19316024e-01
-8.99567127e-01 1.64846331e-01 5.17764986e-01 5.43482721e-01
8.81181240e-01 -1.54797864e+00 -5.22171080e-01 4.49466884e-01
1.53333575e-01 -1.76419124e-01 6.26854375e-02 2.15319842e-01
-2.16357887e-01 -7.17824325e-02 -1.73664749e-01 -9.15867165e-02
-1.02496958e+00 9.45147157e-01 6.21457696e-01 -2.60439783e-01
-3.53893459e-01 1.04623711e+00 5.99911571e-01 -5.22686720e-01
3.98934454e-01 9.26983804e-02 -1.93462539e-02 -3.83302867e-01
6.16743684e-01 2.29675591e-01 -2.37043425e-01 -4.99709308e-01
-1.53185919e-01 -1.33033087e-02 -3.73864949e-01 6.22450300e-02
9.37533617e-01 8.39618742e-02 1.25909805e-01 3.08105916e-01
1.44310534e+00 -2.94490810e-02 -1.33665478e+00 -5.42408526e-01
1.43041695e-02 -7.19383180e-01 -2.75776237e-01 -7.74668753e-01
-1.34202921e+00 9.08606052e-01 6.86405420e-01 5.78005552e-01
1.25495648e+00 -1.44937858e-01 7.48578846e-01 5.08681595e-01
8.66922438e-02 -6.89854860e-01 2.77657777e-01 5.40698707e-01
8.59450102e-01 -1.14342344e+00 -5.70829690e-01 -6.99416548e-02
-8.96716237e-01 8.93261611e-01 8.07494342e-01 -5.38421214e-01
6.48485899e-01 6.85537308e-02 1.88267186e-01 7.76099488e-02
-6.57665670e-01 2.32187405e-01 4.57061350e-01 7.17996478e-01
-1.04837656e-01 2.04099238e-01 3.20786715e-01 6.86455071e-01
-3.06384802e-01 -5.81494749e-01 2.46237352e-01 7.14839160e-01
-4.69329119e-01 -9.87095237e-01 -4.42326158e-01 5.01005173e-01
-6.70626879e-01 -1.99534550e-01 -8.26958776e-01 3.51998597e-01
2.74706427e-02 1.01104629e+00 -1.82608679e-01 -9.96038616e-01
4.85979050e-01 -1.01585314e-02 9.08733308e-02 -5.06732523e-01
-7.37093151e-01 -5.58254063e-01 -3.17383021e-01 -5.85591137e-01
9.79874730e-02 -3.08507830e-01 -9.59721446e-01 -4.89483595e-01
-1.71187073e-01 -5.39738089e-02 2.84453750e-01 1.12606668e+00
-6.20463416e-02 6.53912544e-01 1.32469106e+00 -4.79087710e-01
-1.32200098e+00 -9.76974845e-01 -5.12048483e-01 9.04182255e-01
4.49011177e-01 -2.70420194e-01 -8.87713730e-01 -1.19998246e-01] | [5.689352512359619, 7.883220195770264] |
c22fccaa-7bf0-4978-aaed-0c598a33ae3e | del-dock-molecular-docking-enabled-modeling | 2212.00136 | null | https://arxiv.org/abs/2212.00136v2 | https://arxiv.org/pdf/2212.00136v2.pdf | DEL-Dock: Molecular Docking-Enabled Modeling of DNA-Encoded Libraries | DNA-Encoded Library (DEL) technology has enabled significant advances in hit identification by enabling efficient testing of combinatorially-generated molecular libraries. DEL screens measure protein binding affinity though sequencing reads of molecules tagged with unique DNA-barcodes that survive a series of selection experiments. Computational models have been deployed to learn the latent binding affinities that are correlated to the sequenced count data; however, this correlation is often obfuscated by various sources of noise introduced in its complicated data-generation process. In order to denoise DEL count data and screen for molecules with good binding affinity, computational models require the correct assumptions in their modeling structure to capture the correct signals underlying the data. Recent advances in DEL models have focused on probabilistic formulations of count data, but existing approaches have thus far been limited to only utilizing 2-D molecule-level representations. We introduce a new paradigm, DEL-Dock, that combines ligand-based descriptors with 3-D spatial information from docked protein-ligand complexes. 3-D spatial information allows our model to learn over the actual binding modality rather than using only structured-based information of the ligand. We show that our model is capable of effectively denoising DEL count data to predict molecule enrichment scores that are better correlated with experimental binding affinity measurements compared to prior works. Moreover, by learning over a collection of docked poses we demonstrate that our model, trained only on DEL data, implicitly learns to perform good docking pose selection without requiring external supervision from expensive-to-source protein crystal structures. | ['Theofanis Karaletsos', 'Mohammad M. Sultan', 'Benson Chen', 'Kirill Shmilovich'] | 2022-11-30 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 5.60328841e-01 -5.24033248e-01 -1.94027096e-01 -2.90997207e-01
-1.41871560e+00 -1.02581155e+00 4.07723159e-01 4.18220818e-01
-5.27612984e-01 1.51219308e+00 3.13217849e-01 -2.37571687e-01
-1.65946752e-01 -7.16072142e-01 -1.12239969e+00 -9.97820675e-01
8.28641355e-02 1.11302078e+00 3.95333245e-02 -1.25719711e-01
2.86894381e-01 6.88236773e-01 -1.17539275e+00 5.30394793e-01
6.50724113e-01 4.87935036e-01 3.63829195e-01 6.77231669e-01
-2.07343549e-01 2.55645812e-01 -4.80791152e-01 -3.24272424e-01
-1.23538263e-02 -4.56818521e-01 -3.26328933e-01 -6.31244540e-01
1.52599931e-01 1.64129138e-02 -1.50648072e-01 8.09304655e-01
9.30544794e-01 -1.80309385e-01 1.05687726e+00 -4.16035026e-01
-3.33560467e-01 3.38542499e-02 -2.08492517e-01 7.77476728e-02
7.91522264e-01 3.92015159e-01 1.00441086e+00 -1.27760196e+00
9.36487556e-01 1.03153026e+00 7.21460938e-01 4.71529812e-01
-2.07005763e+00 -5.59981227e-01 -4.40106690e-01 -5.42918369e-02
-1.60189295e+00 -1.84467733e-01 2.67055571e-01 -5.88814914e-01
1.40431345e+00 3.69777173e-01 5.94332755e-01 1.51138616e+00
4.22409058e-01 3.23189020e-01 8.13406527e-01 -2.45712131e-01
5.48186123e-01 -4.22430724e-01 -1.35963798e-01 5.99017441e-01
2.99273103e-01 1.28695041e-01 -8.33519280e-01 -8.23543251e-01
5.06426096e-01 3.31513554e-01 -3.09582353e-01 -8.07893336e-01
-1.02144432e+00 9.63253021e-01 2.70061612e-01 -1.09530345e-01
-6.09995902e-01 -4.84803319e-02 1.85290217e-01 -2.36504301e-01
1.53387010e-01 7.22699225e-01 -8.45769167e-01 -2.67852575e-01
-8.47850144e-01 4.86883700e-01 7.12255955e-01 6.60406768e-01
9.11212981e-01 -5.93596816e-01 -1.11556262e-01 5.65769136e-01
6.96946234e-02 4.75240767e-01 3.13723028e-01 -2.44867682e-01
1.72198728e-01 6.25452816e-01 5.28811634e-01 -5.03481448e-01
-5.45369029e-01 -3.45613092e-01 -4.96135563e-01 -1.35509431e-01
4.42557365e-01 5.87168895e-02 -1.07502317e+00 1.81275260e+00
2.12159261e-01 6.21112287e-02 1.04752675e-01 7.82367170e-01
5.06182909e-01 5.69468319e-01 2.81948090e-01 -2.91636348e-01
1.17823577e+00 -2.14947417e-01 -3.09202552e-01 8.42761546e-02
6.79255128e-01 -5.59618831e-01 9.03925419e-01 4.32804823e-01
-8.06275070e-01 -1.23977019e-02 -1.11055076e+00 5.87119684e-02
-5.59346437e-01 -3.11027095e-02 7.53643513e-01 6.63959682e-01
-5.05840659e-01 6.38871193e-01 -9.09654856e-01 -2.22868860e-01
6.14218533e-01 8.71514976e-01 -6.25741661e-01 -2.73764610e-01
-1.02441394e+00 1.00441849e+00 4.63060826e-01 -1.58527538e-01
-1.34110367e+00 -9.05785263e-01 -4.81699228e-01 -1.51810059e-02
2.88217366e-01 -7.50731587e-01 8.71623755e-01 -2.57553756e-01
-1.32084072e+00 6.28999531e-01 -4.20758516e-01 -2.88133085e-01
5.23213036e-02 -8.78720433e-02 5.85681498e-02 -3.28547098e-02
4.05963361e-02 4.78125334e-01 1.64910242e-01 -9.25713181e-01
-2.55204856e-01 -6.70570552e-01 -1.63798422e-01 2.43161649e-01
5.43448776e-02 -1.68686420e-01 -2.25723013e-01 -3.21957201e-01
1.14815183e-01 -7.76920021e-01 -4.64985967e-01 -4.41979021e-01
-4.66820389e-01 2.30426252e-01 1.91125304e-01 -2.20123395e-01
7.63900578e-01 -1.58902860e+00 6.74203098e-01 3.52780819e-01
2.91122228e-01 3.68223637e-01 -4.07802463e-01 9.34394777e-01
-7.81758726e-02 7.60874078e-02 -4.62598205e-02 2.40681469e-01
-2.07987368e-01 9.27211493e-02 -2.30563775e-01 6.28793776e-01
2.91913390e-01 9.97253835e-01 -9.88755822e-01 4.32557464e-02
4.55709957e-02 9.53141928e-01 -8.17396998e-01 2.69450814e-01
-7.59383917e-01 7.18344688e-01 -6.19321644e-01 6.18666530e-01
6.96230888e-01 -3.65998924e-01 6.82270408e-01 -3.31969202e-01
-2.20256876e-02 3.68681759e-01 -8.29604566e-01 1.89364529e+00
-7.87448436e-02 -2.18447849e-01 -4.13315684e-01 -6.28942430e-01
8.41699362e-01 2.18694299e-01 5.69645405e-01 -5.09330451e-01
-6.33069947e-02 3.98868233e-01 -4.21141200e-02 -1.66745484e-01
-5.48768640e-02 -4.53233570e-01 6.56519691e-03 -5.35085015e-02
2.50416696e-01 2.07046606e-02 -3.48200202e-02 1.49533659e-01
1.50566947e+00 3.25512588e-01 2.75235116e-01 -3.75832133e-02
4.95638162e-01 3.67578328e-01 6.23583674e-01 7.97886133e-01
2.30993599e-01 5.20029902e-01 5.04146755e-01 -5.88717103e-01
-1.31218636e+00 -1.24149466e+00 -3.82939607e-01 9.87278223e-01
-9.59388912e-02 -5.48614621e-01 -5.65398037e-01 -4.87951815e-01
6.56896681e-02 3.54939938e-01 -5.13313949e-01 -2.96956122e-01
-3.82656485e-01 -1.62619698e+00 8.17368388e-01 2.16590121e-01
-4.22052473e-01 -4.73643601e-01 -6.84687197e-02 6.40678883e-01
4.93736267e-02 -5.48234522e-01 -3.12863648e-01 1.01900959e+00
-6.04712605e-01 -1.38647628e+00 -6.31771088e-01 -4.43498105e-01
6.24460280e-01 -1.16551138e-01 9.08430934e-01 -3.87512267e-01
-7.48197019e-01 -8.59360304e-03 -3.52845974e-02 -4.43416476e-01
-1.44257054e-01 -1.57853924e-02 2.93234557e-01 -3.07980418e-01
9.32950974e-01 -6.13185585e-01 -7.57553339e-01 2.48967230e-01
-7.99035490e-01 -1.38998091e-01 8.48648250e-01 1.40399206e+00
1.14638865e+00 -4.08560306e-01 6.64982259e-01 -8.69556606e-01
5.93908191e-01 -4.39035505e-01 -6.26861036e-01 2.33527154e-01
-3.37194413e-01 5.59994578e-01 4.88102078e-01 -3.70511502e-01
-8.84365559e-01 5.12705266e-01 -4.24173534e-01 -8.17227140e-02
-9.48742554e-02 7.32987642e-01 -4.26123798e-01 -1.35518700e-01
8.95844638e-01 4.52889115e-01 -2.12678220e-02 -5.36363482e-01
2.13075235e-01 4.61882681e-01 3.88131380e-01 -8.83258104e-01
2.45315656e-01 4.23528790e-01 4.12517041e-01 -5.52495241e-01
-7.01659679e-01 -6.13172352e-01 -6.10517263e-01 4.67066497e-01
7.68543839e-01 -9.30886865e-01 -1.16867757e+00 1.07856765e-01
-9.90224481e-01 6.96463324e-03 2.00753763e-01 6.92284703e-01
-6.51352882e-01 3.55753690e-01 -7.15898514e-01 -8.13407898e-01
-1.21890455e-01 -1.52946079e+00 1.33483219e+00 -2.14624435e-01
-1.43677965e-01 -6.80918872e-01 7.08195508e-01 1.41247183e-01
9.76072177e-02 2.81654626e-01 1.17447305e+00 -9.24475312e-01
-7.38292217e-01 -5.55235386e-01 -4.40875255e-02 -1.46203741e-01
9.90016386e-02 -2.95071125e-01 -7.49897480e-01 -3.38754237e-01
-3.39762509e-01 -6.28582358e-01 9.74383652e-01 4.45048779e-01
8.52245390e-01 6.77568140e-03 -6.91864431e-01 6.67825043e-01
1.63543344e+00 1.38829380e-01 7.27957010e-01 1.15548864e-01
6.22753024e-01 1.22580573e-01 5.22195637e-01 5.30237198e-01
-2.23275408e-01 9.42290127e-01 5.23956478e-01 -8.83323792e-03
3.85420650e-01 -6.95897758e-01 2.26550117e-01 1.01468831e-01
-2.33946115e-01 -4.42873001e-01 -8.16999078e-01 -9.97053683e-02
-1.70435357e+00 -9.85562742e-01 -1.29496023e-01 2.48235345e+00
1.30634201e+00 3.42978258e-03 1.62595585e-01 -2.34134763e-01
3.89809221e-01 -3.67154062e-01 -7.66986787e-01 1.06388032e-01
-5.19170284e-01 6.39182389e-01 7.20894694e-01 6.79567873e-01
-8.15912604e-01 7.09169507e-01 7.01963902e+00 1.02503121e+00
-9.07691658e-01 -2.25508139e-01 4.38211560e-01 -2.73950964e-01
-4.36092764e-01 1.83984563e-01 -1.16128349e+00 3.68794292e-01
1.09793949e+00 1.80191733e-02 3.15116554e-01 5.57966888e-01
3.43597382e-01 -3.11529133e-02 -1.42269015e+00 9.90074813e-01
-3.18482190e-01 -1.88343942e+00 1.94551110e-01 3.43518287e-01
5.83876550e-01 1.57176837e-01 -1.86510999e-02 -6.68187998e-03
4.81020629e-01 -1.46677935e+00 1.49339184e-01 8.52345884e-01
9.32378292e-01 -8.31939995e-01 9.36205745e-01 5.12056828e-01
-7.61939228e-01 2.16499105e-01 -7.44860947e-01 1.53455893e-02
-5.70146926e-02 7.25298762e-01 -1.49279225e+00 4.90948617e-01
1.25952840e-01 4.59140599e-01 -1.70030385e-01 7.76000261e-01
1.35180935e-01 3.15384328e-01 -4.48272914e-01 -2.52779305e-01
8.44442658e-03 -3.46346468e-01 2.63842791e-01 1.19972801e+00
2.77310729e-01 1.39642939e-01 3.18942338e-01 8.93191516e-01
-1.86334521e-01 5.19607626e-02 -5.24214625e-01 -2.97949642e-01
4.66639459e-01 8.67472529e-01 -3.16049069e-01 -2.85536377e-03
-3.49427164e-01 8.89938116e-01 4.58889008e-01 4.86904621e-01
-7.18492985e-01 8.80659651e-03 9.89360750e-01 2.22098425e-01
2.51642466e-01 -2.20708773e-01 1.05655484e-01 -9.61039603e-01
-4.30952609e-01 -8.84927571e-01 2.64458299e-01 -4.32333350e-01
-1.48200071e+00 1.53745994e-01 -3.26435894e-01 -6.54547811e-01
7.43979514e-02 -1.02358031e+00 5.74322492e-02 1.34184361e+00
-1.14188385e+00 -7.96516836e-01 2.62703180e-01 2.27696329e-01
1.68167368e-01 -8.73420537e-02 1.39418435e+00 2.73261756e-01
-5.51603615e-01 3.23489159e-01 6.80099845e-01 -3.64222556e-01
9.12464082e-01 -1.18637574e+00 6.91567883e-02 1.38516247e-01
-6.38157204e-02 1.14064908e+00 8.29936802e-01 -9.69473779e-01
-1.91901517e+00 -8.67497206e-01 4.86294270e-01 -1.00121319e+00
4.24835920e-01 -6.51360869e-01 -7.80263960e-01 4.79089528e-01
-5.90199411e-01 -1.20171336e-02 1.28366399e+00 1.23193145e-01
-4.85087276e-01 2.28815109e-01 -1.16916060e+00 3.02554846e-01
9.90259171e-01 -5.72803557e-01 -3.27828646e-01 5.62030315e-01
2.88623661e-01 -5.18795192e-01 -7.90569484e-01 4.37818974e-01
5.74352026e-01 -7.19548166e-01 1.43806291e+00 -1.12493014e+00
-7.98241422e-02 -6.38515651e-01 -4.73919749e-01 -1.05497897e+00
-4.55285847e-01 -2.28448853e-01 1.09254204e-01 4.60907549e-01
6.81210458e-01 -3.03059131e-01 1.24302328e+00 2.64932573e-01
-4.00008522e-02 -8.53773773e-01 -9.70416069e-01 -5.10581732e-01
8.23904127e-02 -1.45600855e-01 4.94150460e-01 4.07461196e-01
1.19157307e-01 5.64100623e-01 -3.86779845e-01 2.42495045e-01
5.21884859e-01 -2.59142518e-01 7.44403422e-01 -1.35241365e+00
-4.01591420e-01 -1.04908459e-01 -5.88764668e-01 -1.18012488e+00
-1.44339085e-01 -9.44656134e-01 1.38171958e-02 -1.32144737e+00
7.93949187e-01 -1.06819078e-01 -3.11852872e-01 1.93367630e-01
-1.90226793e-01 3.37779135e-01 -5.45152843e-01 1.67222649e-01
-7.32354999e-01 5.50642014e-01 9.29867089e-01 -2.47478351e-01
-1.88536659e-01 -2.89346248e-01 -5.53758085e-01 3.39042276e-01
3.43153596e-01 -6.54215515e-01 -3.36780012e-01 1.86638787e-01
8.43185782e-01 1.84697974e-02 4.22290415e-01 -7.32138216e-01
9.59844068e-02 -3.56267631e-01 8.86451662e-01 -7.11310148e-01
6.01749897e-01 -5.44834077e-01 7.74147332e-01 5.04917622e-01
-2.97755152e-01 -2.42862463e-01 2.13061467e-01 1.09615397e+00
-4.14235238e-03 -3.10766902e-02 6.41640723e-01 -3.08560908e-01
-1.42549738e-01 3.86586726e-01 -5.41216135e-01 -2.36574173e-01
7.03300357e-01 -1.55422106e-01 -8.10118765e-02 -4.56423983e-02
-8.99953485e-01 -1.75898612e-01 7.67550707e-01 -2.76045382e-01
5.45099616e-01 -1.05132306e+00 -5.70969164e-01 1.61828101e-01
3.67007673e-01 -7.67430142e-02 6.90047294e-02 5.79738081e-01
-7.23363936e-01 9.40078557e-01 -1.78134791e-03 -7.48291850e-01
-1.06145692e+00 7.02526808e-01 4.34135526e-01 -3.43024641e-01
-3.36898528e-02 9.86530960e-01 3.29609782e-01 -4.89736736e-01
1.01843953e-01 4.68069464e-02 1.74822360e-01 -1.58522755e-01
7.05926657e-01 -4.79034446e-02 4.14755404e-01 -3.93446505e-01
-6.24768436e-01 4.14317876e-01 -2.88443416e-01 2.26924524e-01
1.56062198e+00 3.39822322e-01 -1.18419365e-03 1.23403646e-01
1.12562668e+00 1.47352591e-01 -1.26667893e+00 -1.62164360e-01
1.59533188e-01 -4.12122577e-01 -1.97298959e-01 -1.12168860e+00
-5.78649975e-02 5.19755363e-01 6.64924860e-01 -7.89027870e-01
6.16284847e-01 -4.65678088e-02 3.86705607e-01 8.13603461e-01
7.64742076e-01 -7.59698570e-01 2.04784617e-01 4.53337818e-01
5.67090809e-01 -1.09725380e+00 1.14733279e-01 -2.59752274e-01
-1.64894029e-01 1.02896774e+00 1.42638832e-01 3.03805977e-01
1.29043236e-01 3.73755783e-01 -3.11697870e-01 -5.13778865e-01
-7.59992719e-01 2.93281097e-02 3.37183960e-02 8.68146122e-01
7.02711940e-01 -6.56603053e-02 -2.85787404e-01 8.53254378e-01
2.05767557e-01 1.19536750e-01 1.73388839e-01 9.64625478e-01
-5.79669476e-01 -1.67142856e+00 -4.22385484e-01 3.99893433e-01
-6.36230052e-01 -3.91874522e-01 -7.47203231e-01 3.43860954e-01
4.90511842e-02 6.32139802e-01 -3.16194981e-01 -3.48072082e-01
3.43191117e-01 4.43314254e-01 7.90067792e-01 -8.00504863e-01
-5.47272637e-02 2.14587569e-01 -4.00014222e-03 -4.07141000e-01
-1.43748298e-01 -5.01184821e-01 -1.40083015e+00 -2.45065898e-01
-5.39286673e-01 5.48157036e-01 6.67252421e-01 7.19114125e-01
6.67954147e-01 3.58065099e-01 2.60208130e-01 -1.04687011e+00
-5.45481145e-01 -7.39984393e-01 -5.42653203e-01 1.86375111e-01
1.99779004e-01 -8.63436580e-01 -7.97924493e-03 -1.21070951e-01] | [4.855458736419678, 5.597195148468018] |
94dcb244-f054-4e5e-ac73-39062feba9d6 | rethinking-multi-modal-alignment-in-video | 2204.11544 | null | https://arxiv.org/abs/2204.11544v2 | https://arxiv.org/pdf/2204.11544v2.pdf | Rethinking Multi-Modal Alignment in Video Question Answering from Feature and Sample Perspectives | Reasoning about causal and temporal event relations in videos is a new destination of Video Question Answering (VideoQA).The major stumbling block to achieve this purpose is the semantic gap between language and video since they are at different levels of abstraction. Existing efforts mainly focus on designing sophisticated architectures while utilizing frame- or object-level visual representations. In this paper, we reconsider the multi-modal alignment problem in VideoQA from feature and sample perspectives to achieve better performance. From the view of feature,we break down the video into trajectories and first leverage trajectory feature in VideoQA to enhance the alignment between two modalities. Moreover, we adopt a heterogeneous graph architecture and design a hierarchical framework to align both trajectory-level and frame-level visual feature with language feature. In addition, we found that VideoQA models are largely dependent on language priors and always neglect visual-language interactions. Thus, two effective yet portable training augmentation strategies are designed to strengthen the cross-modal correspondence ability of our model from the view of sample. Extensive results show that our method outperforms all the state-of-the-art models on the challenging NExT-QA benchmark, which demonstrates the effectiveness of the proposed method. | ['Jun Xiao', 'Zhimeng Zhang', 'Yi Yang', 'Zhao Wang', 'Kaifeng Gao', 'Long Chen', 'Shaoning Xiao'] | 2022-04-25 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [-1.50792256e-01 -3.27415287e-01 -3.18028986e-01 -3.42145860e-01
-6.85298026e-01 -4.65117395e-01 8.02245080e-01 -1.44783869e-01
-1.54288188e-01 2.34846741e-01 6.29912376e-01 -1.75261468e-01
-7.34794140e-02 -6.34154856e-01 -7.29476571e-01 -4.45081085e-01
1.43764317e-01 -5.06583899e-02 4.72166300e-01 -2.46372283e-01
-8.01831186e-02 1.31925970e-01 -1.48178566e+00 6.40704572e-01
6.73316479e-01 9.53295708e-01 1.40367478e-01 5.57724059e-01
-2.15163052e-01 1.29613674e+00 -1.79591581e-01 -6.66973948e-01
1.51489764e-01 -5.81404209e-01 -8.97609949e-01 3.77844840e-01
6.86442256e-01 -5.56383967e-01 -1.04119658e+00 1.00234294e+00
1.85580492e-01 2.59219974e-01 3.63996983e-01 -1.64245760e+00
-7.52475441e-01 3.65814924e-01 -5.39150119e-01 3.65882784e-01
6.23523653e-01 3.41673523e-01 1.21783614e+00 -7.23146558e-01
7.12748885e-01 1.44712210e+00 3.26103508e-01 4.27836359e-01
-7.96812773e-01 -4.03360248e-01 7.18338192e-01 8.03004682e-01
-1.14156520e+00 -5.32632113e-01 1.00136960e+00 -6.24391139e-01
7.22571611e-01 1.61495224e-01 6.08457386e-01 1.29188776e+00
-1.40319588e-02 1.16505897e+00 7.59294987e-01 -2.77215336e-02
-2.24161502e-02 -2.94169992e-01 1.09926425e-01 1.00215828e+00
-2.14962050e-01 -1.94474995e-01 -8.45902026e-01 -5.85828163e-03
7.84817994e-01 3.14690024e-01 -5.09415030e-01 -6.34976447e-01
-1.59275579e+00 7.80802250e-01 3.63401681e-01 2.74039418e-01
-2.97487140e-01 3.46039891e-01 5.06297052e-01 2.39067271e-01
1.53884441e-01 7.53800869e-02 -1.93862766e-01 -1.89764291e-01
-6.60557449e-01 1.00633614e-01 4.00397986e-01 1.14430523e+00
4.65456843e-01 -4.16236371e-03 -5.21665692e-01 4.75875854e-01
5.04600346e-01 3.57002109e-01 1.98986247e-01 -1.04376018e+00
8.02334189e-01 7.39795327e-01 -5.72286397e-02 -1.41276801e+00
-2.72531599e-01 -1.76636621e-01 -6.17983937e-01 -4.39120948e-01
4.45826560e-01 2.22259551e-01 -8.62313151e-01 1.88352501e+00
4.16775465e-01 5.77433825e-01 -3.95658053e-03 1.21678686e+00
1.04829550e+00 8.45874727e-01 3.82406443e-01 -2.63610095e-01
1.55362177e+00 -1.31703162e+00 -1.01245797e+00 -5.12739830e-02
5.59210598e-01 -4.95298386e-01 1.36564302e+00 4.74472046e-02
-1.23345447e+00 -6.79660618e-01 -9.51134980e-01 -4.06158954e-01
-3.18029374e-01 1.86868012e-02 6.19642496e-01 2.57355183e-01
-8.40576887e-01 1.19750053e-01 -9.19203103e-01 -3.75881433e-01
2.83053666e-01 -3.46064158e-02 -4.16968316e-01 -4.49245960e-01
-1.41782713e+00 6.37040675e-01 1.03608482e-01 1.18063390e-01
-9.36403871e-01 -6.56635225e-01 -1.02083480e+00 9.69969928e-02
6.38802111e-01 -1.10826945e+00 1.07921720e+00 -7.99593091e-01
-1.33261931e+00 4.97740179e-01 -4.00124639e-01 -2.63475329e-01
3.94109458e-01 -1.93912685e-01 -5.36743462e-01 6.71731532e-01
2.83039268e-02 4.89424855e-01 9.35119748e-01 -1.26798260e+00
-7.15937197e-01 -3.15954924e-01 7.51462162e-01 3.49057853e-01
-5.17185211e-01 -4.11867499e-02 -1.08855259e+00 -6.74436569e-01
-5.76120391e-02 -6.69494748e-01 7.18399808e-02 -1.90594289e-02
1.14505403e-01 -3.21700901e-01 9.65986967e-01 -7.87767589e-01
1.37896407e+00 -2.28853488e+00 6.40368581e-01 -2.50568628e-01
4.81370628e-01 4.17162897e-03 -2.57594347e-01 5.59544504e-01
2.71873605e-02 -5.05180210e-02 -1.94283444e-02 -4.02034134e-01
-2.54795719e-02 1.84204876e-01 -3.97889197e-01 4.99445200e-01
2.93011874e-01 1.11475372e+00 -1.30936706e+00 -6.72091722e-01
1.34741873e-01 5.64054191e-01 -7.99436331e-01 3.04802358e-01
-4.07615185e-01 5.81082880e-01 -6.60694838e-01 7.53885448e-01
3.60600919e-01 -6.76697433e-01 1.73637077e-01 -8.32948387e-01
2.25660712e-01 2.23979548e-01 -9.79521930e-01 2.21553278e+00
-3.28135908e-01 4.52922046e-01 -6.95370436e-02 -9.39799011e-01
2.95930117e-01 5.05874276e-01 6.44786477e-01 -8.34567070e-01
6.59732744e-02 -3.02278221e-01 -5.60232513e-02 -9.82212424e-01
4.88931805e-01 1.10308332e-02 6.40377700e-02 1.41625494e-01
3.35733265e-01 2.78535634e-01 2.86789328e-01 5.87275267e-01
9.46968734e-01 4.45077062e-01 1.86199933e-01 1.12518772e-01
6.81050420e-01 -1.05291262e-01 5.78047752e-01 4.71177369e-01
-4.40693259e-01 6.63139105e-01 6.88538611e-01 -2.30844706e-01
-6.32073998e-01 -1.01399803e+00 3.08065593e-01 1.27296889e+00
4.33417857e-01 -8.07135999e-01 -4.91313398e-01 -1.02241063e+00
-3.30134571e-01 5.94705939e-01 -5.88488817e-01 -5.37614673e-02
-6.74492240e-01 -3.15948427e-01 3.22112650e-01 6.31764233e-01
5.31809092e-01 -6.23473227e-01 -3.64958823e-01 5.99943381e-03
-7.22828090e-01 -1.64931679e+00 -6.32669091e-01 -6.61511421e-01
-5.97160876e-01 -1.24592292e+00 -5.78024566e-01 -5.85781217e-01
5.03970981e-01 7.16203988e-01 1.22394753e+00 1.83132634e-01
2.22729549e-01 1.08036745e+00 -7.19451725e-01 4.08995561e-02
-9.73188132e-02 -1.36257336e-01 -1.54096663e-01 3.51031572e-01
1.73872754e-01 -4.54631418e-01 -7.33482599e-01 3.03315699e-01
-1.00869560e+00 2.90831268e-01 3.69737506e-01 7.73396015e-01
4.14146632e-01 -8.63972083e-02 3.57415289e-01 -4.99593377e-01
3.01004440e-01 -7.01509416e-01 -3.47574770e-01 5.75616598e-01
-1.45938685e-02 -7.99688324e-02 6.02522075e-01 -4.39166009e-01
-1.14795005e+00 -1.50765970e-01 9.65609476e-02 -8.58576715e-01
2.72812080e-02 6.82905257e-01 -4.66860116e-01 2.62446076e-01
1.77471340e-01 1.86012775e-01 -1.27254188e-01 -9.96585116e-02
6.88520253e-01 1.93118840e-01 4.11064893e-01 -6.73730850e-01
6.84385538e-01 7.91060269e-01 6.22354336e-02 -7.43283987e-01
-8.41739953e-01 -6.51373327e-01 -5.85985661e-01 -5.12180448e-01
1.13866460e+00 -1.19270551e+00 -8.97117972e-01 1.88039541e-01
-1.15148163e+00 -5.86660132e-02 -1.90858170e-02 4.86218154e-01
-5.00069141e-01 6.89489067e-01 -5.44852972e-01 -4.98509526e-01
8.44786391e-02 -1.35737729e+00 1.26318586e+00 1.15901597e-01
2.87057668e-01 -1.00611424e+00 -1.85815513e-01 7.79850304e-01
1.60187796e-01 8.45147595e-02 8.99598181e-01 -2.90503353e-01
-9.14624751e-01 8.30498710e-02 -4.89119858e-01 -6.16008006e-02
1.31210700e-01 -7.96153992e-02 -7.51567900e-01 -2.77997971e-01
1.18021347e-01 -3.10704052e-01 9.46781576e-01 8.36044550e-02
1.19055283e+00 -1.62751198e-01 -1.72688976e-01 5.61322570e-01
1.11940420e+00 1.54988067e-02 5.50754070e-01 1.55546755e-01
1.19060183e+00 6.65579915e-01 6.82773173e-01 2.11866587e-01
1.04474092e+00 7.71120071e-01 5.98567188e-01 6.92556351e-02
-3.83410484e-01 -5.25855243e-01 5.87939799e-01 1.12428033e+00
-1.42873079e-01 -2.90655941e-01 -8.70671153e-01 6.27379477e-01
-2.33611679e+00 -1.26317227e+00 -1.27127901e-01 1.84167945e+00
5.21684587e-01 -2.37502024e-01 1.90763041e-01 -1.61415637e-01
4.12975878e-01 5.02155006e-01 -2.73701966e-01 3.23414832e-01
5.14990613e-02 -5.17165184e-01 7.19568226e-03 2.63762802e-01
-1.08023953e+00 9.27631080e-01 5.48978615e+00 6.88018441e-01
-9.88114893e-01 2.27416188e-01 3.39207351e-01 -1.39257461e-01
-3.92028630e-01 5.68461493e-02 -5.53825498e-01 3.47238332e-01
7.59245038e-01 7.28993192e-02 4.63310212e-01 4.40909535e-01
1.87598735e-01 1.60387993e-01 -1.32006443e+00 1.04405618e+00
3.55741590e-01 -1.40082073e+00 4.25158501e-01 -1.94721878e-01
4.62314516e-01 -2.62397200e-01 -1.82897858e-02 4.97257262e-01
-5.94076402e-02 -7.64410079e-01 8.14533055e-01 6.73295259e-01
4.47022110e-01 -3.86710465e-01 3.98270935e-01 7.42005408e-02
-1.53826797e+00 -4.21804003e-02 -9.96744931e-02 -1.78225338e-02
5.36620915e-01 8.73895064e-02 -3.73774916e-01 1.02836871e+00
8.36971402e-01 1.06165481e+00 -6.27163231e-01 7.17929363e-01
-6.42265156e-02 4.82576430e-01 5.92441298e-02 3.16672206e-01
3.20126623e-01 -2.29289070e-01 4.29017514e-01 1.03581834e+00
1.81585014e-01 3.24521422e-01 3.24094385e-01 6.09650671e-01
-3.69121544e-02 -5.40876836e-02 -5.67403495e-01 -2.72020161e-01
1.88543037e-01 1.07540345e+00 -4.44218785e-01 -3.14147770e-01
-1.20663488e+00 8.13278258e-01 3.97658408e-01 6.40127003e-01
-1.29080069e+00 3.16411883e-01 7.08887815e-01 -5.55643998e-02
4.57236767e-01 -4.98729944e-01 2.01549485e-01 -1.76033330e+00
1.23453476e-01 -1.09572303e+00 7.09383905e-01 -8.96262348e-01
-1.34465981e+00 4.45791513e-01 1.43996298e-01 -1.35193074e+00
-1.23889692e-01 -6.31306529e-01 -2.45638311e-01 3.73584390e-01
-1.61069798e+00 -1.77802026e+00 -5.32257795e-01 1.14086449e+00
8.40070665e-01 6.99261650e-02 3.96715105e-01 4.84514892e-01
-5.34060061e-01 4.24504817e-01 -3.91535342e-01 2.48124093e-01
6.44694507e-01 -8.64760756e-01 -1.74224507e-02 1.19723248e+00
4.77822781e-01 6.35253370e-01 4.27038461e-01 -5.07161021e-01
-1.98903024e+00 -9.89846706e-01 5.99549890e-01 -6.30738437e-01
1.01973569e+00 -3.53375554e-01 -9.64019120e-01 7.94629872e-01
4.13411468e-01 1.29597336e-01 5.91354609e-01 1.70848534e-01
-6.16218984e-01 1.60168409e-02 -5.78377426e-01 9.68272150e-01
1.29973161e+00 -9.25820947e-01 -7.42389977e-01 2.25280538e-01
1.06849754e+00 -4.09249276e-01 -9.43024337e-01 5.78524411e-01
4.61742759e-01 -9.85917091e-01 1.06170857e+00 -9.47034895e-01
6.21605992e-01 -5.35351455e-01 -3.91602814e-01 -9.00813162e-01
-3.52960825e-01 -5.18990338e-01 -6.17113709e-01 1.41148496e+00
8.33259672e-02 -8.88901204e-02 4.60172117e-01 4.71327990e-01
-5.12575246e-02 -7.52199113e-01 -6.98365390e-01 -5.10040462e-01
-2.72556126e-01 -5.30397475e-01 4.48386729e-01 1.15592635e+00
8.22046697e-02 6.32327437e-01 -6.45547152e-01 4.92612839e-01
3.03037822e-01 2.83584595e-01 7.77853608e-01 -6.99648321e-01
-3.78366202e-01 -4.19482321e-01 -3.86792868e-01 -1.41323256e+00
2.52449989e-01 -6.04388177e-01 -2.19899103e-01 -1.60514796e+00
3.74683559e-01 1.09554254e-01 -5.14628291e-01 3.11958462e-01
-4.08148795e-01 -6.11290485e-02 5.38132787e-01 2.73655057e-01
-1.03788483e+00 8.55727255e-01 1.49483502e+00 -2.59220332e-01
4.18126024e-02 -3.96028996e-01 -5.30718803e-01 6.90391779e-01
3.76142800e-01 -5.51552624e-02 -8.80709469e-01 -9.23162878e-01
2.94869363e-01 4.60032910e-01 5.26356161e-01 -6.17885888e-01
3.96569401e-01 -4.22575891e-01 -1.00803152e-01 -4.88575578e-01
5.17233312e-01 -9.69073772e-01 -9.90682542e-02 -2.69277245e-02
-2.91823715e-01 3.08041632e-01 5.91267645e-02 9.96998429e-01
-5.63389838e-01 2.55599320e-01 3.41575712e-01 -5.97804971e-02
-1.08108127e+00 6.46901429e-01 -2.06156839e-02 1.12427093e-01
1.04497480e+00 3.99881713e-02 -6.03484988e-01 -6.03656948e-01
-5.04950821e-01 6.32330418e-01 3.93652946e-01 8.94886434e-01
6.85025573e-01 -1.44204092e+00 -4.28729355e-01 -7.52910450e-02
3.69288623e-01 -2.20648855e-01 6.30573392e-01 1.25887942e+00
-2.17783287e-01 4.79985356e-01 -6.02435991e-02 -7.24612951e-01
-1.11136281e+00 9.47793484e-01 2.12692201e-01 -2.79762447e-01
-5.82104683e-01 6.56313539e-01 6.67864799e-01 -1.38105135e-02
3.92181009e-01 -2.63517350e-01 -4.10613924e-01 1.37784392e-01
5.37952960e-01 2.17275426e-01 -1.80684537e-01 -9.61636841e-01
-4.18094337e-01 5.31827450e-01 1.00054462e-02 -3.87478317e-03
1.12977087e+00 -4.56152052e-01 3.00867409e-02 5.06500185e-01
1.15255344e+00 -3.02194878e-02 -1.35767066e+00 -3.23751986e-01
-2.92152256e-01 -6.21657729e-01 1.71807762e-02 -4.51819509e-01
-1.24228704e+00 9.35056269e-01 2.67374992e-01 2.35281393e-01
1.22982550e+00 1.28876865e-01 7.78273046e-01 1.98539764e-01
2.55271763e-01 -7.05607355e-01 4.01148528e-01 3.88308257e-01
9.62124407e-01 -1.35752189e+00 4.67774644e-03 -5.71082473e-01
-8.55447829e-01 9.41414177e-01 7.74374962e-01 9.94174778e-02
5.15603304e-01 -3.50460559e-01 -9.74162593e-02 -3.88781399e-01
-8.85387659e-01 -4.53208774e-01 6.43553913e-01 4.38905656e-01
4.25592333e-01 -2.56574661e-01 -2.51824915e-01 6.16246819e-01
3.42688590e-01 -6.29778132e-02 2.59239167e-01 8.35677683e-01
4.58455719e-02 -9.19408798e-01 -1.29022673e-01 3.30470577e-02
-3.55536520e-01 -3.74015421e-02 -6.28387183e-02 9.97636020e-01
-7.94292167e-02 1.06011343e+00 -6.68909177e-02 -4.05860245e-01
3.09614867e-01 7.80783035e-03 6.57304704e-01 -1.89379036e-01
-2.65274793e-01 9.78351012e-02 1.59030840e-01 -1.06625533e+00
-1.01063108e+00 -5.82447469e-01 -1.08508182e+00 -1.57224044e-01
-8.09607729e-02 -4.02948298e-02 2.75866598e-01 1.15698218e+00
6.00148439e-01 6.80985093e-01 4.45594430e-01 -6.21278405e-01
-2.23766342e-01 -5.53953052e-01 -1.96356341e-01 7.24556267e-01
3.83067459e-01 -8.03887248e-01 -1.95001915e-01 3.90237838e-01] | [10.223023414611816, 0.9652472734451294] |
ac6dece1-dd46-4607-8219-f7212f0a8731 | real-time-visual-tracking-by-deep-reinforced | 1702.06291 | null | http://arxiv.org/abs/1702.06291v2 | http://arxiv.org/pdf/1702.06291v2.pdf | Real-time visual tracking by deep reinforced decision making | One of the major challenges of model-free visual tracking problem has been
the difficulty originating from the unpredictable and drastic changes in the
appearance of objects we target to track. Existing methods tackle this problem
by updating the appearance model on-line in order to adapt to the changes in
the appearance. Despite the success of these methods however, inaccurate and
erroneous updates of the appearance model result in a tracker drift. In this
paper, we introduce a novel real-time visual tracking algorithm based on a
template selection strategy constructed by deep reinforcement learning methods.
The tracking algorithm utilizes this strategy to choose the appropriate
template for tracking a given frame. The template selection strategy is
self-learned by utilizing a simple policy gradient method on numerous training
episodes randomly generated from a tracking benchmark dataset. Our proposed
reinforcement learning framework is generally applicable to other confidence
map based tracking algorithms. The experiment shows that our tracking algorithm
runs in real-time speed of 43 fps and the proposed policy network effectively
decides the appropriate template for successful visual tracking. | ['Janghoon Choi', 'Kyoung Mu Lee', 'Junseok Kwon'] | 2017-02-21 | null | null | null | null | ['real-time-visual-tracking'] | ['computer-vision'] | [ 6.01823330e-02 -4.27839816e-01 -9.35395658e-02 -6.95485203e-03
-2.51615167e-01 -6.27923608e-01 4.39575106e-01 -1.66815564e-01
-6.11812532e-01 7.24204063e-01 -5.38654327e-01 9.01421010e-02
1.61778986e-01 -4.58930045e-01 -7.91860282e-01 -8.32583189e-01
1.66277677e-01 3.79059792e-01 7.80103505e-01 4.46335636e-02
2.99439818e-01 7.26710975e-01 -1.64313245e+00 -3.61096293e-01
5.25259376e-01 9.89089727e-01 4.22821999e-01 7.57464051e-01
-6.46761730e-02 6.87965572e-01 -6.68443143e-01 -6.90527409e-02
4.97490346e-01 -2.48879701e-01 -1.82895139e-01 2.64423758e-01
5.22804260e-01 -2.61674315e-01 -1.81024030e-01 1.11654973e+00
5.35554588e-01 1.94671020e-01 3.29467624e-01 -1.30784690e+00
-1.43537879e-01 1.02048844e-01 -6.67599142e-01 4.40889508e-01
2.26332322e-01 3.75794172e-01 3.15912098e-01 -5.80253303e-01
7.94568539e-01 1.07150209e+00 7.76512325e-01 7.59759903e-01
-9.27740753e-01 -8.14413726e-01 4.60010499e-01 2.12212756e-01
-1.24974847e+00 -3.98637235e-01 8.34943175e-01 -5.64685225e-01
4.34294283e-01 -3.15034501e-02 1.00110757e+00 9.59837794e-01
5.11172712e-01 5.36685646e-01 1.05757904e+00 -5.27717531e-01
3.86456579e-01 2.27532461e-02 -2.04449490e-01 7.91877151e-01
4.06640530e-01 6.77659988e-01 -3.46436083e-01 -1.00273237e-01
1.02605379e+00 -1.55760333e-01 -1.28526405e-01 -7.65607655e-01
-1.14168262e+00 4.48123604e-01 4.26217735e-01 2.05729976e-01
-3.96168381e-01 5.68418503e-01 2.58624673e-01 2.29735523e-01
-6.97702076e-03 6.06772453e-02 -3.98939401e-01 -2.02427045e-01
-9.96610403e-01 1.78502798e-01 3.84876877e-01 8.56042981e-01
4.59332079e-01 5.78008771e-01 -3.83076578e-01 4.62022096e-01
6.24784589e-01 7.81903267e-01 3.95836055e-01 -9.29654241e-01
-2.17335939e-01 3.01813632e-01 6.08040333e-01 -1.03423572e+00
-3.46286267e-01 -7.05616236e-01 -2.82120436e-01 8.61213505e-01
6.54143393e-01 -4.40181404e-01 -9.79337275e-01 1.85573006e+00
9.05638218e-01 4.09374207e-01 -1.80109337e-01 8.77916813e-01
3.58339399e-01 3.62874866e-01 2.62641162e-01 -4.25869584e-01
1.00917673e+00 -8.84800732e-01 -8.66036117e-01 1.29919616e-03
-3.57237011e-02 -9.48891819e-01 3.25667977e-01 2.56982893e-01
-8.39342713e-01 -9.17472959e-01 -1.07372463e+00 8.15783918e-01
-8.67849737e-02 3.57454985e-01 2.80776471e-01 6.49715841e-01
-1.02107942e+00 6.15840495e-01 -1.00788033e+00 -5.68107724e-01
2.14404508e-01 5.87188125e-01 2.59559639e-02 3.96139085e-01
-8.16887796e-01 9.04476941e-01 4.99996006e-01 2.95088023e-01
-1.01019788e+00 -3.14835548e-01 -4.91770864e-01 -2.24463701e-01
3.64252299e-01 -7.50943959e-01 1.44668782e+00 -1.39398801e+00
-1.91067970e+00 5.93931317e-01 -1.18699856e-01 -6.40386045e-01
9.60160732e-01 -1.74245581e-01 -4.00992781e-01 -1.33389845e-01
-1.24464273e-01 6.52356148e-01 1.42648363e+00 -1.37558830e+00
-9.33638096e-01 -8.06424096e-02 -3.38649392e-01 1.82322204e-01
3.59283872e-02 -3.26532824e-03 -5.51333964e-01 -5.91042042e-01
-2.00461864e-01 -1.09668052e+00 -2.44068071e-01 3.40786219e-01
1.61003426e-01 -1.14390157e-01 1.18035042e+00 -2.20137760e-01
1.00560236e+00 -1.88739324e+00 -1.62459671e-01 -2.57219505e-02
7.56639242e-02 4.79108989e-01 1.52872801e-01 1.06854849e-01
4.13017303e-01 -5.76980352e-01 3.54820251e-01 2.37742085e-02
-1.79469272e-01 5.45642748e-02 -1.32892326e-01 5.92051804e-01
1.24431131e-02 6.32010281e-01 -1.13165677e+00 -6.93295062e-01
4.84764010e-01 6.01200759e-01 -1.45661369e-01 4.55777586e-01
-3.80040348e-01 7.88140476e-01 -5.59655428e-01 5.67908645e-01
7.04572737e-01 -3.50493222e-01 1.59344330e-01 -1.99767038e-01
-3.92299801e-01 -3.60989213e-01 -1.26168728e+00 1.42148697e+00
-1.83304235e-01 6.42157137e-01 -1.21447980e-01 -4.98471707e-01
1.08808327e+00 2.35485166e-01 6.71969175e-01 -5.31670988e-01
4.86405641e-01 1.81264486e-02 2.85100788e-01 -1.74031511e-01
4.44211483e-01 2.94889081e-02 4.08634663e-01 3.14557344e-01
-1.36680067e-01 2.98719466e-01 1.91065580e-01 -1.91781014e-01
1.10815537e+00 7.43602037e-01 2.66422242e-01 1.07041270e-01
6.93294823e-01 1.60886481e-01 9.57400858e-01 9.51612949e-01
-7.42737055e-01 2.23894104e-01 -3.38269949e-01 -9.24664974e-01
-9.83934879e-01 -9.10995841e-01 -4.26704176e-02 8.80220473e-01
4.33936566e-01 3.48765701e-02 -5.76871037e-01 -6.02926970e-01
7.56959766e-02 2.65317887e-01 -6.13334477e-01 -1.33177355e-01
-7.16599643e-01 -4.13218826e-01 7.67209977e-02 4.80098009e-01
4.90121037e-01 -1.38364828e+00 -1.29397798e+00 5.84343016e-01
3.61600399e-01 -1.15141881e+00 -4.45890337e-01 6.90923557e-02
-7.98784137e-01 -9.71998751e-01 -6.82979882e-01 -6.35157704e-01
8.09358299e-01 2.38642231e-01 8.73650610e-01 2.68582225e-01
-2.26225659e-01 5.11151373e-01 -2.04998955e-01 -4.95932549e-01
-6.40194297e-01 -1.23842329e-01 1.95319206e-01 1.15889624e-01
1.52997509e-01 -5.39196357e-02 -7.92964160e-01 5.30706406e-01
-4.69607085e-01 -8.63423347e-02 4.73842144e-01 7.39581943e-01
7.69711077e-01 -1.20778177e-02 5.17356515e-01 -5.96753001e-01
3.29292417e-01 -1.61505312e-01 -1.54368865e+00 5.21618962e-01
-6.53790951e-01 9.52049792e-02 5.76107800e-01 -8.69920731e-01
-9.91881728e-01 6.69392288e-01 1.13065615e-01 -7.99805939e-01
-3.38461101e-02 -2.08397642e-01 3.27290803e-01 -6.82653844e-01
4.00081694e-01 2.77329296e-01 5.68964109e-02 -6.76017180e-02
1.03945717e-01 1.76511839e-01 5.42828560e-01 -4.44570273e-01
1.12721050e+00 4.28861082e-01 1.31098600e-02 -3.08999807e-01
-6.13476932e-01 -3.57662976e-01 -6.54440880e-01 -9.54222322e-01
6.71875179e-01 -7.93345571e-01 -1.06219172e+00 5.56412220e-01
-9.82009172e-01 -3.40942651e-01 -6.64759055e-02 4.62455332e-01
-5.09887218e-01 3.40269387e-01 -3.08943391e-01 -1.07584107e+00
-5.71496487e-01 -1.11282122e+00 9.87247765e-01 8.10187876e-01
7.00671002e-02 -9.95773077e-01 4.10537899e-01 -3.37222368e-01
6.95189536e-01 4.51878190e-01 9.93397757e-02 -1.90514281e-01
-8.39845061e-01 -3.33809197e-01 3.75944450e-02 -1.66114405e-01
3.36800426e-01 3.66643637e-01 -7.53647685e-01 -6.81335926e-01
-1.87220052e-01 -7.97034875e-02 4.45103258e-01 6.61318064e-01
6.54540300e-01 1.43822134e-01 -5.58789074e-01 5.10359704e-01
1.66532457e+00 6.26143575e-01 1.59561604e-01 7.10157871e-01
4.36313331e-01 -2.06896856e-01 9.92503583e-01 5.69804370e-01
-1.98343154e-02 1.02754176e+00 5.51999032e-01 -2.42340118e-02
-1.85100898e-01 -1.35097221e-01 4.15083677e-01 4.11599845e-01
8.39796290e-02 2.58376412e-02 -6.55973375e-01 3.62216324e-01
-2.01528025e+00 -1.16787779e+00 2.55042076e-01 2.52284193e+00
6.91708148e-01 3.87310624e-01 3.83276641e-01 -3.33977550e-01
1.05166233e+00 -1.23757057e-01 -9.08827305e-01 -1.01536043e-01
2.14002028e-01 -1.89372823e-01 7.25049198e-01 3.59080255e-01
-1.19435596e+00 1.13564801e+00 6.45137978e+00 3.61081153e-01
-1.53625000e+00 -1.66484386e-01 1.81782261e-01 2.16613896e-02
3.54907691e-01 -2.58283224e-02 -1.18096542e+00 6.40419841e-01
7.30400801e-01 -2.44265586e-01 3.68744701e-01 8.79789352e-01
3.06126088e-01 -8.62143263e-02 -7.02528775e-01 1.00141621e+00
-1.09264784e-01 -1.24563992e+00 -2.65530020e-01 -3.28044653e-01
6.45026565e-01 -4.78864312e-02 4.49061133e-02 2.54552484e-01
5.20905077e-01 -4.36721236e-01 9.10177708e-01 6.83732212e-01
4.73010957e-01 -5.16940415e-01 5.01723647e-01 3.12511593e-01
-1.46871901e+00 -7.22375438e-02 -5.39757907e-01 1.72788963e-01
2.02772707e-01 1.28823221e-01 -9.43302333e-01 2.60218829e-01
5.56288242e-01 4.74266738e-01 -5.60565531e-01 1.70874572e+00
1.12180654e-02 4.02956158e-01 -3.67055506e-01 -1.62861139e-01
1.95923313e-01 9.17502958e-03 7.26553738e-01 1.03280318e+00
2.91556150e-01 -3.19337010e-01 4.06618655e-01 6.20467007e-01
2.34720394e-01 -8.86710063e-02 -5.05468071e-01 2.27658615e-01
6.57527268e-01 1.41174352e+00 -1.07619524e+00 -2.68887907e-01
-1.94275051e-01 6.52504265e-01 3.76838923e-01 2.16624126e-01
-1.22793913e+00 1.19452201e-01 1.98738858e-01 8.97300337e-03
8.51776898e-01 -1.65108860e-01 2.78015375e-01 -9.23669815e-01
-1.22771576e-01 -7.46076286e-01 2.36424163e-01 -6.52412772e-01
-1.08066261e+00 9.45888937e-01 -2.19394937e-01 -1.65600097e+00
-3.62360120e-01 -3.73585701e-01 -5.02429724e-01 5.21135926e-01
-1.43555129e+00 -9.24339354e-01 -5.77961385e-01 5.56161106e-01
6.11613572e-01 -3.47056627e-01 4.39447522e-01 1.01912670e-01
-5.53040624e-01 5.53785026e-01 1.63140565e-01 1.16016619e-01
8.26077580e-01 -1.20008385e+00 2.88533896e-01 9.26458180e-01
1.23623230e-01 4.27578211e-01 1.00076020e+00 -8.51102591e-01
-1.40681040e+00 -1.09874129e+00 2.16577694e-01 -2.52581120e-01
5.44519544e-01 -5.89321554e-02 -8.25342834e-01 5.44209063e-01
3.58652920e-01 5.63329101e-01 1.65770486e-01 -4.52464432e-01
5.77068031e-02 -3.44065666e-01 -1.12055886e+00 5.03021955e-01
7.31724322e-01 2.15075284e-01 -1.91372097e-01 2.09268615e-01
2.88667232e-01 -8.48095000e-01 -6.53495431e-01 4.30763960e-01
8.57425869e-01 -7.11067498e-01 7.14667797e-01 -4.55179006e-01
-4.98416066e-01 -8.60392153e-01 3.01979572e-01 -1.14324462e+00
-5.36371887e-01 -8.55375648e-01 -3.39181304e-01 9.10423219e-01
-8.29411522e-02 -3.76503080e-01 1.01008463e+00 4.45788831e-01
3.28255266e-01 -5.95102549e-01 -1.03558064e+00 -9.18271780e-01
-3.40716481e-01 4.26361002e-02 3.79950702e-01 5.48577726e-01
-6.44900084e-01 3.23801078e-02 -6.06284797e-01 2.92011827e-01
1.00540197e+00 2.93493271e-01 9.80179369e-01 -1.25201035e+00
-3.79271239e-01 -2.81774849e-01 -6.08967125e-01 -9.30855393e-01
5.03578596e-03 -3.31620246e-01 4.48160708e-01 -1.11734974e+00
7.06926435e-02 -5.93492150e-01 -5.21683753e-01 2.75333643e-01
-3.79860878e-01 -2.20687911e-02 5.06397545e-01 2.29355052e-01
-9.90097880e-01 4.99531090e-01 1.24517131e+00 -7.76728541e-02
-2.33261228e-01 4.24299806e-01 -7.39541724e-02 5.48268914e-01
7.56757975e-01 -7.96801686e-01 -2.34929532e-01 -3.17665488e-01
-1.25645235e-01 2.17177197e-01 2.57094026e-01 -1.37107384e+00
5.24431229e-01 -1.27362370e-01 8.73575628e-01 -7.56603122e-01
1.48252845e-01 -1.22575259e+00 4.00807679e-01 9.08958852e-01
-6.92219064e-02 5.41917205e-01 5.10717690e-01 8.50854754e-01
7.16457590e-02 -2.37509832e-01 1.05591452e+00 -5.86188119e-03
-9.24142957e-01 3.93978357e-01 -4.21931356e-01 -1.56021833e-01
1.33246779e+00 -5.11360884e-01 -1.74695000e-01 -2.24948972e-01
-7.86332846e-01 1.49107873e-01 6.82405889e-01 5.48817635e-01
4.63002563e-01 -1.50708926e+00 -3.56949925e-01 4.42555211e-02
-7.09275305e-02 -4.64700967e-01 -9.69166234e-02 7.38020062e-01
-5.21567285e-01 1.86472759e-01 -6.57522142e-01 -9.78626013e-01
-1.39001155e+00 7.89670885e-01 7.24334717e-01 -3.05479795e-01
-6.45122111e-01 5.39154232e-01 -1.60173118e-01 8.64216536e-02
5.15608847e-01 -1.12211406e-02 -1.86635152e-01 -5.37527382e-01
5.00393212e-01 6.81945309e-02 -1.29668802e-01 -7.04073191e-01
-4.17895854e-01 9.02665198e-01 -2.30763271e-01 -1.46229658e-03
9.63337064e-01 -1.43962845e-01 4.23463076e-01 4.17204052e-01
5.71327031e-01 -5.56270815e-02 -1.87740278e+00 -2.48829380e-01
6.02779873e-02 -6.98405027e-01 -4.80695181e-02 -8.08248460e-01
-1.31142187e+00 2.78577387e-01 1.36903632e+00 5.70037514e-02
9.81505454e-01 -6.34656847e-01 5.62091172e-01 2.84983695e-01
6.31503522e-01 -1.09578085e+00 7.72576034e-02 3.36351007e-01
5.31343341e-01 -1.38805413e+00 3.14941742e-02 3.66783030e-02
-5.15687048e-01 1.36109114e+00 9.79404151e-01 -2.05003053e-01
5.44922709e-01 4.32571650e-01 4.41162556e-01 2.15896387e-02
-8.09307754e-01 -1.84902981e-01 1.06575415e-01 7.32400894e-01
2.70159632e-01 -2.96076506e-01 -2.59098262e-01 -3.16084653e-01
3.59782994e-01 2.79836059e-01 3.39892745e-01 1.06732583e+00
-5.67050159e-01 -1.22631586e+00 -6.70677900e-01 6.07071035e-02
-3.55680674e-01 5.06790757e-01 -1.59813404e-01 8.09194446e-01
-3.64133082e-02 6.84457898e-01 -7.59891048e-02 -1.53913841e-01
5.42376228e-02 -1.85593456e-01 8.46409678e-01 -1.95469990e-01
-6.94334269e-01 3.85917306e-01 -3.74839514e-01 -4.20007974e-01
-6.48735702e-01 -9.16893601e-01 -1.31955218e+00 -4.17919364e-03
-4.88187790e-01 6.39212318e-03 5.92323303e-01 7.79592097e-01
3.32504034e-01 6.24591708e-01 6.72857642e-01 -8.43465686e-01
-6.54641867e-01 -6.72510207e-01 -1.24573097e-01 3.74357820e-01
3.38924855e-01 -1.19470358e+00 -9.44759417e-03 8.15439299e-02] | [6.396175384521484, -2.0624725818634033] |
927888ec-ab4b-46e1-a8db-3b559dccf544 | transcg-a-large-scale-real-world-dataset-for | 2202.08471 | null | https://arxiv.org/abs/2202.08471v2 | https://arxiv.org/pdf/2202.08471v2.pdf | TransCG: A Large-Scale Real-World Dataset for Transparent Object Depth Completion and a Grasping Baseline | Transparent objects are common in our daily life and frequently handled in the automated production line. Robust vision-based robotic grasping and manipulation for these objects would be beneficial for automation. However, the majority of current grasping algorithms would fail in this case since they heavily rely on the depth image, while ordinary depth sensors usually fail to produce accurate depth information for transparent objects owing to the reflection and refraction of light. In this work, we address this issue by contributing a large-scale real-world dataset for transparent object depth completion, which contains 57,715 RGB-D images from 130 different scenes. Our dataset is the first large-scale, real-world dataset that provides ground truth depth, surface normals, transparent masks in diverse and cluttered scenes. Cross-domain experiments show that our dataset is more general and can enable better generalization ability for models. Moreover, we propose an end-to-end depth completion network, which takes the RGB image and the inaccurate depth map as inputs and outputs a refined depth map. Experiments demonstrate superior efficacy, efficiency and robustness of our method over previous works, and it is able to process images of high resolutions under limited hardware resources. Real robot experiments show that our method can also be applied to novel transparent object grasping robustly. The full dataset and our method are publicly available at www.graspnet.net/transcg | ['Cewu Lu', 'Sheng Xu', 'Hao-Shu Fang', 'Hongjie Fang'] | 2022-02-17 | null | null | null | null | ['transparent-objects', 'depth-completion', 'transparent-object-depth-estimation', 'robotic-grasping'] | ['computer-vision', 'computer-vision', 'computer-vision', 'robots'] | [ 3.20025295e-01 -1.94403023e-01 2.21367687e-01 -4.44121748e-01
-2.62102932e-01 -5.12848377e-01 1.25519782e-01 -2.74787247e-01
-1.50826871e-01 3.52469265e-01 -1.33585751e-01 1.10451810e-01
-8.80730748e-02 -9.33748782e-01 -7.84344375e-01 -6.94847345e-01
2.28116978e-02 5.47675669e-01 6.45598471e-01 -2.40447938e-01
2.37762511e-01 4.70860571e-01 -1.63493347e+00 5.24590909e-01
9.08034682e-01 1.36399829e+00 9.71001387e-01 3.76855284e-01
-8.20080787e-02 2.43058681e-01 -2.95021415e-01 -3.04884583e-01
7.43100464e-01 4.34010118e-01 -6.18865073e-01 1.13053694e-01
4.36515212e-01 -1.04635167e+00 -4.98657942e-01 9.40991879e-01
5.79241872e-01 -1.75433546e-01 5.00432968e-01 -1.06463623e+00
-8.31287622e-01 4.05403584e-01 -5.83805859e-01 -3.97936910e-01
5.25578260e-01 4.21232283e-01 5.57652712e-01 -9.00248408e-01
6.92373395e-01 1.55526495e+00 3.58053237e-01 7.89950013e-01
-7.53538370e-01 -4.69506353e-01 2.78113663e-01 8.24849308e-02
-7.19868004e-01 -1.95623585e-03 8.98166180e-01 -2.52169222e-01
7.21762657e-01 -8.62270370e-02 7.41919696e-01 1.37356830e+00
2.37478733e-01 9.58753228e-01 1.03198290e+00 -1.92147732e-01
2.42888987e-01 -2.93554217e-01 -1.29625961e-01 7.83222139e-01
4.77427512e-01 5.70086241e-02 -4.60397154e-01 5.24950624e-02
1.20898068e+00 4.20332789e-01 -4.37248528e-01 -9.73462462e-01
-1.59751904e+00 4.08108205e-01 8.51123154e-01 -9.08768326e-02
-5.94710827e-01 1.40974805e-01 1.88318104e-01 1.74775869e-01
2.86410213e-01 1.42502502e-01 -5.04901230e-01 -6.25550672e-02
4.37889919e-02 2.39498317e-01 8.80466759e-01 1.44844902e+00
6.71113074e-01 -2.74080843e-01 1.30518720e-01 8.28487635e-01
3.02229702e-01 7.95185149e-01 1.49069428e-01 -1.29926574e+00
7.21622109e-01 6.77322805e-01 4.28911924e-01 -7.56945908e-01
-5.65199435e-01 3.21466088e-01 -6.42466366e-01 6.04336023e-01
6.06860697e-01 2.38298103e-01 -1.13149822e+00 1.24278784e+00
4.35480416e-01 -5.25966763e-01 8.39423612e-02 1.41314590e+00
1.03571010e+00 2.41153389e-01 -3.03065419e-01 2.35554069e-01
1.24287212e+00 -9.37186003e-01 -5.08451343e-01 -3.02285880e-01
3.54468152e-02 -7.47948408e-01 1.36954486e+00 8.05906653e-01
-1.12672484e+00 -3.25944632e-01 -8.99698138e-01 -3.50307941e-01
-2.36562088e-01 2.66846150e-01 1.22160971e+00 2.18475387e-01
-7.11629570e-01 5.12309432e-01 -1.03916693e+00 -4.74554718e-01
6.19224966e-01 3.84583592e-01 -5.08872330e-01 -6.77063763e-01
-6.10176265e-01 8.07372808e-01 3.22844237e-01 5.50102532e-01
-9.62454081e-01 -5.46717405e-01 -6.06559813e-01 -4.39454466e-01
4.98053461e-01 -6.75971985e-01 1.23772156e+00 -3.33533466e-01
-1.77227509e+00 8.11907053e-01 1.35799155e-01 7.55046010e-02
7.78783143e-01 -5.46326816e-01 3.82847399e-01 5.32348394e-01
-7.83978999e-02 7.26914823e-01 7.69593775e-01 -1.73947406e+00
-3.59111398e-01 -7.44195044e-01 4.38218117e-01 2.90731024e-02
-3.36740643e-01 -4.08650726e-01 -4.93401259e-01 -3.78133416e-01
9.07121360e-01 -8.40175867e-01 -1.02949716e-01 7.89393604e-01
-3.26169580e-01 -9.83305648e-02 9.74996805e-01 -4.18831080e-01
8.20341893e-03 -1.86135876e+00 3.36202830e-01 -1.05704755e-01
2.05627128e-01 1.00785814e-01 -1.90002829e-01 4.92160439e-01
4.96952385e-01 -3.68010163e-01 -3.15749973e-01 -1.86500624e-01
9.97736305e-02 2.89590865e-01 -4.08792287e-01 5.16191840e-01
-6.81596622e-02 9.04961109e-01 -9.45911109e-01 -3.31923753e-01
4.12345290e-01 5.02610981e-01 -4.56566513e-01 3.29929680e-01
-5.05156040e-01 7.01757669e-01 -7.49903977e-01 1.20126426e+00
1.02494860e+00 1.47100940e-01 -1.22308612e-01 -4.37916934e-01
-2.56070700e-02 -1.91373825e-01 -9.12565708e-01 2.33500457e+00
-5.08603692e-01 2.26399064e-01 4.32383031e-01 -6.18883073e-01
1.22637808e+00 2.16356497e-02 5.72801411e-01 -5.83711088e-01
2.12634742e-01 5.00049651e-01 -2.47674912e-01 -9.75704849e-01
2.87796259e-01 2.03887850e-01 1.66798890e-01 3.04981500e-01
-3.13373893e-01 -1.03079784e+00 -3.54606546e-02 -1.42836183e-01
1.02688408e+00 6.81665003e-01 -3.73306155e-01 1.03656240e-01
-2.12850235e-03 -1.88628510e-02 3.78550023e-01 5.96093059e-01
-1.47894070e-01 7.47307360e-01 1.22629449e-01 -6.72390223e-01
-1.03856373e+00 -1.45727539e+00 -2.00499579e-01 7.12261915e-01
8.15325141e-01 2.32815742e-01 -4.60065573e-01 -3.38469625e-01
3.90041411e-01 -2.22144946e-02 -3.96224380e-01 -8.63925293e-02
-5.89004099e-01 -4.11315143e-01 -5.60812652e-02 6.80284798e-01
9.37680304e-01 -1.33155119e+00 -1.08158636e+00 1.42361909e-01
-2.79936582e-01 -1.54747546e+00 1.37398377e-01 2.22295225e-02
-1.34770679e+00 -1.18819201e+00 -9.50946510e-01 -7.98872292e-01
7.06528306e-01 7.70933270e-01 7.87769616e-01 -6.05188571e-02
-6.23897791e-01 5.37683368e-01 -6.18965983e-01 -5.56339383e-01
-6.87204301e-02 -1.65901005e-01 2.01378748e-01 -3.37600768e-01
2.24106476e-01 -6.07710958e-01 -9.92008448e-01 4.46864247e-01
-9.15423036e-01 8.63189697e-02 7.56529748e-01 5.36335468e-01
4.23954368e-01 -3.36100936e-01 2.35258505e-01 -3.08220506e-01
3.19403678e-01 -1.56715348e-01 -7.47462094e-01 1.39928773e-01
-2.85538100e-02 -2.99161583e-01 2.64635921e-01 -5.07585347e-01
-1.08511341e+00 2.11733326e-01 1.46221504e-01 -5.77534258e-01
-1.01543091e-01 9.54055320e-03 -2.25707144e-01 -2.00464919e-01
5.80127299e-01 -1.38001800e-01 5.93523420e-02 -6.39682710e-01
2.44974494e-01 7.29958534e-01 6.80642843e-01 -8.43047857e-01
7.50698447e-01 1.05358911e+00 1.90645251e-02 -7.87391722e-01
-5.29494226e-01 -3.49303842e-01 -8.94244730e-01 -2.51414150e-01
6.61867201e-01 -8.99777114e-01 -1.12854302e+00 8.27120721e-01
-1.37503076e+00 -5.65887988e-01 2.36245990e-02 6.97986186e-01
-8.40386450e-01 5.50390780e-01 -7.67004132e-01 -8.56928706e-01
-3.84761810e-01 -1.19845521e+00 1.55235100e+00 8.81335437e-02
3.05713207e-01 -4.35918272e-01 -6.04599655e-01 4.40358341e-01
2.55782694e-01 4.34141308e-01 7.80816913e-01 3.10571432e-01
-1.13947058e+00 -1.92549184e-01 -5.25562882e-01 2.49843836e-01
4.53427255e-01 -9.68784168e-02 -1.06108141e+00 -2.71002859e-01
-1.09037273e-02 -6.89482987e-01 1.04974389e+00 3.05629253e-01
1.33118665e+00 4.54008654e-02 -2.90896833e-01 5.86451948e-01
1.51245761e+00 -5.03502451e-02 6.57909930e-01 3.91260713e-01
8.54779243e-01 9.77920234e-01 9.88299370e-01 6.61555231e-01
3.12629491e-01 6.36819541e-01 1.18975317e+00 -4.01732931e-03
-2.99393330e-02 3.45800295e-02 1.89977378e-01 6.12481058e-01
-5.54036558e-01 -1.02809906e-01 -9.43508625e-01 4.00223017e-01
-1.71791589e+00 -6.66899979e-01 -2.37143442e-01 2.13670325e+00
7.22404718e-01 1.42546818e-01 -1.19733356e-01 1.15028664e-01
3.15093279e-01 -3.27749819e-01 -8.08473170e-01 -1.02364659e-01
-5.40094040e-02 1.09987438e-01 5.04027843e-01 2.25660264e-01
-8.15893710e-01 1.04753745e+00 5.62901926e+00 1.39441550e-01
-1.06013525e+00 -4.55978885e-02 -1.91625208e-01 -1.45161003e-01
-2.54528016e-01 -3.93319368e-01 -3.14926893e-01 1.55433863e-01
-1.07777610e-01 4.89258051e-01 5.43793797e-01 9.97645915e-01
9.83140618e-02 -5.64760983e-01 -1.26704204e+00 1.13659036e+00
-1.17694266e-01 -7.68710732e-01 5.46743302e-03 -1.13426976e-01
6.33520901e-01 2.29158714e-01 4.01341207e-02 -1.12658627e-01
3.28450888e-01 -7.26817071e-01 9.43836272e-01 5.23038566e-01
6.50832176e-01 -1.97456017e-01 6.51590884e-01 3.93716782e-01
-9.65875745e-01 -4.17435110e-01 -8.89515877e-01 -8.34355801e-02
1.69210471e-02 7.48754680e-01 -6.34491384e-01 5.35967588e-01
1.24723589e+00 6.50049627e-01 -1.89386860e-01 1.10742819e+00
-2.63510913e-01 -2.05478817e-01 -3.64775330e-01 -2.02093855e-01
-2.15417735e-04 -2.32738897e-01 3.27779651e-01 6.89677477e-01
2.22145259e-01 1.45869002e-01 2.22622514e-01 1.02230048e+00
1.77456588e-01 -3.70822519e-01 -7.53690481e-01 2.33453616e-01
2.76202589e-01 1.16744828e+00 -8.25098932e-01 -3.48368213e-02
-2.56927043e-01 1.22575581e+00 3.50912064e-01 5.61241150e-01
-6.59852445e-01 -3.56727451e-01 7.19019234e-01 -2.61514246e-01
3.46841604e-01 -6.40248477e-01 -5.29350281e-01 -1.18079221e+00
7.35610068e-01 -4.95464504e-01 -2.35141188e-01 -1.12561190e+00
-1.45577681e+00 5.01925945e-01 -8.88782553e-04 -1.43244529e+00
2.95028299e-01 -1.29638767e+00 -5.96222356e-02 6.84175193e-01
-1.70863497e+00 -1.29151714e+00 -1.14572847e+00 6.56885266e-01
7.30142951e-01 1.69301301e-01 7.58818030e-01 -2.08626892e-02
-2.13786066e-02 -5.22094630e-02 6.38019666e-03 -1.55368164e-01
7.22846508e-01 -9.93832886e-01 1.57856062e-01 3.12896580e-01
-4.73137379e-01 5.62462568e-01 5.85994482e-01 -6.00128412e-01
-2.20422482e+00 -7.50567794e-01 -6.59066811e-02 -6.44486487e-01
3.64706427e-01 -6.81586564e-01 -7.19231725e-01 5.00557661e-01
-2.71919638e-01 1.78598270e-01 -1.94819510e-01 -3.74129176e-01
-3.58730286e-01 -1.67050377e-01 -1.46313977e+00 5.85981190e-01
1.63582993e+00 -2.85738915e-01 -6.07827604e-01 5.68422079e-01
7.74700820e-01 -9.14942801e-01 -9.18260157e-01 7.58062363e-01
9.84279990e-01 -1.05140233e+00 1.19842064e+00 2.46599242e-02
8.82522404e-01 -8.26576799e-02 -3.72758776e-01 -1.04272151e+00
5.04125878e-02 -2.67763317e-01 -9.69444364e-02 8.29786062e-01
-4.59048748e-02 -8.11415911e-01 8.36773753e-01 6.19804323e-01
-4.01385009e-01 -8.65671933e-01 -8.84073794e-01 -8.62970233e-01
-3.14020738e-02 -4.92688537e-01 6.58454239e-01 4.44227010e-01
-8.63940045e-02 -3.64130616e-01 -2.85039786e-02 3.86029512e-01
9.57472801e-01 6.53604448e-01 9.12996292e-01 -1.54317915e+00
5.27544841e-02 -2.26332694e-01 -4.35572594e-01 -1.30477977e+00
-4.60156463e-02 -6.09539747e-01 5.05249500e-01 -2.14378762e+00
1.28209487e-01 -8.32731128e-01 3.84389132e-01 5.73859155e-01
1.39778003e-01 3.28160554e-01 1.50794432e-01 3.94445509e-01
-2.63573170e-01 8.24539900e-01 1.97975516e+00 -3.47298533e-01
-1.95055142e-01 -1.36595201e-02 -1.08432986e-01 7.80827761e-01
7.02845693e-01 8.54130648e-03 -3.08511764e-01 -9.84603822e-01
-9.12816077e-02 -1.04758412e-01 6.42367303e-01 -9.84410226e-01
-7.55269080e-02 -3.05728942e-01 4.50244188e-01 -5.41931510e-01
7.69924283e-01 -1.26339078e+00 -3.12134296e-01 5.39661467e-01
1.25037685e-01 -3.49370658e-01 1.16713397e-01 6.51489735e-01
2.06537738e-01 -1.28845558e-01 4.36159253e-01 -4.77833718e-01
-8.26389968e-01 6.55107796e-01 1.12140276e-01 -3.66238028e-01
1.07040584e+00 -5.30610442e-01 -5.16482949e-01 -7.57517293e-02
-3.87697995e-01 2.05863759e-01 7.53099382e-01 7.39628077e-01
1.06757605e+00 -1.10329604e+00 -5.39985299e-01 2.03723788e-01
3.16394657e-01 6.82467937e-01 1.29677191e-01 6.32799566e-01
-9.37588453e-01 1.95255429e-01 -5.54788232e-01 -1.00239539e+00
-8.77618670e-01 5.28335750e-01 8.85698125e-02 5.08028269e-01
-1.11112309e+00 8.26060653e-01 2.49986336e-01 -5.99403918e-01
5.06300390e-01 -8.83964539e-01 3.07046890e-01 -6.39367938e-01
3.54637355e-01 3.16699535e-01 6.39131740e-02 -1.41271770e-01
-2.00518429e-01 9.70031500e-01 1.70259804e-01 1.39997257e-02
1.56558084e+00 -1.48126975e-01 -2.96267539e-01 4.07528490e-01
9.20095026e-01 -2.59132802e-01 -1.79443479e+00 -1.38538808e-01
-3.99803638e-01 -8.43904614e-01 -3.18426132e-01 -6.25258923e-01
-1.21775448e+00 1.05793798e+00 5.53100467e-01 -1.02627292e-01
1.02129316e+00 2.18909070e-01 9.60600019e-01 8.50008726e-01
1.33625269e+00 -1.01351655e+00 4.17142093e-01 4.43927556e-01
1.39521861e+00 -1.55555248e+00 -1.78922969e-03 -1.07646227e+00
-4.23440784e-01 1.37835371e+00 9.46336687e-01 -1.03013806e-01
2.76568949e-01 2.46235803e-01 1.81153819e-01 -3.00443798e-01
-2.95359731e-01 -6.65090904e-02 -2.04668403e-01 1.07304311e+00
-6.85504600e-02 -4.87185344e-02 5.73043376e-02 2.70176023e-01
-1.78900585e-01 1.33209541e-01 4.13473845e-01 1.40428138e+00
-5.90632260e-01 -7.55298078e-01 -4.51224923e-01 1.38924256e-01
3.26350075e-03 3.40748012e-01 -3.81852537e-01 5.67794085e-01
-1.10653071e-02 1.01988542e+00 -1.48489088e-01 -2.89317101e-01
5.87484181e-01 -3.38687897e-01 1.15095532e+00 -5.39511204e-01
-2.14701146e-01 -3.27633023e-01 -2.95872152e-01 -9.75729048e-01
-5.28289318e-01 -3.11363876e-01 -1.51352727e+00 -2.10395575e-01
-2.51876831e-01 -5.57703912e-01 1.16175520e+00 6.41657770e-01
8.64072070e-02 3.25302541e-01 5.53477407e-01 -1.60063350e+00
-6.63246453e-01 -1.08859849e+00 -5.63988745e-01 4.32688266e-01
3.21651846e-01 -1.04931843e+00 -2.16655850e-01 -8.04365277e-02] | [5.997312545776367, -1.0513148307800293] |
75bd931c-a8d9-414c-a505-d874e3453f46 | vision-language-adaptive-mutual-decoder-for | 2209.00859 | null | https://arxiv.org/abs/2209.00859v1 | https://arxiv.org/pdf/2209.00859v1.pdf | Vision-Language Adaptive Mutual Decoder for OOV-STR | Recent works have shown huge success of deep learning models for common in vocabulary (IV) scene text recognition. However, in real-world scenarios, out-of-vocabulary (OOV) words are of great importance and SOTA recognition models usually perform poorly on OOV settings. Inspired by the intuition that the learned language prior have limited OOV preformence, we design a framework named Vision Language Adaptive Mutual Decoder (VLAMD) to tackle OOV problems partly. VLAMD consists of three main conponents. Firstly, we build an attention based LSTM decoder with two adaptively merged visual-only modules, yields a vision-language balanced main branch. Secondly, we add an auxiliary query based autoregressive transformer decoding head for common visual and language prior representation learning. Finally, we couple these two designs with bidirectional training for more diverse language modeling, and do mutual sequential decoding to get robuster results. Our approach achieved 70.31\% and 59.61\% word accuracy on IV+OOV and OOV settings respectively on Cropped Word Recognition Task of OOV-ST Challenge at ECCV 2022 TiE Workshop, where we got 1st place on both settings. | ['Bing Yin', 'Jiajia Wu', 'Fengli yu', 'Xuyang Zhu', 'Qiandong Yan', 'Chenyu Liu', 'Jinshui Hu'] | 2022-09-02 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 1.30054042e-01 -6.99974373e-02 -3.24126601e-01 -2.47507811e-01
-6.79659247e-01 -3.77200931e-01 9.64236856e-01 -1.61709622e-01
-6.22011960e-01 9.93481055e-02 4.14881319e-01 -5.93887687e-01
6.01988018e-01 -4.46856320e-01 -7.96793044e-01 -3.71885687e-01
7.01983511e-01 5.11554122e-01 2.40848124e-01 -1.36895880e-01
1.52043402e-01 7.83332437e-02 -1.17934775e+00 5.19935012e-01
5.49708724e-01 8.27268839e-01 7.09353149e-01 1.09546649e+00
-7.49440014e-01 1.07837546e+00 -3.66558403e-01 -4.24419254e-01
7.29894936e-02 -7.77394548e-02 -6.98723972e-01 3.15341532e-01
7.59230077e-01 -3.96173537e-01 -7.58103907e-01 7.96991110e-01
5.71268857e-01 -4.51537780e-02 1.09173191e+00 -1.08963418e+00
-1.18595123e+00 7.81581581e-01 -8.07781875e-01 2.70468980e-01
6.76181316e-02 3.92887920e-01 1.18281555e+00 -1.32661557e+00
4.85585779e-01 1.52417397e+00 3.13428909e-01 6.51855469e-01
-1.03783405e+00 -5.59803843e-01 4.59493965e-01 2.14794502e-01
-1.43631995e+00 -7.18390405e-01 4.33949977e-01 -4.17204350e-01
1.55273283e+00 5.58476597e-02 3.22925031e-01 1.45094168e+00
3.63632888e-01 1.32599330e+00 8.32895398e-01 -6.14704251e-01
-2.82686651e-01 5.48474729e-01 4.70742881e-01 6.38091445e-01
-1.78148672e-02 -1.59296423e-01 -4.12218928e-01 2.18419537e-01
4.59644437e-01 1.38347633e-02 -3.31801683e-01 -2.75724858e-01
-1.10345280e+00 1.00244272e+00 1.84272185e-01 3.11759263e-01
-6.70809820e-02 5.99273324e-01 5.25771081e-01 4.67594624e-01
1.91190869e-01 -2.32127443e-01 -3.21300119e-01 2.45101482e-01
-9.21369791e-01 -1.50413588e-01 5.90029001e-01 1.33941698e+00
3.53317678e-01 4.10494864e-01 -5.68120897e-01 1.17373252e+00
1.00854731e+00 9.17386293e-01 6.75993621e-01 -2.61121929e-01
5.64850628e-01 1.59981355e-01 -3.74832660e-01 -7.13949859e-01
-8.08750167e-02 -4.43410695e-01 -9.54040289e-01 -2.30060071e-01
2.20022406e-02 3.18888128e-01 -1.58004558e+00 1.59719050e+00
-2.55571246e-01 6.06182106e-02 2.38937914e-01 6.49607599e-01
1.41782582e+00 9.84727502e-01 3.60467851e-01 3.32557820e-02
1.56613791e+00 -1.23656666e+00 -7.60593534e-01 -6.07733488e-01
8.70583534e-01 -9.29941297e-01 1.35802889e+00 2.23745361e-01
-8.97435248e-01 -7.21401870e-01 -9.64733481e-01 -4.32814777e-01
-3.05448920e-01 2.76257008e-01 2.69407809e-01 6.34010553e-01
-1.43943596e+00 -2.90021211e-01 -5.48328102e-01 -5.57753265e-01
3.28893334e-01 1.68747485e-01 -1.49188831e-01 -3.34487289e-01
-1.05867732e+00 7.21949935e-01 3.06935519e-01 -5.29974364e-02
-1.35453069e+00 -2.67477840e-01 -1.04047954e+00 -9.54184532e-02
3.55814099e-01 -6.92619443e-01 1.08050811e+00 -9.43928123e-01
-1.12703776e+00 1.21609545e+00 -5.66629648e-01 -6.37172043e-01
3.23324472e-01 -1.48804322e-01 -2.96727479e-01 -1.58646107e-01
2.43549934e-03 9.54366148e-01 1.26535368e+00 -1.22956717e+00
-3.32082570e-01 -3.37540627e-01 -2.12699652e-01 3.46017778e-01
-4.36895460e-01 -3.21260057e-02 -1.02586412e+00 -6.87603652e-01
-1.20309986e-01 -6.70853198e-01 6.30187616e-02 -7.01612756e-02
-3.89004439e-01 -6.08091950e-01 9.67995346e-01 -7.34864056e-01
1.14657831e+00 -2.23505998e+00 1.79961935e-01 -2.29001954e-01
3.67394120e-01 3.51137608e-01 -4.91832674e-01 3.13978881e-01
1.31736711e-01 3.09890229e-02 9.01992023e-02 -7.15024650e-01
1.61407888e-01 3.01419526e-01 -8.25309753e-01 2.19658315e-01
-1.23732984e-02 1.37829435e+00 -3.69289339e-01 -5.86078942e-01
3.23223054e-01 5.65313220e-01 -4.82412368e-01 1.47673190e-01
-3.82626534e-01 -8.82225856e-02 -2.42885858e-01 7.39535570e-01
4.23585325e-01 -4.47210312e-01 -1.96028531e-01 -1.65421516e-01
1.52409807e-01 -6.88420748e-03 -9.26629663e-01 1.53478479e+00
-6.93427384e-01 9.82313514e-01 4.77919988e-02 -1.06701469e+00
8.55302155e-01 2.86237597e-01 -2.49200836e-01 -9.59081888e-01
5.05368650e-01 -1.40055260e-02 -1.20290838e-01 -4.54825848e-01
5.73907077e-01 7.06861988e-02 1.28285348e-01 -8.45141634e-02
3.90830636e-01 -3.23055148e-01 -2.15448305e-01 5.69935501e-01
7.99509466e-01 -1.34709477e-02 4.03311670e-01 -8.83509293e-02
7.27902293e-01 -1.42806247e-01 -4.54820581e-02 1.15952909e+00
-3.74619931e-01 6.77066207e-01 2.38623962e-01 -1.41479196e-02
-1.05022299e+00 -8.89062226e-01 -2.02434305e-02 1.36373627e+00
1.06936648e-01 -4.47378218e-01 -3.75582129e-01 -5.02316475e-01
-1.76541746e-01 1.15415847e+00 -4.64395851e-01 -1.52104601e-01
-3.84472519e-01 -4.86830324e-01 6.98089242e-01 6.38039410e-01
3.46337199e-01 -1.14625883e+00 -3.34076695e-02 8.72047916e-02
-4.15561013e-02 -1.49855769e+00 -7.95320570e-01 2.88688511e-01
-4.86186802e-01 -4.72355455e-01 -1.16166997e+00 -1.21808290e+00
9.44657326e-02 7.08456576e-01 1.13069713e+00 -1.65669754e-01
-2.63219148e-01 1.03090870e+00 -3.63612056e-01 -5.80779910e-01
-4.29560333e-01 -4.71111340e-03 -5.40291630e-02 -5.32164052e-02
7.83004880e-01 -1.43777281e-01 -1.89392567e-01 -8.83962493e-03
-7.93529391e-01 1.33043140e-01 6.55699849e-01 9.88073885e-01
6.06035113e-01 -5.53178430e-01 1.08264558e-01 -3.44420999e-01
4.92858469e-01 -3.15472335e-01 -5.85599005e-01 4.13711876e-01
-4.81782734e-01 1.35712609e-01 3.90612274e-01 -6.55825436e-01
-7.51688778e-01 -4.17315215e-02 -3.08914840e-01 -8.25913846e-01
-1.58005729e-01 2.14167580e-01 -2.41959810e-01 1.87431827e-01
2.95787305e-01 9.16945636e-01 -3.26199502e-01 -4.33859259e-01
6.30767703e-01 1.03134048e+00 3.29595268e-01 -9.59366933e-02
6.13931000e-01 3.30242217e-01 -6.70801461e-01 -1.50847960e+00
-5.17099023e-01 -6.55066371e-01 -2.98657805e-01 1.65386543e-01
1.46793973e+00 -1.42420757e+00 -4.89929408e-01 6.54142201e-01
-1.38242137e+00 -5.76160848e-01 3.79001945e-02 3.56524974e-01
-2.40665391e-01 7.52701938e-01 -4.65046287e-01 -8.77919137e-01
-5.79447448e-01 -1.47006416e+00 1.51198065e+00 -2.30221823e-02
1.04144588e-01 -1.04801226e+00 -4.83377911e-02 5.39189458e-01
4.40861136e-01 -7.69330740e-01 8.38442087e-01 -7.02032685e-01
-4.86779749e-01 2.70157531e-02 -7.23267138e-01 4.19450939e-01
-2.74375319e-01 -4.01052326e-01 -1.07902122e+00 -4.57028210e-01
-1.31266668e-01 -6.02478325e-01 1.33547211e+00 4.80845481e-01
1.04208398e+00 2.13594854e-01 -4.60956842e-01 7.21614599e-01
1.41284072e+00 4.10935879e-02 5.28138936e-01 5.79137057e-02
1.27200842e+00 2.49889910e-01 1.75643057e-01 1.57562837e-01
7.22714722e-01 8.48979056e-01 3.74132603e-01 -6.82591274e-02
-7.41184115e-01 -2.79060245e-01 7.20932961e-01 1.07048357e+00
5.84736466e-01 -9.97905791e-01 -1.04027772e+00 8.16117704e-01
-1.48420787e+00 -5.21002471e-01 -4.24241126e-01 1.85620618e+00
4.87914830e-01 1.99152812e-01 -2.63166547e-01 -3.85794520e-01
4.77684587e-01 5.59984446e-01 -4.85166997e-01 -5.78519762e-01
-4.65265155e-01 1.15795135e-01 7.22160637e-01 7.37482250e-01
-1.11385930e+00 1.55049121e+00 5.57876682e+00 1.27940822e+00
-1.15250528e+00 4.06435907e-01 6.59177423e-01 2.28641815e-02
-4.89606589e-01 -9.32752714e-02 -1.31536853e+00 1.51580498e-01
7.80824780e-01 1.66282907e-01 7.39160553e-02 8.48857105e-01
2.57198438e-02 1.82005897e-01 -9.17673647e-01 1.56868517e+00
5.62138975e-01 -1.13108611e+00 6.38979316e-01 1.52542621e-01
4.18384641e-01 6.79762244e-01 9.61605236e-02 7.38304079e-01
3.04543912e-01 -1.44288278e+00 6.91430688e-01 3.53800565e-01
1.02781677e+00 -3.71210217e-01 6.78148150e-01 5.57075143e-01
-1.34468317e+00 1.44384772e-01 -6.37132287e-01 3.04976910e-01
2.55300462e-01 1.09861486e-01 -6.99568331e-01 9.24054459e-02
6.60704792e-01 6.38766587e-01 -5.34438312e-01 4.69750911e-01
-7.04991966e-02 7.46236444e-01 -2.13082001e-01 -2.68879622e-01
6.10569537e-01 8.84617344e-02 6.79796576e-01 1.36696208e+00
-3.75169739e-02 -1.62878096e-01 2.53056973e-01 6.18257880e-01
-2.64830530e-01 2.89800763e-01 -8.61087143e-01 -2.00092345e-01
1.42635211e-01 8.53740692e-01 -4.23119038e-01 -5.98565638e-01
-7.74224222e-01 1.40978897e+00 3.13089490e-01 4.68986779e-01
-1.03795421e+00 -1.20293312e-01 3.89238864e-01 -1.82216927e-01
7.34501123e-01 -3.97166908e-01 -5.04668541e-02 -1.53662062e+00
-1.70248672e-01 -9.92794275e-01 2.00854972e-01 -9.25592482e-01
-1.16022444e+00 6.72511220e-01 -2.58600354e-01 -8.42690647e-01
-1.60418283e-02 -9.37930822e-01 -3.52150977e-01 8.66025746e-01
-1.60324991e+00 -1.57578635e+00 -2.78541923e-01 5.90766549e-01
1.50761104e+00 -5.05645394e-01 5.80582798e-01 1.71810329e-01
-4.91826683e-01 8.98866892e-01 1.36645481e-01 2.23834530e-01
7.97127604e-01 -9.94043887e-01 5.68461478e-01 7.92099357e-01
6.67977512e-01 3.16991210e-01 4.78115261e-01 -5.90916693e-01
-1.93622911e+00 -1.21394730e+00 9.81893659e-01 -7.28107810e-01
6.76620066e-01 -7.80840814e-01 -9.33779776e-01 9.13087428e-01
5.50502241e-01 -1.19801380e-01 2.75039017e-01 -4.69648361e-01
-5.03651857e-01 3.83909911e-01 -5.18163025e-01 8.02240610e-01
9.84711587e-01 -6.72347069e-01 -7.75614321e-01 4.51460272e-01
1.23157394e+00 -1.86155930e-01 -2.52771556e-01 1.91013396e-01
3.47479373e-01 -4.14287955e-01 1.13540447e+00 -5.24981618e-01
2.92202830e-01 -1.22153178e-01 -7.31362402e-01 -7.19846666e-01
-2.23635659e-01 -2.49536887e-01 -2.40345493e-01 1.27140677e+00
4.05043900e-01 -4.53723907e-01 5.04721522e-01 2.93982737e-02
-6.61524618e-03 -5.06764710e-01 -6.43835604e-01 -7.14525104e-01
2.97103345e-01 -9.89384711e-01 -7.14466348e-02 7.86336541e-01
-5.57325780e-01 8.96955132e-01 -7.78886437e-01 1.52733937e-01
6.70993805e-01 -1.98057845e-01 9.17594612e-01 -6.44854486e-01
-5.54971159e-01 -4.34319228e-01 -3.13288361e-01 -2.00895643e+00
1.71056464e-01 -1.06071472e+00 -5.57743981e-02 -1.63776672e+00
4.67031300e-01 9.34742987e-02 -7.09939674e-02 3.11580688e-01
-1.31926551e-01 4.56410289e-01 4.31600302e-01 1.75749883e-01
-6.73605323e-01 8.14811409e-01 1.00121868e+00 -7.03855515e-01
-1.36930242e-01 -4.15287971e-01 -5.34679770e-01 6.70657873e-01
3.24826062e-01 -2.12121174e-01 -6.28695548e-01 -1.01847780e+00
-1.91701546e-01 -3.70111354e-02 3.77431542e-01 -4.91499305e-01
2.27069348e-01 1.04033485e-01 3.17946136e-01 -8.64317715e-01
3.17638516e-01 -6.22878909e-01 -4.54779714e-01 4.87868607e-01
-4.40822959e-01 -2.34901592e-01 2.35283107e-01 6.60700738e-01
-2.02198714e-01 -2.22115710e-01 7.35600412e-01 -1.66718870e-01
-1.22390997e+00 3.48140717e-01 -5.88513792e-01 2.15832144e-01
6.74203098e-01 -3.96053314e-01 -7.10541522e-03 -6.60300434e-01
-7.73491442e-01 5.09855628e-01 2.29965463e-01 8.49412024e-01
9.25818503e-01 -1.03308511e+00 -1.08039296e+00 2.23068714e-01
5.16208887e-01 -2.59489477e-01 2.04136267e-01 8.35870326e-01
-2.82767981e-01 7.50645041e-01 4.62785274e-01 -1.07060945e+00
-1.46315694e+00 6.79695845e-01 2.40790457e-01 -2.74428248e-01
-7.70076036e-01 1.22513092e+00 1.02145517e+00 -4.12605673e-01
5.80617905e-01 -2.99152017e-01 -4.84536856e-01 1.77248362e-02
5.76448441e-01 -1.24727018e-01 -1.62479177e-01 -9.63583589e-01
-5.03617585e-01 7.71815121e-01 -5.12954414e-01 -2.12570772e-01
8.61794412e-01 -5.42979300e-01 3.36966902e-01 5.60089707e-01
1.43719101e+00 -6.33089319e-02 -7.59544253e-01 -5.71028531e-01
-2.73916364e-01 -1.46149909e-02 4.76754129e-01 -5.94939590e-01
-8.88712466e-01 1.40463030e+00 8.37882102e-01 -1.45754576e-01
8.02133143e-01 2.61638969e-01 8.64466548e-01 4.58300710e-01
2.40790546e-02 -7.55460560e-01 7.43839592e-02 1.04822588e+00
9.34471965e-01 -1.58876419e+00 -3.33792597e-01 -2.27717966e-01
-9.75374997e-01 9.12569165e-01 6.30113304e-01 -1.01650521e-01
4.94212657e-01 2.32259572e-01 1.45710930e-01 -2.40193427e-01
-9.66458917e-01 -3.41556221e-01 6.71548188e-01 5.72725832e-01
4.97830123e-01 -1.68541238e-01 -8.42584148e-02 4.96704400e-01
7.61585906e-02 -2.97170311e-01 3.66962969e-01 5.13715386e-01
-5.38821816e-01 -6.74870193e-01 -2.95042247e-01 3.40213746e-01
-2.82860845e-01 -7.43713319e-01 -4.00970012e-01 7.42659807e-01
-2.27485120e-01 8.38120222e-01 1.14015751e-01 -1.95417494e-01
1.31648049e-01 2.19156235e-01 2.81375349e-01 -6.36120021e-01
-1.91482887e-01 5.70336163e-01 -1.02969646e-01 -4.47824150e-01
1.32422715e-01 -2.15992227e-01 -1.04102719e+00 2.79823830e-03
-4.55944896e-01 -2.47354001e-01 6.64133310e-01 1.21361721e+00
2.38255844e-01 6.24012351e-01 1.56044707e-01 -5.12286723e-01
-5.90107322e-01 -9.88156617e-01 -3.76723200e-01 1.27300397e-01
4.40398097e-01 -2.58473307e-01 -3.75218064e-01 1.19363464e-01] | [11.529302597045898, 1.952925443649292] |
3145e8a2-b28a-40d0-bb49-250a19248a23 | emerging-properties-in-self-supervised-vision | 2104.14294 | null | https://arxiv.org/abs/2104.14294v2 | https://arxiv.org/pdf/2104.14294v2.pdf | Emerging Properties in Self-Supervised Vision Transformers | In this paper, we question if self-supervised learning provides new properties to Vision Transformer (ViT) that stand out compared to convolutional networks (convnets). Beyond the fact that adapting self-supervised methods to this architecture works particularly well, we make the following observations: first, self-supervised ViT features contain explicit information about the semantic segmentation of an image, which does not emerge as clearly with supervised ViTs, nor with convnets. Second, these features are also excellent k-NN classifiers, reaching 78.3% top-1 on ImageNet with a small ViT. Our study also underlines the importance of momentum encoder, multi-crop training, and the use of small patches with ViTs. We implement our findings into a simple self-supervised method, called DINO, which we interpret as a form of self-distillation with no labels. We show the synergy between DINO and ViTs by achieving 80.1% top-1 on ImageNet in linear evaluation with ViT-Base. | ['Armand Joulin', 'Piotr Bojanowski', 'Julien Mairal', 'Hervé Jégou', 'Ishan Misra', 'Hugo Touvron', 'Mathilde Caron'] | 2021-04-29 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Caron_Emerging_Properties_in_Self-Supervised_Vision_Transformers_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Caron_Emerging_Properties_in_Self-Supervised_Vision_Transformers_ICCV_2021_paper.pdf | iccv-2021-1 | ['self-supervised-image-classification', 'single-object-discovery'] | ['computer-vision', 'computer-vision'] | [ 3.39823246e-01 7.18485534e-01 -1.68890238e-01 -5.39889038e-01
-3.05313051e-01 -6.66997254e-01 8.23397756e-01 -1.47078574e-01
-6.58029497e-01 5.78698695e-01 1.47772536e-01 -1.48672476e-01
-5.87256625e-02 -6.89374685e-01 -1.04657793e+00 -6.60258174e-01
9.04241577e-02 2.18065947e-01 4.50547367e-01 -2.43464649e-01
-1.34617519e-02 2.24607870e-01 -1.54700720e+00 2.75277644e-01
8.02533150e-01 1.18986130e+00 2.29451597e-01 6.92740560e-01
-2.38990691e-02 1.12739849e+00 -5.59490442e-01 -5.51226556e-01
3.57461870e-01 -2.89184123e-01 -9.06501412e-01 1.82941914e-01
9.99742389e-01 -2.06616148e-01 -3.20992410e-01 9.62991953e-01
2.67353743e-01 -2.03288808e-01 8.15368652e-01 -1.06867504e+00
-8.86995494e-01 9.24619377e-01 -3.44851673e-01 2.27379024e-01
-3.63881737e-01 6.65950239e-01 1.25255728e+00 -6.87637746e-01
6.38103604e-01 1.00989926e+00 1.06415904e+00 4.43437368e-01
-1.27442145e+00 -3.67468446e-01 8.51552188e-02 7.70595670e-02
-9.48441386e-01 -4.54689920e-01 5.15592098e-01 -4.87988681e-01
9.68175054e-01 -1.04650110e-01 7.41588414e-01 1.13109612e+00
-4.53711953e-03 1.04287255e+00 1.50309026e+00 -5.03282905e-01
1.72136754e-01 5.39450407e-01 2.09688619e-01 7.47014046e-01
2.33690053e-01 4.38044935e-01 -4.99905258e-01 4.32619780e-01
7.28115261e-01 -1.80559441e-01 9.12697613e-02 -4.00198489e-01
-1.10602653e+00 7.22463608e-01 8.80751431e-01 2.75386930e-01
-2.92217970e-01 3.69017959e-01 4.42061305e-01 3.90982568e-01
5.08276999e-01 6.85602009e-01 -7.92420328e-01 -2.31049191e-02
-1.13174367e+00 -3.07443559e-01 6.13843203e-01 9.29049611e-01
1.14201224e+00 3.66315156e-01 -2.02568904e-01 6.48517370e-01
-7.77174607e-02 4.98223096e-01 5.37304759e-01 -9.64722872e-01
-2.44582240e-02 6.34453654e-01 -4.07958090e-01 -4.43732709e-01
-4.35755461e-01 -7.68638730e-01 -8.93059433e-01 4.52486217e-01
5.38902700e-01 -3.27431798e-01 -1.19220150e+00 1.71467793e+00
-2.20888048e-01 5.78161478e-02 3.91335487e-01 5.46415687e-01
9.39842582e-01 3.13553602e-01 1.48334488e-01 1.44380137e-01
1.10894549e+00 -1.39265573e+00 -1.90967023e-01 -5.60624957e-01
5.77975750e-01 -4.68765199e-01 1.11840999e+00 2.96422869e-01
-1.01739085e+00 -8.97033989e-01 -9.77259755e-01 1.25692517e-01
-5.64339995e-01 3.49822193e-01 9.16870832e-01 6.59997940e-01
-1.59497368e+00 9.49477136e-01 -5.95880687e-01 -6.30964994e-01
9.24647331e-01 2.85049319e-01 -2.55659968e-01 2.72990435e-01
-8.39783430e-01 1.10139298e+00 3.74357641e-01 -1.06064640e-01
-1.12139750e+00 -6.36655629e-01 -8.13602328e-01 8.96366835e-02
4.07722145e-01 -4.99898583e-01 1.31261194e+00 -1.67934918e+00
-1.33202839e+00 1.05471385e+00 -8.92216042e-02 -9.08020318e-01
4.53755945e-01 -9.22031254e-02 -3.87210469e-03 2.94681668e-01
1.16597407e-01 1.23308516e+00 1.10622275e+00 -1.19273770e+00
-5.96813381e-01 -1.58082172e-01 2.55829424e-01 1.09821655e-01
-4.14396673e-01 -3.62029135e-01 -1.35669857e-01 -4.38807935e-01
-7.53772780e-02 -7.34123647e-01 -2.77052164e-01 6.20132200e-02
-4.38775510e-01 -3.64815623e-01 5.58283448e-01 -2.07515419e-01
6.16754532e-01 -2.23054886e+00 -2.05567300e-01 1.48576936e-02
4.02094603e-01 4.53031033e-01 -2.79109240e-01 1.19708382e-01
-2.41209567e-01 4.98601384e-02 -4.47745830e-01 -3.48918349e-01
-5.26423194e-02 4.05428767e-01 -1.63308948e-01 2.34455004e-01
9.18322682e-01 1.47725403e+00 -8.99943948e-01 -5.15579641e-01
2.62250513e-01 2.48020634e-01 -3.18047196e-01 -1.21935278e-01
-3.38113964e-01 2.70556182e-01 -1.31332770e-01 4.94073749e-01
5.78000069e-01 -4.93999213e-01 5.71792945e-02 -3.21452022e-01
-3.85330200e-01 2.31538862e-01 -6.42265260e-01 1.54347408e+00
-1.65640727e-01 9.29490089e-01 -2.52264172e-01 -1.26882887e+00
8.37412417e-01 -2.50531603e-02 1.93770915e-01 -9.14981127e-01
5.38500994e-02 1.32631704e-01 -9.69728380e-02 -4.27305758e-01
4.26570088e-01 -1.71418060e-02 3.01991135e-01 2.89285034e-01
8.80368233e-01 -1.63873568e-01 3.55653644e-01 2.49958292e-01
1.01266897e+00 3.88554484e-01 -3.29730357e-03 -5.34020960e-01
2.22344920e-01 2.09141374e-01 2.22510576e-01 1.05701029e+00
-1.73567891e-01 6.07573867e-01 6.96245968e-01 -2.42951289e-01
-1.13047528e+00 -9.58398640e-01 -2.24980772e-01 1.06454372e+00
-1.04182862e-01 -2.13231534e-01 -8.72854650e-01 -9.33910847e-01
-1.46268643e-02 3.59500170e-01 -9.77887332e-01 -2.63053715e-01
-2.89144397e-01 -6.30216897e-01 8.03748190e-01 7.90479481e-01
7.28731930e-01 -1.24679983e+00 -7.29968905e-01 -1.10938065e-01
3.53911310e-01 -1.21846306e+00 3.65908593e-02 9.74759817e-01
-1.03453636e+00 -1.20551038e+00 -9.19355392e-01 -1.02586508e+00
6.98246539e-01 2.29776248e-01 1.36512220e+00 -7.68329874e-02
-2.28173241e-01 5.00171542e-01 -3.70469362e-01 -5.11321664e-01
-3.95765871e-01 3.92746747e-01 -2.05779448e-01 -1.27030358e-01
2.32969195e-01 -5.90369284e-01 -5.41471064e-01 1.54133409e-01
-7.95249701e-01 1.09401412e-01 9.66996849e-01 9.48580086e-01
4.79867995e-01 9.93582420e-03 5.34289181e-01 -1.16029739e+00
6.38456121e-02 -2.40535036e-01 -3.34621459e-01 2.69619823e-01
-8.69965494e-01 2.44320497e-01 6.77035213e-01 -3.33570451e-01
-8.92240942e-01 2.14363560e-01 -7.65912384e-02 -2.86772192e-01
-4.40240353e-01 3.02924603e-01 2.40897015e-01 -3.13496053e-01
1.10941720e+00 2.59852499e-01 2.53232956e-01 -3.03226352e-01
6.39408350e-01 5.17005324e-01 6.15917444e-01 -3.04320484e-01
8.97809088e-01 5.79240441e-01 -1.97235659e-01 -7.88970590e-01
-1.38886154e+00 -5.34763277e-01 -8.21575046e-01 -2.63063796e-02
9.60920751e-01 -1.07356441e+00 -6.07894719e-01 6.51411772e-01
-8.11308265e-01 -8.61149907e-01 -1.06363535e+00 2.28520811e-01
-6.75522447e-01 2.11089253e-01 -6.32320225e-01 -6.44180238e-01
-2.29257420e-01 -8.70491683e-01 6.86643898e-01 5.04165888e-01
2.49691755e-02 -1.08650923e+00 -2.11173922e-01 1.38963267e-01
6.50808215e-01 -1.68315358e-02 5.26438594e-01 -5.13018847e-01
-4.91696596e-01 1.63847342e-01 -5.86292684e-01 9.49356973e-01
8.87089968e-03 -7.15900511e-02 -1.53598440e+00 -1.12174369e-01
-2.34294295e-01 -1.00070977e+00 1.43103075e+00 5.49464822e-01
9.72620606e-01 -1.47291780e-01 5.04517332e-02 7.70834625e-01
1.53877687e+00 -2.99934804e-01 6.82634175e-01 4.16573316e-01
6.59050822e-01 5.37744105e-01 1.83522463e-01 -6.73611090e-02
5.16207874e-01 1.46754861e-01 5.49410284e-01 -6.16420507e-01
-5.32767415e-01 -2.69367397e-01 3.69420320e-01 5.60882986e-01
-1.67601496e-01 1.79014578e-01 -6.73758626e-01 5.98174810e-01
-1.71078718e+00 -6.55069470e-01 -2.04571515e-01 1.81321096e+00
8.22691560e-01 4.74285871e-01 1.77519366e-01 5.58168180e-02
3.67444783e-01 1.98432848e-01 -6.46442592e-01 -2.72338241e-01
-7.36483395e-01 5.95973909e-01 1.00234139e+00 1.30233064e-01
-1.31394160e+00 1.37656438e+00 7.16278982e+00 6.43543124e-01
-1.22143447e+00 9.13794041e-02 7.38276482e-01 3.43677789e-01
5.16173802e-03 1.07110091e-01 -6.78149462e-01 1.67287037e-01
8.92428219e-01 2.71715254e-01 3.17843676e-01 9.83508945e-01
-1.89014450e-01 -2.99741685e-01 -1.07725871e+00 6.83431804e-01
7.11266845e-02 -1.40221822e+00 -6.05122931e-02 -2.53875107e-01
1.00550342e+00 5.97574234e-01 2.43934855e-01 4.38564926e-01
5.72719157e-01 -1.13872576e+00 8.98559153e-01 3.59521478e-01
9.90541756e-01 -3.77929509e-01 6.91092551e-01 1.03378460e-01
-7.32770026e-01 -1.30772769e-01 -7.33790338e-01 -1.11216053e-01
-1.62663326e-01 5.56662261e-01 -7.55922318e-01 3.79371136e-01
8.70917022e-01 1.16053212e+00 -1.09090042e+00 1.03279400e+00
-5.24435818e-01 1.18544245e+00 -2.84751147e-01 1.56751662e-01
6.41544759e-01 1.13048606e-01 2.40580112e-01 1.34784913e+00
-2.46044159e-01 -3.32771361e-01 -1.77608728e-01 8.69928658e-01
-2.14382634e-02 -2.58314013e-01 -5.41173041e-01 -1.47159547e-01
-1.15268983e-01 1.35050178e+00 -8.61112952e-01 -5.52371860e-01
-4.52708602e-01 1.07828224e+00 4.49351192e-01 3.92340183e-01
-3.56273770e-01 -6.96603358e-02 2.88923115e-01 -1.55133400e-02
6.33199513e-01 4.34090979e-02 -6.17351890e-01 -1.08784294e+00
-9.09618884e-02 -6.91029727e-01 1.06671624e-01 -9.14867163e-01
-1.46145391e+00 6.90962195e-01 -2.43198022e-01 -1.07911742e+00
-4.61979657e-02 -9.89276111e-01 -6.36964083e-01 4.89797205e-01
-1.99150336e+00 -1.32044852e+00 -2.99316734e-01 4.65826601e-01
3.57364029e-01 -3.48145097e-01 7.99086332e-01 -7.85580575e-02
-3.86912942e-01 7.20630884e-01 1.70678228e-01 3.44525605e-01
6.86017990e-01 -1.56657791e+00 7.12807655e-01 6.87314808e-01
3.89757335e-01 4.40727800e-01 3.83292645e-01 -2.66881794e-01
-8.51814568e-01 -1.01922297e+00 6.90860271e-01 -5.08249581e-01
7.26430774e-01 -2.04748526e-01 -7.34535277e-01 6.72475278e-01
5.95787585e-01 1.65791780e-01 2.45250389e-01 3.18738580e-01
-6.67363882e-01 -1.92530826e-01 -1.02804649e+00 4.05882448e-01
1.17316866e+00 -5.20324409e-01 -5.13802707e-01 3.55582893e-01
7.91187525e-01 -1.49137110e-01 -5.29854298e-01 2.72622287e-01
2.92793155e-01 -1.19290972e+00 9.00782585e-01 -5.67308605e-01
7.82212198e-01 -1.11858360e-02 1.04162917e-01 -1.41685569e+00
-3.62541407e-01 -3.12065154e-01 1.19337924e-01 1.04813337e+00
6.63138688e-01 -6.66659355e-01 1.10666680e+00 1.40815988e-01
-2.14602649e-01 -6.76400363e-01 -5.95265090e-01 -9.04890060e-01
2.87900865e-01 -3.81220996e-01 2.35367343e-01 1.10378003e+00
-3.46234083e-01 4.69445765e-01 -1.12616584e-01 -2.69807130e-01
6.91693962e-01 -2.99909294e-01 5.90031326e-01 -1.36040604e+00
-3.78790855e-01 -5.13640881e-01 -5.66311955e-01 -1.03517699e+00
1.68877110e-01 -1.10898578e+00 4.30071019e-02 -1.53499758e+00
2.78391927e-01 -6.13970399e-01 -4.29937989e-01 9.19976294e-01
-1.15153808e-02 5.88174045e-01 1.62638932e-01 2.63402462e-01
-6.73836410e-01 3.79895002e-01 1.26747751e+00 -3.62767488e-01
3.24536748e-02 -1.18056133e-01 -1.00527084e+00 7.91123271e-01
9.09607232e-01 -3.42342049e-01 -3.02264571e-01 -5.88832021e-01
1.80922270e-01 -5.86976886e-01 6.79975033e-01 -1.14673638e+00
2.28726208e-01 1.78098753e-01 5.17121553e-01 -2.10542917e-01
1.47311851e-01 -5.39879858e-01 -5.59583545e-01 3.74897480e-01
-3.83656025e-01 -2.22999603e-01 3.03187966e-01 3.56375545e-01
-1.81329370e-01 -4.56674963e-01 8.63648951e-01 -4.35598910e-01
-1.06913733e+00 -2.18681823e-02 -2.94777662e-01 2.25899264e-01
7.00538218e-01 -5.77829301e-01 -4.59590137e-01 -2.97760218e-01
-6.87882721e-01 2.61841863e-01 4.05280709e-01 2.05886394e-01
2.87455618e-01 -7.98305929e-01 -6.89216077e-01 1.98445365e-01
2.35208005e-01 2.52008498e-01 5.53029366e-02 7.52931774e-01
-3.73251170e-01 4.07003731e-01 -3.29516262e-01 -8.19974899e-01
-7.94435501e-01 2.27371007e-01 3.82976592e-01 -1.10503651e-01
-7.30514646e-01 1.19072354e+00 2.81306177e-01 -5.56367576e-01
3.68811369e-01 -2.97177911e-01 -2.89609104e-01 8.41857195e-02
1.68203071e-01 -6.15210161e-02 -1.96481142e-02 -3.74346107e-01
-1.57345578e-01 4.86487448e-01 -4.61851619e-02 -6.53856026e-04
1.56733525e+00 8.01714361e-02 -5.04285283e-02 5.71013391e-01
9.73189116e-01 -3.93837392e-01 -1.80952752e+00 -3.69463205e-01
7.53475502e-02 2.34422833e-01 1.05226211e-01 -1.11733830e+00
-1.32388544e+00 8.65277827e-01 6.12966120e-01 2.99030188e-02
9.46744502e-01 2.26940095e-01 3.77700061e-01 5.87748766e-01
1.32806569e-01 -1.17961347e+00 3.08634996e-01 7.52425313e-01
4.24625278e-01 -1.42186463e+00 -1.57695830e-01 -1.42843038e-01
-9.36657667e-01 9.03925478e-01 7.66869128e-01 -4.45726186e-01
5.70287824e-01 2.19450668e-01 3.03009897e-01 -1.90564603e-01
-7.10912228e-01 -8.22701216e-01 7.75934085e-02 8.89580190e-01
3.34051788e-01 -5.38805537e-02 7.03789443e-02 2.40578696e-01
-4.13972914e-01 1.42439470e-01 4.97443885e-01 7.56319046e-01
-4.91734326e-01 -8.72228920e-01 4.93678600e-02 7.17930675e-01
-2.67368138e-01 -4.32381272e-01 -5.32444715e-01 7.00653434e-01
3.07311952e-01 7.52285838e-01 2.84189165e-01 -4.95737106e-01
1.97660923e-01 1.32914677e-01 5.26654005e-01 -7.41342127e-01
-9.35120523e-01 -3.44173878e-01 -9.72167850e-02 -3.10478151e-01
-6.50087178e-01 -3.93855125e-01 -8.21683466e-01 2.17462610e-02
-4.21295762e-01 -3.97462659e-02 7.27550745e-01 1.04420400e+00
7.78150260e-02 6.31003797e-01 5.69307148e-01 -7.66044736e-01
-6.30280256e-01 -1.10073066e+00 -4.48394001e-01 2.23867998e-01
4.34436381e-01 -3.62486154e-01 -4.60695446e-01 2.46072859e-01] | [9.636892318725586, 2.060929298400879] |
2eed535f-8cae-4c81-a76d-709d83451147 | high-precision-automated-reconstruction-of | null | null | https://doi.org/10.1038/s41592-018-0049-4 | https://www.biorxiv.org/content/biorxiv/early/2017/10/09/200675.full-text.pdf | High-Precision Automated Reconstruction of Neurons with Flood-filling Networks | Reconstruction of neural circuits from volume electron microscopy data requires the tracing of complete cells including all their neurites. Automated approaches have been developed to perform the tracing, but without costly human proofreading their error rates are too high to obtain reliable circuit diagrams. We present a method for automated segmentation that, like the majority of previous efforts, employs convolutional neural networks, but contains in addition a recurrent pathway that allows the iterative optimization and extension of the reconstructed shape of individual neural processes. We used this technique, which we call flood-filling networks, to trace neurons in a data set obtained by serial block-face electron microscopy from a male zebra finch brain. Our method achieved a mean error-free neurite path length of 1.1 mm, an order of magnitude better than previously published approaches applied to the same dataset. Only 4 mergers were observed in a neurite test set of 97 mm path length. | ['Jeremy Maitin-Shepard', 'Jörgen Kornfeld', 'Michał Januszewski', 'Winfried Denk', 'Art Pope', 'Viren Jain', 'Peter H. Li', 'Larry Lindsey', 'Tim Blakely', 'Mike Tyka'] | 2017-10-09 | null | null | null | nature-methods-2017-10 | ['electron-microscopy-image-segmentation'] | ['computer-vision'] | [ 4.00381386e-01 3.05838734e-01 7.80183434e-01 -1.98600605e-01
-4.31700379e-01 -7.53590882e-01 3.66087496e-01 2.18729988e-01
-9.74359155e-01 1.05879450e+00 -7.12675571e-01 -6.05602324e-01
3.23353037e-02 -6.02463007e-01 -7.32469916e-01 -6.82920754e-01
2.78334487e-02 9.18060958e-01 5.41305006e-01 2.49308690e-01
6.14186704e-01 1.04653370e+00 -1.08856153e+00 -2.11590528e-02
4.84423757e-01 7.72290289e-01 5.79707026e-01 6.49764955e-01
-3.23746115e-01 5.54485381e-01 -3.36309940e-01 -4.41350460e-01
1.07476555e-01 -3.63196433e-01 -1.10893619e+00 1.96341559e-01
1.26567781e-01 -3.17326188e-01 2.63403729e-02 6.49472892e-01
2.60294288e-01 -2.90251791e-01 7.04194486e-01 -6.00495517e-01
-1.97735429e-01 4.13095653e-01 -2.98600942e-01 3.16072911e-01
-1.64738387e-01 -1.48897897e-02 6.16245091e-01 -5.72614968e-01
1.13138926e+00 7.40665078e-01 9.34791744e-01 7.15695381e-01
-1.73566353e+00 -8.03329110e-01 -4.62410808e-01 -3.19444478e-01
-1.21874082e+00 -3.42706144e-01 3.58582735e-01 -7.32805550e-01
9.81095552e-01 -3.11147571e-01 1.23467040e+00 5.34195483e-01
5.49102128e-01 2.64686078e-01 1.43410242e+00 -2.66994357e-01
3.94574642e-01 -3.26413482e-01 8.59243423e-02 7.16647327e-01
2.31995627e-01 -2.27233201e-01 1.89912707e-01 5.70659786e-02
1.24164784e+00 -1.17189936e-01 6.83741346e-02 7.75091276e-02
-1.05351424e+00 6.73671305e-01 5.02075404e-02 4.83610779e-01
5.38036274e-03 3.84748846e-01 3.75569075e-01 -1.11680441e-01
1.66385755e-01 4.48772937e-01 -3.76971185e-01 1.29268959e-01
-1.16002643e+00 3.90881270e-01 7.50196815e-01 7.51341164e-01
7.72419930e-01 -9.15466323e-02 5.14716446e-01 5.38479686e-01
2.99490362e-01 -1.71788871e-01 6.88226297e-02 -1.50691295e+00
-5.38324332e-03 9.15822506e-01 -1.73672214e-01 -4.42567527e-01
-8.80798161e-01 -1.83232069e-01 -6.76304400e-01 9.29887533e-01
1.05330789e+00 -3.24318349e-01 -1.03625143e+00 1.17369163e+00
1.47321239e-01 -4.10478503e-01 -3.24786693e-01 5.62896550e-01
5.19410253e-01 2.26492226e-01 -1.93448082e-01 -1.84186041e-01
1.04254198e+00 -3.45705301e-01 -3.94820482e-01 6.63105994e-02
6.03065133e-01 -6.94062948e-01 5.99832594e-01 5.80108047e-01
-1.32731080e+00 9.52523351e-02 -1.29752898e+00 -7.56338462e-02
-4.05277520e-01 -8.03696811e-02 6.05127096e-01 4.63419139e-01
-1.15862203e+00 1.14510584e+00 -1.00427878e+00 -4.31036621e-01
1.10862124e+00 8.18747938e-01 -6.48481250e-01 3.90541673e-01
-1.34925291e-01 7.95066833e-01 2.18510643e-01 -8.40223432e-02
-9.63190198e-01 -5.49576759e-01 -2.78653771e-01 -2.59962715e-02
-2.37031039e-02 -5.00377059e-01 1.32529497e+00 -3.90013725e-01
-1.50285172e+00 1.26765561e+00 -6.96216300e-02 -6.70019150e-01
7.04034746e-01 3.27285141e-01 2.51913577e-01 4.29214001e-01
-2.11042657e-01 9.81181026e-01 2.21973121e-01 -1.42767763e+00
-4.27535027e-01 -6.45504177e-01 -1.77698895e-01 -4.76365745e-01
3.06356877e-01 8.18412900e-02 -2.32465357e-01 -2.49601305e-01
4.37750250e-01 -6.88827753e-01 -6.08386755e-01 4.63323683e-01
-2.82529861e-01 1.59248620e-01 7.23132670e-01 -6.69285178e-01
6.06960475e-01 -1.57324326e+00 -8.99545774e-02 3.98182988e-01
5.67366004e-01 6.53719380e-02 2.97528744e-01 3.38136137e-01
2.07787618e-01 2.67005742e-01 -6.98303878e-01 -5.62103212e-01
-4.81412590e-01 2.48432726e-01 1.19140819e-01 9.01683867e-01
-1.40681509e-02 7.32390702e-01 -5.38643062e-01 -7.61446118e-01
-3.10691260e-02 4.72788930e-01 -3.36764365e-01 -2.90408015e-01
-6.63772449e-02 6.87507212e-01 -6.66143671e-02 6.29511237e-01
5.15157759e-01 -3.79859000e-01 2.34202489e-01 3.03159356e-01
-5.34600556e-01 2.52695400e-02 -7.13319480e-01 1.69754553e+00
-3.42537835e-02 9.22500968e-01 4.43373144e-01 -6.29174829e-01
8.97285283e-01 2.16461897e-01 5.48282623e-01 -5.34187555e-01
5.96286535e-01 4.07877088e-01 2.72107899e-01 -7.47379735e-02
2.94118170e-02 -6.00208282e-01 5.20355582e-01 5.53138793e-01
3.97816211e-01 -3.19433868e-01 4.98909920e-01 1.77251026e-01
1.25978601e+00 3.63910168e-01 5.63141815e-02 -4.92105603e-01
2.35895112e-01 3.23615015e-01 5.88800073e-01 2.56394863e-01
-3.64250946e-03 1.04150391e+00 7.23518789e-01 -5.02216876e-01
-1.70648563e+00 -8.31961155e-01 -5.52293718e-01 2.09382743e-01
-8.51731747e-02 -7.41930455e-02 -1.25201249e+00 -2.47256815e-01
-3.43439221e-01 1.69486701e-01 -6.67916894e-01 6.84492648e-01
-6.20243728e-01 -6.75318003e-01 9.09278035e-01 3.88890624e-01
2.36761659e-01 -1.37198305e+00 -1.07000661e+00 6.76880300e-01
3.50505769e-01 -1.03499389e+00 3.70019734e-01 7.03546703e-01
-1.18908811e+00 -1.31565917e+00 -8.62561524e-01 -7.67615438e-01
1.06221533e+00 -5.41040063e-01 7.94975281e-01 4.64824498e-01
-6.98437214e-01 -3.01025659e-01 4.87873331e-02 -3.21080297e-01
-3.24227333e-01 -5.74894994e-03 -3.36204439e-01 -6.00423753e-01
-2.38891393e-02 -9.80456889e-01 -4.78703648e-01 2.73292661e-01
-7.98482060e-01 1.01796174e-02 5.52488506e-01 7.10131228e-01
1.13479805e+00 -1.60778239e-01 4.58453536e-01 -1.14649510e+00
3.86971295e-01 -2.04477578e-01 -1.03966188e+00 -2.48868927e-01
-8.52934718e-01 -1.63306236e-01 7.26908565e-01 -1.89846784e-01
-6.49486482e-01 3.50878090e-01 -6.06177509e-01 -4.25022207e-02
-5.70643425e-01 3.31780538e-02 2.07081258e-01 -6.65103316e-01
4.26013231e-01 1.15528576e-01 3.76617938e-01 -3.97837609e-01
-2.08037168e-01 1.32068425e-01 6.15294814e-01 -2.96169877e-01
4.22790825e-01 1.06036592e+00 4.98099595e-01 -7.73991585e-01
-1.05424784e-01 -3.43649447e-01 -1.05804324e+00 -3.56880695e-01
9.04216170e-01 -6.92018941e-02 -9.87897158e-01 5.23172736e-01
-1.39123869e+00 -7.22959578e-01 -2.01451797e-02 2.77354866e-01
-8.83190691e-01 3.13452482e-01 -1.06499493e+00 -5.48634112e-01
-4.80200440e-01 -1.05513978e+00 6.53092444e-01 1.82239592e-01
-3.90940249e-01 -8.73958170e-01 2.79145300e-01 1.05299652e-01
1.44978061e-01 4.87416714e-01 1.07816362e+00 -4.06021833e-01
-5.20098269e-01 -1.92462578e-01 -2.55501688e-01 5.79019450e-02
-2.32456967e-01 3.68384719e-01 -8.66381645e-01 1.28644677e-02
-1.68697461e-01 -3.00991684e-01 7.61376619e-01 5.17883956e-01
1.10214925e+00 -6.77246526e-02 -6.74233913e-01 7.68878996e-01
1.63113248e+00 5.74798763e-01 9.80634034e-01 5.76961756e-01
4.65101123e-01 8.12023580e-01 -6.11133203e-02 1.48943052e-01
-1.55076236e-01 4.08293493e-02 5.01332700e-01 -1.69587821e-01
-1.33632362e-01 2.85939783e-01 -3.21984977e-01 3.90549392e-01
-5.14262199e-01 8.62927362e-03 -1.05224597e+00 6.59115911e-01
-1.20856893e+00 -8.30728233e-01 -5.17865121e-01 1.75218511e+00
6.54965401e-01 3.70212734e-01 2.13724181e-01 3.92982662e-01
4.93153334e-01 -5.83250463e-01 -6.38194680e-01 -4.64843482e-01
-3.39203142e-02 3.23972493e-01 6.99229479e-01 4.82769996e-01
-6.12366378e-01 8.44329476e-01 7.68928337e+00 4.41135317e-01
-7.98335016e-01 -6.67966977e-02 6.71000540e-01 -9.65292379e-02
-1.48718640e-01 1.94358364e-01 -9.66581285e-01 3.46670806e-01
7.54719555e-01 2.20093474e-01 5.87854445e-01 4.25700665e-01
-2.73568574e-02 -4.38134253e-01 -8.61144781e-01 7.42627978e-01
-3.51092219e-01 -1.84306002e+00 -3.31977546e-01 6.44005597e-01
5.45557976e-01 2.21148759e-01 -4.74150479e-01 -4.11949426e-01
3.47476482e-01 -1.25257552e+00 7.95496941e-01 6.48764431e-01
9.14983094e-01 -9.89443660e-01 7.09635317e-01 3.83230180e-01
-8.03453982e-01 1.20594233e-01 -3.68883520e-01 -1.88379973e-01
3.90945017e-01 4.24506754e-01 -1.06898081e+00 -2.95563862e-02
8.02502751e-01 1.17675260e-01 -5.13336182e-01 1.33314681e+00
1.86662585e-01 4.48757052e-01 -4.83383656e-01 -7.47875571e-02
2.62270182e-01 -4.26581532e-01 3.54293346e-01 1.23056984e+00
2.46922389e-01 5.78066818e-02 -6.57836676e-01 1.31111228e+00
-2.00842515e-01 2.19853241e-02 -6.09525144e-01 -2.51013577e-01
2.16100588e-01 1.70825160e+00 -1.84534168e+00 8.43293294e-02
-2.12215006e-01 6.02168679e-01 5.64269781e-01 1.12930976e-01
-3.18000406e-01 -5.81782520e-01 1.79206192e-01 4.87819374e-01
3.74752730e-01 -5.13918638e-01 -8.87740910e-01 -1.43118441e-01
-2.41413534e-01 -1.53940096e-01 -7.12613389e-02 -5.11556506e-01
-9.31293905e-01 4.94545758e-01 -3.58193547e-01 -6.01348281e-01
-1.02530137e-01 -9.19684708e-01 -8.06088567e-01 6.58883214e-01
-9.09089208e-01 -8.50946724e-01 7.59341568e-02 1.30053520e-01
3.13942879e-01 3.14582395e-03 8.23201895e-01 1.82607144e-01
-3.59998405e-01 6.38075545e-02 1.21031813e-01 1.28450185e-01
4.43494326e-04 -1.29050136e+00 4.02539998e-01 4.82080251e-01
-1.74975693e-01 7.06826806e-01 6.32412374e-01 -7.24272490e-01
-8.92458260e-01 -9.77961898e-01 8.09047520e-01 -1.38026506e-01
6.06120467e-01 -3.70135128e-01 -9.16901886e-01 8.66312861e-01
2.54475772e-01 -2.23738477e-01 7.80940175e-01 -2.47077256e-01
3.72245282e-01 4.45264131e-01 -1.44810045e+00 5.80702603e-01
7.85474658e-01 -4.61004674e-02 -4.53173310e-01 -8.02698433e-02
6.88313097e-02 -1.04104176e-01 -9.80645061e-01 1.88550815e-01
9.54498231e-01 -1.00141311e+00 6.63368464e-01 -1.10024318e-01
5.54741740e-01 -2.84752816e-01 2.31912464e-01 -8.10244739e-01
-3.26963872e-01 -4.28306580e-01 2.13243544e-01 1.21319532e+00
4.83711064e-01 -4.88553971e-01 1.10512209e+00 3.03520322e-01
-4.20254976e-01 -1.18781960e+00 -1.07047462e+00 -6.94602311e-01
5.39395928e-01 -1.27632245e-01 2.31421262e-01 3.71544212e-01
-1.88905075e-02 2.69461900e-01 4.33743298e-01 -3.83171111e-01
9.15792525e-01 -8.58427137e-02 2.22544730e-01 -1.56842482e+00
1.71689004e-01 -7.23531783e-01 -4.98445630e-01 -6.38709247e-01
2.94764042e-01 -9.81919944e-01 2.04966277e-01 -1.88948047e+00
1.90050080e-01 -3.06419998e-01 3.50243390e-01 4.28903610e-01
6.76541209e-01 6.19095027e-01 -3.06875467e-01 3.51808041e-01
-2.73895562e-01 7.88087174e-02 1.51070321e+00 1.86382562e-01
4.52375785e-02 -3.17741364e-01 -4.90007430e-01 1.28798842e+00
9.29228544e-01 -7.54039824e-01 -4.43770401e-02 -3.61884683e-01
2.47061923e-01 -9.95584279e-02 4.13876712e-01 -1.31503844e+00
6.01417303e-01 2.37462506e-01 9.22881901e-01 -9.71049845e-01
3.69615078e-01 -8.37820411e-01 3.16573173e-01 5.10074973e-01
-2.21404478e-01 -2.37556249e-02 2.12873042e-01 2.68136948e-01
1.92880318e-01 -5.71481109e-01 1.06604445e+00 -6.96596146e-01
-2.26408705e-01 2.03259170e-01 -1.20422637e+00 -4.20445681e-01
9.69705224e-01 -6.05726600e-01 -3.13953906e-01 1.68370813e-01
-8.05182397e-01 -1.12555560e-03 1.00967014e+00 -7.95163870e-01
8.19163680e-01 -9.06411052e-01 -3.10610563e-01 -5.17635569e-02
-3.89653683e-01 2.18373358e-01 1.07592726e-02 9.04906034e-01
-1.25074112e+00 2.82134235e-01 -7.73228884e-01 -4.56531793e-01
-1.13753724e+00 3.03574979e-01 3.87439251e-01 -1.70083791e-01
-9.26689148e-01 7.69239664e-01 -7.28834495e-02 -5.71542025e-01
-1.00270271e-01 -2.23550886e-01 -3.71379942e-01 -2.31683522e-01
4.85575348e-01 6.22190714e-01 7.20573813e-02 -5.58465838e-01
-1.87896013e-01 6.84978068e-01 1.68460906e-02 -3.70512336e-01
1.56302488e+00 -2.83998102e-01 -5.08954763e-01 5.88685036e-01
9.99945223e-01 -3.10313046e-01 -1.52445912e+00 5.59391677e-01
8.46823230e-02 4.87056421e-03 -3.80513282e-03 -7.20389903e-01
-9.74456787e-01 9.97797668e-01 3.50540757e-01 1.23508405e-02
6.69101119e-01 6.72369972e-02 9.01990056e-01 4.46281731e-01
5.53030968e-01 -9.92326021e-01 -3.28728944e-01 6.71431303e-01
6.01903081e-01 -8.23789954e-01 -2.97877956e-02 -3.46852481e-01
1.56371653e-01 1.38986433e+00 4.40079629e-01 -4.37314719e-01
4.50464129e-01 9.67100978e-01 -1.18779637e-01 -6.84980750e-01
-6.00816488e-01 -7.81890899e-02 -3.29037637e-01 6.50785267e-01
4.02403206e-01 -4.06250477e-01 -4.32287425e-01 4.92307127e-01
-1.34641573e-01 2.98056334e-01 6.23859584e-01 1.12142813e+00
-6.82402372e-01 -9.74407792e-01 -2.21364334e-01 6.85301304e-01
-9.17620838e-01 2.15674251e-01 -7.28380442e-01 9.40550208e-01
1.87177792e-01 5.30843377e-01 3.98591459e-01 -2.14291081e-01
5.12289032e-02 2.55625565e-02 9.39665735e-01 -4.90043849e-01
-5.57783544e-01 3.16576391e-01 -1.35287605e-02 -3.57669920e-01
-1.88296080e-01 -6.19814456e-01 -2.19556236e+00 -4.44393426e-01
-2.13858813e-01 -1.51341453e-01 9.26156282e-01 1.26716661e+00
-1.83073580e-01 6.34828925e-01 -1.92550141e-02 -1.13447464e+00
4.74650711e-01 -8.62730920e-01 -1.08592737e+00 -1.19394407e-01
1.23775072e-01 -4.96474147e-01 -3.50911826e-01 4.66331810e-01] | [14.2518892288208, -3.134556531906128] |
a9ef8075-6241-46e4-93b3-ccf86015558f | crosslingual-generalization-through-multitask | 2211.01786 | null | https://arxiv.org/abs/2211.01786v2 | https://arxiv.org/pdf/2211.01786v2.pdf | Crosslingual Generalization through Multitask Finetuning | Multitask prompted finetuning (MTF) has been shown to help large language models generalize to new tasks in a zero-shot setting, but so far explorations of MTF have focused on English data and models. We apply MTF to the pretrained multilingual BLOOM and mT5 model families to produce finetuned variants called BLOOMZ and mT0. We find finetuning large multilingual language models on English tasks with English prompts allows for task generalization to non-English languages that appear only in the pretraining corpus. Finetuning on multilingual tasks with English prompts further improves performance on English and non-English tasks leading to various state-of-the-art zero-shot results. We also investigate finetuning on multilingual tasks with prompts that have been machine-translated from English to match the language of each dataset. We find training on these machine-translated prompts leads to better performance on human-written prompts in the respective languages. Surprisingly, we find models are capable of zero-shot generalization to tasks in languages they have never intentionally seen. We conjecture that the models are learning higher-level capabilities that are both task- and language-agnostic. In addition, we introduce xP3, a composite of supervised datasets in 46 languages with English and machine-translated prompts. Our code, datasets and models are freely available at https://github.com/bigscience-workshop/xmtf. | ['Colin Raffel', 'Edward Raff', 'Albert Webson', 'Zaid Alyafeai', 'Samuel Albanie', 'Khalid Almubarak', 'Alham Fikri Aji', 'Dragomir Radev', 'Xiangru Tang', 'Hailey Schoelkopf', 'Zheng-Xin Yong', 'Sheng Shen', 'M Saiful Bari', 'Teven Le Scao', 'Stella Biderman', 'Adam Roberts', 'Lintang Sutawika', 'Thomas Wang', 'Niklas Muennighoff'] | 2022-11-03 | null | null | null | null | ['coreference-resolution', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.07267909e-02 -1.12919874e-01 -2.31484026e-01 -4.52255398e-01
-1.17352688e+00 -8.40698957e-01 7.59850383e-01 -1.94276534e-02
-1.01061225e+00 8.84591341e-01 3.47088099e-01 -8.04074764e-01
4.38752882e-02 -4.59177405e-01 -8.81057739e-01 -7.01276287e-02
2.19330013e-01 8.14227164e-01 1.80725202e-01 -6.72396779e-01
-1.89258292e-01 -6.78768307e-02 -1.13915718e+00 6.06439948e-01
9.42284703e-01 4.02117908e-01 8.18643272e-01 7.79365599e-01
-1.58698589e-01 4.62476790e-01 -3.88076991e-01 -2.87312806e-01
3.67684871e-01 -2.57690549e-02 -1.06389177e+00 -3.83443832e-01
8.25525463e-01 -2.59619001e-02 -2.35620901e-01 7.40824044e-01
6.53633356e-01 4.72564787e-01 5.02451777e-01 -8.30350697e-01
-1.04424262e+00 8.85880649e-01 -1.99430034e-01 6.21230185e-01
2.82605171e-01 4.07961518e-01 7.96716273e-01 -1.18394375e+00
7.69524395e-01 1.24339652e+00 6.69536591e-01 9.48875487e-01
-1.41300952e+00 -8.60830009e-01 6.89812824e-02 -2.30695982e-03
-1.10105491e+00 -7.15392172e-01 8.76278877e-02 -4.03086990e-01
1.72058511e+00 1.80837765e-01 -6.70791697e-03 1.62943137e+00
4.02442694e-01 5.98058224e-01 1.55948734e+00 -7.02977300e-01
-1.48561344e-01 2.98250526e-01 1.98067710e-01 5.39692998e-01
-1.91616967e-01 3.80117029e-01 -7.99493074e-01 -1.16171325e-02
4.22633737e-01 -2.73705810e-01 -2.86618441e-01 2.18673751e-01
-1.58794546e+00 7.87450194e-01 1.85343921e-01 4.53879327e-01
-1.88213170e-01 -9.09516662e-02 6.00532830e-01 9.09819722e-01
8.80108297e-01 9.65797782e-01 -1.32783842e+00 -1.45053372e-01
-7.14570701e-01 1.39923066e-01 7.88204670e-01 1.20200598e+00
1.04969299e+00 4.66232859e-02 -4.81416643e-01 1.16768014e+00
-3.36221337e-01 6.84291005e-01 1.02264953e+00 -6.93364084e-01
6.23154938e-01 2.50079483e-01 -1.06358498e-01 5.33735659e-03
-7.22146332e-01 -6.38241887e-01 -4.15489286e-01 -2.29080454e-01
5.51025867e-01 -4.69792306e-01 -9.20258403e-01 2.08083963e+00
-1.83937058e-03 -7.69599974e-02 3.17530632e-01 5.66238523e-01
8.95636618e-01 6.87544763e-01 4.72817451e-01 -1.24001905e-01
1.43134594e+00 -1.05020535e+00 -3.85718793e-01 -5.88059187e-01
1.21040809e+00 -1.07218921e+00 2.18469763e+00 3.07407379e-01
-8.17783713e-01 -8.25414121e-01 -5.88192105e-01 -4.23747897e-01
-7.05692649e-01 -1.53984949e-01 5.27040601e-01 3.96846652e-01
-1.34728730e+00 3.45267475e-01 -4.44504052e-01 -8.52925777e-01
-1.65183485e-01 1.33559033e-01 -4.07365441e-01 -3.46724898e-01
-1.64619625e+00 1.21745121e+00 5.59931815e-01 -8.20509553e-01
-1.12096226e+00 -1.04079664e+00 -8.50172520e-01 -8.48165676e-02
4.67839688e-01 -6.98773146e-01 1.88170934e+00 -7.75779843e-01
-1.11821651e+00 1.13073182e+00 -1.79183796e-01 -4.75007474e-01
3.35540146e-01 -1.76037088e-01 -5.68635404e-01 -5.13546824e-01
5.54709017e-01 1.02384734e+00 5.70114553e-01 -6.97162211e-01
-8.14634323e-01 -1.04054093e-01 1.01740018e-01 3.62423569e-01
-5.83189070e-01 2.46153474e-01 -1.43311575e-01 -7.47558177e-01
-4.60876197e-01 -8.74191165e-01 -1.64958779e-02 -7.03074038e-01
-9.73137468e-02 -6.89981580e-01 3.95478576e-01 -6.36260390e-01
9.99864697e-01 -1.95996010e+00 -2.47391596e-01 -4.87496793e-01
-8.73200223e-02 1.23550750e-01 -7.48219550e-01 4.43664342e-01
1.18224313e-02 1.69046700e-01 6.13060519e-02 -3.64423186e-01
3.61673199e-02 2.39661396e-01 -4.11511272e-01 3.45286876e-02
-4.20586858e-03 1.16991186e+00 -1.09886646e+00 -1.76931337e-01
-4.96608466e-02 7.73080215e-02 -5.23449898e-01 5.93745708e-02
-6.34572864e-01 4.32810932e-01 5.51626086e-02 4.54487294e-01
2.42002651e-01 -2.35074416e-01 -1.97523490e-01 1.21120386e-01
-1.82076722e-01 5.63974977e-01 -5.16560793e-01 2.18556929e+00
-1.00843179e+00 4.69110638e-01 -2.12130949e-01 -4.82546866e-01
8.54833841e-01 4.65412050e-01 -3.20835249e-03 -1.08100843e+00
-2.58710027e-01 6.21971488e-01 2.33978912e-01 -4.44119126e-01
7.09236383e-01 -4.90894556e-01 -3.98024559e-01 7.58799493e-01
6.33236408e-01 -1.74033031e-01 5.19052088e-01 3.18164706e-01
1.07718098e+00 1.24053545e-01 3.22161108e-01 -7.60983050e-01
9.72080082e-02 2.56128967e-01 2.42215663e-01 1.24188375e+00
-2.28939742e-01 1.47778347e-01 -2.64468938e-01 -4.56809551e-01
-1.09046626e+00 -1.09208131e+00 -3.38610411e-01 2.36548495e+00
-4.80850428e-01 -4.11583453e-01 -5.50326228e-01 -8.81841063e-01
-3.62361409e-02 1.24739397e+00 -5.52760959e-01 -1.55548409e-01
-3.71377617e-01 -5.07596612e-01 6.29095733e-01 3.99378628e-01
1.75189584e-01 -1.32441330e+00 -4.01547492e-01 3.27747315e-01
-4.22078073e-01 -1.21064651e+00 -9.65483606e-01 8.26886237e-01
-6.87572241e-01 -6.61123693e-01 -8.28101933e-01 -9.52297747e-01
2.00193554e-01 1.97988138e-01 1.52643919e+00 -3.08521897e-01
2.77869049e-02 4.56402868e-01 -4.31941807e-01 -6.41638458e-01
-7.42000103e-01 7.44653642e-01 6.01309836e-01 -5.04167676e-01
7.34607339e-01 -2.50802189e-01 2.91192811e-02 3.14952105e-01
-9.33506906e-01 2.29690596e-01 5.30091763e-01 9.25110877e-01
4.09986854e-01 -4.57832187e-01 7.86766112e-01 -1.24181855e+00
9.75557983e-01 -6.54638290e-01 -3.62390697e-01 4.52921659e-01
-6.05159998e-01 2.79949427e-01 8.00344408e-01 -8.06829453e-01
-1.22078586e+00 -2.87716925e-01 -6.39270842e-02 -2.90792644e-01
-3.70132089e-01 5.73360443e-01 3.38239074e-01 5.64594269e-02
1.41675699e+00 2.88573951e-01 -5.96235573e-01 -7.12022483e-01
6.31535649e-01 6.65983737e-01 5.59760451e-01 -1.07785213e+00
4.28134322e-01 -1.12690240e-01 -6.22703552e-01 -7.76051342e-01
-1.02290595e+00 -5.36893010e-01 -4.33549196e-01 1.35623291e-01
6.14187896e-01 -1.07740319e+00 -2.19078556e-01 2.99181372e-01
-1.18269384e+00 -1.16211987e+00 -4.59438235e-01 4.50578332e-01
-5.30834138e-01 -2.09960327e-01 -8.68702531e-01 -2.80611753e-01
-4.79320973e-01 -1.09974849e+00 1.03540003e+00 5.41273803e-02
-5.77504635e-01 -1.40920508e+00 3.47370774e-01 9.30546373e-02
7.26541400e-01 -6.75475419e-01 1.15557659e+00 -1.24201977e+00
2.43856050e-02 3.43240887e-01 -1.16425827e-01 1.58372462e-01
9.27688554e-02 -6.92027688e-01 -1.15263832e+00 -5.83640158e-01
6.24025008e-04 -9.54856813e-01 1.00244486e+00 2.42919669e-01
6.34926617e-01 -4.21008885e-01 -2.36314967e-01 7.79737115e-01
1.17464244e+00 -1.11278892e-01 -1.50515407e-01 5.92589319e-01
3.97951037e-01 6.39098465e-01 6.02016449e-01 5.25235943e-02
4.92608458e-01 5.53195715e-01 -2.52949566e-01 -1.93628632e-02
-4.22267437e-01 -4.36009496e-01 5.85193574e-01 9.92294729e-01
3.20050418e-01 -1.47358403e-01 -1.20249522e+00 7.80477405e-01
-1.50215793e+00 -5.14972866e-01 5.65948673e-02 2.00794196e+00
1.44282353e+00 1.67784929e-01 -9.41991210e-02 -7.63529241e-01
3.83625716e-01 -3.49113233e-02 -7.24044919e-01 -6.84395850e-01
-3.34505081e-01 4.59064156e-01 5.76477945e-01 9.05530691e-01
-9.82604086e-01 1.75023413e+00 5.90964890e+00 1.06082928e+00
-1.20939457e+00 7.81297684e-01 5.31069279e-01 -2.85474122e-01
-5.66337764e-01 -1.68396831e-02 -1.22371781e+00 1.70715004e-01
1.21748579e+00 -4.93702054e-01 8.18402410e-01 6.74055755e-01
1.11320280e-02 3.34634632e-02 -1.28831947e+00 6.78093314e-01
-9.55536962e-03 -9.83965755e-01 2.34034494e-01 -1.87432081e-01
1.17606127e+00 8.96125853e-01 6.94805533e-02 1.17923713e+00
8.03153217e-01 -9.27682757e-01 6.83225989e-01 3.38272005e-01
1.35227203e+00 -3.76824617e-01 3.39307994e-01 8.53578985e-01
-7.42610633e-01 9.16457921e-03 -6.75582886e-01 -3.09209168e-01
-3.26289833e-02 2.38880798e-01 -1.23016930e+00 2.27592364e-01
8.96005332e-01 3.28127623e-01 -8.29642832e-01 5.42727590e-01
-2.73111075e-01 6.32419109e-01 -1.81841522e-01 2.38326564e-02
3.88993084e-01 3.51963997e-01 5.51199973e-01 1.52771449e+00
2.92740941e-01 -2.25591138e-01 4.94598657e-01 5.14437914e-01
-3.10709417e-01 4.64926541e-01 -8.66258681e-01 -7.35902712e-02
4.54339504e-01 1.21541834e+00 -2.34827384e-01 -5.84163129e-01
-6.58089936e-01 8.75745714e-01 7.97235727e-01 7.07852960e-01
-3.02841067e-01 -1.61497355e-01 4.81258929e-01 2.26952627e-01
-4.29949552e-01 -2.00635552e-01 -1.91749468e-01 -1.34684932e+00
-3.47163171e-01 -1.25578749e+00 5.52846313e-01 -8.52274477e-01
-1.52885318e+00 8.22607756e-01 2.94027459e-02 -6.36690736e-01
-5.61895609e-01 -7.65702486e-01 -3.48355234e-01 1.27324545e+00
-1.41623890e+00 -1.18624043e+00 3.97064462e-02 8.88129950e-01
1.29801250e+00 -5.63629448e-01 1.15479243e+00 1.67028382e-01
-1.82476193e-01 7.64784694e-01 8.55575949e-02 -1.20539322e-01
1.59583211e+00 -1.22938406e+00 8.96495461e-01 8.46624255e-01
3.63502175e-01 8.11440468e-01 7.61917830e-01 -8.61398876e-01
-9.22336161e-01 -1.14541936e+00 1.56302130e+00 -8.34002495e-01
9.93049562e-01 -5.29883265e-01 -9.47476983e-01 1.25278831e+00
5.82404554e-01 -2.37491786e-01 4.98806626e-01 6.27700508e-01
-5.48864901e-01 7.62753189e-02 -6.64830208e-01 6.59550309e-01
1.22726762e+00 -8.58541548e-01 -7.81156838e-01 7.87967026e-01
1.20611298e+00 -3.50018740e-01 -6.71950161e-01 2.91942567e-01
1.97416708e-01 -4.70759451e-01 6.16521001e-01 -1.09343112e+00
2.37194851e-01 2.64198929e-01 -3.74955386e-01 -1.99993932e+00
-5.94874442e-01 -5.06809950e-01 2.48797685e-01 9.64693606e-01
8.28151345e-01 -9.24590051e-01 1.05884574e-01 3.13554376e-01
-4.63231862e-01 -4.60429311e-01 -8.24774802e-01 -1.20398235e+00
6.13082409e-01 -6.52146518e-01 3.15418392e-01 1.29633510e+00
8.41439217e-02 9.17388797e-01 -2.66909391e-01 -3.80066991e-01
3.98012817e-01 -3.08296412e-01 6.79223776e-01 -1.01467812e+00
-5.57323813e-01 -3.87064338e-01 5.74154735e-01 -9.43391323e-01
5.05881786e-01 -1.54300523e+00 1.51748598e-01 -1.27587497e+00
4.91642088e-01 -5.90903997e-01 -2.99426883e-01 9.57825243e-01
-4.40037459e-01 1.23925917e-01 2.04774395e-01 2.95702219e-01
-5.49573839e-01 2.42029324e-01 1.30255437e+00 -2.23813236e-01
-1.81014448e-01 -1.71661556e-01 -8.20474803e-01 5.03778756e-01
9.39794838e-01 -6.38310790e-01 -4.63658988e-01 -9.77762938e-01
2.98877478e-01 -2.74068028e-01 -6.90074861e-02 -8.51212859e-01
3.22497606e-01 -1.71299279e-01 2.42852941e-01 -1.16741098e-03
8.86374936e-02 -2.52129614e-01 -2.96154529e-01 3.87108862e-01
-6.05105639e-01 4.21022058e-01 6.42601132e-01 4.25915048e-02
1.26512572e-01 -2.30630919e-01 6.47797883e-01 -5.68891048e-01
-1.07425761e+00 3.39166284e-01 -2.84544885e-01 7.01582849e-01
6.29714847e-01 4.18525130e-01 -7.41159022e-01 -2.29171291e-01
-7.69059539e-01 3.10322016e-01 4.67939019e-01 9.24717724e-01
2.23292872e-01 -1.12771583e+00 -1.18254125e+00 2.90003985e-01
5.14931560e-01 -3.42770100e-01 2.59901732e-01 8.34534943e-01
-2.88427733e-02 8.65808785e-01 -1.31799802e-01 -4.95874494e-01
-9.44736838e-01 7.73180664e-01 1.70096070e-01 -4.91300553e-01
-1.94076791e-01 1.13107061e+00 5.94981790e-01 -1.28789735e+00
7.37596005e-02 -3.31347793e-01 1.97883472e-01 9.43097007e-03
4.26364303e-01 1.97393641e-01 2.46597722e-01 -3.26393932e-01
-1.93433538e-01 2.16474488e-01 -4.73003000e-01 -4.78867143e-01
9.55842912e-01 -6.22459315e-02 2.61041105e-01 8.08686614e-01
1.01047659e+00 7.22680613e-02 -1.05690277e+00 -8.96056533e-01
2.24458829e-01 -2.59215315e-03 -1.38087943e-01 -1.38850391e+00
-3.41430664e-01 9.63213801e-01 4.23356891e-01 -1.68113664e-01
7.92604387e-01 3.46055448e-01 8.28795433e-01 7.21596003e-01
6.97307825e-01 -1.20712852e+00 5.60452119e-02 1.29371691e+00
8.46937120e-01 -1.49661803e+00 -6.54418528e-01 2.92445123e-01
-6.78116560e-01 7.74379432e-01 1.01086366e+00 2.61122137e-01
3.13622534e-01 2.83291936e-01 4.01972055e-01 1.47632807e-01
-1.22104514e+00 -2.08663017e-01 3.76672477e-01 6.50154054e-01
8.65454257e-01 4.03834552e-01 -1.59711838e-01 8.39000702e-01
-5.65336823e-01 -1.46176992e-02 1.46410182e-01 6.09207213e-01
-6.04274750e-01 -1.12793565e+00 -4.09502119e-01 4.78774309e-01
-3.33605319e-01 -9.65274572e-01 -2.91995943e-01 7.36392975e-01
1.31282851e-01 8.69052768e-01 -1.21968441e-01 -2.58512557e-01
4.85195488e-01 6.97460711e-01 5.21107078e-01 -1.26229465e+00
-8.82997870e-01 -2.09421720e-02 2.96331614e-01 -4.11169589e-01
3.51125181e-01 -5.35268426e-01 -9.92958009e-01 -2.36076713e-01
1.09181985e-01 7.49705508e-02 3.92041057e-01 1.01429796e+00
2.36064881e-01 5.00446141e-01 5.83166964e-02 -6.53334022e-01
-9.12553608e-01 -1.62205923e+00 -3.46432745e-01 4.58591223e-01
2.62139082e-01 -3.87008250e-01 -9.87612158e-02 -1.19425217e-02] | [11.067535400390625, 9.590680122375488] |
e63f114d-f01d-4690-90bb-697bad77a2f8 | taco-temporal-latent-action-driven | 2306.13229 | null | https://arxiv.org/abs/2306.13229v1 | https://arxiv.org/pdf/2306.13229v1.pdf | TACO: Temporal Latent Action-Driven Contrastive Loss for Visual Reinforcement Learning | Despite recent progress in reinforcement learning (RL) from raw pixel data, sample inefficiency continues to present a substantial obstacle. Prior works have attempted to address this challenge by creating self-supervised auxiliary tasks, aiming to enrich the agent's learned representations with control-relevant information for future state prediction. However, these objectives are often insufficient to learn representations that can represent the optimal policy or value function, and they often consider tasks with small, abstract discrete action spaces and thus overlook the importance of action representation learning in continuous control. In this paper, we introduce TACO: Temporal Action-driven Contrastive Learning, a simple yet powerful temporal contrastive learning approach that facilitates the concurrent acquisition of latent state and action representations for agents. TACO simultaneously learns a state and an action representation by optimizing the mutual information between representations of current states paired with action sequences and representations of the corresponding future states. Theoretically, TACO can be shown to learn state and action representations that encompass sufficient information for control, thereby improving sample efficiency. For online RL, TACO achieves 40% performance boost after one million environment interaction steps on average across nine challenging visual continuous control tasks from Deepmind Control Suite. In addition, we show that TACO can also serve as a plug-and-play module adding to existing offline visual RL methods to establish the new state-of-the-art performance for offline visual RL across offline datasets with varying quality. | ['Furong Huang', 'Hal Daumé III', 'Huazhe Xu', 'Jieyu Zhao', 'Shuang Ma', 'Yanchao Sun', 'Xiyao Wang', 'Ruijie Zheng'] | 2023-06-22 | null | null | null | null | ['contrastive-learning', 'contrastive-learning', 'continuous-control'] | ['computer-vision', 'methodology', 'playing-games'] | [ 2.74011999e-01 2.36627180e-02 -5.97806931e-01 -9.54457093e-03
-7.98806787e-01 -5.92126846e-01 1.10741723e+00 5.32485470e-02
-5.11850893e-01 9.00837958e-01 3.62692118e-01 -3.01578879e-01
3.02167907e-02 -3.44072878e-01 -7.56964862e-01 -8.30771506e-01
-4.45591748e-01 3.99001926e-01 1.33667346e-02 -1.43855065e-01
1.58242106e-01 5.52316725e-01 -1.69098139e+00 1.55298963e-01
4.28302526e-01 8.02278519e-01 3.66365999e-01 7.98255086e-01
9.84972343e-02 1.36705101e+00 -5.65818310e-01 6.63461030e-01
4.08207953e-01 -6.76317275e-01 -7.11290896e-01 1.75018132e-01
2.82111704e-01 -6.19254172e-01 -6.43369496e-01 5.67321539e-01
5.67458689e-01 5.50955474e-01 4.53265846e-01 -1.40619969e+00
-4.93265271e-01 2.88512200e-01 -4.68065977e-01 1.91812649e-01
2.46122926e-01 1.25252616e+00 1.04511201e+00 -2.35957801e-01
8.82691085e-01 1.38890111e+00 1.52874768e-01 1.01068437e+00
-1.82134140e+00 -4.98958886e-01 7.31710970e-01 1.83495134e-01
-5.69289327e-01 -6.20594144e-01 5.97965717e-01 -3.25754583e-01
1.52209771e+00 -1.21634409e-01 1.13080800e+00 1.55847692e+00
8.57590884e-02 1.28780675e+00 1.37494886e+00 -1.14852600e-01
5.52370250e-01 -3.49078745e-01 -2.90604234e-01 8.59815061e-01
-3.04026157e-01 8.99054468e-01 -6.25114083e-01 4.03923355e-02
1.01383257e+00 -8.45000595e-02 -7.19335079e-02 -8.99573028e-01
-1.40087616e+00 7.24195123e-01 5.48120379e-01 -1.86696753e-01
-4.26892161e-01 9.42901194e-01 5.76639056e-01 4.70532298e-01
1.28319323e-01 7.57145703e-01 -5.50464988e-01 -4.92369980e-01
-3.66905332e-01 6.03138208e-01 2.58170187e-01 7.31537938e-01
6.48514092e-01 6.26885772e-01 -6.66799545e-01 3.38125229e-01
2.04574078e-01 5.90527475e-01 4.72768068e-01 -1.45776093e+00
3.36872220e-01 3.69883090e-01 4.36235160e-01 -2.93016225e-01
-3.85087729e-01 -2.19866022e-01 -4.45999026e-01 1.11014414e+00
4.61967736e-01 -3.27542275e-01 -1.25488770e+00 2.12224650e+00
3.03883165e-01 1.89918682e-01 2.43868962e-01 7.36438870e-01
5.41868396e-02 8.61427546e-01 3.99437636e-01 -3.13517302e-01
8.01368058e-01 -1.30322146e+00 -6.57427251e-01 -7.55195379e-01
4.87956554e-01 -1.31825343e-01 1.33217621e+00 2.64671177e-01
-1.11532545e+00 -5.49273789e-01 -9.78941798e-01 1.16872415e-01
-1.84338719e-01 8.99485201e-02 9.86822069e-01 -6.01992980e-02
-9.85106230e-01 8.11933637e-01 -1.33949602e+00 -1.11823328e-01
5.89571416e-01 3.43012571e-01 -3.12842906e-01 2.95351446e-01
-7.70520747e-01 1.08905005e+00 3.49419355e-01 -1.69597313e-01
-1.84492052e+00 -7.92370260e-01 -9.33820844e-01 -4.13833465e-03
8.55369270e-01 -5.40095270e-01 1.64970386e+00 -1.17064738e+00
-2.00058746e+00 3.00173879e-01 6.86805323e-03 -8.33576202e-01
5.06081760e-01 -2.79525429e-01 7.57703185e-02 1.35711238e-01
6.35871738e-02 1.01077795e+00 1.17564976e+00 -1.19018507e+00
-6.16262138e-01 -2.58132041e-01 3.60921264e-01 5.45804977e-01
2.31354848e-01 -4.25996691e-01 -2.88020194e-01 -5.12069821e-01
-4.79358643e-01 -9.34675753e-01 -6.75857544e-01 5.80087721e-01
8.88221189e-02 -2.90929258e-01 7.74278164e-01 -3.75481397e-01
8.59823346e-01 -2.08264995e+00 5.26958108e-01 -1.99056491e-01
1.78736299e-01 3.59642625e-01 -4.87595767e-01 4.01563883e-01
-6.26995116e-02 -3.01309586e-01 -2.16462806e-01 -4.24449235e-01
2.41768360e-01 4.17060167e-01 -5.14847636e-01 5.71474493e-01
3.60871136e-01 1.26631451e+00 -1.27461004e+00 -2.47010350e-01
6.87503159e-01 3.50946248e-01 -4.45428938e-01 4.89136755e-01
-8.78407598e-01 8.49926949e-01 -5.03820240e-01 2.28101447e-01
-9.39874128e-02 -2.09885016e-01 4.70257312e-01 3.27670932e-01
-1.60847336e-01 3.50385725e-01 -9.24009502e-01 1.84208882e+00
-5.43049872e-01 6.88123941e-01 1.38085619e-01 -1.00947821e+00
6.13633871e-01 2.42050484e-01 7.51905441e-01 -1.16804135e+00
-5.59471361e-02 -3.27771366e-01 -2.29139045e-01 -3.02584201e-01
1.77931041e-01 -9.75817293e-02 -7.95281306e-02 5.70909321e-01
1.69525340e-01 -3.11622709e-01 1.68430939e-01 1.20307975e-01
1.14861846e+00 1.04165518e+00 5.24529099e-01 1.64523005e-01
4.63764630e-02 1.92877844e-01 3.73554140e-01 7.44551778e-01
-5.06946683e-01 4.75542918e-02 7.64910400e-01 -4.45223838e-01
-1.09603345e+00 -1.24175000e+00 5.03890455e-01 1.20981395e+00
-1.40859917e-01 -3.36375803e-01 -2.69716591e-01 -8.37099850e-01
5.07727750e-02 7.69787669e-01 -9.04420853e-01 -4.52522427e-01
-6.79445267e-01 -1.99648097e-01 2.49157116e-01 7.64042437e-01
3.36983144e-01 -1.59468305e+00 -1.34924269e+00 3.27144504e-01
3.52914035e-01 -7.40887284e-01 -4.17310923e-01 6.30091548e-01
-8.70231092e-01 -1.15623796e+00 -5.88831067e-01 -4.89273667e-01
4.59640175e-01 6.27867803e-02 1.12613916e+00 -2.09960073e-01
-3.80504519e-01 8.26936603e-01 -5.36330342e-02 -2.11837515e-02
-4.39054221e-01 -3.02869827e-01 2.96880715e-02 -3.80370915e-01
-1.87824637e-01 -5.25283813e-01 -6.28769040e-01 -1.45267621e-01
-6.85769796e-01 3.21181417e-01 4.98372316e-01 1.06700420e+00
6.15457892e-01 -4.98355478e-01 4.51545984e-01 -3.71378213e-01
5.60612082e-01 -2.35459998e-01 -1.06560981e+00 1.49770886e-01
-7.91867673e-01 7.91659176e-01 8.05457950e-01 -6.74888015e-01
-1.11280465e+00 4.21326458e-01 2.38905638e-01 -7.26066232e-01
-1.22425027e-01 1.96209192e-01 3.26439053e-01 2.91504771e-01
7.30691135e-01 4.64478165e-01 4.69160080e-01 -2.64620095e-01
8.31578434e-01 -5.13712987e-02 4.57358688e-01 -6.87922895e-01
5.40294826e-01 4.08543825e-01 1.76812023e-01 -4.37492430e-01
-7.41142690e-01 -3.12455326e-01 -3.25589806e-01 -2.94963598e-01
7.73559034e-01 -8.86960924e-01 -1.24138594e+00 2.79482037e-01
-5.78060806e-01 -1.30305040e+00 -8.62344921e-01 3.46232355e-01
-1.23550117e+00 1.31135046e-01 -3.81698519e-01 -1.05071902e+00
-2.03679856e-02 -1.20914972e+00 1.04627049e+00 1.22616656e-01
-1.30004093e-01 -9.41008031e-01 3.99391204e-01 -8.37459639e-02
3.33893925e-01 3.98248255e-01 7.31068909e-01 -5.86411320e-02
-5.96273959e-01 2.87431628e-01 1.04503430e-01 2.51109689e-01
1.58900186e-01 -9.31161419e-02 -8.05446148e-01 -7.41702020e-01
-5.00079274e-01 -1.01919556e+00 9.83031571e-01 4.99653429e-01
1.04086328e+00 -3.91816229e-01 -1.92816958e-01 4.06062126e-01
1.20251274e+00 4.08139855e-01 4.80367333e-01 2.80397415e-01
4.69094336e-01 3.18100274e-01 7.74850905e-01 7.21548378e-01
1.08366258e-01 8.28521430e-01 6.29247606e-01 -7.84856733e-03
-2.88031250e-01 -7.11390376e-01 9.32081938e-01 -2.97387056e-02
6.56334171e-03 7.12739304e-02 -6.60269797e-01 5.20327032e-01
-2.06320000e+00 -1.16485095e+00 6.92760766e-01 2.14756060e+00
9.68166351e-01 6.84655756e-02 3.56282830e-01 -2.80465961e-01
1.57481909e-01 5.82007766e-01 -1.31935501e+00 -1.94644377e-01
2.58819431e-01 4.53144193e-01 3.70620131e-01 6.13116026e-01
-1.20361006e+00 1.19145310e+00 6.62477112e+00 6.76871479e-01
-1.20036817e+00 -1.81893349e-01 4.56392944e-01 -5.01288652e-01
-8.74690339e-02 -6.06701151e-03 -4.43778902e-01 1.05904803e-01
1.03509223e+00 -6.34008646e-02 9.80629563e-01 9.54709351e-01
3.99275243e-01 -2.22267479e-01 -1.23685634e+00 8.38729978e-01
-3.30840051e-01 -1.46753120e+00 -1.38238162e-01 1.19586423e-01
8.45117986e-01 4.72779572e-02 4.21167463e-01 8.56595814e-01
8.84218574e-01 -1.19155729e+00 8.09088409e-01 4.84784544e-01
8.46772134e-01 -7.02096283e-01 -4.74296398e-02 3.29886973e-01
-1.20747852e+00 -4.78855103e-01 -6.63977116e-02 -3.37840408e-01
3.87447514e-02 -5.26468635e-01 -7.44014263e-01 -5.63125126e-02
4.35418755e-01 1.09275615e+00 -4.78520691e-01 5.48126221e-01
-4.33230430e-01 3.63424778e-01 -3.76379527e-02 -2.11600631e-01
8.53514612e-01 -6.40664250e-02 4.53461975e-01 6.00316644e-01
-2.02700362e-01 -2.31528699e-01 4.62125868e-01 8.62153292e-01
2.43645534e-01 -4.39996719e-01 -8.52420211e-01 -4.72262502e-01
1.14298403e-01 9.13676322e-01 -5.00619888e-01 -4.70774204e-01
-1.46884248e-01 1.02260292e+00 7.69403219e-01 5.91004074e-01
-9.35177147e-01 2.64176399e-01 1.12334514e+00 -2.49375671e-01
4.85807836e-01 -6.01113856e-01 1.78943187e-01 -1.05415010e+00
-2.99774200e-01 -1.13669145e+00 1.61227629e-01 -6.70264602e-01
-8.22370946e-01 9.15688872e-02 1.39406189e-01 -1.30259371e+00
-8.32075298e-01 -6.60469294e-01 -3.09310734e-01 4.62808102e-01
-1.30791020e+00 -1.04735839e+00 5.00161648e-02 7.23353684e-01
9.03474391e-01 -1.98477119e-01 8.93541157e-01 -3.63262266e-01
-4.50225443e-01 1.56618401e-01 2.16542393e-01 -2.57893592e-01
4.82260317e-01 -1.56548524e+00 5.01515508e-01 5.73536992e-01
1.97801918e-01 3.66345912e-01 6.90415978e-01 -4.89270091e-01
-1.74557745e+00 -9.88169014e-01 -5.98390736e-02 -4.05187607e-01
6.85826540e-01 -2.70253450e-01 -4.75518346e-01 9.55810249e-01
3.21605772e-01 2.62254566e-01 2.39648391e-02 -1.93009134e-02
-3.18171024e-01 2.77932212e-02 -8.14888537e-01 1.09481788e+00
1.13849664e+00 -6.59554243e-01 -4.44360465e-01 2.27192059e-01
7.16051996e-01 -3.36322486e-01 -4.20667619e-01 -4.04271670e-03
5.49167216e-01 -8.43353927e-01 1.17641342e+00 -1.16173148e+00
2.97405273e-01 -3.90728235e-01 4.66422848e-02 -1.49382699e+00
-3.02170008e-01 -9.28409815e-01 -5.66201866e-01 5.13535500e-01
1.78432330e-01 -3.90969455e-01 7.59494483e-01 3.81674767e-01
-7.57690370e-02 -8.32784295e-01 -9.04451787e-01 -8.15116405e-01
5.72854020e-02 -2.84077734e-01 1.97639510e-01 4.91683483e-01
-2.96916831e-02 4.87868398e-01 -6.73215628e-01 -1.79458171e-01
5.04116058e-01 1.74196556e-01 9.09338593e-01 -6.21309161e-01
-5.92255235e-01 -5.36401629e-01 -3.61306481e-02 -1.37713325e+00
5.74891508e-01 -5.79002619e-01 3.14646691e-01 -1.54872715e+00
-8.50022247e-04 -1.39694780e-01 -3.17698300e-01 8.53573442e-01
-8.99755731e-02 -2.83655256e-01 5.42941153e-01 -8.27350318e-02
-7.73257136e-01 1.17019796e+00 1.56609094e+00 -4.66470450e-01
-4.30519968e-01 -1.74669445e-01 -2.69924402e-01 4.23389167e-01
8.06021690e-01 -1.00754552e-01 -9.12096143e-01 -1.90799519e-01
-5.34309298e-02 3.52650374e-01 6.73730373e-01 -1.03208065e+00
-6.89550117e-02 -6.63377583e-01 3.66869509e-01 -1.66964203e-01
6.96272016e-01 -7.79354036e-01 -1.22031912e-01 8.63225937e-01
-8.77174616e-01 2.98228636e-02 3.73654544e-01 9.91488814e-01
1.35115877e-01 1.79219335e-01 7.77225852e-01 -3.98196995e-01
-1.16962349e+00 4.20353442e-01 -7.01524019e-01 6.59852549e-02
1.19378233e+00 1.01172119e-01 -3.36742312e-01 -5.59641778e-01
-9.11807001e-01 4.55476999e-01 3.21989745e-01 4.30177331e-01
6.66163862e-01 -1.24903464e+00 -3.31293941e-01 1.49880260e-01
9.10532251e-02 -3.07923228e-01 2.47843355e-01 4.63511437e-01
-1.47717908e-01 2.70434499e-01 -6.37384295e-01 -4.66267675e-01
-1.02246201e+00 9.78831887e-01 4.10441756e-01 -6.02787375e-01
-9.81037080e-01 4.54764515e-01 1.73097536e-01 -3.48927051e-01
4.14071143e-01 -4.40412611e-01 -1.22959018e-01 -1.99696999e-02
4.74478394e-01 2.21418425e-01 -6.69425070e-01 -1.67426482e-01
-1.88521426e-02 2.54116774e-01 1.75045520e-01 -5.19554079e-01
1.39931881e+00 5.66596631e-03 4.79934484e-01 6.37480497e-01
9.68777120e-01 -5.76471925e-01 -2.49945617e+00 -1.08427472e-01
-1.88596517e-01 -2.81176418e-01 1.38592273e-01 -9.54122901e-01
-9.90778565e-01 1.00318515e+00 8.49729300e-01 -1.67757273e-01
9.90746319e-01 -9.61431190e-02 2.53662169e-01 7.11552918e-01
4.59542751e-01 -1.22016335e+00 7.26176679e-01 5.51579654e-01
1.09598756e+00 -1.31673074e+00 -1.55281052e-02 5.06723583e-01
-1.08440244e+00 8.55041921e-01 8.46266508e-01 -2.47184351e-01
1.32887632e-01 2.01971784e-01 -9.69440490e-02 -1.28072619e-01
-1.40867567e+00 -6.36186481e-01 6.29211515e-02 8.74218524e-01
7.07431585e-02 6.83444887e-02 4.00507957e-01 -1.42946631e-01
3.41467530e-01 -8.70596245e-02 1.18938230e-01 1.21539426e+00
-3.84582102e-01 -1.08037496e+00 1.32921234e-01 2.60521203e-01
6.92537725e-02 1.27632141e-01 -6.61305487e-02 8.81515622e-01
-4.03287441e-01 6.04865432e-01 1.57624841e-01 -5.04674986e-02
1.68443665e-01 -3.97025049e-03 8.47581029e-01 -5.69244683e-01
-5.52939057e-01 4.10216264e-02 8.30287412e-02 -1.25497520e+00
-2.91800261e-01 -6.77474201e-01 -1.41966581e+00 -3.09789963e-02
3.75483513e-01 -2.57131934e-01 2.67901361e-01 8.09501112e-01
4.16312307e-01 8.60166669e-01 5.24205148e-01 -1.17783427e+00
-9.71917748e-01 -7.79321432e-01 -2.12819636e-01 3.69434983e-01
8.73501778e-01 -9.96052861e-01 -1.37396127e-01 -3.02924179e-02] | [4.245301246643066, 1.507858157157898] |
0e6892e1-436c-4f4b-8eb3-5a84c7814405 | faster-stochastic-first-order-method-for | 2211.12880 | null | https://arxiv.org/abs/2211.12880v1 | https://arxiv.org/pdf/2211.12880v1.pdf | Faster Stochastic First-Order Method for Maximum-Likelihood Quantum State Tomography | In maximum-likelihood quantum state tomography, both the sample size and dimension grow exponentially with the number of qubits. It is therefore desirable to develop a stochastic first-order method, just like stochastic gradient descent for modern machine learning, to compute the maximum-likelihood estimate. To this end, we propose an algorithm called stochastic mirror descent with the Burg entropy. Its expected optimization error vanishes at a $O ( \sqrt{ ( 1 / t ) d \log t } )$ rate, where $d$ and $t$ denote the dimension and number of iterations, respectively. Its per-iteration time complexity is $O ( d^3 )$, independent of the sample size. To the best of our knowledge, this is currently the computationally fastest stochastic first-order method for maximum-likelihood quantum state tomography. | ['Yen-Huan Li', 'Hao-Chung Cheng', 'Chung-En Tsai'] | 2022-11-23 | null | null | null | null | ['quantum-state-tomography'] | ['medical'] | [ 2.18614265e-02 2.91113257e-02 1.20206647e-01 -3.56395394e-01
-1.20099092e+00 -3.76920730e-01 2.22848818e-01 8.45718384e-02
-9.78985071e-01 9.18759823e-01 -4.93258506e-01 -7.99597442e-01
5.33418022e-02 -7.56784379e-01 -5.07137418e-01 -8.42989802e-01
-3.60141903e-01 6.93175018e-01 7.15750530e-02 -5.67648187e-02
5.60225248e-01 3.55652988e-01 -9.41319823e-01 -6.96898103e-01
1.02299047e+00 1.03670347e+00 -1.44301221e-01 9.03364718e-01
6.48693964e-02 5.51017344e-01 -9.39260945e-02 -6.02172256e-01
2.36595422e-01 -1.04017138e+00 -9.78541672e-01 -2.89026588e-01
-8.31157193e-02 -3.88609111e-01 -7.25762308e-01 1.52572846e+00
4.54669386e-01 1.32724836e-01 5.38403749e-01 -4.61561203e-01
-2.61845328e-02 5.37931800e-01 -3.77138883e-01 2.57990479e-01
2.33398184e-01 5.14473431e-02 1.02901649e+00 -6.75022721e-01
6.55765593e-01 8.67132425e-01 2.57579327e-01 4.69991028e-01
-1.37506974e+00 -8.37164342e-01 -4.77528453e-01 7.96366856e-02
-1.65994167e+00 -4.43588436e-01 8.28654230e-01 -3.09654683e-01
9.40796077e-01 2.93881651e-02 6.88157618e-01 3.60659510e-01
1.03777967e-01 5.45823395e-01 1.44054294e+00 -7.71147609e-01
6.84058309e-01 2.99011767e-01 -1.10268891e-02 1.24840760e+00
3.83106232e-01 1.65771231e-01 -6.57047868e-01 -3.26158822e-01
7.27335334e-01 -4.61846530e-01 -1.92039624e-01 -3.84008676e-01
-1.11824441e+00 1.12246192e+00 -1.30820572e-02 2.63720065e-01
-2.57824808e-01 8.05267155e-01 2.80876011e-01 5.12502253e-01
4.39383149e-01 4.02941048e-01 -3.20324928e-01 -5.94448864e-01
-1.00976479e+00 6.67833090e-02 1.04394686e+00 6.61980271e-01
1.23425329e+00 -3.84508222e-02 5.07731616e-01 1.70105383e-01
3.66854966e-01 1.11688948e+00 -3.44422832e-02 -9.67499137e-01
4.15489435e-01 1.10638484e-01 5.14488459e-01 -2.79211253e-01
-2.20720261e-01 -4.31841671e-01 -9.46901798e-01 1.51526928e-01
7.36563146e-01 -5.33602834e-01 -5.74032187e-01 1.66338038e+00
3.59636009e-01 -8.93230289e-02 -4.02064025e-02 8.46939921e-01
-6.39167279e-02 4.63307261e-01 -5.40662825e-01 -5.46513736e-01
1.07710969e+00 -2.71228313e-01 -6.71870887e-01 -2.68200040e-01
9.30537760e-01 -7.48670518e-01 6.76726222e-01 3.82465512e-01
-1.21616483e+00 1.90003842e-01 -1.19613802e+00 1.56094879e-01
1.65974826e-01 -6.74860403e-02 9.66544688e-01 1.30963433e+00
-8.86663735e-01 9.99258876e-01 -1.09274864e+00 9.84519199e-02
2.44700443e-02 5.48907816e-01 -1.07864723e-01 -3.74579579e-02
-9.88423049e-01 7.82795250e-01 3.07130337e-01 1.68955147e-01
-7.57928789e-01 -1.76338609e-02 -5.93782365e-01 -1.63135111e-01
3.06527823e-01 -5.81595242e-01 1.18779492e+00 3.56321409e-02
-2.01126456e+00 8.50208402e-01 -6.65730417e-01 -6.92617536e-01
1.94628760e-01 1.49573237e-01 6.25660196e-02 2.73876876e-01
-5.72356023e-02 1.39123797e-01 6.56895876e-01 -5.22198319e-01
-2.88696647e-01 -6.58286691e-01 -1.52560160e-01 -1.96839288e-01
1.11855576e-02 -4.34086323e-02 -2.87648104e-02 5.44576228e-01
9.07324135e-01 -1.35976148e+00 -5.02021909e-01 -1.84081912e-01
-3.56045216e-01 -9.31915420e-04 4.40889336e-02 -2.94128150e-01
9.80267942e-01 -1.96624279e+00 2.83694357e-01 3.84438753e-01
2.36983061e-01 1.18964188e-01 3.66720885e-01 5.51626384e-01
3.31936985e-01 2.60340758e-02 -4.69828814e-01 -4.53406870e-01
1.36603162e-01 -1.54002041e-01 -6.90795109e-02 9.50214684e-01
-6.85396269e-02 6.80025935e-01 -9.87872839e-01 -2.50743747e-01
5.01301363e-02 1.95831299e-01 -6.10896230e-01 -2.31400311e-01
1.14927851e-01 5.30648232e-01 -7.58672833e-01 3.53258729e-01
8.09932888e-01 -5.97452641e-01 4.06636506e-01 2.51211748e-02
-3.75236452e-01 8.61433029e-01 -1.45356321e+00 1.87797081e+00
-2.80454904e-01 5.49244702e-01 1.64244205e-01 -1.02011561e+00
7.84986913e-01 3.68469179e-01 2.84706831e-01 -5.97699821e-01
4.18538719e-01 1.04115558e+00 4.10950705e-02 -2.81665444e-01
4.61265802e-01 -8.16090822e-01 -3.40426743e-01 8.97632182e-01
-3.66753675e-02 -5.49257100e-01 4.08217877e-01 3.43992084e-01
1.12895799e+00 -3.18652093e-01 3.26281577e-01 -3.25915933e-01
3.23217958e-01 -1.63931653e-01 4.24955994e-01 1.02989721e+00
-4.20824945e-01 -4.56829257e-02 7.95356810e-01 -2.69862980e-01
-1.23815584e+00 -1.08276808e+00 -2.38350600e-01 5.08492053e-01
3.71264488e-01 -4.60182697e-01 -8.61435235e-01 -2.01093957e-01
-2.44085282e-01 6.85152888e-01 -1.55109793e-01 -2.17438117e-01
-5.64529300e-01 -9.73677576e-01 6.05233729e-01 -1.33874968e-01
7.66050100e-01 -7.26525486e-01 -4.86386865e-01 4.55196530e-01
6.49891794e-02 -1.02526712e+00 -2.00954646e-01 5.75700521e-01
-1.18020773e+00 -7.15217471e-01 -4.36004132e-01 -4.24409956e-01
6.44162238e-01 -1.18917592e-01 7.17302501e-01 -3.32657188e-01
-2.33355358e-01 -1.27728088e-02 -2.08516702e-01 -2.37328652e-02
-4.99541789e-01 -1.21806543e-02 2.90807664e-01 -2.99238980e-01
5.13380706e-01 -7.21902728e-01 -7.51230419e-01 -3.76774579e-01
-5.76344728e-01 -2.42093816e-01 6.05425596e-01 9.87048388e-01
5.93984485e-01 -1.44403912e-02 -2.96268463e-01 -7.11895704e-01
5.61731517e-01 -8.04131571e-03 -1.22380233e+00 -3.85365412e-02
-7.22643554e-01 7.08577991e-01 6.98796868e-01 -1.64878760e-02
-6.17503464e-01 1.46666706e-01 -2.54923165e-01 2.69582391e-01
1.90409452e-01 8.38879228e-01 6.10747039e-01 -5.13396740e-01
6.56375647e-01 7.28272021e-01 -2.01319233e-01 -5.02647698e-01
3.18014771e-01 6.13739371e-01 1.51466772e-01 -5.34917116e-01
9.10968006e-01 7.48366594e-01 6.68810010e-01 -7.04056263e-01
-9.36710000e-01 -2.86568224e-01 -5.99802971e-01 2.23969951e-01
6.15550816e-01 -6.53944910e-01 -1.07772374e+00 4.79831576e-01
-1.06384146e+00 -1.82074264e-01 -1.88667804e-01 1.11749446e+00
-6.44875050e-01 6.85654938e-01 -7.10092366e-01 -1.53413343e+00
-5.26112735e-01 -1.22500789e+00 8.71218562e-01 8.19740295e-02
3.07126701e-01 -7.03196824e-01 7.36546516e-02 1.93406835e-01
3.98076683e-01 -1.88802391e-01 6.29353285e-01 -9.74042043e-02
-1.06637275e+00 -4.72576708e-01 -2.26063550e-01 2.44324252e-01
-2.49366075e-01 -5.54508209e-01 -6.11257970e-01 -4.40352291e-01
5.71672559e-01 -4.21819001e-01 8.97878230e-01 2.02500850e-01
5.93267083e-01 -1.87851369e-01 -6.03646077e-02 6.59746945e-01
1.53741729e+00 7.33302087e-02 5.99493742e-01 -4.11798283e-02
2.17908621e-01 -6.52910993e-02 5.57651699e-01 7.42928505e-01
8.94354731e-02 2.60132074e-01 1.34561121e-01 3.46383065e-01
3.67883533e-01 -1.79405779e-01 3.08039337e-01 1.18785858e+00
-8.50980505e-02 -1.49971694e-02 -8.52406561e-01 3.00108761e-01
-1.33748281e+00 -9.49272275e-01 -2.24958003e-01 2.63006115e+00
1.01842809e+00 3.94392371e-01 -1.89488396e-01 1.29151374e-01
4.08652216e-01 7.48740062e-02 -7.08628237e-01 -4.81061846e-01
7.89996237e-02 7.92378485e-01 9.27775383e-01 7.34122574e-01
-7.33628809e-01 1.02571440e+00 6.03888416e+00 9.64972019e-01
-1.10792446e+00 3.12232703e-01 4.28474426e-01 -2.69311577e-01
-1.96575776e-01 6.65222704e-01 -8.06429565e-01 5.56263626e-01
1.23436272e+00 -6.48945943e-02 8.23453605e-01 5.66794693e-01
7.80636072e-02 -5.96665144e-01 -8.18083465e-01 1.25193429e+00
-3.32436919e-01 -1.19169712e+00 -5.75668335e-01 4.67400700e-01
7.70517290e-01 4.30187285e-01 -3.23929340e-02 8.89845341e-02
1.07272349e-01 -6.89974010e-01 3.59935462e-01 1.54072687e-01
8.86295140e-01 -8.93261909e-01 6.87674165e-01 7.12050855e-01
-1.09918547e+00 1.94516256e-01 -5.83656371e-01 -1.83944136e-01
6.85145378e-01 1.04157007e+00 -7.61866152e-01 5.29503040e-02
2.71892637e-01 -6.62494153e-02 1.35347590e-01 9.06581700e-01
-4.06449676e-01 6.21876359e-01 -9.93486047e-01 -7.53025889e-01
5.65037251e-01 -9.25841749e-01 6.51015520e-01 9.13558841e-01
5.98996580e-01 3.44579011e-01 -1.96902707e-01 8.44149590e-01
-1.45362541e-01 -8.57647061e-02 -2.99292117e-01 -5.59117913e-01
6.30764306e-01 7.82516360e-01 -7.38433778e-01 -2.83748895e-01
-1.41440451e-01 1.24998057e+00 3.60425860e-01 1.05117112e-01
-5.30701637e-01 -7.50746906e-01 4.12607878e-01 -1.31548047e-01
6.08744383e-01 -1.11297405e+00 -1.74378887e-01 -1.38675308e+00
2.99318284e-01 -3.30594122e-01 -9.30934474e-02 -4.11573499e-02
-9.23857152e-01 2.36446172e-01 -2.95197994e-01 -8.25351000e-01
-4.02641892e-01 -5.59949934e-01 -1.31245643e-01 1.00703120e+00
-1.46407104e+00 -3.15747768e-01 5.01907885e-01 1.92545846e-01
-2.64426500e-01 4.99094762e-02 1.07248139e+00 -6.05016313e-02
-4.05054152e-01 5.40296376e-01 9.49383140e-01 7.91616738e-02
1.36905670e-01 -1.23602188e+00 5.82147896e-01 8.63943398e-01
4.38147373e-02 8.80974472e-01 1.06652904e+00 -4.67920035e-01
-1.90234077e+00 -2.06768036e-01 1.24010980e+00 1.90263897e-01
1.06952941e+00 -6.15003109e-01 -5.17633557e-01 3.54426473e-01
-2.79564530e-01 1.50296176e-02 8.16698074e-01 1.53945982e-01
-2.87589192e-01 -8.05497766e-02 -1.28976846e+00 5.19559860e-01
9.13403988e-01 -1.25903392e+00 7.00860396e-02 6.28049314e-01
2.23803669e-01 -4.09217834e-01 -1.07195079e+00 1.19646132e-01
4.61987168e-01 -1.11044371e+00 5.16377091e-01 -4.98503856e-02
-3.02600473e-01 -1.38827696e-01 -2.06513360e-01 -7.95953274e-01
4.23563272e-02 -1.33400297e+00 -2.55722225e-01 2.85248250e-01
6.73923314e-01 -9.42687273e-01 1.23003662e+00 7.06573844e-01
3.72565866e-01 -6.63910568e-01 -1.62291014e+00 -9.10737574e-01
3.52200091e-01 -4.68499273e-01 1.10316224e-01 4.03237522e-01
3.42486531e-01 4.15852219e-01 -5.31098247e-01 1.54403113e-02
9.93526995e-01 3.69698018e-01 5.53623974e-01 -9.19345796e-01
-7.10641444e-01 -2.47764856e-01 -4.53713685e-01 -1.73206151e+00
-1.81533188e-01 -6.93596661e-01 6.70112744e-02 -9.35270190e-01
4.46906805e-01 -4.17327613e-01 -2.01967254e-01 -2.20390797e-01
1.83064416e-02 8.00701082e-02 -1.05904035e-01 3.21780294e-02
-7.45083809e-01 8.48825514e-01 1.10869122e+00 2.04745904e-01
-1.61488861e-01 5.33380620e-02 -2.64148563e-01 5.66799402e-01
9.46981490e-01 -1.01157188e+00 1.27878606e-01 -3.26148629e-01
6.97772920e-01 6.09050214e-01 7.70803243e-02 -1.01091921e+00
4.44508225e-01 -5.15630692e-02 -3.16205651e-01 -5.66476226e-01
7.57375598e-01 2.81460350e-03 -1.55518442e-01 8.43414783e-01
-5.07834889e-02 -3.39474171e-01 -2.94874400e-01 5.77470958e-01
-8.40756446e-02 -9.35458243e-01 9.73795474e-01 -4.15741593e-01
-1.07942067e-01 2.82501161e-01 -5.78444719e-01 -1.35576218e-01
4.42939818e-01 5.79189323e-02 -5.02947494e-02 -6.44249201e-01
-4.04246122e-01 -2.30943486e-01 4.77311552e-01 -7.36054838e-01
5.42906463e-01 -9.90843475e-01 -5.99355996e-01 2.99879070e-02
-3.29215050e-01 -1.17909767e-01 2.84963548e-01 1.23075330e+00
-7.74594605e-01 7.18765438e-01 5.55142045e-01 -4.51910228e-01
-8.24025214e-01 1.26060128e-01 4.14382577e-01 -5.20129800e-01
-2.48294994e-01 1.29786718e+00 -4.41026330e-01 -3.30436975e-01
-2.19186753e-01 8.31079781e-02 7.78553963e-01 -5.05623996e-01
5.36284149e-01 4.13088053e-01 -1.60904184e-01 -4.78077382e-01
-2.30262145e-01 4.85452265e-01 -1.99098751e-01 -7.41240561e-01
1.03449035e+00 -1.35661736e-01 -4.99878675e-01 4.32775497e-01
1.59968901e+00 6.84268726e-03 -9.98898089e-01 -4.79576111e-01
1.15929857e-01 -2.42778391e-01 2.00082019e-01 -1.66362837e-01
-5.18972337e-01 1.10226238e+00 6.80219412e-01 3.85948360e-01
6.56487763e-01 1.19155698e-01 1.12110436e+00 1.17532837e+00
1.07253182e+00 -1.17484593e+00 -1.99815586e-01 8.66741776e-01
1.73771568e-02 -1.32546473e+00 1.62695989e-01 -2.23849773e-01
-8.44057743e-03 1.04770780e+00 -8.83927345e-02 -3.11142206e-01
6.15597188e-01 1.54416580e-02 -3.82403314e-01 -2.67920732e-01
-3.86997253e-01 -1.99443445e-01 -7.24862963e-02 4.37301444e-03
5.11225879e-01 2.96805501e-01 -7.45897055e-01 -1.75043419e-01
-5.62435567e-01 1.19620457e-03 5.71084142e-01 1.09069812e+00
-8.94149065e-01 -1.49280918e+00 -1.10625580e-01 3.95013332e-01
-4.83867586e-01 -3.01090002e-01 -2.61228055e-01 4.15914714e-01
-3.49661469e-01 1.09964204e+00 -3.14609230e-01 -1.03790641e-01
-1.52256504e-01 1.90420508e-01 9.98767793e-01 -4.00089890e-01
5.04810885e-02 -2.27502100e-02 2.29600798e-02 -4.19671893e-01
-2.84647554e-01 -7.44627655e-01 -1.43467653e+00 -7.82924354e-01
-7.74745762e-01 6.76132560e-01 1.22797906e+00 1.15210712e+00
1.34415865e-01 -3.08347642e-01 7.57633209e-01 -5.58769464e-01
-1.17134964e+00 -8.98129404e-01 -9.83061731e-01 -1.07640132e-01
1.51093021e-01 -2.79544890e-01 -7.71204531e-01 -7.46058762e-01] | [5.7243170738220215, 4.8523478507995605] |
a7518356-ac26-41a3-8b53-2b1376008e2f | very-fast-streaming-submodular-function | 2010.10059 | null | https://arxiv.org/abs/2010.10059v5 | https://arxiv.org/pdf/2010.10059v5.pdf | Very Fast Streaming Submodular Function Maximization | Data summarization has become a valuable tool in understanding even terabytes of data. Due to their compelling theoretical properties, submodular functions have been in the focus of summarization algorithms. These algorithms offer worst-case approximations guarantees to the expense of higher computation and memory requirements. However, many practical applications do not fall under this worst-case, but are usually much more well-behaved. In this paper, we propose a new submodular function maximization algorithm called ThreeSieves, which ignores the worst-case, but delivers a good solution in high probability. It selects the most informative items from a data-stream on the fly and maintains a provable performance on a fixed memory budget. In an extensive evaluation, we compare our method against $6$ other methods on $8$ different datasets with and without concept drift. We show that our algorithm outperforms current state-of-the-art algorithms and, at the same time, uses fewer resources. Last, we highlight a real-world use-case of our algorithm for data summarization in gamma-ray astronomy. We make our code publicly available at https://github.com/sbuschjaeger/SubmodularStreamingMaximization. | ['Lukas Pfahler', 'Katharina Morik', 'Philipp-Jan Honysz', 'Sebastian Buschjäger'] | 2020-10-20 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 2.10120622e-02 5.89914024e-02 -4.31382209e-01 -3.96466166e-01
-1.02653170e+00 -5.71377635e-01 8.59324262e-02 5.82188189e-01
-3.87923360e-01 9.85360980e-01 1.76448092e-01 -1.92544591e-02
-4.92289960e-01 -6.66426480e-01 -8.85065138e-01 -8.62238824e-01
-9.78596658e-02 9.65322077e-01 2.33877137e-01 -2.87936274e-02
3.15673947e-01 2.89904952e-01 -1.42920899e+00 4.80854101e-02
9.71780539e-01 1.08551466e+00 1.66176751e-01 5.46667218e-01
-1.06269941e-01 4.42522138e-01 -7.12359071e-01 -4.17339236e-01
3.71678710e-01 -3.48608822e-01 -6.49657607e-01 3.52189302e-01
2.58138716e-01 -2.10026875e-01 -3.57492656e-01 9.36879575e-01
7.42333651e-01 -1.27096772e-02 3.10840189e-01 -1.33093095e+00
-1.49668708e-01 6.66533411e-01 -7.54206121e-01 3.69124830e-01
2.99239039e-01 -1.04230672e-01 1.20373559e+00 -5.90070546e-01
6.50120795e-01 1.01617002e+00 4.10403758e-01 2.62110382e-01
-1.00799358e+00 -5.22958398e-01 3.71274799e-01 1.69727847e-01
-1.20797777e+00 -3.30682725e-01 5.68166077e-01 2.76798159e-02
9.00968492e-01 7.13176906e-01 6.96921647e-01 4.07353967e-01
2.22125396e-01 1.28433108e+00 5.57736158e-01 -7.28386194e-02
4.88814384e-01 -2.78160572e-02 4.47737664e-01 4.61099833e-01
8.71791005e-01 -3.61604780e-01 -8.98031950e-01 -7.63884425e-01
1.42472699e-01 1.84690669e-01 -5.88179469e-01 -4.62944329e-01
-1.03842092e+00 8.47403646e-01 3.28868292e-02 -1.47591121e-02
-5.62049985e-01 4.86203343e-01 2.51601964e-01 2.59361386e-01
5.79632461e-01 2.96155244e-01 -2.22093165e-01 -3.32965106e-01
-1.10220706e+00 7.50504375e-01 1.00714219e+00 1.11233985e+00
4.52024192e-01 -2.57843714e-02 -1.86683983e-01 6.84725702e-01
-2.68425737e-02 6.34364724e-01 3.33593518e-01 -8.99223864e-01
5.32211602e-01 5.53467155e-01 2.64619291e-01 -7.30407953e-01
-5.89416623e-01 -6.76454306e-01 -6.97571039e-01 -2.55708754e-01
1.58410788e-01 -1.51443526e-01 -6.08787835e-01 1.59523833e+00
4.69027370e-01 -1.03053071e-01 -5.34027703e-02 7.89118052e-01
4.87631947e-01 6.60877347e-01 -6.03134692e-01 -9.12283540e-01
1.04288614e+00 -6.13313019e-01 -6.90707266e-01 -4.34747525e-02
3.45195323e-01 -6.00914776e-01 6.15830898e-01 8.92306507e-01
-1.37977099e+00 2.37889633e-01 -1.01726508e+00 1.41150877e-01
3.53511602e-01 -5.16677499e-01 8.47609997e-01 7.78124034e-01
-9.89833176e-01 3.74381721e-01 -9.38908994e-01 -3.46315712e-01
5.02498329e-01 4.03687924e-01 1.33254841e-01 -1.18026614e-01
-5.45983970e-01 3.40408444e-01 3.36949289e-01 -4.13453311e-01
-7.78521299e-01 -8.03147972e-01 -4.11618561e-01 1.86836198e-01
8.23483169e-01 -1.06278884e+00 1.55433393e+00 -3.33870053e-01
-1.00751364e+00 5.34694850e-01 -5.89753866e-01 -8.06350589e-01
6.82771087e-01 -3.73920739e-01 6.81305304e-03 1.69478625e-01
-1.58330157e-01 2.20619321e-01 5.37283063e-01 -1.20577323e+00
-8.82710397e-01 -4.80310768e-01 -1.49586603e-01 3.23978126e-01
-5.40779710e-01 -1.83484286e-01 -5.86705565e-01 -6.40756905e-01
3.59096706e-01 -7.24691749e-01 -4.55498457e-01 -4.44588691e-01
-4.93230700e-01 -4.18441117e-01 4.38521326e-01 -3.12768131e-01
1.74216807e+00 -1.86628425e+00 2.46872157e-01 1.31576195e-01
3.58926773e-01 -1.80101916e-01 3.78766805e-01 6.18010223e-01
2.84289420e-01 2.96359789e-02 -7.22465694e-01 -6.25181973e-01
5.96780553e-02 1.60620883e-01 -3.90771240e-01 6.61600411e-01
-4.09631312e-01 6.68670714e-01 -9.28866565e-01 -2.21680254e-01
-2.38107830e-01 -2.14777082e-01 -8.31072032e-01 -1.62836656e-01
-5.95870078e-01 4.07718308e-03 -3.88114423e-01 7.99638510e-01
7.71828413e-01 -3.38961959e-01 2.44545549e-01 4.02045906e-01
-2.48201303e-02 3.51279564e-02 -1.42094576e+00 1.63938951e+00
-1.14090480e-01 4.41620409e-01 8.90770555e-02 -1.19441521e+00
7.73297787e-01 1.04884669e-01 9.80125904e-01 -4.45351809e-01
-1.42803891e-02 4.28128809e-01 -3.36921692e-01 -2.54548281e-01
9.17971313e-01 -1.01663120e-01 -2.32397258e-01 8.73101711e-01
-3.98926795e-01 -2.55034357e-01 5.16223431e-01 5.39887190e-01
1.21619284e+00 -5.23568392e-01 5.19670784e-01 -2.91162461e-01
-2.12058332e-02 1.01108260e-01 7.76021302e-01 9.91292834e-01
7.85686150e-02 9.58833575e-01 7.32813179e-01 -2.82324642e-01
-8.68473172e-01 -9.09620643e-01 -1.77553549e-01 8.06284726e-01
2.31258973e-01 -6.10732555e-01 -5.78635871e-01 -4.91154522e-01
3.88753623e-01 8.85967374e-01 -2.94279367e-01 8.28686953e-02
-5.54780126e-01 -1.37525976e+00 2.11287066e-01 1.56839341e-01
2.06657127e-01 -6.22130156e-01 -7.98843443e-01 2.83237815e-01
-2.74661034e-01 -8.88179660e-01 -4.25971329e-01 -3.72350030e-03
-1.07961178e+00 -9.13761079e-01 -7.76904881e-01 -3.70985538e-01
5.89096725e-01 6.13905191e-01 1.27492225e+00 -8.09897482e-02
-3.33691537e-01 5.96156657e-01 -3.25486600e-01 -9.76203620e-01
-3.56080644e-02 4.12443057e-02 1.30162522e-01 -7.41132349e-02
2.91755676e-01 -6.26152992e-01 -6.78147912e-01 1.48881003e-01
-1.16826272e+00 -4.46094930e-01 2.97866046e-01 6.25391543e-01
1.04253149e+00 2.46121034e-01 8.76130998e-01 -9.43392456e-01
6.12327099e-01 -8.81748319e-01 -8.35009873e-01 2.33627111e-01
-6.78990304e-01 7.54921660e-02 3.30172449e-01 -2.25598998e-02
-7.70319521e-01 -7.64760002e-03 8.18174928e-02 -2.68930286e-01
4.45830733e-01 5.80066621e-01 -2.73949560e-02 4.59885836e-01
6.66807413e-01 4.61069703e-01 6.94150031e-02 -4.61897761e-01
1.86542228e-01 7.47186005e-01 4.23159182e-01 -4.76559669e-01
6.06816709e-01 1.04829037e+00 1.07753448e-01 -9.38057184e-01
-8.26372802e-01 -8.63680124e-01 1.10175647e-01 -1.01371117e-01
2.41262645e-01 -7.70954609e-01 -5.85934460e-01 4.09460753e-01
-9.61068332e-01 8.32654089e-02 -5.68503737e-01 1.92343667e-01
-6.90282285e-01 5.69198370e-01 1.08478986e-01 -9.75860357e-01
-6.24723256e-01 -7.85826087e-01 8.37093949e-01 1.23344228e-01
-1.14797026e-01 -5.50430596e-01 8.00101236e-02 2.94649154e-01
1.76712796e-01 3.90252918e-01 5.77079356e-01 -6.74995720e-01
-9.28126931e-01 -3.20982099e-01 1.65166587e-01 1.44842952e-01
-6.03101663e-02 -3.07431430e-01 -6.57345474e-01 -5.32854617e-01
2.41888285e-01 -5.06259389e-02 1.18359184e+00 7.82414556e-01
1.44808471e+00 -3.94975305e-01 -3.21779758e-01 6.74673617e-01
1.32038689e+00 6.23906292e-02 5.14023185e-01 2.40504026e-01
2.85979241e-01 4.90192235e-01 9.38223958e-01 1.34843254e+00
4.48961586e-01 5.45932353e-01 6.80716395e-01 3.91133577e-01
1.92786902e-01 1.34366840e-01 3.56885463e-01 6.57503784e-01
9.96080264e-02 -8.40346634e-01 -7.74458408e-01 8.15691411e-01
-2.26999164e+00 -9.91741657e-01 -2.56154418e-01 2.59962583e+00
7.30649650e-01 2.02374294e-01 5.18194556e-01 1.33603588e-01
5.37662208e-01 1.30802095e-01 -8.74913990e-01 -2.49616876e-01
-4.06851918e-01 -5.76140964e-03 8.27233851e-01 4.09647077e-01
-7.64030516e-01 4.59669679e-01 5.97848415e+00 8.18959415e-01
-7.72226036e-01 9.90884826e-02 6.39693379e-01 -1.10133755e+00
-8.18561554e-01 -1.98915809e-01 -9.89549875e-01 4.93912876e-01
7.52891541e-01 -8.16779077e-01 3.64921749e-01 8.09321702e-01
1.68625325e-01 -4.82689738e-01 -1.13061476e+00 1.26212740e+00
3.48076552e-01 -1.61647749e+00 1.94906835e-02 2.33658507e-01
9.21075046e-01 1.00235008e-01 9.08372998e-02 -8.77349302e-02
1.85392261e-01 -7.85382688e-01 6.65246010e-01 3.04870069e-01
4.31510329e-01 -1.07545877e+00 7.13498890e-01 4.65521187e-01
-7.04990089e-01 -2.37313285e-01 -4.93833393e-01 8.22671726e-02
5.26856601e-01 1.27882075e+00 -9.24786150e-01 9.86370385e-01
7.46317506e-01 3.67796689e-01 -1.99031577e-01 1.43496346e+00
2.63486058e-01 6.34632289e-01 -8.65089417e-01 -2.54082173e-01
7.13662803e-02 -7.89875016e-02 1.05361152e+00 1.04876363e+00
6.50533915e-01 1.44058391e-01 7.49475658e-02 2.91689306e-01
-3.61770332e-01 1.29733130e-01 -5.02449393e-01 -3.73727009e-02
6.58072114e-01 8.46842468e-01 -6.71587586e-01 -2.79123336e-01
-2.92642295e-01 6.48987174e-01 1.55228242e-01 1.82813436e-01
-8.02293181e-01 -3.19676548e-01 6.98574007e-01 4.11806762e-01
3.89794737e-01 -1.91252276e-01 -6.52628839e-01 -9.98042107e-01
4.58905429e-01 -8.29044104e-01 9.45571303e-01 -1.29597500e-01
-1.02885902e+00 2.10236087e-01 2.45126501e-01 -1.10726595e+00
-2.97718525e-01 -9.73366648e-02 -3.36840183e-01 2.83037663e-01
-1.25794208e+00 -3.49885732e-01 -3.21682692e-01 3.26933533e-01
5.91176927e-01 -1.20013915e-01 2.82844305e-01 1.59080535e-01
-5.03621399e-01 5.64403951e-01 5.85532427e-01 -7.08377659e-01
6.02074742e-01 -1.31301117e+00 3.00593108e-01 9.64801610e-01
1.31150649e-03 3.41644943e-01 1.08729410e+00 -6.24902666e-01
-1.71351016e+00 -8.62382174e-01 8.50311220e-01 -3.30188990e-01
3.93457025e-01 -2.24988222e-01 -7.39814520e-01 6.59903884e-01
7.55131394e-02 -3.20895165e-01 6.82860255e-01 1.56968102e-01
9.26213041e-02 -4.14569408e-01 -1.15896821e+00 3.85791451e-01
1.22736275e+00 2.90331274e-01 -4.48572844e-01 7.40290940e-01
8.12855780e-01 -4.39316899e-01 -6.00821555e-01 4.16859895e-01
2.40170121e-01 -1.10777473e+00 8.87446582e-01 -3.24028730e-01
2.48867199e-01 -2.11212516e-01 -2.51641989e-01 -1.11829174e+00
2.86183544e-02 -1.19552600e+00 -4.99448597e-01 9.60654080e-01
3.30846369e-01 -7.66114235e-01 9.55141485e-01 4.39127296e-01
-2.44104251e-01 -1.08648705e+00 -1.15949893e+00 -1.19263732e+00
-1.32336468e-01 -6.43297076e-01 6.30244076e-01 5.19236982e-01
3.28542739e-02 6.02833834e-03 -5.07136345e-01 1.28347188e-01
9.36824083e-01 5.66399217e-01 9.37225521e-01 -1.03849483e+00
-5.83131075e-01 -4.89484131e-01 -1.65055409e-01 -1.34485126e+00
-2.95857012e-01 -9.06976223e-01 -1.02413848e-01 -1.70049429e+00
5.66952229e-01 -3.70199591e-01 -2.35177532e-01 1.21526994e-01
-2.30644673e-01 -3.53672095e-02 1.84292585e-01 2.05350995e-01
-1.00778592e+00 6.35861337e-01 8.21259618e-01 -4.00405086e-04
-4.60273385e-01 3.88035417e-01 -1.10224140e+00 5.02312899e-01
1.02346909e+00 -6.47957563e-01 -4.43532825e-01 -4.52841997e-01
5.20327151e-01 1.04317956e-01 -1.25921205e-01 -8.19235325e-01
4.59882617e-01 -2.36566871e-01 -2.07649991e-01 -1.00948334e+00
2.09065840e-01 -4.38030452e-01 1.91685900e-01 6.72467291e-01
8.13823845e-03 1.05650373e-01 1.14839733e-01 8.93586576e-01
-2.16101766e-01 -5.09988964e-01 7.17256486e-01 -6.51297672e-03
-3.27774763e-01 5.61208367e-01 -8.25424865e-02 3.68027747e-01
1.30636454e+00 -4.79709394e-02 -4.53516155e-01 -6.53947055e-01
-2.88534135e-01 7.47907102e-01 4.26656157e-01 3.77522223e-02
7.15603590e-01 -9.24723923e-01 -1.14951777e+00 -3.05704057e-01
5.69558218e-02 5.09638190e-01 4.07035410e-01 9.91417944e-01
-5.68032265e-01 4.75784421e-01 2.42008328e-01 -7.82478094e-01
-1.27417386e+00 5.26230693e-01 -6.79746419e-02 -4.17923987e-01
-5.86588562e-01 9.94723082e-01 1.39505967e-01 1.04836814e-01
4.42831784e-01 -1.43485725e-01 2.60012239e-01 4.26207781e-01
5.97518802e-01 7.91237712e-01 2.14765191e-01 6.21740408e-02
-4.03782994e-01 1.09072179e-01 -2.44503811e-01 -2.01717570e-01
1.59804976e+00 -1.41888231e-01 -1.42802894e-01 3.47441107e-01
8.84451866e-01 2.33414337e-01 -9.38046932e-01 -3.58059615e-01
-1.22986650e-02 -5.55014491e-01 -7.40211606e-02 -4.31127936e-01
-9.53259528e-01 2.71303743e-01 1.13104984e-01 5.96849322e-01
1.46395433e+00 2.67537653e-01 1.10094750e+00 4.87086356e-01
4.97476369e-01 -1.13592768e+00 1.39140055e-01 1.48449957e-01
1.01942420e+00 -1.08913994e+00 5.76233864e-01 -4.87247437e-01
-6.37950599e-01 8.23477983e-01 1.61876157e-01 -1.69255450e-01
4.53350365e-01 2.76847631e-01 -4.14876103e-01 -2.96576679e-01
-1.11127484e+00 -6.45071128e-03 -8.71496946e-02 2.36123890e-01
9.70031042e-03 1.76281855e-01 -7.91522801e-01 7.41688073e-01
-4.47473019e-01 -1.29063070e-01 7.51708508e-01 1.10019660e+00
-8.61597717e-01 -1.10337055e+00 -5.93322456e-01 7.23626375e-01
-6.27311230e-01 1.87404789e-02 -2.24753127e-01 6.07140839e-01
-4.67514008e-01 1.04927933e+00 9.43438038e-02 -6.95317164e-02
3.19341391e-01 -1.80128545e-01 3.86712700e-01 -5.58126807e-01
-4.52009946e-01 1.64920822e-01 1.70579001e-01 -6.36544168e-01
-2.34334588e-01 -1.19601417e+00 -1.30673766e+00 -6.22706234e-01
7.24464580e-02 1.19562641e-01 6.03479147e-01 6.09377205e-01
4.23519313e-01 4.78118271e-01 7.62379467e-01 -3.74346763e-01
-1.04035389e+00 -6.07556224e-01 -5.72745681e-01 1.82516336e-01
3.45862806e-01 -3.05570334e-01 -2.96425283e-01 -2.21530676e-01] | [6.599205493927002, 4.9251484870910645] |
a2aaa5cf-ef4b-4aba-86c5-5660978e09fa | deltaedit-exploring-text-free-training-for | 2303.06285 | null | https://arxiv.org/abs/2303.06285v1 | https://arxiv.org/pdf/2303.06285v1.pdf | DeltaEdit: Exploring Text-free Training for Text-Driven Image Manipulation | Text-driven image manipulation remains challenging in training or inference flexibility. Conditional generative models depend heavily on expensive annotated training data. Meanwhile, recent frameworks, which leverage pre-trained vision-language models, are limited by either per text-prompt optimization or inference-time hyper-parameters tuning. In this work, we propose a novel framework named \textit{DeltaEdit} to address these problems. Our key idea is to investigate and identify a space, namely delta image and text space that has well-aligned distribution between CLIP visual feature differences of two images and CLIP textual embedding differences of source and target texts. Based on the CLIP delta space, the DeltaEdit network is designed to map the CLIP visual features differences to the editing directions of StyleGAN at training phase. Then, in inference phase, DeltaEdit predicts the StyleGAN's editing directions from the differences of the CLIP textual features. In this way, DeltaEdit is trained in a text-free manner. Once trained, it can well generalize to various text prompts for zero-shot inference without bells and whistles. Code is available at https://github.com/Yueming6568/DeltaEdit. | ['Tieniu Tan', 'Jing Dong', 'Dongliang He', 'Fu Li', 'Tianwei Lin', 'Yueming Lyu'] | 2023-03-11 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lyu_DeltaEdit_Exploring_Text-Free_Training_for_Text-Driven_Image_Manipulation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lyu_DeltaEdit_Exploring_Text-Free_Training_for_Text-Driven_Image_Manipulation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-manipulation'] | ['computer-vision'] | [ 2.71268874e-01 -2.23668709e-01 -1.05745882e-01 -6.18992865e-01
-8.00445676e-01 -6.59275413e-01 7.78095841e-01 -5.66546977e-01
-1.67041928e-01 3.77262682e-01 6.22970089e-02 -9.07527879e-02
2.37304419e-01 -5.48921347e-01 -9.25398469e-01 -6.76881492e-01
5.71908951e-01 4.00693089e-01 1.30808214e-03 -1.74947426e-01
3.12011212e-01 -1.09233730e-01 -1.31728828e+00 3.83459538e-01
7.95233488e-01 1.00895822e+00 5.40368676e-01 1.05919743e+00
-2.85279989e-01 7.09903836e-01 -4.36355531e-01 -4.59797114e-01
2.20011801e-01 -6.87411904e-01 -1.40037253e-01 1.38624981e-01
5.33292353e-01 -5.31209230e-01 -6.10202789e-01 1.25052464e+00
6.16971731e-01 7.54101798e-02 6.21765018e-01 -1.34102857e+00
-1.02059042e+00 7.23457217e-01 -5.77158451e-01 -5.59610203e-02
1.75494447e-01 6.33768559e-01 8.57567310e-01 -1.31182635e+00
7.17390299e-01 1.18518150e+00 3.78242522e-01 6.42521560e-01
-1.18732429e+00 -7.90210009e-01 2.46249452e-01 3.60186309e-01
-1.26403046e+00 -6.54886246e-01 1.08705604e+00 -6.48177683e-01
5.53534150e-01 3.03558916e-01 6.12322748e-01 1.65320218e+00
1.93750694e-01 9.75184262e-01 9.18953776e-01 -4.50310081e-01
1.71476185e-01 2.79696643e-01 -3.26943457e-01 7.01775432e-01
-2.58137763e-01 -3.05587519e-02 -8.10260236e-01 4.38755423e-01
7.11990297e-01 7.99443871e-02 -4.26665246e-01 -2.51194924e-01
-1.31523335e+00 8.49488616e-01 6.37862459e-02 1.51171491e-01
-5.68209440e-02 2.11360052e-01 3.77431720e-01 1.84287801e-01
5.37453353e-01 1.14491649e-01 -2.70634174e-01 -3.17053735e-01
-1.23599052e+00 1.11203924e-01 7.37480998e-01 1.35338247e+00
8.30393553e-01 2.91652292e-01 -4.51545894e-01 6.03399873e-01
3.63321275e-01 8.49520266e-01 4.36509907e-01 -7.04375207e-01
7.16277838e-01 3.77483040e-01 -2.08656162e-01 -1.11685443e+00
1.41948223e-01 -1.10535093e-01 -9.55080807e-01 1.38318539e-01
3.67814720e-01 -4.51272696e-01 -9.51562226e-01 1.59080720e+00
1.54367208e-01 1.37558579e-01 -1.48050100e-01 8.33035350e-01
6.23077929e-01 9.71986473e-01 -3.83100808e-01 -1.51480541e-01
1.32726657e+00 -9.92506802e-01 -8.58097196e-01 -4.35212851e-01
2.26451635e-01 -8.95611584e-01 1.46540594e+00 2.30903700e-01
-1.15130854e+00 -6.70763016e-01 -9.84728873e-01 -4.06395316e-01
-2.91949779e-01 6.41656995e-01 1.46385014e-01 2.20172867e-01
-8.48380923e-01 3.03524375e-01 -8.32370877e-01 -2.66606361e-01
3.66651714e-01 -8.64035413e-02 1.31324396e-01 7.54522393e-04
-1.12990320e+00 6.16205633e-01 3.92613888e-01 3.65456641e-01
-1.03076184e+00 -7.53317416e-01 -9.72462475e-01 4.27632034e-02
7.16551304e-01 -6.36344314e-01 1.30126452e+00 -1.15871561e+00
-1.93447542e+00 7.20501661e-01 -5.38619719e-02 -2.49229953e-01
8.39609087e-01 -2.55528688e-01 -3.39338273e-01 1.15189724e-01
4.54008430e-02 7.01714396e-01 1.46526110e+00 -1.00870347e+00
-4.50783044e-01 -2.57609874e-01 -1.56553015e-01 2.27102578e-01
-2.94518143e-01 -1.55233905e-01 -9.54706252e-01 -8.06601703e-01
-3.02644372e-01 -1.03706217e+00 1.26096979e-01 3.46970499e-01
-8.12968850e-01 -1.69468857e-02 9.69527602e-01 -5.95697045e-01
1.13640416e+00 -2.25120234e+00 2.19559476e-01 -1.76095605e-01
8.11607242e-02 2.70805508e-01 -1.56827062e-01 4.62516397e-01
2.02006981e-01 -3.52983661e-02 -1.16450906e-01 -3.57960343e-01
3.62687528e-01 3.98810394e-02 -6.15056932e-01 3.10994327e-01
3.97816658e-01 1.12871873e+00 -6.48437917e-01 -7.70155549e-01
5.27054489e-01 5.30037642e-01 -5.50236285e-01 5.64949930e-01
-7.15902448e-01 5.11007488e-01 -4.25639600e-01 4.51612443e-01
6.46312535e-01 -2.45925426e-01 5.58458939e-02 -4.56328183e-01
-3.22051346e-01 -1.82654291e-01 -8.17395270e-01 2.05414486e+00
-3.64461988e-01 1.03694594e+00 -1.28810897e-01 -9.89476025e-01
1.13711357e+00 1.44815013e-01 -3.91935520e-02 -6.29821420e-01
3.19125473e-01 -1.29855290e-01 -3.28727007e-01 -8.67585599e-01
4.84596848e-01 1.36324763e-01 -5.24442643e-02 3.24355811e-01
2.34343708e-01 -2.93279737e-01 1.93307817e-01 1.40565172e-01
4.26231861e-01 5.84630072e-01 3.78175378e-02 4.58841771e-02
3.04887235e-01 -2.72821039e-01 3.92894447e-01 6.89780414e-01
8.45086947e-02 7.52883852e-01 5.93688726e-01 -2.32531607e-01
-1.25737250e+00 -9.30335522e-01 1.46187752e-01 1.38722658e+00
8.20722207e-02 -4.62406963e-01 -8.44643295e-01 -5.12259126e-01
-3.10541600e-01 1.00225532e+00 -8.48687649e-01 -1.56166449e-01
-4.76837605e-01 -2.19535023e-01 4.89027053e-01 3.54889423e-01
6.57701433e-01 -8.89487028e-01 -5.81638336e-01 -7.98603222e-02
-1.44069627e-01 -1.09679723e+00 -1.10808241e+00 5.24461307e-02
-4.53311205e-01 -7.65530229e-01 -8.78167927e-01 -7.56562829e-01
6.75425231e-01 3.34265381e-02 8.84263813e-01 -2.97701746e-01
-4.00267869e-01 2.32515410e-01 -3.21590215e-01 -3.43097150e-01
-2.57400900e-01 7.21158285e-04 -3.36101234e-01 2.58049577e-01
3.11721802e-01 -6.13142312e-01 -7.75767207e-01 1.23218738e-01
-8.98853779e-01 5.79888821e-01 5.78870654e-01 9.45341349e-01
5.63157022e-01 -3.66132051e-01 8.08044598e-02 -8.82600486e-01
4.02597457e-01 -4.29108977e-01 -7.48293638e-01 4.89853054e-01
-4.74927723e-01 2.66995966e-01 8.64814281e-01 -7.85999775e-01
-1.27078068e+00 6.09650183e-03 9.03743356e-02 -9.99866545e-01
-1.18573345e-01 3.95430893e-01 -2.91667253e-01 3.99927437e-01
5.02613902e-01 7.34480798e-01 4.83522541e-04 -2.90412128e-01
6.84961021e-01 6.04242504e-01 8.21312308e-01 -6.59758270e-01
1.04069328e+00 3.47799957e-01 -4.42016542e-01 -8.99492979e-01
-9.28410828e-01 1.06861681e-01 -6.67755842e-01 -2.20617980e-01
1.10718608e+00 -7.71865189e-01 -5.56270719e-01 6.01952076e-01
-1.13534212e+00 -5.69157541e-01 -1.64539456e-01 3.37015629e-01
-6.94869041e-01 2.26985991e-01 -4.03898478e-01 -6.38221920e-01
-4.28969741e-01 -1.18785691e+00 1.16380656e+00 4.74177510e-01
1.03554390e-01 -1.15417504e+00 -4.84760245e-03 9.71946120e-02
4.19240117e-01 2.21628055e-01 7.88430870e-01 -4.94783252e-01
-6.51135504e-01 -8.15901980e-02 -2.16269836e-01 3.42029631e-01
-1.27445059e-02 4.18258846e-01 -8.40362430e-01 -2.34419614e-01
8.03199112e-02 -3.97762001e-01 6.48467302e-01 4.72165942e-01
1.28650808e+00 -5.29508471e-01 -1.48943979e-02 1.15925729e+00
1.39964581e+00 2.44690552e-01 5.17363906e-01 2.06678346e-01
7.88636029e-01 2.27352768e-01 6.17482781e-01 6.31930709e-01
4.35935766e-01 7.02390492e-01 3.10007602e-01 5.12897633e-02
-1.09274603e-01 -7.09111571e-01 4.52712566e-01 8.66437912e-01
2.33948901e-01 -3.97683203e-01 -6.16266668e-01 2.91896492e-01
-1.84682059e+00 -1.06724858e+00 2.41854519e-01 1.89071226e+00
1.10452557e+00 1.83465451e-01 -3.45425755e-01 -4.32665855e-01
8.56920362e-01 4.31655377e-01 -9.31255400e-01 -2.93735266e-01
-1.71800684e-02 -7.31655955e-02 1.26957804e-01 4.06543553e-01
-8.41171920e-01 1.02411258e+00 4.70305204e+00 8.85321379e-01
-1.43090081e+00 -6.48256987e-02 5.87091088e-01 -2.51858234e-01
-3.18220288e-01 1.61676750e-01 -8.99489403e-01 8.27734768e-01
6.08134389e-01 -4.10497636e-01 5.98573148e-01 8.29444289e-01
3.10084671e-01 4.50081602e-02 -1.34753346e+00 1.18296838e+00
4.00527626e-01 -1.37519145e+00 1.08269885e-01 -2.70404190e-01
7.03385413e-01 -2.28101015e-01 3.88357967e-01 5.23602486e-01
9.09940898e-02 -8.64428043e-01 8.44365180e-01 8.78475010e-01
1.14658821e+00 -4.77420300e-01 2.54819512e-01 4.08202648e-01
-9.92380261e-01 2.33953506e-01 -3.32188725e-01 2.47700468e-01
5.79792820e-02 4.05193061e-01 -8.79166067e-01 2.97187150e-01
4.87004519e-01 8.25614512e-01 -5.95584512e-01 4.50527966e-01
-4.01719958e-01 6.00848377e-01 4.61867228e-02 -1.33780912e-01
1.27860501e-01 -3.14695656e-01 6.29437268e-01 1.43277252e+00
5.50341368e-01 -1.18279727e-02 2.25267172e-01 1.24642444e+00
-2.70166863e-02 -1.24546558e-01 -5.92783809e-01 -2.68159568e-01
6.10093236e-01 1.32575893e+00 -3.97513151e-01 -4.72012520e-01
-3.16607028e-01 1.30430746e+00 1.10166252e-01 6.36698842e-01
-1.20140624e+00 -6.70791209e-01 3.87330264e-01 -2.00450882e-01
6.48433447e-01 -2.57748663e-01 -8.86885002e-02 -1.50014329e+00
-1.37996674e-02 -9.52355981e-01 7.12837279e-02 -1.19783080e+00
-1.09372103e+00 5.26905775e-01 6.02307059e-02 -1.31674349e+00
-3.75732303e-01 -5.11737704e-01 -8.95835340e-01 7.62202978e-01
-1.21268725e+00 -1.34682608e+00 -4.95692134e-01 8.53626788e-01
1.07575083e+00 -2.51971543e-01 5.38321495e-01 6.25183061e-02
-7.90486455e-01 8.96786153e-01 2.50768691e-01 3.40237826e-01
8.85538340e-01 -1.24494731e+00 4.22113210e-01 9.34098959e-01
1.36646286e-01 4.87257153e-01 9.00640190e-01 -4.20317560e-01
-1.71434832e+00 -1.18832314e+00 4.27968353e-01 -1.79624081e-01
7.61034966e-01 -6.67872488e-01 -8.04948449e-01 9.97394025e-01
5.81928432e-01 7.10186409e-03 2.73166925e-01 -2.59447724e-01
-4.30792063e-01 -2.96575934e-01 -5.32060027e-01 8.99733186e-01
7.31169641e-01 -6.85846925e-01 -3.85377228e-01 3.56341094e-01
7.59043813e-01 -7.32673824e-01 -6.66855514e-01 -7.70487338e-02
4.70577419e-01 -9.23770726e-01 6.53860867e-01 -3.43718380e-01
8.02543759e-01 -2.37528160e-01 -1.53257892e-01 -1.27364397e+00
-7.14346096e-02 -9.37337875e-01 -1.36025578e-01 1.47726774e+00
2.90992647e-01 -3.74498308e-01 4.92325872e-01 4.72240686e-01
-5.43794297e-02 -5.61218381e-01 -6.40675783e-01 -4.52552974e-01
-4.60942015e-02 -3.26624691e-01 3.73082459e-01 7.99919844e-01
-2.00340986e-01 5.81246674e-01 -7.14458466e-01 1.09378658e-02
6.33929789e-01 3.53042036e-01 1.03643107e+00 -6.07935071e-01
-4.37800467e-01 -4.29050624e-01 -8.84544328e-02 -1.31541169e+00
1.92497186e-02 -8.55286062e-01 3.64506394e-01 -1.21112132e+00
2.42810547e-01 -7.49551803e-02 1.47830799e-01 4.89316016e-01
-3.04163873e-01 7.21975639e-02 3.96738023e-01 9.47670862e-02
-6.48254335e-01 8.02464187e-01 1.47644222e+00 -2.41988242e-01
-3.66613790e-02 -1.83451146e-01 -5.21430433e-01 5.44499993e-01
9.11673129e-01 -2.64288276e-01 -5.99075377e-01 -7.30524719e-01
1.75040483e-01 2.39652842e-01 5.49346983e-01 -7.19560385e-01
4.77110028e-01 -4.25629526e-01 5.27250230e-01 -7.92420328e-01
3.90071541e-01 -5.68321049e-01 4.90687042e-02 1.49460867e-01
-7.08316565e-01 -1.81360394e-02 -3.32888030e-02 5.03649354e-01
-9.41283554e-02 -2.27122679e-01 6.75701499e-01 -1.30699322e-01
-6.83489859e-01 5.44326723e-01 -1.28363609e-01 3.52147073e-01
9.67241406e-01 -2.64978498e-01 -4.16439116e-01 -5.42602658e-01
-5.40336013e-01 2.26058647e-01 3.58857125e-01 5.93681276e-01
6.41229451e-01 -1.34413850e+00 -7.72795260e-01 3.64060998e-01
1.17086291e-01 2.23247886e-01 6.29650891e-01 8.72858107e-01
-2.59223223e-01 2.60835111e-01 -2.22797096e-01 -8.56501758e-01
-1.09046364e+00 6.29058719e-01 2.01758221e-01 1.58533733e-02
-7.50594199e-01 8.70167494e-01 4.60410744e-01 -2.08353698e-01
2.57281005e-01 -3.19152564e-01 2.81487048e-01 7.74652585e-02
5.80156207e-01 9.86671001e-02 -4.06320930e-01 -2.99578160e-01
1.05271466e-01 7.05991626e-01 -3.25080276e-01 -2.79919684e-01
1.11151123e+00 -3.00000578e-01 9.69550163e-02 7.06911623e-01
1.39672816e+00 -1.22530557e-01 -1.85355306e+00 -3.00614476e-01
-2.54458100e-01 -5.87246954e-01 -1.21492468e-01 -6.18346751e-01
-1.12770832e+00 1.16648102e+00 5.70464969e-01 -7.38369226e-02
1.16574728e+00 -1.30588815e-01 6.34771049e-01 3.00782979e-01
-2.32411683e-01 -1.19880569e+00 3.63575369e-01 4.48285133e-01
1.02683127e+00 -1.25769830e+00 -2.69837290e-01 5.24961092e-02
-9.58379447e-01 1.33278489e+00 7.50017643e-01 -1.24650687e-01
4.16254938e-01 3.47053796e-01 1.60893381e-01 -7.77141079e-02
-9.90248919e-01 1.47852078e-01 2.75463641e-01 5.80606997e-01
3.21092874e-01 -8.22406933e-02 1.10127039e-01 3.89965713e-01
-3.76605213e-01 1.12625845e-01 3.07051688e-01 6.72617316e-01
-4.37393874e-01 -7.59102464e-01 -1.88603297e-01 2.27393121e-01
-2.85455529e-02 -2.42211178e-01 -1.68474793e-01 6.89122617e-01
7.16641098e-02 6.82565153e-01 1.47267893e-01 -4.20941889e-01
1.10407747e-01 1.16048023e-01 4.24476922e-01 -3.79224598e-01
-6.53136075e-02 2.53060162e-01 -3.14894646e-01 -3.82079005e-01
-3.85485478e-02 -5.57852387e-01 -9.70033586e-01 -2.48670712e-01
-2.39099607e-01 -6.64165244e-02 7.06382096e-01 8.50241959e-01
2.98636615e-01 5.13376534e-01 8.44047785e-01 -9.70005393e-01
-6.75085247e-01 -9.82609332e-01 -4.16507423e-01 2.41863385e-01
4.03034151e-01 -2.62143224e-01 -1.88401476e-01 5.36625028e-01] | [11.280091285705566, -0.2179359644651413] |
ccb00935-242a-4420-9f9c-c463fb9ccdb0 | exploring-the-power-of-generative-deep | 2303.09012 | null | https://arxiv.org/abs/2303.09012v1 | https://arxiv.org/pdf/2303.09012v1.pdf | Exploring the Power of Generative Deep Learning for Image-to-Image Translation and MRI Reconstruction: A Cross-Domain Review | Deep learning has become a prominent computational modeling tool in the areas of computer vision and image processing in recent years. This research comprehensively analyzes the different deep-learning methods used for image-to-image translation and reconstruction in the natural and medical imaging domains. We examine the famous deep learning frameworks, such as convolutional neural networks and generative adversarial networks, and their variants, delving into the fundamental principles and difficulties of each. In the field of natural computer vision, we investigate the development and extension of various deep-learning generative models. In comparison, we investigate the possible applications of deep learning to generative medical imaging problems, including medical image translation, MRI reconstruction, and multi-contrast MRI synthesis. This thorough review provides scholars and practitioners in the areas of generative computer vision and medical imaging with useful insights for summarizing past works and getting insight into future research paths. | ['Yuda Bi'] | 2023-03-16 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 6.34337842e-01 2.70395458e-01 -1.03909457e-02 -2.99990028e-01
-6.55094028e-01 -1.04331262e-01 4.64107037e-01 -5.67532897e-01
-1.63733035e-01 5.20246148e-01 2.56837428e-01 -2.42856532e-01
-8.19063038e-02 -9.36032951e-01 -5.74244916e-01 -1.07564020e+00
1.97066426e-01 4.86204654e-01 -3.96105111e-01 -1.14203587e-01
-1.67510048e-01 5.57412386e-01 -7.61576593e-01 2.67794400e-01
7.62443781e-01 6.94000483e-01 -3.15456837e-02 8.10156584e-01
-9.16341767e-02 1.08972907e+00 -6.30024910e-01 -6.54256940e-01
-2.40597814e-01 -9.36007261e-01 -9.30756867e-01 3.49355876e-01
-4.83919866e-02 -4.58574951e-01 -8.22728932e-01 1.02468598e+00
9.66980517e-01 -1.93679959e-01 7.72434473e-01 -1.13539481e+00
-1.32424235e+00 6.46252573e-01 -4.31172282e-01 4.13855225e-01
-3.40311299e-03 1.53353095e-01 3.70804816e-01 -8.38391721e-01
6.05095685e-01 1.06464100e+00 7.35244334e-01 7.97910929e-01
-9.98080909e-01 -4.63001847e-01 -3.11294168e-01 3.14723343e-01
-1.09423935e+00 -2.41015658e-01 1.00525260e+00 -6.56989634e-01
4.05433118e-01 1.28201619e-01 5.89378417e-01 1.31985271e+00
7.62427509e-01 9.74439681e-01 1.14615571e+00 -4.26163167e-01
-6.27252534e-02 -2.06424758e-01 -2.97984153e-01 7.73036718e-01
1.61145665e-02 4.09499407e-01 -2.20979288e-01 1.00153863e-01
1.31153119e+00 2.12321490e-01 -1.62462622e-01 -2.54447721e-02
-1.15889394e+00 1.15924072e+00 7.36245334e-01 7.06919849e-01
-6.42375946e-01 2.57480234e-01 2.57881165e-01 8.57079774e-02
4.24934179e-01 2.27847859e-01 2.32693329e-01 4.13375586e-01
-1.01963758e+00 1.81023896e-01 2.86957771e-01 8.42457592e-01
3.32171679e-01 9.12745118e-01 -2.70169616e-01 7.52150536e-01
1.86979547e-01 4.47573006e-01 8.65574896e-01 -8.91904712e-01
-2.46356830e-01 -4.64253277e-02 -5.15242636e-01 -1.07281315e+00
-3.64016891e-01 -7.97755599e-01 -1.54253161e+00 3.83504145e-02
-2.47194648e-01 -2.25184381e-01 -9.88948047e-01 1.40802968e+00
1.29460469e-01 1.98107764e-01 4.84019704e-02 6.39687181e-01
1.60190248e+00 5.35728216e-01 2.05788746e-01 -2.37359554e-01
1.19649780e+00 -1.02370727e+00 -8.90025258e-01 -3.92181695e-01
2.81573180e-03 -8.67971182e-01 5.93654513e-01 1.04688294e-01
-1.35400307e+00 -7.81713128e-01 -7.74229288e-01 -3.02410841e-01
-1.03463575e-01 -9.96782184e-02 7.69635797e-01 5.91946483e-01
-1.16243112e+00 4.26955462e-01 -1.12732601e+00 -1.13253281e-01
7.90795326e-01 9.53358561e-02 -1.06622861e-03 -3.06978047e-01
-1.11394930e+00 8.52668345e-01 1.22380055e-01 1.30864605e-01
-1.20360327e+00 -7.53679931e-01 -9.57333982e-01 -1.61843508e-01
-6.31165504e-02 -1.41870332e+00 1.28091037e+00 -1.00588262e+00
-1.26278770e+00 1.21692204e+00 2.10960582e-01 -4.84156877e-01
6.10848427e-01 -3.27173956e-02 -3.60130101e-01 7.71039873e-02
1.12238027e-01 7.19874144e-01 9.44367349e-01 -1.18683815e+00
-2.99105532e-02 -4.14493173e-01 -2.24309668e-01 -4.63234186e-02
9.98171642e-02 -6.57414421e-02 -1.73349872e-01 -1.16599536e+00
1.13673635e-01 -9.22571242e-01 -5.31571388e-01 -1.01394311e-01
-5.09258747e-01 1.92567155e-01 4.25746679e-01 -6.95014894e-01
7.68168747e-01 -1.92926753e+00 3.45295012e-01 -2.07985416e-01
5.33182740e-01 2.02773914e-01 -2.12796688e-01 3.14664721e-01
-2.77574271e-01 -2.96440404e-02 -3.86540532e-01 -2.24384576e-01
-2.98711509e-01 4.40141976e-01 -3.35460275e-01 3.98661971e-01
6.57997057e-02 1.79517961e+00 -8.97997797e-01 -4.94678110e-01
3.69419396e-01 9.82280314e-01 -3.69145930e-01 3.31855863e-01
1.53479889e-01 1.03700173e+00 -4.76970524e-01 6.43771946e-01
4.19030696e-01 -4.05766755e-01 5.63584641e-02 -3.35235983e-01
4.07789409e-01 -8.64671692e-02 -3.44149381e-01 1.59071887e+00
-6.44725621e-01 6.86358213e-01 -2.18134731e-01 -1.29954505e+00
6.26132309e-01 5.67769110e-01 7.82265723e-01 -7.23498583e-01
4.43729937e-01 3.17804441e-02 2.11669784e-02 -6.77735150e-01
1.18477002e-01 -7.14660406e-01 1.02052741e-01 4.99606371e-01
2.03752875e-01 -4.19470072e-01 -2.11159945e-01 -4.30243611e-02
6.90284848e-01 -1.42250005e-02 4.08968002e-01 7.89959878e-02
3.16896826e-01 -2.39969611e-01 1.70456022e-01 6.84633911e-01
-9.89312604e-02 8.29584539e-01 -6.86897850e-03 -7.51119792e-01
-1.22380888e+00 -1.31358469e+00 -1.26979575e-01 8.46454799e-01
-1.78798318e-01 1.15042053e-01 -9.46623683e-01 -1.91625655e-01
-5.19243360e-01 4.65351224e-01 -8.73104751e-01 -4.54522729e-01
-7.55220115e-01 -1.38659322e+00 6.10901475e-01 7.49912620e-01
5.86922944e-01 -1.57429194e+00 -3.27260375e-01 3.15470517e-01
-5.44268191e-01 -1.09608853e+00 -3.51563603e-01 -2.23663613e-01
-1.09080815e+00 -9.75812793e-01 -1.24999225e+00 -1.12959719e+00
6.79889679e-01 5.08716181e-02 1.43927264e+00 7.23496996e-05
-6.97281480e-01 4.20031875e-01 -1.97033390e-01 -4.06179965e-01
-1.06521225e+00 -6.80299550e-02 -3.41481119e-01 -1.63938642e-01
-3.26285847e-02 -8.03924024e-01 -6.80568635e-01 -3.69563773e-02
-1.37167394e+00 2.27024138e-01 8.83212924e-01 1.17885363e+00
8.21504295e-01 -2.30141040e-02 5.31501830e-01 -1.24159706e+00
8.40413451e-01 -4.27455872e-01 -9.87077504e-02 2.23996326e-01
-5.92420876e-01 -4.16997895e-02 3.74994785e-01 -2.88170308e-01
-8.35284114e-01 -2.86399275e-01 -7.05028534e-01 -3.92245442e-01
-1.05084434e-01 6.47430182e-01 7.78741390e-02 -2.85363793e-01
7.47161329e-01 7.95080721e-01 1.16146684e-01 -1.31626591e-01
5.67842007e-01 1.86128348e-01 9.56989288e-01 -2.08298877e-01
6.08145654e-01 5.63146889e-01 5.49225323e-02 -7.03709543e-01
-7.49392450e-01 2.65915275e-01 -5.41481256e-01 -1.39934063e-01
1.15000534e+00 -6.64265096e-01 -2.29965016e-01 8.07711720e-01
-1.17400610e+00 -2.76291132e-01 -4.26246911e-01 3.07189912e-01
-1.00762975e+00 3.37696373e-01 -1.14362454e+00 -1.11562498e-01
-7.58114696e-01 -1.57321250e+00 9.97887850e-01 3.37865263e-01
-6.54750988e-02 -1.53013051e+00 2.52329081e-01 4.78299886e-01
6.07659638e-01 7.27671266e-01 1.10770178e+00 -1.70320764e-01
-5.07623196e-01 -1.37840554e-01 7.82696977e-02 4.97772098e-01
2.49449089e-01 -2.95385718e-01 -9.40920591e-01 -1.31515488e-01
4.12697673e-01 -8.32995027e-02 8.18894148e-01 1.01110518e+00
1.47751927e+00 -3.16583306e-01 -2.15314522e-01 9.89432096e-01
1.32234979e+00 4.28307712e-01 1.03659368e+00 7.27239624e-02
7.19199300e-01 2.75822163e-01 -2.70770907e-01 6.81383014e-02
2.56114483e-01 3.62508893e-01 2.31623769e-01 -7.13197529e-01
-5.78695953e-01 -6.70542717e-02 -9.59559679e-02 1.04115129e+00
-3.63062650e-01 -1.57780260e-01 -7.42136538e-01 3.84099573e-01
-1.40724683e+00 -1.02655721e+00 9.13745239e-02 1.48415518e+00
8.02945971e-01 -2.11752340e-01 -1.12215452e-01 9.08305198e-02
6.93576097e-01 2.23522648e-01 -4.79768425e-01 -7.59141222e-02
-2.05531389e-01 6.97978139e-01 1.17971562e-01 2.93521285e-01
-1.10735798e+00 8.22022557e-01 7.67626333e+00 7.31712162e-01
-1.33790278e+00 4.18971241e-01 1.04402173e+00 3.10265601e-01
-3.50959361e-01 -6.73808455e-01 -9.33149159e-02 2.48154044e-01
7.62638748e-01 -8.46678093e-02 4.48193938e-01 8.05358529e-01
6.50760829e-02 3.52403373e-01 -9.91691470e-01 1.12270248e+00
4.08573091e-01 -1.68213594e+00 2.04735160e-01 1.89085752e-02
1.06076539e+00 8.19883645e-02 6.09157145e-01 1.60459280e-02
3.21218163e-01 -1.38988745e+00 4.60174114e-01 6.30081773e-01
8.50029767e-01 -6.50531769e-01 7.91693032e-01 5.06243035e-02
-5.07113993e-01 2.84697682e-01 -1.75197616e-01 4.64109123e-01
4.06347513e-01 5.69818199e-01 -5.74348152e-01 5.65143585e-01
5.20127952e-01 6.91607058e-01 -1.43868640e-01 7.91578174e-01
-3.22961539e-01 4.87975091e-01 4.83657122e-01 4.52398181e-01
2.04917684e-01 -1.66824535e-01 3.88674051e-01 1.12639022e+00
2.89054990e-01 1.14870943e-01 -1.01109752e-02 1.40581441e+00
-1.78999826e-01 -1.78693563e-01 -4.92152482e-01 -1.72771081e-01
-1.30498722e-01 1.16152728e+00 -8.81235480e-01 -3.89446884e-01
-2.93483555e-01 9.62345779e-01 -3.55637938e-01 3.82200062e-01
-1.09534502e+00 -1.93235934e-01 2.57944584e-01 2.21596912e-01
-5.64067028e-02 -1.84397146e-01 -5.63793361e-01 -1.02196157e+00
-5.39902508e-01 -8.83827150e-01 1.61173165e-01 -1.01402771e+00
-1.34326673e+00 1.14607656e+00 4.67872173e-02 -9.01969969e-01
-5.49557090e-01 -4.90772754e-01 -6.24936998e-01 7.22279727e-01
-1.43241882e+00 -1.55563653e+00 -4.54898208e-01 6.40871525e-01
6.60959303e-01 -3.69785994e-01 8.40866864e-01 4.42021906e-01
-3.46171856e-01 4.32959765e-01 2.06537962e-01 4.83937800e-01
2.73656189e-01 -7.95541406e-01 6.13765359e-01 7.27090716e-01
4.09107834e-01 4.45037842e-01 5.24630964e-01 -3.99758518e-01
-9.58989322e-01 -1.27426386e+00 7.06054568e-01 -1.02552831e-01
2.96633869e-01 5.50938472e-02 -7.59229720e-01 8.37443471e-01
5.39927125e-01 2.63469461e-02 8.84217560e-01 -5.33081114e-01
1.23093359e-01 1.49974167e-01 -1.18810749e+00 6.06046379e-01
7.39264965e-01 -5.24595916e-01 -3.07240069e-01 6.08854353e-01
5.08586526e-01 -6.96203172e-01 -9.31276858e-01 4.08362597e-01
4.43959683e-01 -9.74665940e-01 1.49268496e+00 -6.55421674e-01
9.59942222e-01 2.86434799e-01 2.39620268e-01 -1.40399098e+00
-7.04270303e-01 -4.74312127e-01 1.99886516e-01 6.12216413e-01
-6.65503815e-02 -4.67722774e-01 6.29240870e-01 3.11668187e-01
-4.49293464e-01 -9.94243205e-01 -7.02396631e-01 -2.40377486e-01
5.51249683e-01 -3.24569076e-01 4.93112028e-01 8.73577952e-01
-7.64448643e-01 3.45422775e-01 -5.15488207e-01 -2.81259358e-01
5.78362346e-01 1.65302336e-01 4.43551183e-01 -8.61467123e-01
-3.95605862e-01 -6.29706502e-01 -4.40263778e-01 -6.75047815e-01
2.21233517e-01 -1.01824999e+00 -9.86029208e-02 -1.84252262e+00
3.74520630e-01 -4.98619080e-02 -1.00463383e-01 2.88195878e-01
-5.82257025e-02 7.85366416e-01 5.97275160e-02 2.43973091e-01
1.40608018e-02 5.04277766e-01 1.87881839e+00 -4.74463493e-01
2.23937452e-01 2.90337682e-01 -9.65573549e-01 7.60378599e-01
6.77189529e-01 -3.62180084e-01 -4.38185930e-01 -7.41799057e-01
-7.06507964e-03 1.63132131e-01 6.58607066e-01 -7.29492426e-01
-9.44835022e-02 -1.50119707e-01 7.10878491e-01 -3.41062546e-01
1.29264638e-01 -4.46648747e-01 4.91360098e-01 7.36065924e-01
-4.02974755e-01 2.28244692e-01 1.05495416e-02 1.29862756e-01
-4.10285562e-01 -1.48507252e-01 1.29958212e+00 -7.02692389e-01
-5.34551620e-01 6.64733291e-01 -4.99499857e-01 1.29476368e-01
9.16046679e-01 -7.76150674e-02 -1.25524074e-01 -7.12517977e-01
-1.31615019e+00 -4.43390250e-01 -5.55010960e-02 3.19508165e-01
9.97429192e-01 -1.43955266e+00 -9.55434084e-01 2.84101069e-01
-3.02763581e-01 -6.39758855e-02 6.03558838e-01 8.79703045e-01
-7.92191327e-01 4.26523119e-01 -5.46923876e-01 -4.78137344e-01
-8.84076774e-01 6.28666282e-01 7.12677360e-01 -3.16467494e-01
-5.82118571e-01 7.67779410e-01 7.14141250e-01 -9.16720405e-02
-1.43020898e-01 -1.48968056e-01 -8.31323490e-02 -5.07357359e-01
4.51371312e-01 2.92613506e-01 1.31787643e-01 -6.98223174e-01
-9.37466472e-02 4.73109782e-01 4.06877883e-02 1.27912313e-01
1.58249521e+00 -5.85366823e-02 -2.83157647e-01 4.10483079e-03
1.08876359e+00 -4.82276648e-01 -7.09910512e-01 -3.32585067e-01
-5.30719638e-01 -5.81382811e-02 1.83465779e-01 -7.18964159e-01
-1.74982893e+00 1.22872055e+00 6.72002494e-01 6.32314757e-03
1.41497648e+00 2.73245037e-01 9.83682990e-01 -2.86748052e-01
1.76740840e-01 -5.67737818e-01 5.04288316e-01 2.57949531e-01
1.17473769e+00 -1.11373067e+00 -6.02108389e-02 -1.92803293e-01
-6.33602321e-01 1.05224526e+00 2.66889334e-01 -3.08989108e-01
8.30233157e-01 5.00134528e-01 1.74997196e-01 -2.69503564e-01
-1.27150878e-01 8.35642889e-02 4.10824805e-01 8.87629628e-01
7.55089462e-01 1.79524422e-01 -2.29690418e-01 5.10328054e-01
-4.36638951e-01 3.30270529e-01 3.75928283e-01 6.95835888e-01
6.92590103e-02 -1.12969291e+00 -3.39988172e-01 3.08226228e-01
-6.38212264e-01 -3.13297123e-01 -6.08878508e-02 5.06790996e-01
2.38061979e-01 3.81948203e-01 1.74902156e-01 -2.09959283e-01
-7.72903338e-02 -1.64853543e-01 9.48765934e-01 -4.69148040e-01
-4.20992553e-01 4.07023698e-01 -5.55655479e-01 -1.82797477e-01
-8.03517938e-01 -4.13531214e-01 -8.95460367e-01 -3.66263777e-01
4.00084183e-02 -2.86629438e-01 5.59961736e-01 8.96189272e-01
1.80159241e-01 1.13431382e+00 3.76637280e-01 -7.74902165e-01
-2.58191347e-01 -9.91763592e-01 -5.03736913e-01 3.31936508e-01
2.60635406e-01 -4.32264119e-01 2.81747639e-01 5.86783528e-01] | [14.053439140319824, -2.0098636150360107] |
e272515f-c17a-443e-b5a4-c2cc4eb8725a | influence-of-color-spaces-for-deep-learning | 2204.02850 | null | https://arxiv.org/abs/2204.02850v1 | https://arxiv.org/pdf/2204.02850v1.pdf | Influence of Color Spaces for Deep Learning Image Colorization | Colorization is a process that converts a grayscale image into a color one that looks as natural as possible. Over the years this task has received a lot of attention. Existing colorization methods rely on different color spaces: RGB, YUV, Lab, etc. In this chapter, we aim to study their influence on the results obtained by training a deep neural network, to answer the question: "Is it crucial to correctly choose the right color space in deep-learning based colorization?". First, we briefly summarize the literature and, in particular, deep learning-based methods. We then compare the results obtained with the same deep neural network architecture with RGB, YUV and Lab color spaces. Qualitative and quantitative analysis do not conclude similarly on which color space is better. We then show the importance of carefully designing the architecture and evaluation protocols depending on the types of images that are being processed and their specificities: strong/small contours, few/many objects, recent/archive images. | ['Patricia Vitoria', 'Lara Raad', 'Rémi Giraud', 'Michaël Clément', 'Hernan Carrillo', 'Aurélie Bugeau', 'Coloma Ballester'] | 2022-04-06 | null | null | null | null | ['colorization'] | ['computer-vision'] | [-2.95930147e-01 -4.42693561e-01 2.46803477e-01 -3.84827197e-01
-8.07846338e-02 -7.26601660e-01 5.05873680e-01 2.53066868e-02
-8.70748162e-01 7.29452133e-01 -2.26240277e-01 -4.24981296e-01
2.06083685e-01 -8.40720236e-01 -4.45174813e-01 -8.63082111e-01
2.82067537e-01 2.82121271e-01 1.23992100e-01 -3.37456971e-01
4.39715743e-01 9.48964894e-01 -1.50097799e+00 1.10196240e-01
4.98924345e-01 9.35892045e-01 -2.27228895e-01 8.55313718e-01
-5.41867733e-01 5.23033679e-01 -7.08596826e-01 -3.26413184e-01
3.80700767e-01 -7.83899665e-01 -8.25632632e-01 -3.52010876e-02
4.65907335e-01 -1.72935754e-01 1.11499898e-01 1.21541917e+00
4.33938801e-01 5.45415320e-02 7.62094736e-01 -1.19261348e+00
-9.57779288e-01 3.08744401e-01 -5.61587691e-01 2.18464568e-01
8.54777396e-02 1.41544431e-01 3.77848119e-01 -7.02397406e-01
6.80705845e-01 1.08728147e+00 5.19795477e-01 6.42352760e-01
-1.27042949e+00 -5.54903388e-01 -2.27634460e-02 3.76456290e-01
-1.28716671e+00 2.38046050e-03 1.04670548e+00 -4.64887321e-01
3.51458520e-01 2.31445223e-01 9.24488783e-01 7.91614056e-01
3.11136782e-01 5.23312986e-01 1.74301898e+00 -8.89482737e-01
4.90823090e-01 3.47574055e-01 1.05611570e-01 6.31178141e-01
4.58400965e-01 1.28340395e-02 -1.49133429e-01 2.79009044e-01
7.66253829e-01 -1.48606211e-01 -2.56445318e-01 -5.00921965e-01
-7.82727361e-01 7.18556345e-01 5.63327491e-01 8.76908243e-01
-3.52287114e-01 3.05168211e-01 2.19405025e-01 2.93787390e-01
2.90627509e-01 5.77209413e-01 -4.00362194e-01 9.82316583e-02
-1.03810561e+00 -1.37772352e-01 6.00618899e-01 4.98311430e-01
1.11440134e+00 2.69822866e-01 1.46612227e-01 7.50275135e-01
2.14123651e-01 4.57053691e-01 2.98755527e-01 -1.05003560e+00
-2.23920897e-01 5.90581119e-01 2.23482475e-01 -8.14946651e-01
-5.36295176e-01 -2.56748013e-02 -7.17520773e-01 1.17662227e+00
8.78868699e-01 -5.21576941e-01 -1.15000927e+00 1.35871887e+00
6.68406934e-02 -5.40555954e-01 6.42111748e-02 1.12667012e+00
7.41443813e-01 4.12867725e-01 1.29004851e-01 2.99719810e-01
1.35640478e+00 -6.28117681e-01 -5.59952021e-01 -6.54499233e-02
1.86151847e-01 -1.04754794e+00 1.25092602e+00 5.87343693e-01
-1.03702712e+00 -6.64523184e-01 -1.21654630e+00 -4.99035344e-02
-1.15887189e+00 3.06543440e-01 6.66806281e-01 1.10387743e+00
-1.54409945e+00 5.92428982e-01 -5.74710906e-01 -6.27703309e-01
-6.87552541e-02 8.98377299e-02 -3.99857163e-01 1.58732727e-01
-1.13396907e+00 1.05920088e+00 5.54626048e-01 3.40635568e-01
-3.01457107e-01 -2.08550036e-01 -3.19151163e-01 1.28141614e-02
-1.15916230e-01 -3.37196589e-01 9.39583719e-01 -1.69693875e+00
-1.64726925e+00 1.24626017e+00 1.85417742e-01 -1.57862246e-01
7.16427088e-01 -1.47155091e-01 -3.20769817e-01 3.75452906e-01
-4.60946232e-01 8.17366362e-01 6.43235147e-01 -1.76733279e+00
-7.12110162e-01 -3.61412883e-01 1.74082294e-01 -2.31905475e-01
-3.70667949e-02 6.19692132e-02 -6.39572322e-01 -4.59413499e-01
1.33974310e-02 -1.00939000e+00 -3.24122459e-02 2.09031865e-01
-4.27829653e-01 -1.48809850e-01 6.90807164e-01 -5.05638480e-01
7.63787746e-01 -2.15004873e+00 -7.30200559e-02 3.53143692e-01
5.39236777e-02 3.83157432e-01 9.05825756e-03 3.53917986e-01
-3.67775768e-01 3.04775119e-01 -1.20569646e-01 -5.25720455e-02
3.60752195e-02 9.22376662e-02 2.34369457e-01 4.82533962e-01
8.21532756e-02 5.97480953e-01 -5.11315465e-01 -6.75651610e-01
4.73135680e-01 7.85881519e-01 -1.33222416e-01 2.16153607e-01
7.44010136e-02 3.31148982e-01 -2.45001987e-01 6.54733598e-01
1.04188108e+00 4.59285267e-02 1.27742216e-01 -5.47257662e-01
-5.30198038e-01 -6.81676149e-01 -9.79295194e-01 1.27923548e+00
-4.59247053e-01 1.16356993e+00 1.14677563e-01 -7.24633873e-01
1.10665905e+00 -2.69342177e-02 3.09417158e-01 -1.08457494e+00
5.77398479e-01 2.53219634e-01 -1.13778837e-01 -3.69370043e-01
5.86807489e-01 -3.29217613e-01 1.75516888e-01 3.82944942e-01
-1.14318468e-01 -1.85123175e-01 5.03205538e-01 -1.93217576e-01
3.47027510e-01 2.19850466e-01 -1.09649450e-01 -3.13657731e-01
5.58562517e-01 3.10741216e-02 2.47679800e-01 4.39280778e-01
-3.95504862e-01 9.46709454e-01 9.35968459e-01 -7.04197109e-01
-9.73138869e-01 -1.08779943e+00 7.64619783e-02 1.12821674e+00
2.78613120e-01 4.50001746e-01 -9.80868578e-01 -4.34836417e-01
-1.10578083e-01 8.46762419e-01 -1.13914120e+00 -1.75582930e-01
-5.74058235e-01 -7.94447660e-01 3.22134852e-01 3.54450226e-01
6.18417978e-01 -1.49623477e+00 -1.02243030e+00 -1.08131349e-01
1.42138466e-01 -7.01566160e-01 6.88961148e-02 4.62581903e-01
-9.35635209e-01 -1.22085142e+00 -1.12574148e+00 -6.85312986e-01
7.76204348e-01 1.57250017e-01 1.35630524e+00 1.53722912e-01
-2.56636113e-01 4.29310143e-01 -8.11604023e-01 -3.03642392e-01
-3.35840911e-01 -3.38815637e-02 -5.48440218e-01 1.44625306e-01
5.42930961e-01 -2.03780532e-01 -8.95129085e-01 -1.10173151e-01
-1.21580327e+00 5.45446984e-02 8.11557293e-01 4.08472866e-01
4.10117656e-01 -2.53404945e-01 -9.17867944e-02 -1.04648471e+00
6.72108829e-01 1.27986237e-01 -6.18514061e-01 5.38250566e-01
-7.09676504e-01 2.99576610e-01 6.97810650e-01 -7.23245146e-04
-1.07290471e+00 1.39910176e-01 -2.05458179e-01 -2.28798881e-01
-5.65499246e-01 2.29814991e-01 1.13405265e-01 -3.40403110e-01
6.57147348e-01 2.37827182e-01 -2.27168277e-02 -4.96265799e-01
6.41442239e-01 2.30938315e-01 3.78805697e-01 -3.57623369e-01
7.13211656e-01 7.26336956e-01 -2.56315231e-01 -8.21009219e-01
-1.14736691e-01 -1.41236305e-01 -8.50896299e-01 -5.71935594e-01
1.43109250e+00 -2.07452118e-01 -5.34105659e-01 6.46585107e-01
-1.10987341e+00 -6.06755435e-01 -1.24096863e-01 3.38614643e-01
-3.46507102e-01 2.31586233e-01 -5.02935469e-01 -7.67624915e-01
-4.00088817e-01 -1.10239613e+00 7.18057990e-01 8.84557009e-01
1.07445478e-01 -1.10253096e+00 2.16074899e-01 -1.89271942e-01
6.05286360e-01 5.07021546e-01 1.09270942e+00 -6.18053451e-02
-2.91008383e-01 -2.65746832e-01 -7.60568559e-01 4.76310819e-01
1.36837587e-01 7.37002134e-01 -9.96735215e-01 6.84478357e-02
-4.28732991e-01 -1.16223983e-01 9.90117788e-01 5.49607217e-01
1.08401501e+00 2.99023479e-01 1.53613180e-01 7.83216655e-01
2.02560019e+00 6.06850028e-01 8.91449749e-01 7.29259253e-01
5.54039240e-01 7.17389345e-01 3.71835947e-01 2.31650904e-01
-3.34159508e-02 2.22403377e-01 5.59598148e-01 -9.52482939e-01
-2.97526032e-01 2.61400878e-01 1.99336167e-02 2.65112638e-01
-3.41496736e-01 -1.01380691e-01 -8.36725712e-01 2.37123072e-01
-1.34874105e+00 -5.48453510e-01 -1.10389538e-01 1.90402234e+00
6.46684289e-01 6.81948364e-02 2.26753786e-01 4.13413346e-02
6.63157642e-01 -1.99295543e-02 -2.86649466e-01 -8.50540519e-01
-4.18631375e-01 4.09974813e-01 6.20980203e-01 3.52927893e-01
-1.12892163e+00 9.59765077e-01 6.14252567e+00 2.25233540e-01
-2.02600431e+00 -1.84873149e-01 9.87980604e-01 3.20883214e-01
-1.71690449e-01 -1.61154181e-01 -1.51931703e-01 4.33484256e-01
5.70169508e-01 2.58829713e-01 3.25219065e-01 6.54281020e-01
2.02658087e-01 -5.61231017e-01 -7.59291410e-01 9.16404009e-01
8.04621503e-02 -1.06150007e+00 1.64278522e-02 -2.94141740e-01
5.45170009e-01 -6.35764450e-02 2.39265874e-01 2.69782301e-02
4.06463921e-01 -8.70207369e-01 9.58886266e-01 8.76782060e-01
7.59984732e-01 -5.97495556e-01 8.24702621e-01 -5.54628313e-01
-9.23438966e-01 2.39167437e-01 -3.06494594e-01 4.36890453e-01
-1.03363782e-01 2.45653495e-01 -3.24462444e-01 3.58398825e-01
9.78268743e-01 8.79991055e-03 -1.09312987e+00 1.18915308e+00
-3.12302083e-01 4.09739673e-01 -1.16744064e-01 -3.74661475e-01
5.14272928e-01 -5.93825996e-01 -2.28795961e-01 1.59108245e+00
2.74326533e-01 -1.60018831e-01 -2.90728867e-01 9.84097004e-01
2.55010962e-01 1.22135125e-01 -3.07078183e-01 -5.57465740e-02
1.61025643e-01 1.43921661e+00 -1.44568396e+00 -3.67489308e-01
-5.14995396e-01 1.00886679e+00 2.00877637e-02 9.98544455e-01
-6.22911870e-01 -7.51242995e-01 3.76289964e-01 -6.79912865e-02
1.61824688e-01 -3.38732302e-01 -5.14321268e-01 -8.95606935e-01
-3.95825416e-01 -5.68804622e-01 2.72938222e-01 -1.22961342e+00
-9.91292417e-01 8.12030792e-01 -6.55551776e-02 -1.05504286e+00
3.94652411e-02 -1.14057398e+00 -8.42985988e-01 9.22370970e-01
-1.56545973e+00 -8.80586386e-01 -6.52547836e-01 4.03215796e-01
6.90006837e-02 5.81582077e-02 7.02867806e-01 3.03496808e-01
-5.77529311e-01 2.57032394e-01 3.24536234e-01 4.32728499e-01
9.01315928e-01 -1.62946093e+00 3.67186554e-02 8.14914465e-01
-5.48498295e-02 3.27445060e-01 8.46688628e-01 -4.02109995e-02
-1.20129001e+00 -5.66720366e-01 4.23540026e-01 -2.48297509e-02
2.30567470e-01 3.23471501e-02 -7.58152902e-01 1.61101893e-01
6.89419985e-01 -1.04117937e-01 4.96275246e-01 -1.64582077e-02
-4.06176746e-01 -4.75890279e-01 -1.11153018e+00 6.71178639e-01
4.49494980e-02 -3.75326753e-01 -2.81898409e-01 -1.04396388e-01
2.04214379e-01 5.76324090e-02 -4.84054655e-01 -1.29144177e-01
6.95685208e-01 -1.67341125e+00 9.12294924e-01 -4.07996893e-01
3.28058660e-01 -3.75009686e-01 1.69394836e-01 -1.42617762e+00
-3.58409196e-01 1.90023512e-01 7.77294874e-01 8.40835512e-01
6.16750300e-01 -6.27175391e-01 1.18479717e+00 7.77977288e-01
8.99636820e-02 -3.58313292e-01 -6.87070549e-01 -4.18827832e-01
5.37360072e-01 -8.45517665e-02 4.89279926e-01 8.26700151e-01
-5.37552118e-01 -2.49854580e-01 -3.89122851e-02 -2.60026813e-01
3.76159072e-01 3.31950754e-01 7.68788576e-01 -1.13690186e+00
2.15686262e-01 -1.04407382e+00 -2.09626943e-01 -1.44273564e-02
-3.10472757e-01 -3.69617641e-01 -3.51705067e-02 -1.69436193e+00
-1.71980113e-01 -4.19274062e-01 -5.04048705e-01 4.28603053e-01
-1.57332093e-01 6.78975165e-01 4.45100754e-01 -1.25376970e-01
-4.03151542e-01 2.50131577e-01 1.17965698e+00 -1.85734391e-01
-2.32135981e-01 -3.39755207e-01 -5.79069674e-01 7.61314571e-01
1.09777975e+00 -6.79725036e-03 1.23394262e-02 -6.52682483e-01
3.25305015e-01 -2.95443416e-01 2.23837093e-01 -1.06994903e+00
7.63396770e-02 -2.23724574e-01 9.06369150e-01 -3.72603983e-01
2.09786892e-01 -1.00500560e+00 1.01411521e-01 6.15633249e-01
-2.10666999e-01 3.21520448e-01 2.91849136e-01 1.13517307e-01
-2.53905267e-01 -4.77900475e-01 1.29829180e+00 -3.47977638e-01
-1.25979984e+00 -7.99236894e-02 -5.11321962e-01 -1.62310660e-01
9.82405543e-01 -4.57640558e-01 -6.80661649e-02 -3.82851362e-01
-8.38818371e-01 -3.69133383e-01 6.63750589e-01 2.86919326e-01
4.55952674e-01 -1.16108286e+00 -5.18379450e-01 4.40073311e-02
1.06778204e-01 -5.98058462e-01 1.89303786e-01 6.96043253e-01
-1.49782979e+00 1.88752592e-01 -8.16516519e-01 -2.20785007e-01
-1.21311510e+00 6.27641141e-01 5.44637501e-01 1.11178324e-01
-3.16595763e-01 7.14377284e-01 5.39137796e-02 -1.77503973e-02
7.55069479e-02 -4.86118317e-01 -4.43040311e-01 2.28128910e-01
1.79330871e-01 4.47862029e-01 9.79550481e-02 -4.76458818e-01
-3.33476484e-01 8.83383930e-01 2.41986319e-01 -2.94500321e-01
1.08036697e+00 -8.10852498e-02 -2.67294377e-01 4.70009774e-01
1.28233612e+00 -2.07428098e-01 -1.16807151e+00 1.61440015e-01
-7.48512968e-02 -4.03045505e-01 1.15044057e-01 -9.13957536e-01
-1.71170640e+00 1.30386472e+00 1.30809450e+00 7.35283375e-01
1.41622865e+00 -2.65586674e-01 3.11690569e-01 1.94270909e-01
6.96750283e-02 -1.51216006e+00 6.37093559e-02 3.19516391e-01
8.42776179e-01 -1.17968357e+00 -7.65805021e-02 8.14951956e-02
-6.68489277e-01 1.68101192e+00 6.68400407e-01 -4.30016339e-01
7.28868961e-01 -4.59676161e-02 8.35943937e-01 -2.48384520e-01
-4.19329852e-02 -4.92576301e-01 2.23693222e-01 6.81286931e-01
8.96794915e-01 1.47701412e-01 -4.35271680e-01 -9.64326113e-02
-1.26250163e-01 -3.51708173e-03 6.45252526e-01 7.15095222e-01
-5.64393818e-01 -1.13584018e+00 -7.43437052e-01 1.76672265e-01
-3.00161481e-01 1.42517626e-01 -1.00893545e+00 1.26690400e+00
3.20343703e-01 8.33076239e-01 2.07687452e-01 -3.29978168e-01
2.82407582e-01 -6.46526227e-04 4.91554409e-01 -5.13301566e-02
-6.62501514e-01 9.39370990e-02 -1.72169566e-01 -3.62523496e-01
-5.66382229e-01 -3.99866104e-01 -1.21350956e+00 -4.97158468e-01
1.68304816e-02 2.15043351e-01 1.01851833e+00 6.77853405e-01
-5.92828318e-02 4.92055923e-01 3.82898420e-01 -9.59276259e-01
2.09447742e-01 -7.47502565e-01 -7.11696148e-01 4.84503776e-01
2.31891498e-01 -4.10349399e-01 -2.33123690e-01 2.06585735e-01] | [10.443824768066406, -2.3916285037994385] |
9e67dcd2-0421-4c31-817c-2216fcecea4a | leveraging-wikidata-s-edit-history-in | 2210.15495 | null | https://arxiv.org/abs/2210.15495v1 | https://arxiv.org/pdf/2210.15495v1.pdf | Leveraging Wikidata's edit history in knowledge graph refinement tasks | Knowledge graphs have been adopted in many diverse fields for a variety of purposes. Most of those applications rely on valid and complete data to deliver their results, pressing the need to improve the quality of knowledge graphs. A number of solutions have been proposed to that end, ranging from rule-based approaches to the use of probabilistic methods, but there is an element that has not been considered yet: the edit history of the graph. In the case of collaborative knowledge graphs (e.g., Wikidata), those edits represent the process in which the community reaches some kind of fuzzy and distributed consensus over the information that best represents each entity, and can hold potentially interesting information to be used by knowledge graph refinement methods. In this paper, we explore the use of edit history information from Wikidata to improve the performance of type prediction methods. To do that, we have first built a JSON dataset containing the edit history of every instance from the 100 most important classes in Wikidata. This edit history information is then explored and analyzed, with a focus on its potential applicability in knowledge graph refinement tasks. Finally, we propose and evaluate two new methods to leverage this edit history information in knowledge graph embedding models for type prediction tasks. Our results show an improvement in one of the proposed methods against current approaches, showing the potential of using edit information in knowledge graph refinement tasks and opening new promising research lines within the field. | ['Daniel Gayo-Avello', 'Alejandro Gonzalez-Hevia'] | 2022-10-27 | null | null | null | null | ['type-prediction', 'knowledge-graph-embedding'] | ['computer-code', 'graphs'] | [ 4.81941178e-02 5.54707468e-01 -4.51522022e-01 -1.02529399e-01
8.54824334e-02 -5.21651685e-01 8.74966145e-01 8.20081174e-01
-3.25630933e-01 7.70435691e-01 3.10420364e-01 7.01229647e-02
-7.97607183e-01 -1.33689833e+00 -5.25098860e-01 -4.54171449e-01
-7.54444525e-02 6.19367898e-01 6.52520478e-01 -3.20836097e-01
3.22231382e-01 2.20290542e-01 -1.73252177e+00 2.15227336e-01
8.57722044e-01 8.61183405e-01 3.20113450e-02 1.46781892e-01
-4.40629840e-01 6.99298322e-01 -4.53296691e-01 -8.07470977e-01
2.65028719e-02 -4.39132303e-02 -1.04066479e+00 -3.64069283e-01
2.37455308e-01 4.17161405e-01 -1.84244916e-01 1.03819680e+00
2.49944270e-01 3.30941528e-01 6.03200078e-01 -1.43713129e+00
-7.34798431e-01 1.23320365e+00 -1.65920198e-01 1.61986910e-02
4.22174722e-01 -3.29809666e-01 1.36131549e+00 -5.95112503e-01
1.17615855e+00 1.13107252e+00 8.02242875e-01 1.33421943e-01
-1.09748089e+00 -3.85817796e-01 1.89900905e-01 8.83924961e-01
-1.40056992e+00 -7.00184852e-02 9.47557807e-01 -7.26905644e-01
9.56300557e-01 1.69380724e-01 8.53801847e-01 7.99743354e-01
-8.94592628e-02 5.00153840e-01 1.16035271e+00 -5.75204492e-01
2.79450685e-01 3.73493165e-01 5.00914931e-01 7.13693559e-01
6.40257299e-01 -3.75108987e-01 -6.42463505e-01 -2.87242264e-01
2.46335968e-01 -1.54776484e-01 -3.79024953e-01 -6.39362335e-01
-8.66004229e-01 7.47042358e-01 5.22691250e-01 5.57424247e-01
-3.59283596e-01 -4.89828661e-02 2.85459727e-01 1.64602056e-01
4.29490477e-01 6.58584177e-01 -4.55434352e-01 -1.23869866e-01
-7.24153757e-01 1.04712978e-01 1.12919307e+00 7.29151487e-01
9.55861151e-01 -4.39463049e-01 -1.46982878e-01 7.20752418e-01
5.08179963e-01 -6.66686743e-02 6.24271072e-02 -5.66394746e-01
2.98645288e-01 1.36036944e+00 -1.72617018e-01 -1.33961284e+00
-2.36862764e-01 -2.85424143e-01 -3.21812242e-01 7.45477602e-02
3.60823572e-01 2.38830224e-01 -7.13906109e-01 1.45285594e+00
5.78763187e-01 3.42633545e-01 -1.92230523e-01 5.20731270e-01
7.96738267e-01 3.58684540e-01 -3.22175995e-02 -7.35878199e-02
1.27469432e+00 -5.88617384e-01 -6.18591249e-01 1.89322472e-01
4.57240164e-01 -4.57411021e-01 6.42641842e-01 5.43178380e-01
-6.10430658e-01 -2.12581232e-01 -9.35546219e-01 1.25086784e-01
-9.28650200e-01 -2.96676636e-01 5.92558861e-01 7.22597897e-01
-9.60952878e-01 7.95285225e-01 -6.99887574e-01 -5.49187303e-01
4.70264673e-01 1.68503881e-01 -4.60151076e-01 -2.34353825e-01
-1.47417986e+00 1.06570864e+00 8.77524674e-01 -9.46969446e-03
-3.21645737e-01 -6.98261023e-01 -6.45375609e-01 2.60907471e-01
9.04746771e-01 -7.47773349e-01 4.19571787e-01 -6.69379652e-01
-9.66747820e-01 5.30084550e-01 1.56313494e-01 -5.30898273e-01
3.17058861e-01 -2.60488112e-02 -4.60986912e-01 9.14707023e-04
-1.94213465e-01 1.32508412e-01 6.43550754e-01 -1.37387061e+00
-7.29166508e-01 -5.42286456e-01 3.93586904e-01 -8.66245702e-02
-7.86995649e-01 -1.41526610e-01 -7.19441533e-01 -4.40036505e-01
-2.07244545e-01 -1.05063570e+00 1.90307535e-02 -2.01819628e-01
-4.68297452e-01 -7.23167598e-01 5.81926703e-01 -6.87030733e-01
1.74340641e+00 -1.53288198e+00 6.47807539e-01 6.43459082e-01
5.35947025e-01 4.71246421e-01 1.78894445e-01 8.72479200e-01
3.54675412e-01 5.03026485e-01 -1.29280925e-01 -6.82571158e-02
1.78440526e-01 3.34152848e-01 8.21826011e-02 9.44488794e-02
-7.33214989e-02 8.55020344e-01 -9.95969117e-01 -6.39141798e-01
1.02604769e-01 6.07339263e-01 -4.22482103e-01 -8.99915993e-02
-4.67013687e-01 -8.17387551e-03 -6.55150950e-01 4.03928399e-01
3.12702566e-01 -2.67563492e-01 6.12891138e-01 -4.82657373e-01
-5.54944277e-02 2.65161246e-01 -1.40187991e+00 1.31952703e+00
-2.06661642e-01 4.04622346e-01 -4.00328338e-01 -9.55844998e-01
8.17452073e-01 3.57561298e-02 5.74007869e-01 -2.77554750e-01
-8.64310339e-02 1.23037368e-01 2.52807364e-02 -4.58809853e-01
7.40659475e-01 2.12909356e-01 2.33449101e-01 3.35961521e-01
8.85134190e-02 2.39500962e-02 5.84452152e-01 4.49480057e-01
1.24444115e+00 1.39029101e-01 4.97526050e-01 -1.26438975e-01
5.72453022e-01 1.60939589e-01 5.87361813e-01 6.87486887e-01
2.15944141e-01 2.06643194e-01 5.26924670e-01 -1.58811003e-01
-7.84442663e-01 -6.94126904e-01 -8.47112462e-02 8.54694188e-01
1.25820652e-01 -1.00139821e+00 -4.44293052e-01 -1.05700099e+00
4.02676493e-01 7.53801405e-01 -8.67257893e-01 -2.64766872e-01
-4.04218346e-01 -3.61513674e-01 6.00743890e-01 3.76987904e-01
2.42210418e-01 -1.14510858e+00 -1.75186977e-01 3.33391905e-01
-1.62215650e-01 -9.71843481e-01 8.39978978e-02 -7.29961097e-02
-6.89474225e-01 -1.60163891e+00 -1.48081735e-01 -4.35465991e-01
4.68531370e-01 -1.55357318e-02 1.07863104e+00 3.14504892e-01
-8.14454705e-02 7.31651962e-01 -8.64758551e-01 -4.15653676e-01
-4.39916223e-01 3.44032168e-01 -4.44473661e-02 1.09373845e-01
3.52327347e-01 -5.00518858e-01 -2.38225371e-01 2.87505627e-01
-9.37361479e-01 -2.04575405e-01 3.79264444e-01 6.68184102e-01
5.13850629e-01 4.54429686e-01 5.29068291e-01 -1.26639414e+00
6.47787809e-01 -6.60325885e-01 -5.96608758e-01 5.91204107e-01
-1.21564782e+00 3.55583042e-01 6.11963212e-01 -2.58045703e-01
-9.84202564e-01 -3.21457386e-01 1.32168382e-01 -4.06329215e-01
2.37180054e-01 1.35052383e+00 -1.15400940e-01 -1.47435322e-01
5.77611983e-01 -2.02368218e-02 -3.07867751e-02 -6.04289532e-01
5.25449753e-01 3.89797658e-01 8.95761475e-02 -5.10382056e-01
8.48380327e-01 1.87546656e-01 1.30680665e-01 -6.88991249e-01
-7.52557397e-01 -5.34076989e-01 -6.51904106e-01 -3.27758402e-01
5.61982691e-01 -4.26284552e-01 -5.18191993e-01 2.82143325e-01
-8.01053703e-01 -1.06447190e-01 -2.71463394e-01 3.76048207e-01
-1.79032147e-01 6.18857622e-01 -2.64532506e-01 -6.79516673e-01
-1.86264008e-01 -7.73466945e-01 4.41921860e-01 2.18820587e-01
-1.38128713e-01 -1.33464658e+00 3.41979176e-01 5.70987463e-01
3.87155801e-01 3.26117933e-01 1.20264900e+00 -1.01002049e+00
-8.54723871e-01 -1.75164476e-01 4.56995703e-03 2.61165589e-01
2.82228202e-01 2.61454940e-01 -6.31065190e-01 -1.23133771e-01
-5.35234988e-01 -1.17571931e-03 9.28637326e-01 5.08734211e-02
7.43315697e-01 -2.90475637e-01 -7.16133058e-01 1.84338436e-01
1.41701591e+00 -1.34107992e-01 6.85822785e-01 4.61391330e-01
9.92362082e-01 7.24650145e-01 6.92687869e-01 4.83001590e-01
9.15071726e-01 9.48899150e-01 4.98107046e-01 4.19950843e-01
-2.75103301e-01 -2.80294180e-01 1.46711648e-01 9.36451733e-01
-5.57585716e-01 -4.19654638e-01 -9.85051632e-01 6.89251184e-01
-2.12252831e+00 -1.07799077e+00 -3.92568588e-01 2.27205586e+00
9.74210620e-01 1.42849088e-01 7.63345063e-02 2.73547113e-01
7.62172997e-01 1.26506642e-01 -1.66760847e-01 -2.21070752e-01
-4.80476990e-02 2.26280004e-01 3.39458674e-01 4.02286768e-01
-7.93625414e-01 9.22238350e-01 5.07776070e+00 7.50199437e-01
-6.55110717e-01 -1.10675640e-01 -2.09567502e-01 4.61771607e-01
-5.11973321e-01 5.49377561e-01 -8.19212615e-01 4.79703039e-01
7.05000639e-01 -4.61520046e-01 4.89245594e-01 7.44151115e-01
-1.33463308e-01 -2.05405951e-01 -9.95467186e-01 5.58985889e-01
2.16477275e-01 -1.44080150e+00 1.42521501e-01 3.23858470e-01
6.48613691e-01 -9.02692229e-02 -4.32256073e-01 3.27231467e-01
4.90792334e-01 -8.18222165e-01 4.51964200e-01 1.03402734e+00
2.02147737e-01 -6.74771428e-01 7.79672563e-01 2.15659395e-01
-1.27193761e+00 -7.15231001e-02 -2.07841307e-01 1.70794621e-01
1.31500997e-02 9.48901594e-01 -1.28857231e+00 1.18215704e+00
6.45120502e-01 9.00531232e-01 -9.72160816e-01 1.27071214e+00
-4.68070835e-01 7.00702727e-01 -4.09951270e-01 -2.87503332e-01
-1.53582439e-01 -3.16779196e-01 6.60036743e-01 1.03857768e+00
2.72683322e-01 -1.73257098e-01 1.50135189e-01 8.43847752e-01
-1.96016863e-01 3.17722827e-01 -6.40560150e-01 -3.40824783e-01
7.37885773e-01 1.39726782e+00 -6.79909885e-01 -2.71389902e-01
-4.15762961e-01 4.41014856e-01 4.92908388e-01 7.02672824e-03
-6.44155562e-01 -4.37058002e-01 4.60843921e-01 3.59914184e-01
4.93090659e-01 -8.97880122e-02 1.47417346e-02 -1.01929617e+00
1.78868368e-01 -6.97223365e-01 8.65697086e-01 -6.47549093e-01
-1.38505089e+00 4.40505832e-01 1.83034062e-01 -8.34592283e-01
-1.50381133e-01 -4.49546188e-01 -3.05225074e-01 4.64773953e-01
-1.55824649e+00 -1.36060107e+00 -4.58251268e-01 4.71409947e-01
1.96624678e-02 -7.55467415e-02 6.64505661e-01 2.51476079e-01
-4.11253721e-01 4.40760523e-01 2.97248214e-02 8.57349113e-02
7.14197040e-01 -1.45742607e+00 1.26245439e-01 8.18113983e-01
5.41253388e-01 7.63217330e-01 7.56267428e-01 -1.15686679e+00
-1.59522665e+00 -1.08213246e+00 9.94235158e-01 -7.72091508e-01
8.78256977e-01 -7.82907754e-02 -1.26504409e+00 6.54909611e-01
4.37072478e-02 -1.52961254e-01 5.33109486e-01 5.41138470e-01
-4.67415333e-01 -3.77698429e-02 -9.75515485e-01 3.65553617e-01
1.20253694e+00 -3.71974021e-01 -8.00450325e-01 -6.13330416e-02
3.88573706e-01 8.75955224e-02 -1.36796439e+00 3.91649425e-01
4.92318600e-01 -8.33778858e-01 8.75332832e-01 -5.63799262e-01
1.42347038e-01 -5.47060490e-01 1.51453778e-01 -1.62576163e+00
-3.30239892e-01 -3.94936681e-01 -5.88298857e-01 1.54006505e+00
4.59609658e-01 -7.05247164e-01 7.50616312e-01 4.92202550e-01
-1.61089897e-01 -6.70336664e-01 -7.09775031e-01 -7.27679193e-01
-2.04150945e-01 -3.51253629e-01 5.50189972e-01 1.29616332e+00
3.29989731e-01 1.83251828e-01 -1.98275819e-01 1.32385850e-01
4.79313254e-01 2.29228973e-01 8.72501969e-01 -1.94692588e+00
-1.65152848e-01 -4.99786377e-01 -7.71284997e-01 -2.97816634e-01
1.73774526e-01 -1.49689174e+00 -4.32572097e-01 -2.10471368e+00
2.50414282e-01 -5.59232175e-01 -3.80330652e-01 7.20490754e-01
-4.04340565e-01 -3.92179228e-02 1.43197283e-01 1.11831188e-01
-7.71523356e-01 3.93741101e-01 8.46279979e-01 -2.70027518e-01
-2.16724977e-01 -2.12329015e-01 -6.24934077e-01 5.87431788e-01
4.75434214e-01 -6.31298482e-01 -4.36958969e-01 -9.54112224e-03
8.24861765e-01 -4.27820861e-01 3.23927552e-01 -7.96791136e-01
5.40866196e-01 -1.93938538e-01 -1.10634960e-01 -2.58133143e-01
2.00799122e-01 -8.39642704e-01 6.23374224e-01 1.80796340e-01
-1.30618095e-01 -1.97697356e-01 -2.07224652e-01 9.27822292e-01
-2.16387108e-01 -3.87745798e-01 3.26963574e-01 1.15069598e-01
-9.61344481e-01 2.41912305e-01 3.85347642e-02 1.53380036e-01
9.83439922e-01 -4.76328969e-01 -4.57199991e-01 -4.89003025e-02
-8.95910621e-01 2.58826584e-01 6.56321943e-01 7.45374918e-01
4.68520582e-01 -1.12010396e+00 -5.79591036e-01 -1.98224306e-01
5.41788042e-01 -3.20474327e-01 -3.76665071e-02 9.10041630e-01
-1.62708551e-01 1.63807765e-01 -1.12751968e-01 -2.52235889e-01
-1.34921527e+00 5.71551442e-01 2.52623651e-02 -5.36223471e-01
-5.80065072e-01 4.00438637e-01 -5.69351852e-01 -2.85588443e-01
-2.59591974e-02 -1.02193013e-01 -8.83988798e-01 5.71701169e-01
1.80919439e-01 6.79784954e-01 3.01769108e-01 -5.56126595e-01
-4.68757778e-01 4.52223986e-01 -8.03100318e-02 1.88890845e-01
1.61873138e+00 -1.87863596e-02 -5.63469768e-01 4.63253945e-01
7.03268409e-01 3.64867359e-01 -7.42122710e-01 -4.01218653e-01
6.03462100e-01 -4.20130223e-01 6.61390871e-02 -9.78264451e-01
-1.15084791e+00 3.30142885e-01 2.27261573e-01 5.93694806e-01
7.72699714e-01 1.80466071e-01 4.03405279e-01 4.26817358e-01
7.28869140e-01 -1.22903192e+00 -2.25613460e-01 5.10571361e-01
6.39537692e-01 -1.04702818e+00 2.19849110e-01 -7.36618996e-01
-5.16013086e-01 1.07160461e+00 3.25484931e-01 1.43069550e-01
8.04249465e-01 -1.10645078e-01 -4.77003932e-01 -5.58234990e-01
-8.22961926e-01 -5.88611603e-01 7.60665119e-01 7.18773067e-01
2.74027824e-01 6.56474680e-02 -6.49112105e-01 7.25711286e-01
6.41325563e-02 1.86597198e-01 5.90848565e-01 8.59620690e-01
-3.39941502e-01 -1.46520400e+00 -9.47146267e-02 7.85149872e-01
-1.98241800e-01 -4.70134020e-02 -8.55243921e-01 9.04602289e-01
1.68238088e-01 9.74606156e-01 -5.01788259e-01 -4.64057297e-01
3.75825316e-01 4.05466467e-01 4.89669532e-01 -7.72430241e-01
-7.83406734e-01 -5.50336421e-01 5.71148932e-01 -3.09091538e-01
-5.34940481e-01 -6.29283249e-01 -1.09217870e+00 -2.38146693e-01
-7.04925418e-01 2.28889778e-01 6.09178782e-01 8.90911281e-01
5.72492957e-01 4.52514380e-01 2.70393580e-01 -2.86697000e-01
-1.10774010e-01 -7.85302162e-01 -4.89140540e-01 7.46519089e-01
-3.94194365e-01 -1.09760380e+00 -2.66878366e-01 -7.17941895e-02] | [8.945575714111328, 7.892004489898682] |
4703c1ad-f154-4664-ab11-8e1338c3f5ac | improved-dynamic-memory-network-for-dialogue | 1811.05021 | null | http://arxiv.org/abs/1811.05021v1 | http://arxiv.org/pdf/1811.05021v1.pdf | Improved Dynamic Memory Network for Dialogue Act Classification with Adversarial Training | Dialogue Act (DA) classification is a challenging problem in dialogue
interpretation, which aims to attach semantic labels to utterances and
characterize the speaker's intention. Currently, many existing approaches
formulate the DA classification problem ranging from multi-classification to
structured prediction, which suffer from two limitations: a) these methods are
either handcrafted feature-based or have limited memories. b) adversarial
examples can't be correctly classified by traditional training methods. To
address these issues, in this paper we first cast the problem into a question
and answering problem and proposed an improved dynamic memory networks with
hierarchical pyramidal utterance encoder. Moreover, we apply adversarial
training to train our proposed model. We evaluate our model on two public
datasets, i.e., Switchboard dialogue act corpus and the MapTask corpus.
Extensive experiments show that our proposed model is not only robust, but also
achieves better performance when compared with some state-of-the-art baselines. | ['Yao Wan', 'Philip S. Yu', 'Jian Wu', 'Zhou Zhao', 'Wenqiang Yan', 'Jianwei Gao'] | 2018-11-12 | null | null | null | null | ['dialogue-act-classification', 'dialogue-interpretation'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.95049781e-01 4.01988834e-01 3.86278890e-02 -7.52323806e-01
-8.11450362e-01 -4.79189605e-01 8.63806069e-01 -2.15344667e-01
-3.33178729e-01 9.60733235e-01 5.56497812e-01 -1.86450601e-01
4.90016669e-01 -5.29954195e-01 -2.84471869e-01 -5.67741811e-01
3.86568159e-01 7.79470861e-01 3.52506220e-01 -6.30059600e-01
1.90432921e-01 -5.99558651e-02 -9.59377170e-01 5.06426513e-01
9.30345535e-01 1.18140268e+00 -9.55036730e-02 5.87487340e-01
-2.64646590e-01 1.53941894e+00 -9.11833584e-01 -6.87407732e-01
-2.29388908e-01 -6.02584124e-01 -1.51306581e+00 2.34619081e-01
1.06141776e-01 -8.50217164e-01 -4.59048599e-01 1.06672418e+00
4.39289540e-01 3.45430315e-01 7.09221244e-01 -1.25363064e+00
-6.76811755e-01 4.89800006e-01 -1.62433952e-01 -4.79107015e-02
6.50954008e-01 1.16255812e-01 1.19699752e+00 -8.14399600e-01
2.75813192e-01 1.61291218e+00 3.82333338e-01 1.15524936e+00
-9.78587389e-01 -4.44318414e-01 2.69491583e-01 3.87941062e-01
-8.90413523e-01 -8.22302043e-01 1.13248074e+00 -2.59737611e-01
1.02241743e+00 3.86321455e-01 2.14234851e-02 1.54498625e+00
-1.54825985e-01 1.35604095e+00 1.33497274e+00 -3.35246235e-01
2.16865897e-01 1.43943101e-01 3.91314715e-01 7.59498179e-01
-9.14089680e-01 -3.79850298e-01 -2.80199945e-01 -3.26222390e-01
4.49214429e-01 -2.55885005e-01 -3.31834882e-01 1.14374399e-01
-9.73334491e-01 1.27866256e+00 8.17329511e-02 2.42969215e-01
-8.47180933e-02 -2.91661680e-01 8.44503164e-01 5.62175810e-01
6.20475113e-01 3.95037979e-01 -4.73255306e-01 -2.31466025e-01
-1.10466994e-01 3.30857933e-01 1.10741353e+00 7.70861387e-01
5.22540689e-01 1.39900908e-01 -2.36637533e-01 1.24872148e+00
2.38699272e-01 3.16855013e-01 8.67847860e-01 -8.70130122e-01
8.45281780e-01 5.36920011e-01 1.28420919e-01 -8.96235287e-01
-5.86768687e-01 3.63931477e-01 -1.03524745e+00 -1.69853568e-01
4.20905471e-01 -5.73012352e-01 -4.16838795e-01 1.80696368e+00
2.34520033e-01 2.15721205e-01 5.52000701e-01 8.76008213e-01
1.26218450e+00 9.97970164e-01 1.16147809e-01 -3.85065019e-01
1.32058215e+00 -1.53822589e+00 -1.22006738e+00 -4.19061482e-01
5.37244558e-01 -4.76933509e-01 1.14142942e+00 2.78953999e-01
-1.11292899e+00 -5.58941662e-01 -1.02249742e+00 -1.03517599e-01
-2.18160897e-01 -3.59525741e-03 5.10547578e-01 5.72345674e-01
-7.65763700e-01 1.66741431e-01 -6.65902197e-01 4.73230965e-02
4.91771586e-02 3.10611218e-01 -1.21298045e-01 3.23910207e-01
-1.60393715e+00 1.05016565e+00 4.04728502e-01 -2.30361484e-02
-1.06610560e+00 8.86025429e-02 -1.06625533e+00 3.35467309e-02
4.59632874e-01 -3.34936738e-01 2.03456903e+00 -1.23410261e+00
-2.40999174e+00 7.01117098e-01 -2.37475246e-01 -6.81201100e-01
2.86886513e-01 -2.53289789e-01 -4.19495434e-01 6.75805137e-02
-1.14603326e-01 4.28592980e-01 8.22409749e-01 -1.12926745e+00
-6.76830828e-01 -3.62300605e-01 6.33784413e-01 5.17342389e-01
-5.44313669e-01 1.11360863e-01 -8.03839788e-02 -5.20838499e-01
-4.05690968e-02 -8.93527746e-01 -2.17578784e-01 -6.15622938e-01
-5.93090773e-01 -6.59104884e-01 9.99466836e-01 -8.44899893e-01
1.27244127e+00 -1.87304747e+00 4.77532983e-01 -4.42791432e-01
1.89981028e-01 5.31735063e-01 7.05830455e-02 3.73328656e-01
2.65120059e-01 -9.81036052e-02 -2.58382082e-01 -5.79514802e-01
1.32442117e-01 3.51077527e-01 -6.61470413e-01 2.20156863e-01
1.54804513e-01 7.57670522e-01 -8.20257246e-01 -4.64763820e-01
2.37843782e-01 -5.51155545e-02 -3.16679478e-01 9.71657336e-01
-6.22045994e-01 7.18725383e-01 -6.77822828e-01 4.50404942e-01
3.90328944e-01 -1.65724844e-01 3.31579506e-01 9.50271636e-03
3.52804244e-01 6.74963534e-01 -7.55425751e-01 1.73632073e+00
-7.46030331e-01 3.87462288e-01 2.05996513e-01 -1.26476407e+00
9.26141202e-01 6.33893013e-01 8.88973325e-02 -5.99144161e-01
1.37022763e-01 8.05014670e-02 -4.96433824e-02 -6.13283098e-01
5.77219546e-01 -3.94129962e-01 -4.42083597e-01 4.77068514e-01
2.63658702e-01 -1.28639027e-01 -3.68352294e-01 -7.64138773e-02
9.57983375e-01 -9.24351960e-02 4.77999449e-01 -3.71276736e-02
1.24397016e+00 -1.95319787e-01 6.97911739e-01 4.80630606e-01
-5.01710474e-01 2.68523306e-01 6.74139857e-01 -5.15340745e-01
-5.29692769e-01 -7.38118112e-01 5.56931365e-03 1.51072717e+00
2.24402681e-01 -1.77223504e-01 -9.19719577e-01 -1.29018664e+00
-4.27796781e-01 9.41771686e-01 -4.77433652e-01 -3.39103378e-02
-8.65037024e-01 -6.73468709e-01 9.26630557e-01 4.17711914e-01
1.15809786e+00 -1.16775513e+00 -2.08856300e-01 2.94062108e-01
-6.00944400e-01 -1.25052309e+00 -5.17338336e-01 -8.17791298e-02
-4.72516686e-01 -8.80953908e-01 -5.02292633e-01 -1.02689624e+00
3.90293777e-01 8.31750333e-02 1.16523123e+00 -7.15408474e-02
5.25556743e-01 3.08451802e-01 -4.77088809e-01 -2.58551031e-01
-9.09351766e-01 2.51070410e-01 3.46392542e-02 2.74426818e-01
4.52645481e-01 -2.98742324e-01 -2.55226135e-01 5.48211396e-01
-6.78738356e-01 2.43708506e-01 2.67328709e-01 1.38862967e+00
1.20946333e-01 -2.36335874e-01 1.17112899e+00 -1.16578710e+00
1.05855560e+00 -4.97462720e-01 -2.46947959e-01 4.24670786e-01
-3.17588955e-01 4.83998545e-02 9.91175234e-01 -1.85890555e-01
-1.60314989e+00 1.67478677e-02 -6.89713538e-01 -7.83328712e-02
-5.19849241e-01 3.82915288e-01 -5.56444347e-01 1.75159350e-01
4.65966314e-01 3.95775884e-01 2.46074140e-01 -4.32525367e-01
3.39086652e-01 1.24648571e+00 3.57352346e-01 -5.94841182e-01
3.02858770e-01 1.42415054e-02 -5.23354232e-01 -7.45514810e-01
-1.05540156e+00 -2.66675234e-01 -6.08396649e-01 -1.59619272e-01
1.10355878e+00 -7.85866797e-01 -9.25775826e-01 7.06948221e-01
-1.46810579e+00 -4.28722084e-01 4.14582372e-01 -6.38644677e-03
-8.31793368e-01 6.44185126e-01 -1.14257777e+00 -9.45519745e-01
-5.41262507e-01 -1.37013590e+00 7.68045664e-01 1.90181866e-01
-1.36844009e-01 -1.27055538e+00 1.18883662e-01 1.02254999e+00
3.87806088e-01 -2.01323899e-04 9.95452642e-01 -1.31784022e+00
-1.05344197e-02 6.05355874e-02 4.60977405e-02 5.67023814e-01
2.36001104e-01 -5.86266816e-01 -1.16944075e+00 -2.21747190e-01
6.15760744e-01 -1.25006700e+00 5.81914306e-01 -2.19438687e-01
1.18999946e+00 -9.25902843e-01 -8.68341699e-02 -2.70364918e-02
6.98943615e-01 4.07304645e-01 5.71622849e-01 1.05751440e-01
6.04500830e-01 7.16575742e-01 9.19229448e-01 3.39902222e-01
7.87251115e-01 9.82809961e-01 4.35637951e-01 2.00951755e-01
2.61737615e-01 -1.86048120e-01 5.38563192e-01 1.24084485e+00
2.30347484e-01 -3.52061719e-01 -9.25762415e-01 2.70667017e-01
-2.16928411e+00 -9.11487222e-01 2.13983431e-01 1.63625956e+00
1.22790468e+00 2.65588969e-01 2.43849963e-01 -1.69207156e-01
5.25038242e-01 6.18369937e-01 -5.18786311e-01 -7.06501007e-01
5.35009941e-03 -1.08346708e-01 -2.27635652e-01 6.63596570e-01
-1.58161354e+00 1.12579954e+00 5.51821995e+00 7.50037730e-01
-9.64720666e-01 4.34710532e-01 9.43802595e-01 3.59223455e-01
5.05092777e-02 -4.77237433e-01 -7.58043885e-01 4.95202750e-01
1.20987964e+00 -1.11379333e-01 2.48078987e-01 9.82844830e-01
-1.83416530e-01 3.37506086e-01 -1.17635083e+00 8.66632402e-01
4.86420691e-01 -1.03211951e+00 -1.29880160e-02 -3.29426318e-01
4.50452328e-01 -2.97000915e-01 -1.55409444e-02 9.67306435e-01
5.15235603e-01 -1.23316658e+00 2.74312496e-01 2.07913145e-01
4.14553821e-01 -6.95155919e-01 9.43745077e-01 8.15358043e-01
-7.28440762e-01 -8.58100802e-02 -3.41955096e-01 -2.35212669e-01
1.34232193e-01 -2.14409977e-01 -1.04853010e+00 4.08906877e-01
2.20516145e-01 5.27667165e-01 -2.69579023e-01 1.94829613e-01
-3.51776689e-01 8.91325176e-01 1.12666346e-01 -3.64008307e-01
5.50525904e-01 -1.76456586e-01 3.43202919e-01 1.04986489e+00
-2.88509130e-01 2.96795547e-01 4.95766252e-01 3.95687163e-01
-2.92221218e-01 3.07357252e-01 -6.12284660e-01 1.52384117e-01
4.40575570e-01 1.03276587e+00 8.26946795e-02 -4.17974979e-01
-8.65856051e-01 1.32435691e+00 6.36759520e-01 5.67547232e-02
-8.64287853e-01 -2.32542872e-01 5.81476331e-01 -4.64879632e-01
1.35736950e-02 -5.42764254e-02 -1.70635357e-02 -1.42491305e+00
-1.41595408e-01 -1.43854702e+00 4.10146385e-01 -3.46565574e-01
-1.38177943e+00 8.83469105e-01 -3.17839950e-01 -1.12918103e+00
-8.35166097e-01 -5.47196448e-01 -7.41783261e-01 5.43701947e-01
-1.23740554e+00 -1.42791045e+00 -1.06821746e-01 6.67696953e-01
1.31624866e+00 -6.63523257e-01 1.53864992e+00 2.19556883e-01
-6.39158785e-01 6.84924483e-01 1.38328299e-01 5.89729726e-01
6.77215159e-01 -1.41270947e+00 2.56190181e-01 3.93998235e-01
3.78635824e-02 2.12219745e-01 5.34087360e-01 -2.77125925e-01
-1.04657423e+00 -9.34941649e-01 6.97518706e-01 -5.19881010e-01
8.36511254e-01 -3.79680693e-01 -1.13339603e+00 9.70088959e-01
7.38993227e-01 -4.25195873e-01 1.11257005e+00 2.11520851e-01
-1.37715921e-01 2.09182382e-01 -1.20259976e+00 5.47631443e-01
6.37432873e-01 -5.79941392e-01 -1.13167715e+00 6.04115486e-01
7.90973783e-01 -7.37722814e-01 -8.82162035e-01 2.27659613e-01
1.42697722e-01 -9.28884923e-01 8.15021157e-01 -1.02038288e+00
5.79618573e-01 2.25801453e-01 -2.63257861e-01 -1.47338772e+00
-5.81748085e-03 -6.37233734e-01 -2.93834507e-01 1.45283592e+00
3.89784396e-01 -5.34992218e-01 5.51703393e-01 7.73184597e-01
-4.35201824e-01 -9.15603101e-01 -1.25458395e+00 -3.76821637e-01
3.35952997e-01 -6.55364990e-02 5.84263980e-01 1.16464245e+00
3.77915233e-01 1.15740931e+00 -9.23107266e-01 -8.43721852e-02
1.49685726e-01 1.51208550e-01 8.70557785e-01 -9.25828516e-01
-5.50354183e-01 -2.84474373e-01 -2.44710997e-01 -1.67660570e+00
9.53892648e-01 -6.25873983e-01 2.60935515e-01 -1.08009124e+00
-5.49143702e-02 -2.71016419e-01 6.75411001e-02 4.86592323e-01
-3.93404812e-01 -3.55572194e-01 -2.01760102e-02 2.40885884e-01
-8.84979725e-01 9.80702519e-01 9.67576563e-01 -4.62262869e-01
-1.90285772e-01 3.76605481e-01 -4.16638613e-01 9.22290921e-01
1.06088161e+00 -4.43698745e-03 -5.54240823e-01 -4.92148787e-01
-4.41578358e-01 6.93125010e-01 8.69542062e-02 -7.32918441e-01
9.95223373e-02 -3.30451518e-01 -1.85889512e-01 -4.34019148e-01
7.73216605e-01 -5.35750687e-01 -6.98831022e-01 2.96922177e-01
-7.75819123e-01 -1.08853526e-01 -5.24706505e-02 5.79626977e-01
-5.45386970e-01 -5.87610722e-01 7.49157488e-01 -2.48230338e-01
-7.56858289e-01 3.10522653e-02 -4.50958341e-01 2.05942899e-01
9.23123479e-01 5.75718999e-01 -5.90854049e-01 -8.75975311e-01
-5.98236620e-01 5.88415444e-01 -1.06823720e-01 5.29506207e-01
6.33942068e-01 -1.28825057e+00 -7.36824095e-01 -1.18730448e-01
-7.94414431e-03 2.61429064e-02 2.84096956e-01 4.96102452e-01
-3.03580225e-01 6.16948247e-01 -8.50606859e-02 -3.41633499e-01
-1.32410800e+00 4.41050231e-01 4.70062733e-01 -5.39341986e-01
-3.63517046e-01 7.79162049e-01 2.81176329e-01 -9.30018783e-01
4.60885495e-01 1.85927048e-01 -6.23052478e-01 -8.60489067e-03
6.78757012e-01 1.02805525e-01 -2.95418650e-01 -9.47220623e-01
-2.51392245e-01 -4.88183126e-02 -5.12183607e-01 -9.36274976e-02
1.02946150e+00 -3.13351393e-01 -2.65784506e-02 6.31018698e-01
1.16883862e+00 -2.68886894e-01 -9.51037824e-01 -6.64266527e-01
2.01010499e-02 -2.56091028e-01 -2.67237067e-01 -7.38927364e-01
-7.87053823e-01 1.04491961e+00 3.10855806e-01 5.62572896e-01
8.89973104e-01 -2.71605141e-02 1.23642313e+00 9.18861091e-01
2.60328084e-01 -1.29445314e+00 6.38005912e-01 9.06704724e-01
1.19574654e+00 -1.58372283e+00 -3.33865851e-01 -3.47813725e-01
-1.27134228e+00 1.09948325e+00 1.33920288e+00 1.83626458e-01
3.73408854e-01 -2.07041293e-01 4.29540396e-01 1.28022842e-02
-9.20185566e-01 6.45513907e-02 -7.71915615e-02 4.37011391e-01
5.03136039e-01 -2.77168658e-02 -2.74123371e-01 1.09118354e+00
-1.22127451e-01 -3.79839122e-01 4.87500846e-01 7.26689696e-01
-5.10994136e-01 -1.22483158e+00 -5.55576570e-02 4.09248561e-01
-5.44708788e-01 6.97386041e-02 -6.85299873e-01 4.58682984e-01
-4.42126930e-01 1.28614056e+00 -1.77756622e-01 -6.84868932e-01
2.74759978e-01 5.62563837e-01 -1.95547994e-02 -7.04294086e-01
-7.32479215e-01 -2.16280475e-01 6.77341878e-01 -1.30639166e-01
-4.43498135e-01 -5.37305772e-01 -1.24556947e+00 -6.19182251e-02
-4.20490563e-01 3.17921191e-01 2.48800948e-01 1.32128394e+00
8.98381546e-02 4.15337354e-01 1.03668904e+00 -4.95228887e-01
-1.24199140e+00 -1.50100756e+00 -2.00916290e-01 6.30003750e-01
2.72456735e-01 -6.22595727e-01 -2.41866559e-01 -1.13164030e-01] | [12.737732887268066, 7.7289252281188965] |
c6fb18b4-328e-4e74-9407-e03f644bfd8c | automatic-milp-solver-configuration-by | 2307.00670 | null | https://arxiv.org/abs/2307.00670v1 | https://arxiv.org/pdf/2307.00670v1.pdf | Automatic MILP Solver Configuration By Learning Problem Similarities | A large number of real-world optimization problems can be formulated as Mixed Integer Linear Programs (MILP). MILP solvers expose numerous configuration parameters to control their internal algorithms. Solutions, and their associated costs or runtimes, are significantly affected by the choice of the configuration parameters, even when problem instances have the same number of decision variables and constraints. On one hand, using the default solver configuration leads to suboptimal solutions. On the other hand, searching and evaluating a large number of configurations for every problem instance is time-consuming and, in some cases, infeasible. In this study, we aim to predict configuration parameters for unseen problem instances that yield lower-cost solutions without the time overhead of searching-and-evaluating configurations at the solving time. Toward that goal, we first investigate the cost correlation of MILP problem instances that come from the same distribution when solved using different configurations. We show that instances that have similar costs using one solver configuration also have similar costs using another solver configuration in the same runtime environment. After that, we present a methodology based on Deep Metric Learning to learn MILP similarities that correlate with their final solutions' costs. At inference time, given a new problem instance, it is first projected into the learned metric space using the trained model, and configuration parameters are instantly predicted using previously-explored configurations from the nearest neighbor instance in the learned embedding space. Empirical results on real-world problem benchmarks show that our method predicts configuration parameters that improve solutions' costs by up to 38% compared to existing approaches. | ['Sherief Reda', 'Abdelrahman Hosny'] | 2023-07-02 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [-6.68146238e-02 -2.69928783e-01 -3.22511792e-01 -4.04936492e-01
-9.70020533e-01 -9.23840821e-01 -4.71564569e-02 4.17813301e-01
-1.75000519e-01 9.81056929e-01 -1.84391901e-01 -5.10649383e-02
-7.74814248e-01 -1.02234638e+00 -7.40579367e-01 -8.23882580e-01
-1.12119600e-01 1.02579308e+00 -4.27387774e-01 -6.99748099e-02
5.15586257e-01 4.04859751e-01 -1.43249214e+00 3.16549510e-01
1.01200533e+00 1.16353774e+00 3.78208607e-02 2.73978293e-01
-3.55122268e-01 3.22110415e-01 -7.19100237e-01 -3.68628919e-01
6.92932904e-01 -4.79836687e-02 -8.74717236e-01 -1.43927680e-02
4.26622570e-01 1.03084005e-01 3.99035923e-02 1.08707166e+00
2.19603300e-01 3.33381742e-01 4.08106208e-01 -1.47481620e+00
-3.44598413e-01 6.71122074e-01 -2.92768687e-01 -5.12265079e-02
4.87386763e-01 3.22833925e-01 1.17787385e+00 -6.74933374e-01
6.87794447e-01 9.00078416e-01 4.53641087e-01 1.63803503e-01
-1.63492036e+00 -4.70792592e-01 3.75260532e-01 2.43063465e-01
-1.64684200e+00 5.32449596e-03 7.67722964e-01 -3.34299594e-01
1.05968404e+00 4.44630802e-01 5.36885381e-01 7.81951725e-01
1.73892200e-01 1.63243920e-01 8.13591957e-01 -5.43393716e-02
7.30047047e-01 3.51427197e-01 4.20471542e-02 6.14540100e-01
4.55289662e-01 -8.47899392e-02 -1.79965720e-01 -2.64802545e-01
-4.66339616e-03 -1.15051143e-01 -2.10758314e-01 -6.14060640e-01
-1.25203431e+00 1.02322543e+00 4.51619625e-01 3.65204841e-01
-2.79530615e-01 1.57884419e-01 1.83374524e-01 4.00064588e-01
1.35893419e-01 1.35287845e+00 -6.34652078e-01 -2.16415107e-01
-8.35480034e-01 5.88626206e-01 1.13904321e+00 8.43381107e-01
1.04170680e+00 -2.98341334e-01 -3.27509880e-01 6.19268775e-01
-2.88408965e-01 1.54184535e-01 3.09127301e-01 -9.45780993e-01
1.17459428e+00 9.46046114e-01 2.50391126e-01 -1.42250466e+00
-3.96211147e-01 -6.14849806e-01 -3.19982260e-01 1.56939089e-01
4.09102350e-01 -1.89832047e-01 -5.56966126e-01 1.66023886e+00
3.73494089e-01 2.37634763e-01 6.66141063e-02 9.90655303e-01
1.77638635e-01 9.68874276e-01 -1.65361375e-01 -2.31940255e-01
9.41221774e-01 -1.17048538e+00 -2.12011531e-01 -3.82370353e-01
9.64313745e-01 -6.19094074e-01 9.20300961e-01 4.95579511e-01
-9.69527006e-01 -9.11089778e-02 -1.14908993e+00 3.75541300e-01
-6.18400097e-01 4.50713979e-03 5.96277714e-01 6.24477088e-01
-6.89936817e-01 1.09060562e+00 -6.44329786e-01 1.33455573e-02
1.16062783e-01 6.41475618e-01 -2.90979981e-01 -3.41038555e-01
-9.32717264e-01 1.04397011e+00 7.03461349e-01 2.28697568e-01
-7.16907620e-01 -1.10938823e+00 -7.42365301e-01 3.69363874e-01
7.08701551e-01 -5.31547189e-01 7.16469526e-01 -7.04040587e-01
-1.35689509e+00 2.94954985e-01 -5.84254824e-02 -2.03080863e-01
5.11808813e-01 9.71405581e-02 -3.14042836e-01 -2.22819045e-01
-3.30734327e-02 1.40235588e-01 6.73022389e-01 -1.16242695e+00
-5.06251216e-01 -1.80090576e-01 5.13069451e-01 2.75919914e-01
-3.21129024e-01 -2.92614937e-01 -3.31224024e-01 -1.07308201e-01
2.05426574e-01 -9.60691392e-01 -6.37589931e-01 -4.56296861e-01
-6.37641609e-01 1.30944774e-01 2.49328092e-01 -3.62791270e-01
1.47233784e+00 -1.90134263e+00 7.93533325e-01 7.98091412e-01
1.42228976e-01 -1.65158287e-02 -5.13012707e-01 4.52378213e-01
-1.34363994e-01 3.30967218e-01 -3.02929729e-01 -1.60267502e-01
4.37274992e-01 2.97025472e-01 -1.78517655e-01 3.43881965e-01
1.73352331e-01 7.19099045e-01 -9.48002279e-01 -4.43990290e-01
4.08872843e-01 6.13947771e-02 -9.90093291e-01 2.01682061e-01
-4.61907238e-01 3.46809745e-01 -5.00902832e-01 5.49362183e-01
6.93415761e-01 -2.84732610e-01 4.49925601e-01 -3.67212236e-01
5.05654961e-02 2.32772335e-01 -1.76746655e+00 1.59192967e+00
-8.62147272e-01 2.55526453e-01 -4.24029499e-01 -1.41836607e+00
7.82648444e-01 -1.66817039e-01 5.41310966e-01 -7.46357679e-01
1.00887150e-01 4.14010704e-01 -1.19556449e-01 -4.73501712e-01
4.77429092e-01 1.82907775e-01 -3.13290209e-01 4.41273242e-01
-9.71388817e-02 -1.86457083e-01 7.65476465e-01 -2.76327997e-01
1.26432300e+00 -1.78789675e-01 9.09677669e-02 -1.26230419e-01
6.11460090e-01 2.65182704e-01 9.32267249e-01 5.65819621e-01
2.92167217e-01 4.77860838e-01 8.54832709e-01 -9.19302583e-01
-9.46727395e-01 -1.06472075e+00 -2.73766845e-01 6.88818276e-01
4.02893066e-01 -5.59982181e-01 -6.49300337e-01 -6.91162229e-01
1.84950009e-01 9.48397160e-01 -5.36068201e-01 -2.30274796e-01
-8.46500635e-01 -9.07360137e-01 1.16141755e-02 2.75027722e-01
8.18965882e-02 -6.20995700e-01 -6.05253160e-01 5.38502455e-01
-1.76219061e-01 -1.12544680e+00 -3.16046089e-01 1.86114937e-01
-8.14750552e-01 -1.19409275e+00 -2.07621872e-01 -5.02622724e-01
1.12118745e+00 -3.15195382e-01 1.31917465e+00 1.33083150e-01
-7.24628031e-01 -2.40839683e-02 -3.94555815e-02 2.39476979e-01
-1.38225257e-01 3.45135838e-01 6.47224486e-02 -9.41386893e-02
8.61881077e-02 -4.29346502e-01 -4.12910461e-01 4.10690099e-01
-6.78680003e-01 -1.32753447e-01 3.77043962e-01 8.69244277e-01
8.42642665e-01 5.11305273e-01 1.80114746e-01 -6.92692280e-01
6.41362309e-01 -7.15559185e-01 -1.23347855e+00 7.03657866e-01
-7.66397297e-01 6.06080115e-01 9.94097829e-01 -5.39467454e-01
-6.27698720e-01 1.46041170e-01 4.78357434e-01 -5.73095441e-01
1.77983791e-01 8.46771240e-01 -1.99343741e-01 -1.56288937e-01
6.48275495e-01 -1.81251884e-01 -4.46950853e-01 -2.78527498e-01
2.61905760e-01 7.08164722e-02 7.42483139e-02 -1.22137237e+00
9.87588465e-01 -9.19227675e-03 2.71172494e-01 -1.01910606e-01
-9.65883434e-01 7.78403953e-02 -3.14335376e-01 1.06160127e-01
4.92137671e-01 -2.72960037e-01 -7.20484972e-01 -1.98112428e-01
-9.76281703e-01 -1.13660119e-01 -2.92762816e-01 4.63912874e-01
-4.96120602e-01 -2.19740778e-01 -3.55594084e-02 -3.49181652e-01
8.19525197e-02 -1.62731957e+00 6.46407068e-01 1.45650908e-01
-2.83055842e-01 -1.05767882e+00 2.71199763e-01 4.14103687e-01
4.78106469e-01 6.29347205e-01 1.30125594e+00 -6.39305055e-01
-8.78461659e-01 -2.47439936e-01 -3.40389386e-02 1.95322976e-01
5.10050207e-02 6.22784197e-02 -4.37117428e-01 -3.70372176e-01
-1.65387183e-01 8.49294811e-02 1.76051080e-01 2.48704717e-01
1.46048534e+00 -7.72584081e-01 -3.69589210e-01 1.05058360e+00
1.97355509e+00 7.10468441e-02 3.85374278e-01 5.00340164e-01
4.89673525e-01 5.73030174e-01 8.08983564e-01 4.39645588e-01
3.64007242e-02 7.45028734e-01 4.93346512e-01 2.73349643e-01
4.86604065e-01 -4.73593809e-02 1.59070596e-01 3.85804355e-01
8.40168167e-03 -6.23461753e-02 -1.04616952e+00 5.02735615e-01
-1.80019820e+00 -7.13065326e-01 3.94492596e-01 2.41943049e+00
8.80642891e-01 7.91592076e-02 -1.65615126e-01 4.84877974e-02
7.07512796e-01 9.35850199e-04 -6.18268609e-01 -7.93678939e-01
2.09405109e-01 4.42500919e-01 6.26219988e-01 6.16694748e-01
-8.50321412e-01 7.18813837e-01 5.70790339e+00 6.38879061e-01
-1.26064622e+00 -2.19394878e-01 7.28595793e-01 -9.16105390e-01
-2.80290008e-01 1.65257514e-01 -7.75302112e-01 6.63198471e-01
9.57818449e-01 -6.91976607e-01 1.23926783e+00 1.12438178e+00
-1.74236268e-01 -2.07485054e-02 -1.70545173e+00 1.09669411e+00
-7.40035921e-02 -1.65566516e+00 -1.73159763e-01 8.77926499e-02
1.12073112e+00 -3.75833750e-01 2.15803966e-01 4.36045051e-01
2.08264455e-01 -1.30678391e+00 5.28265774e-01 1.64680123e-01
4.51588869e-01 -1.14757216e+00 8.16232860e-01 -2.49920469e-02
-1.09975600e+00 -5.38453341e-01 -4.99989390e-01 1.41592398e-02
1.13383032e-01 8.11876953e-01 -7.63188124e-01 5.30557752e-01
5.33665717e-01 3.62113744e-01 -3.36684495e-01 1.11837006e+00
-7.72034377e-02 1.27064943e-01 -4.91491884e-01 -3.22908983e-02
4.17643666e-01 -5.06387353e-01 4.50023800e-01 7.07607985e-01
5.92988312e-01 -1.38122633e-01 2.61403710e-01 1.39755070e+00
9.12386328e-02 2.09535927e-01 -4.38321114e-01 -1.61007121e-01
6.90221727e-01 1.20479846e+00 -5.65389276e-01 -9.72396135e-03
-6.24181069e-02 5.75530231e-01 5.86087286e-01 5.38752258e-01
-1.37300861e+00 -4.01369750e-01 1.08851302e+00 -1.48570493e-01
2.01402724e-01 -7.27686584e-02 -5.30680001e-01 -1.06170619e+00
5.11213601e-01 -8.07873726e-01 3.62731844e-01 -1.72230408e-01
-1.25333238e+00 4.88115340e-01 5.53363711e-02 -1.31151211e+00
-2.17425361e-01 -6.34699523e-01 -4.70770150e-01 7.84405112e-01
-1.42663133e+00 -4.26047176e-01 -2.68797845e-01 3.50370467e-01
1.64379478e-01 5.09497430e-03 8.68885756e-01 3.51246148e-01
-1.04570639e+00 6.83003068e-01 2.45055810e-01 -2.28098705e-01
3.29617441e-01 -1.07467020e+00 -1.28841072e-01 5.34691453e-01
8.43510777e-02 6.43171608e-01 8.78973424e-01 -1.65485010e-01
-1.73221850e+00 -1.09833050e+00 8.89505565e-01 -3.94187510e-01
7.37205029e-01 -1.45523697e-01 -6.27795517e-01 5.10444820e-01
-3.22284520e-01 1.46147326e-01 5.09855807e-01 3.15152168e-01
-3.50808591e-01 -4.69474256e-01 -1.52892721e+00 6.61981583e-01
8.53683352e-01 -2.15940997e-01 -5.01594663e-01 6.38513684e-01
5.61332583e-01 -6.58443570e-01 -1.09545755e+00 3.26178104e-01
1.75953224e-01 -6.80322528e-01 1.24052227e+00 -8.94013762e-01
6.46349311e-01 -3.18309039e-01 -4.48408455e-01 -1.60524344e+00
-2.68318892e-01 -3.47951114e-01 -1.04690775e-01 9.57205355e-01
6.36773229e-01 -8.37371528e-01 6.87338471e-01 1.31024969e+00
1.62128627e-01 -1.24099028e+00 -1.17687058e+00 -1.06289339e+00
1.20031171e-01 -3.21817607e-01 1.34123886e+00 9.88838911e-01
1.19145192e-01 -1.22480094e-01 8.35149437e-02 4.70239699e-01
5.64711452e-01 6.64924443e-01 6.01674736e-01 -9.99895751e-01
-3.93974602e-01 -4.88288790e-01 -5.47059059e-01 -1.95072025e-01
4.24662709e-01 -9.73861992e-01 -1.83110699e-01 -1.36484981e+00
2.44797673e-03 -9.57943380e-01 -5.16280949e-01 4.84385788e-01
1.72422267e-02 -1.41635939e-01 2.83019632e-01 -1.75004110e-01
-6.36485696e-01 2.72329092e-01 8.29307377e-01 -3.73169303e-01
-4.76284087e-01 -2.19949618e-01 -5.86161733e-01 5.03003776e-01
8.43810916e-01 -6.77841842e-01 -4.37764049e-01 -5.68867981e-01
7.62268364e-01 7.45688602e-02 1.43690854e-02 -1.15602839e+00
1.29375041e-01 -8.29467058e-01 7.79127181e-02 -1.54321671e-01
2.95089871e-01 -1.13622868e+00 5.62069297e-01 3.99330139e-01
-3.84843737e-01 8.85644704e-02 1.95004866e-01 1.17362007e-01
-8.71206224e-02 -6.15040839e-01 6.50224984e-01 3.13242059e-03
-4.57933694e-01 2.50858843e-01 9.69000831e-02 2.93380678e-01
1.60678947e+00 -1.60332426e-01 -3.75309795e-01 1.68251067e-01
-7.07394481e-01 5.38402438e-01 5.44278145e-01 3.65741968e-01
3.91844690e-01 -1.33056223e+00 -4.58020598e-01 7.72820367e-03
1.50846571e-01 2.48754621e-01 1.12009495e-01 6.92336440e-01
-6.55047953e-01 4.71752852e-01 -1.80159092e-01 -4.50797945e-01
-8.14104855e-01 7.72125542e-01 4.60363269e-01 -7.94304669e-01
-3.18180740e-01 6.19066060e-01 -2.82794118e-01 -6.74141586e-01
1.39687940e-01 -5.09304464e-01 2.76651889e-01 1.10082455e-01
4.58428532e-01 6.93528891e-01 2.78552622e-01 -1.02693744e-01
-4.30146545e-01 6.71635687e-01 5.06092212e-04 2.04149291e-01
1.70403302e+00 3.39900523e-01 -2.37009943e-01 4.39286791e-02
1.47547042e+00 -1.15692534e-01 -9.12960708e-01 -9.11983252e-02
-1.15454858e-02 -1.01823115e+00 2.46690493e-02 -8.31762254e-01
-1.56886494e+00 4.32940692e-01 3.78642350e-01 -1.42154917e-01
9.74690020e-01 -2.57883668e-01 5.57535052e-01 7.18079329e-01
6.89632475e-01 -1.39284217e+00 7.31263822e-03 3.25429797e-01
8.57981801e-01 -1.11004853e+00 7.49923512e-02 -3.32437664e-01
-4.85406846e-01 1.21236873e+00 7.14127481e-01 -1.99898943e-01
2.77017623e-01 3.82117808e-01 -3.45559716e-01 -1.10249706e-01
-7.84785867e-01 3.66796285e-01 1.10581927e-01 3.30205858e-01
-1.28816992e-01 2.99388826e-01 -2.22576186e-01 6.20387912e-01
-4.93764400e-01 -1.44849360e-01 2.80125648e-01 7.00431108e-01
-1.05631789e-02 -1.12999475e+00 -4.97756720e-01 4.86898094e-01
-3.99360284e-02 -5.22695258e-02 2.38414500e-02 4.46194142e-01
1.52769461e-01 8.86353731e-01 1.21869765e-01 -3.79139632e-01
3.64670247e-01 8.93278271e-02 5.48065066e-01 -6.57890439e-01
-7.39108682e-01 -7.85497487e-01 1.86866865e-01 -1.06286526e+00
3.13101888e-01 -4.90877092e-01 -1.33118701e+00 -4.97678697e-01
-6.96206987e-02 2.40409076e-01 9.18244183e-01 8.54858220e-01
5.21932960e-01 5.38592458e-01 8.37427020e-01 -8.02586019e-01
-9.28985000e-01 -1.99327350e-01 -3.59558672e-01 4.09461826e-01
5.11074103e-02 -8.32727909e-01 -6.29574478e-01 -3.82932156e-01] | [5.190420150756836, 2.969623327255249] |
b71b02c0-1920-4842-af14-7d83eeaf6c41 | cross-modality-sub-image-retrieval-using | 2201.03597 | null | https://arxiv.org/abs/2201.03597v2 | https://arxiv.org/pdf/2201.03597v2.pdf | Cross-Modality Sub-Image Retrieval using Contrastive Multimodal Image Representations | In tissue characterization and cancer diagnostics, multimodal imaging has emerged as a powerful technique. Thanks to computational advances, large datasets can be exploited to discover patterns in pathologies and improve diagnosis. However, this requires efficient and scalable image retrieval methods. Cross-modality image retrieval is particularly challenging, since images of similar (or even the same) content captured by different modalities might share few common structures. We propose a new application-independent content-based image retrieval (CBIR) system for reverse (sub-)image search across modalities, which combines deep learning to generate representations (embedding the different modalities in a common space) with classical feature extraction and bag-of-words models for efficient and reliable retrieval. We illustrate its advantages through a replacement study, exploring a number of feature extractors and learned representations, as well as through comparison to recent (cross-modality) CBIR methods. For the task of (sub-)image retrieval on a (publicly available) dataset of brightfield and second harmonic generation microscopy images, the results show that our approach is superior to all tested alternatives. We discuss the shortcomings of the compared methods and observe the importance of equivariance and invariance properties of the learned representations and feature extractors in the CBIR pipeline. Code is available at: \url{https://github.com/MIDA-group/CrossModal_ImgRetrieval}. | ['Nataša Sladoje', 'Joakim Lindblad', 'Elisabeth Wetzer', 'Eva Breznik'] | 2022-01-10 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 3.78432959e-01 -4.35385883e-01 -1.02138050e-01 -1.45389259e-01
-1.65699875e+00 -7.83041775e-01 1.00831258e+00 4.50206399e-01
-5.49622834e-01 4.09690350e-01 3.57910067e-01 -2.99779065e-02
-6.32857442e-01 -4.86417413e-01 -2.86709040e-01 -1.15815997e+00
9.66910943e-02 4.87052768e-01 8.36954415e-02 -5.15506528e-02
2.86118865e-01 8.95951569e-01 -1.62946010e+00 6.31171644e-01
1.62054032e-01 1.11443806e+00 4.83366370e-01 8.13723862e-01
-8.55732039e-02 5.21578491e-01 -3.79217565e-01 -2.08561793e-01
8.71770382e-02 -3.41466963e-01 -1.20776153e+00 -2.22906142e-01
4.20644820e-01 -1.73559368e-01 -5.10385334e-01 7.73177505e-01
9.84735608e-01 -7.08728954e-02 8.98085952e-01 -9.22255874e-01
-8.20979238e-01 -9.89895985e-02 -4.94138509e-01 5.35355985e-01
4.06037569e-01 -2.49831900e-02 8.80280614e-01 -9.17796314e-01
1.18096662e+00 9.12147880e-01 1.58490822e-01 6.58098340e-01
-1.34213185e+00 -3.80208433e-01 -5.35594225e-01 4.19242918e-01
-1.37670147e+00 -5.72793067e-01 4.18033600e-01 -4.79831487e-01
9.04235184e-01 5.89721978e-01 4.00388271e-01 9.26005960e-01
3.48459095e-01 7.39203036e-01 1.31595659e+00 -4.53155786e-01
-2.06588686e-01 1.86330661e-01 -2.76541971e-02 6.46157324e-01
-1.96678843e-02 7.37143904e-02 -6.35553718e-01 -5.04286289e-01
4.81352895e-01 3.28443319e-01 -5.35531640e-01 -3.80908966e-01
-1.65577507e+00 7.82990515e-01 5.88412046e-01 9.22907531e-01
-2.67320573e-01 7.53961280e-02 5.22282898e-01 4.29122984e-01
8.80382359e-02 5.28981686e-01 -1.31889179e-01 2.50706375e-01
-9.13516223e-01 1.43653080e-01 3.40916276e-01 5.22069156e-01
5.02009869e-01 -5.17560422e-01 -4.34175998e-01 1.02078891e+00
1.62938952e-01 6.31236911e-01 1.02630246e+00 -7.89422452e-01
-2.36793250e-01 4.13610399e-01 -3.17863315e-01 -1.03092360e+00
-5.08152962e-01 -3.25156182e-01 -1.02898777e+00 9.46266726e-02
2.94788688e-01 7.19805956e-01 -8.75029922e-01 1.46365964e+00
2.04824314e-01 -1.82932064e-01 -1.65740345e-02 9.58960414e-01
1.33380711e+00 3.74090672e-01 8.80021527e-02 -1.16913244e-01
1.81626415e+00 -7.00188816e-01 -4.93603706e-01 2.61337996e-01
6.04159117e-01 -1.06461656e+00 6.38143480e-01 1.65048316e-01
-9.80552852e-01 -3.05117726e-01 -6.06964588e-01 -4.92983699e-01
-8.58289838e-01 9.22060609e-02 4.41175044e-01 2.05875933e-01
-1.47089827e+00 3.35704982e-01 -5.07025778e-01 -5.74204504e-01
4.09258485e-01 4.46516246e-01 -1.01121628e+00 -4.74470377e-01
-9.79292691e-01 9.47937429e-01 1.65135905e-01 -1.56652868e-01
-8.69831204e-01 -7.80230284e-01 -7.69922078e-01 -1.29882619e-01
-1.68951645e-01 -8.02189827e-01 6.50685668e-01 -5.69286644e-01
-1.01998055e+00 1.59434104e+00 -2.00144112e-01 -6.14056662e-02
2.16297090e-01 2.59543240e-01 -1.59680709e-01 8.29737306e-01
1.62666440e-01 8.33120584e-01 7.30865180e-01 -1.15175664e+00
-2.74982452e-01 -5.65836012e-01 -2.27632210e-01 8.40335526e-03
-3.08653891e-01 2.27848604e-01 -5.78256309e-01 -5.47868729e-01
2.08922103e-01 -9.23420727e-01 -2.48608831e-03 3.14851135e-01
-3.28714669e-01 -3.45656015e-02 5.86899936e-01 -7.29738355e-01
8.05043578e-01 -2.13574481e+00 5.42237103e-01 2.61091769e-01
3.34286273e-01 3.82576138e-02 -5.30319691e-01 6.58990026e-01
-3.41826588e-01 1.36393696e-01 -1.00553639e-01 -3.95241201e-01
-3.04129004e-01 4.20813411e-02 -8.71751904e-02 8.09618652e-01
1.12723269e-01 1.12224007e+00 -8.22016537e-01 -7.75721669e-01
3.75943333e-01 8.63958180e-01 -7.15381727e-02 1.66747108e-01
4.67198342e-01 7.47040033e-01 -3.64722967e-01 9.80679870e-01
5.12936234e-01 -6.03005886e-01 1.37927055e-01 -6.06506765e-01
4.05557826e-02 -1.26238912e-01 -5.78069448e-01 1.95150363e+00
-4.81110334e-01 6.21234477e-01 -1.59270838e-01 -1.20645738e+00
4.66901571e-01 3.83486211e-01 9.41924512e-01 -1.15088153e+00
9.54413414e-02 4.74726707e-01 -2.73989081e-01 -6.35941684e-01
2.36777633e-01 -2.59568989e-01 6.47782758e-02 3.56744736e-01
5.38844109e-01 -1.43437117e-01 1.74123704e-01 2.00386330e-01
1.10832775e+00 -3.12117606e-01 4.42581713e-01 -2.39352062e-01
7.32068837e-01 -9.29931924e-02 -2.35821128e-01 8.20773482e-01
-1.43457949e-01 1.04642379e+00 2.42322832e-01 -3.53483111e-01
-8.85819435e-01 -1.03451645e+00 -6.11075997e-01 9.58079278e-01
8.82951319e-02 -2.62906194e-01 -4.31739874e-02 -4.01025355e-01
3.00336676e-03 -1.23053484e-01 -8.64315093e-01 -7.50752613e-02
-3.63657653e-01 -8.80539477e-01 4.99824375e-01 6.34213462e-02
-6.90095425e-02 -9.87382352e-01 -5.65745354e-01 -2.28934884e-01
-2.96370417e-01 -9.13382471e-01 -1.50300577e-01 -3.47589217e-02
-7.47438550e-01 -1.18319881e+00 -1.36078489e+00 -7.71830559e-01
7.04142869e-01 6.20375395e-01 1.05292475e+00 3.80587488e-01
-1.05551326e+00 9.74162042e-01 -3.53818685e-01 1.40853211e-01
-3.91739815e-01 -7.65986443e-02 -2.52540886e-01 -1.12203725e-01
-9.90431217e-05 -2.77924001e-01 -1.10322058e+00 2.51764636e-02
-1.53372657e+00 -8.81713331e-02 8.59523594e-01 1.25699508e+00
1.03398967e+00 -4.92241532e-01 2.86224782e-01 -6.04012370e-01
4.73054826e-01 -5.56659877e-01 -3.56834978e-01 5.23330927e-01
-4.88375276e-01 -9.88771021e-03 1.06746495e-01 -2.16766492e-01
-5.26527941e-01 -1.01200633e-01 -1.02112927e-01 -5.32472134e-01
-1.81086570e-01 6.20498955e-01 3.34534615e-01 -4.51657474e-01
5.17251670e-01 6.15271926e-01 4.23578054e-01 -3.49176407e-01
4.27342594e-01 5.10204315e-01 4.57169503e-01 -2.53030598e-01
5.42898536e-01 9.50043023e-01 3.12846035e-01 -8.56019676e-01
-4.48856384e-01 -1.00177348e+00 -5.13849139e-01 -1.25546917e-01
8.61497641e-01 -9.19797242e-01 -6.90999746e-01 3.32998335e-01
-9.58723605e-01 1.93164781e-01 -2.95641929e-01 4.51884657e-01
-6.05463982e-01 4.85760212e-01 -7.34985113e-01 -3.52423877e-01
-5.78053057e-01 -1.35154319e+00 1.65832007e+00 1.29754737e-01
-3.26679349e-02 -1.00354648e+00 3.55288118e-01 5.62264502e-01
6.25293612e-01 2.15283856e-01 9.88150418e-01 -7.72872865e-01
-4.95100141e-01 -4.28778023e-01 -5.34928799e-01 3.17523330e-02
2.75410086e-01 -1.69433236e-01 -1.07175493e+00 -5.39876342e-01
-2.80862063e-01 -4.06090409e-01 1.17810190e+00 3.27851564e-01
1.30604649e+00 -1.06944270e-01 -4.51582313e-01 5.60004890e-01
1.64023447e+00 -2.38774195e-01 7.83595204e-01 4.39863771e-01
3.59113544e-01 7.77295232e-01 2.60592997e-01 2.39274070e-01
2.80416589e-02 8.04904461e-01 2.02358827e-01 -3.23634773e-01
-4.37582582e-01 3.15749705e-01 -1.10804975e-01 7.96658158e-01
-6.10633083e-02 -7.92732909e-02 -8.89376819e-01 7.81943321e-01
-1.49248743e+00 -1.02865958e+00 1.83853015e-01 2.27365923e+00
7.62006879e-01 -6.17099822e-01 -4.11529355e-02 -7.64414817e-02
3.30343604e-01 1.00412980e-01 -2.30241701e-01 6.82140291e-02
-4.32006001e-01 4.41615492e-01 2.04720348e-01 3.67265075e-01
-1.04575634e+00 3.91703784e-01 5.94632244e+00 1.15358925e+00
-1.39641857e+00 3.88365418e-01 7.34029412e-01 -5.40440269e-02
-4.64823276e-01 -3.37377697e-01 -4.14669603e-01 9.15754437e-02
7.30065346e-01 -8.64013508e-02 1.60374731e-01 3.10912460e-01
-3.51043165e-01 -1.27179340e-01 -9.48892415e-01 1.27524376e+00
4.70858127e-01 -1.62529469e+00 3.21153492e-01 1.81894839e-01
4.93155986e-01 3.19947839e-01 4.36245859e-01 -7.41838142e-02
-4.20754254e-01 -1.12058890e+00 6.46856204e-02 8.92231405e-01
1.07023430e+00 -2.10699052e-01 8.92654181e-01 -2.54798263e-01
-8.89271200e-01 4.83216718e-02 -2.72445440e-01 8.01128924e-01
-1.56954870e-01 5.42587996e-01 -5.60989201e-01 9.17668521e-01
8.04453075e-01 6.81143463e-01 -8.42080176e-01 1.11912203e+00
4.71787304e-01 -1.39942124e-01 -8.62339363e-02 3.29193950e-01
8.68216855e-04 6.30641580e-02 4.75890130e-01 1.39843941e+00
5.91352046e-01 -1.39952317e-01 -2.21934214e-01 6.55818522e-01
8.24612603e-02 3.62811953e-01 -7.91615069e-01 -2.00010851e-01
-4.21213284e-02 1.68438423e+00 -8.65553796e-01 -2.48296678e-01
-3.90694857e-01 1.02408731e+00 1.20200180e-01 3.09595913e-01
-3.73163015e-01 -5.94493560e-02 3.98484528e-01 1.08244516e-01
1.03636079e-01 8.72152448e-02 3.55294049e-01 -1.09412849e+00
-1.95520431e-01 -8.43884230e-01 8.12240303e-01 -8.54373157e-01
-1.59645891e+00 7.65728951e-01 1.14549465e-01 -1.41892672e+00
-3.24557364e-01 -7.42756903e-01 -3.76812890e-02 8.46452594e-01
-1.87067103e+00 -1.52021861e+00 -2.44989991e-01 8.39395523e-01
1.77404955e-01 -2.51639247e-01 1.24410510e+00 4.70959067e-01
4.08723168e-02 4.46291745e-01 6.10307157e-01 -4.20257449e-02
9.42583919e-01 -9.14881706e-01 -5.14848769e-01 1.03215784e-01
2.01970622e-01 6.73825443e-01 3.55738372e-01 -1.59424588e-01
-1.65564013e+00 -7.01683223e-01 7.90504992e-01 -4.32641298e-01
7.39934564e-01 -1.03778262e-02 -8.20276320e-01 2.13345557e-01
3.35300952e-01 4.69922870e-01 1.13471651e+00 -2.62417525e-01
-5.18566847e-01 -1.00538716e-01 -1.15658653e+00 3.83828580e-01
5.19374669e-01 -8.38621080e-01 -1.31616145e-01 6.49823487e-01
3.70730907e-01 -2.54642725e-01 -1.16297054e+00 4.50085908e-01
9.49561894e-01 -9.74289775e-01 1.38286507e+00 -5.33898950e-01
5.01164496e-01 -1.15074694e-01 -3.49397510e-01 -8.11097682e-01
-2.93460548e-01 -1.61687717e-01 2.34118626e-01 8.10095072e-01
1.33792341e-01 -6.84401155e-01 2.27965415e-01 1.70210570e-01
1.19899899e-01 -8.97525966e-01 -1.34068298e+00 -4.61086214e-01
1.82829887e-01 -6.24825992e-03 1.23801537e-01 9.14528906e-01
8.18360224e-03 -1.77762181e-01 -7.61050060e-02 -1.12956971e-01
4.54699129e-01 4.48821366e-01 3.91381145e-01 -9.58210528e-01
-3.17047387e-01 -6.39120936e-01 -8.43282878e-01 -3.26328099e-01
1.26174286e-01 -1.38196170e+00 -2.77530491e-01 -1.44007695e+00
7.71369100e-01 -2.28433669e-01 -8.75433385e-01 4.10001934e-01
8.85099620e-02 8.79256845e-01 2.41471305e-01 6.70176923e-01
-7.42315888e-01 1.98055655e-01 1.31420040e+00 -4.47634816e-01
3.70169014e-01 -3.08854491e-01 -6.52210534e-01 5.21243475e-02
5.96410871e-01 -4.12828088e-01 -7.32750148e-02 -2.14908943e-01
2.73077190e-01 2.05425501e-01 7.77380943e-01 -8.14196587e-01
2.47562855e-01 2.47641578e-01 5.27562201e-01 -2.80429572e-01
5.69273293e-01 -6.76318705e-01 3.71277452e-01 5.37256658e-01
-4.87276286e-01 1.69239398e-02 2.74402887e-01 4.31728303e-01
-5.76108813e-01 -2.85274774e-01 7.32274950e-01 -3.28174561e-01
-4.87708628e-01 3.55330288e-01 -3.64135325e-01 -3.27182949e-01
7.45015264e-01 5.15545383e-02 -5.81282616e-01 -2.85208374e-01
-8.47179294e-01 -7.77796432e-02 3.08229744e-01 3.84303391e-01
9.39075053e-01 -1.32981634e+00 -7.95164943e-01 5.18516377e-02
6.55471802e-01 -4.12946761e-01 7.87147999e-01 1.34172618e+00
-4.42942172e-01 5.69817722e-01 -3.41370285e-01 -8.44669819e-01
-1.49581647e+00 6.12366676e-01 4.35441762e-01 -6.33014023e-01
-3.44567060e-01 6.59820974e-01 4.03467029e-01 -5.09766698e-01
-2.50460505e-01 2.09108889e-01 -3.61466557e-01 3.02569389e-01
7.05821097e-01 8.47309753e-02 3.24484468e-01 -9.11155879e-01
-4.82514709e-01 8.85374010e-01 -3.20001602e-01 -6.74019232e-02
1.42568803e+00 -1.09778762e-01 -5.45773327e-01 3.47122401e-01
1.63866222e+00 -1.77320391e-01 -1.55331165e-01 -3.34295064e-01
-1.79899961e-01 -4.86734897e-01 5.82029782e-02 -7.90826261e-01
-1.26830041e+00 9.13833141e-01 1.20863163e+00 1.73455849e-01
1.34008956e+00 5.91729343e-01 5.08887351e-01 2.12769270e-01
2.53360152e-01 -5.56724906e-01 1.89034924e-01 2.95402505e-03
1.22829390e+00 -1.54352582e+00 2.83600569e-01 -1.12789080e-01
-3.85431111e-01 1.16064513e+00 -9.50653628e-02 6.07828200e-02
7.29469419e-01 -3.65012529e-04 1.72880158e-01 -6.23647273e-01
-5.84822774e-01 -3.52702171e-01 6.34551466e-01 5.13189614e-01
7.07943738e-01 -1.48347721e-01 -3.83364618e-01 1.04748599e-01
3.39292169e-01 -1.72891334e-01 1.09810762e-01 9.65169847e-01
1.19876280e-01 -1.30222774e+00 -3.30675066e-01 5.59025943e-01
-7.54550576e-01 -9.98154059e-02 -2.76312828e-01 6.86811686e-01
-1.60038978e-01 6.49240136e-01 -1.61363229e-01 -1.07681178e-01
-2.22694371e-02 -1.10844746e-01 6.42992079e-01 -2.78982908e-01
-5.16330004e-01 2.30778322e-01 -4.94692951e-01 -8.17261994e-01
-9.78582025e-01 -6.32992804e-01 -9.08527911e-01 1.43275142e-01
-2.31977805e-01 -9.71354693e-02 7.12436974e-01 6.72787070e-01
6.13922834e-01 4.27752316e-01 4.85617220e-01 -9.92469192e-01
-6.62416890e-02 -7.35770702e-01 -5.67727447e-01 8.23959947e-01
5.08469522e-01 -6.50668740e-01 -3.79400015e-01 1.23505518e-01] | [14.363710403442383, -1.5736665725708008] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.